Ninein domains required for its localization, association with partners dynein and ensconsin, and microtubule organization

Abstract

Ninein (Nin) is a microtubule (MT) anchor at the subdistal appendages of mother centrioles and the pericentriolar material (PCM) of centrosomes that also functions to organize MTs at noncentrosomal MT-organizing centers (ncMTOCs). In humans, the NIN gene is mutated in Seckel syndrome, an inherited developmental disorder. Here, we dissect the protein domains involved in Nin’s localization and interactions with dynein and ensconsin (ens/MAP7) and show that the association with ens cooperatively regulates MT assembly in Drosophila fat body cells. We define domains of Nin responsible for its localization to the ncMTOC on the fat body cell nuclear surface, localization within the nucleus, and association with Dynein light intermediate chain (Dlic) and ens, respectively. We show that Nin’s association with ens synergistically regulates MT assembly. Together, these findings reveal novel features of Nin function and its regulation of a ncMTOC.

• Ninein is an important regulator of MTs, but its mechanism of action is mostly unknown.

• This work shows that Ninein works together with ensconsin to promote MT assembly. A predicted alpha-helical domain in Ninein associates with ensconsin and, together, Ninein and ensconsin synergistically promote MT assembly. The N-terminal domain associates with dynein light intermediate chain, consistent with human Ninein. Domains that target Ninein to the MT-organizing center and to the nucleus were also mapped.

• This work advances our understanding of Ninein’s role in MT regulation through its association with ensconsin.

INTRODUCTION

Ninein (Nin) is a microtubule (MT)-anchoring protein that localizes to subdistal appendages on the mother centriole, within the pericentriolar material (PCM) at the centrosome, and at subcellular sites where noncentrosomal MT-organizing centers (ncMTOCs) are organized (Bouckson-Castaing et al., 1996; Mogensen et al., 2000; Dammermann and Merdes, 2002; Ou et al., 2002; Delgehyr et al., 2005; Dyachuk et al., 2016; Kowanda et al., 2016; Zheng et al., 2016; Muroyama and Lechler, 2017; Sanchez and Feldman, 2017; Wu and Akhmanova, 2017; Paz and Luders, 2018; Tillery et al., 2018; Vineethakumari and Luders, 2022). Mutations in NIN, also known as SCKL7, are linked to Seckel syndrome, a type of congenital microcephalic primordial dwarfism disorder (Dauber et al., 2012). Underscoring its role in growth and development, Nin is essential in neural progenitor cells (NPCs) where it is needed for the asymmetric segregation of mother and daughter centrosomes (Wang et al., 2009), for cell cycle-dependent nuclear movement and MT aster formation at centrosomes (Shinohara et al., 2013), and for NPC maintenance and segregation behavior of the older centrosome (Royall et al., 2023). During epidermal progenitor cell division, Nin is required for mitotic spindle orientation. A Nin null mutation is semilethal in mice, and survivors show disruption of desmosomes and lamellar body secretion in keratinocytes, resulting in a thin-skin phenotype (Lecland et al., 2019). Nin mutant mice also have significant defects in bone development, with associated deficits in MT assembly from centrosomes and centrosome cohesion and clustering defects in osteoclasts (Gilbert et al., 2024). In macrophages, Nin is important with dynein for phagocytosis and for phagosome trafficking (Omer et al., 2024).

Ninein-like protein (Nlp), a Nin paralogue, is also a centrosomal protein (Casenghi et al., 2003) with indirect links to ciliopathies (van Wijk et al., 2009). Like Nin, Nlp associates with dynein (Redwine et al., 2017) and γ-tubulin (Casenghi et al., 2003). Nlp was shown to be an essential component of the antiviral innate immune response. NINL human knockout cells showed enhanced viral replication, making them more susceptible to infection (Stevens et al., 2022). In contrast, Drosophila has just one Nin orthologue (also known as Bsg25D), and Nin null mutants are viable and fertile (Kowanda et al., 2016; Zheng et al., 2016; Rosen et al., 2019). Despite these findings establishing the importance of Nin and Nlp in health and development, little is understood about the molecular functions of Nin and how it organizes an MTOC.

In addition to its localization at centrosomes, Nin is also a component of ncMTOCs where it also functions as an MT anchor (Mogensen et al., 2000; Casenghi et al., 2003; Delgehyr et al., 2005; Wang et al., 2015; Zheng et al., 2016; Sanchez and Feldman, 2017; Tillery et al., 2018). Examples of Nin’s involvement at ncMTOCs include localizing apically in mammalian cochlear cells (Mogensen et al., 2000), at the cell cortex in the murine epidermis (Lechler and Fuchs, 2007), and perinuclearly in mammalian and Drosophila myotubes (Bugnard et al., 2005; Rosen et al., 2019) and Drosophila larval fat body cells (Zheng et al., 2020). In differentiating keratinocytes, Nin is necessary for the cortical organization of MTs and the relocalization of MT-organizing proteins to the cell cortex (Lecland et al., 2019). In epithelial cells, the development of an apical-basal polar array of MTs involves a switch from centrosomal MTs that requires Nin and its trafficking by CLIP-170 (Goldspink et al., 2017). In the developing vasculature, Nin is required to control tubular morphogenesis of angiogenic endothelial cells (Matsumoto et al., 2008). In the mouse brain, alternative splicing of the Nin transcript results in expression of a noncentrosomal isoform, implicating a noncentrosomal role in neurons as well (Zhang et al., 2016).

Recent findings indicate functional requirements for Nin at ncMTOCs. When centrioles are experimentally eliminated from interphase human cell culture, a single acentriolar MTOC forms from the assembly of PCM proteins CDK5RAP2, Pericentrin, Nin, and γ-tubulin. Nin is required for the assembly of this ncMTOC, acting late in the assembly process to promote a coalescence of smaller PCM clusters into a compact MTOC and for the formation of the radial MT network (Chen et al., 2022).

In Caenorhabditis elegans larval epidermis, the Nin orthologue NOCA-1 functions in parallel with the MT stabilizer and Patronin/CAMSAP orthologue PTRN-1 to organize a ncMTOC critical for growth and morphogenesis. NOCA-1 works with γ-tubulin, while PTRN-1 does not (Wang et al., 2015). In C. elegans neurons, dendritic ncMTOCs on RAB-11-positive vesicles require parallel activities of PTRN-1 and NOCA-2, a closer orthologue to Nin than NOCA-1 (He et al., 2022).

In Drosophila embryonic muscle, Nin is not essential for nuclear positioning, but its loss sensitizes cells to heterozygous loss of ens to affect nuclear positioning. Additionally, Nin overexpression impairs nuclear positioning, but cooverexpression of ens suppresses this (Rosen et al., 2019). In Drosophila larval fat body cells, a perinuclear MTOC requires the parallel activities of Nin and Patronin (Zheng et al., 2020). The Nesprin Muscle-specific protein 300 kDa (Msp300) organizes the fat body perinuclear MTOC by recruiting Patronin and the MT polymerase mini spindles (msps; Zheng et al., 2020). This ncMTOC controls MT assembly, nuclear positioning, and retrograde endocytic trafficking (Zheng et al., 2020). The role of Nin at the fat body MTOC, however, is unclear. Although progress has been made in understanding Nin’s localization to various MTOCs and its roles in cell function, development, and disease, we lack a clear understanding of how Nin functions in MT organization and MTOC control.

Human NIN binds to dynein motor (Redwine et al., 2017; Celestino et al., 2019; Lee et al., 2020); Drosophila Nin binds to MTs (Kowanda et al., 2016) and associates with ensconsin (ens/MAP7; Rosen et al., 2019); and human, Drosophila, and C. elegans Nin orthologues all associate with γ-tubulin (Casenghi et al., 2003; Delgehyr et al., 2005; Lin et al., 2006; Wang et al., 2015; Zheng et al., 2016). It is likely that these interactions are conserved across species that express Nin orthologues. Nin and Nlp associate with the dynein-dynactin complex and are activating dynein adapters (Redwine et al., 2017; Reck-Peterson et al., 2018; Celestino et al., 2019) that bind directly to DLIC (Celestino et al., 2019; Lee et al., 2020). Nin transport to the centrosome is dynein-dependent as inhibiting dynein through excess p50 (Dynamitin) or p150^glued CC1 causes a reduction of Nin at the centrosome (Dammermann and Merdes, 2002; Casenghi et al., 2005), and Nin may be trafficked along MTs via the dynein complex (Moss et al., 2007). Furthermore, loss of NINL leads to a reduction in dynein-dependent transport of intracellular cargoes (Stevens et al., 2022). In Drosophila embryonic muscle, Nin interacts and colocalizes with ens to control myonuclear positioning (Rosen et al., 2019). Currently, the functional relationships between Nin and its partners are poorly understood.

To understand how Nin functions in MT assembly, we investigated the respective contributions its protein domains make to its localization at the nuclear surface and its relationships with ens and dynein. These genetic and cell biological findings indicate that multiple domains contribute to Nin localization to the nuclear surface while a central domain is responsible for localizing Nin inside the nucleus. Furthermore, we confirm that Drosophila Nin binds to Dlic through the N-terminus of Nin. And finally, we map the ens-binding domain to Nin and show that Nin cooperates with ens to organize MTs.

RESULTS

Localization of Nin to the MTOC involves N- and C-terminal domains

We conducted a structure-function analysis of Nin (Figure 1A) in the Drosophila larval fat body, a tissue analogous to human liver or adipocytes that features large, monolayered cells with a perinuclear ncMTOC (Zheng et al., 2020). We generated a series of transgenic constructs that express regions of Nin fused with C-terminal TagRFP and Myc tags (Figure 1, B and C). We based the design of these constructs on previously annotated domains of Nin that include the N-terminal γ-tubulin-binding (Casenghi et al., 2003; Delgehyr et al., 2005; Zheng et al., 2016), MT-binding (Kowanda et al., 2016), and dynein-binding regions of Nin (Redwine et al., 2017; Celestino et al., 2019) and a central region of Nin reported to bind ens (Rosen et al., 2019).

FIGURE 1:

Nin transgenic constructs for in vivo structure-function analysis. (A) The full-length 1091-amino acid Drosophila Nin protein (isoform B), showing regions that bind γ-tubulin (aa 1-241) and MTs (aa 1-352). Solid black boxes represent predicted coiled-coil (CC) regions. (B) A full-length, transgenic construct (yellow box), in addition to 11 other constructs that divide Nin into domains (purple box) were generated for structure-function analysis. Constructs were tagged with TagRFP (red star) and Myc tags (blue diamond) at the C-terminus. Numbers reflect amino acids in Nin-PB. (C) Western blot of whole larval lysates probed with an antibody against Myc (red). Nin transgenes driven with SPARC-GAL4 express proteins of molecular weights consistent with their predicted sizes: 1-1091 = 156 kDa; 1-100 = 43.6 kDa; 1-242 = 59.2 kDa; 1-353 = 71.7 kDa; 1-451 = 83.4 kDa; 454-571 = 46.2 kDa; 1-571 = 97.1 kDa; Δ451-567 = 142.8 kDa; 454-1091 = 105.2 kDa; 568-1091 = 92.3 kDa; 550-924 = 74.1 kDa; and 778-1091 = 67.9 kDa. α-Tubulin (green, 50 kDa) served as a loading control. Asterisk denotes a nonspecific band. Black frames indicate lanes from a single gel.

Taking advantage of the Drosophila UAS-GAL4 binary expression system, we individually expressed these transgenes using SPARC-GAL4, a fat body driver, to evaluate their subcellular localizations in fat body cells (Figure 2; Supplemental Figure 1A). Co-IP of full-length Nin (Nin^1-1091-TagRFP-Myc) with Nin-GFP indicated that Nin can multimerize (Supplemental Figure 1B), so we compared Nin localization in a Nin¹ null mutant background (Figure 2) to a wild-type background (Supplemental Figure 1A). Localization patterns, however, did not differ whether endogenous Nin was absent (Figure 2) or present (Supplemental Figure 1A).

FIGURE 2:

Multiple Nin domains localize to the nuclear surface or inside the nucleus. IF staining of DNA (DAPI, green) in Nin¹ null mutant larval fat body cells expressing Nin transgenic constructs (TagRFP fluorescence, red). Scale bar = 10 μm in this and all subsequent figures. (A) Localization patterns of TagRFP-Myc-labeled Nin transgenes expressed in Nin¹ null mutant fat body cells. Boxed yellow regions are a magnified view of the nuclear surface. Multiple domains of Nin localize to the MTOC. Nin^1-100 and fragments containing amino acids 454–567 localize predominantly intranuclearly. (B) Quantification of results in (A). Six larvae from three experiments involving ∼100 cells per construct were measured for the relative localization of Nin at the periphery or inside the nucleus. Bars indicate the relative (percent) localization of Nin to the perinuclear surface versus inside the nucleus.

Nin^1-1091 localized to the nuclear surface in fat body cells (Figure 2; Supplemental Figure 1A). Localization was most prominent at the nuclear surface with a weaker signal present inside the nucleus (Figure 2). Perinuclear localization in punctate aggregates was variably present, which appeared to be due to overexpression as it was more prominent with two copies of the transgene than with one (Supplemental Figure 2A). Intranuclear localization also increased with two copies of Nin^1-1091 (Supplemental Figure 2A). Expression of two Nin transgenes that express at higher levels, Eos-Nin and Nin-GFP (Supplemental Figure 2B), resulted in large intranuclear and cytoplasmic punctate aggregates (Supplemental Figure 2C), indicating that these intranuclear and aggregate patterns correlate with a high dosage effect of Nin overexpression.

High levels of ubiquitous Nin overexpression are lethal ([Zheng et al., 2016] and Supplemental Figure 2D) and toxic in the fat body and muscle (Supplemental Figure 2D). Levels of Nin^1-1091 overexpression used in this study were not lethal in the fat body or muscle. However, ubiquitous overexpression of one copy of Nin^1-1091 with tub-GAL4 was semilethal (Supplemental Figure 2D), and its overexpression in the fat body with SPARC-GAL4 was not lethal but increased nuclear mispositioning (Supplemental Figure 3A). Overexpression with SPARC-GAL4 of two copies of Nin^1-1091 or other Nin transgenes that express at higher levels was lethal (Supplemental Figure 2B). Overexpression of each of the Nin transgenes shown in Figure 1 was not lethal in the fat body using SPARC-GAL4 (unpublished data).

Expression of Nin-TagRFP-Myc constructs containing N- (Nin^1-353 and Nin^1-451) or C-terminal (Nin^568-1091, Nin^550-924, and Nin^Δ451-567) domains localized predominantly to the nuclear surface (Figure 2; Supplemental Figure 1). An N-terminal fragment, Nin^1-100, localized mostly inside the nucleus (Figure 2; Supplemental Figure 1) and was the only construct to affect nuclear morphology by increasing nuclear circularity in the Nin¹ background, albeit to a small degree (Supplemental Figure 3B). Interestingly, a central region of Nin (Nin^454-571) also localized intranuclearly but, unlike the N-terminal 100 amino acids, drove all fragments containing it into the nucleus (Figure 2; Supplemental Figure 1). We used two programs to identify a potential nuclear localization signal (NLS) in these regions (see Materials and Methods), but none were predicted. While amino acids 454–567 target all fragments that contain it into the nucleus, this region is less efficient at targeting the full-length protein in the nucleus, possibly due to the presence of competing domains that localize Nin to the nuclear surface. Large fragments that lack this domain, including Nin^∆451-567, were primarily perinuclear (Figure 2; Supplemental Figure 1). These data indicate that at least two domains of Nin contribute to its localization to the nuclear surface while amino acids 454–567 drive the protein into the nucleus and Nin1-100 can passively localize there.

MT-dependent and -independent modes of Nin localization to the MTOC

Because Nin binds MTs and is an MT anchor (Mogensen et al., 2000; Abal et al., 2002; Delgehyr et al., 2005; Kowanda et al., 2016), we next sought to determine whether MTs play a role in Nin localization at the nuclear surface. Overexpression of Nin-TagRFP-Myc constructs variably affected MT organization in null or wild-type backgrounds with varied significance (Supplemental Figure 3C). However, disrupting MTs by overexpressing the MT-severing enzyme spastin or by knocking down the MT component α-tubulin by RNAi reduced, but did not eliminate, perinuclear localization of Nin^1-1091 while increasing the incidence of punctae formation (Figure 3). Thus, there appears to be MT-dependent and -independent modes of Nin localization to the nuclear surface.

FIGURE 3:

Localization of Nin to the fat body ncMTOC is partially dependent upon MTs. IF staining of larval fat body cells with DNA (DAPI, green), Nin^1-1091 (TagRFP fluorescence, red), F-actin (488 Phalloidin, cyan), and MTs (YL1/2, white). Insets show an enlarged view of the nuclear surface. Nin^1-1091 localizes to the perinuclear MTOC and forms aggregates when MTs are disrupted by the overexpression (OE) of CFP-spastin or by the RNAi-mediated KD of α-tubulin. Experiment was repeated at least nine times and included >18 larvae.

We next sought to determine which proteins anchor or recruit Nin to the nuclear surface. Our previous work showed that the Linker of Nucleoskeleton and Cytoskeleton (LINC) complex protein Msp300/Nesprin was a major structural component of the fat body MTOC required to recruit msps and Patronin to the nuclear surface to organize and assemble MTs there (Zheng et al., 2020). Knockdown of Msp300 in the fat body increased nuclear mispositioning due to MT disruption (Zheng et al., 2020). Loss of Msp300 (Figure 4A) or Patronin (Figure 4B) reduced perinuclear Nin and increased its intranuclear localization and cytoplasmic aggregate accumulation (Figure 4, A and B, arrowheads). However, there was no overt effect on localization of Nin to the nuclear surface when msps (Figure 4C) or the other Drosophila Nesprin klarsicht (klar) was knocked down (Figure 4D). Therefore, Nin localization to the nuclear surface relies partly on MTs, but also on a MT-independent anchor that involves Msp300 and Patronin. Whether these interactions are direct or indirect remains to be determined.

FIGURE 4:

Fat body MTOC proteins Msp300 and Patronin are necessary for perinuclear Nin localization. IF staining of larval fat body cells expressing Nin^1-1091 using SPARC-GAL4 in a wild-type background with RNAi KD of the genes indicated. DNA (DAPI, green), Nin^1-1091 (TagRFP fluorescence, red), F-actin (488 Phalloidin, cyan), and MTs (YL1/2, white). The right panels show an enlarged view of the nuclear surface. Experiments were repeated at least nine times and included >18 larvae. (A and B) KD of Msp300 or Patronin reduces Nin localization to the nuclear surface and increases the formation of cytoplasmic aggregates (arrowheads). (C and D) KD of msps or klar using either of two RNAi lines (GD9271 above, HMS01612 below) does not affect Nin localization.

Ensconsin but not dynein is required for Nin localization to the MTOC

We next analyzed whether Nin requires dynein and/or ens for localization. Loss of Dynein heavy chain (Dhc) or Dynein light intermediate chain (Dlic), a direct partner of NIN (Celestino et al., 2019; Lee et al., 2020), blocks the trafficking of Rab5 vesicles from the cell membrane to the nuclear periphery ((Zheng et al., 2020) and Supplemental Figure 4A). Furthermore, depletion of dynein subunits or dynein inactivation from the overexpression of DCTN2-p50/Dynamitin did not alter MT organization (Zheng et al., 2020), nor did it visibly reduce perinuclear Nin localization (Figure 5A). Loss of ens, on the other hand, did disrupt Nin localization to the nuclear surface, resulting in perinuclear Nin aggregate formation (Figure 5B).

FIGURE 5:

Nin localization is dependent upon ensconsin, not dynein. IF staining of larval fat body cells expressing Nin^1-1091 using SPARC-GAL4 in a wild-type background with RNAi KD of the genes indicated. DNA (DAPI, green), Nin^1-1091 (TagRFP fluorescence, red), F-actin (488 Phalloidin, cyan), and MTs (YL1/2, white). Insets show an enlarged view of the nuclear surface. Experiments were repeated at least 6 times and included >12 larvae. (A) KD of Dynein heavy chain (Dhc) or Dlic or inactivation of the dynein motor with DCTN2-p50/Dynamitin overexpression (OE) does not affect Nin localization to the MTOC. (B) Loss of ensconsin (ens) perturbs Nin localization, causing it to aggregate near the nucleus.

Nin interacts with ens and Dlic

We next aimed to determine how Nin coordinates MT organization and if its partners play a role by first confirming whether Nin associates with ens and/or dynein in fat body cells. We overexpressed Nin-GFP in fat body cells using SPARC-GAL4, treated lysates with nocodazole on ice to depolymerize MTs, and pulled down Nin-GFP using GFP nanobody beads. We detected an association with endogenous ens (Figure 6A). Using the Nin constructs (Figure 1), we further mapped the interaction domain by overexpressing ens-GFP together with various Nin constructs in fat bodies, pulling down ens-GFP, and probing for the Myc tag on Nin fragments. The C-terminus of Nin (Nin^568-1091 and Nin^550-924) coimmunoprecipitated with ens (Figure 6B, green arrows), but a large N-terminal portion (Nin^1-571) and a C-terminal portion (Nin^778-1091) did not (Figure 6B, red arrows). We infer from the co-IP data that the minimum interaction domain between Nin and ens is within amino acids 568–777 of Nin (Figure 6B, see also Figure 7).

FIGURE 6:

Mapping of Nin’s ens- and Dlic-interaction domains. Western blot analysis of GFP immunoprecipitates from whole larval lysates expressing the indicated proteins in fat bodies. Unbnd = Unbound fraction in lysate after IP. Green arrows indicate a positive co-IP, and red arrows, a negative. co-IPs were repeated at least three times. (A) co-IP of ens with Nin-GFP. The membrane was probed for endogenous ens (∼150 and 100 kDa) and GFP. Endogenous ens (∼150 kDa) coimmunoprecipitated with Nin. (B) co-IP of Nin-TagRFP-Myc fragments with ens-GFP. The membrane was probed for GFP (ens-GFP; ∼180 kDa) and Myc (Nin; various sizes). (C) co-IP of Nin-TagRFP-Myc fragments with Dlic-GFP. The membrane was probed for GFP (Dlic-GFP; ∼81 kDa) and Myc (Nin; various sizes). Nin^550-924 + ens-GFP was used as a positive control.

FIGURE 7:

Nin and ensconsin synergize to form an ectopic MTOC that disrupts nuclear morphology. DNA (DAPI, blue), Nin^1-1091 (TagRFP fluorescence, red), ens (GFP fluorescence, green), and MTs (YL1/2, white) or LamC (LC28.26, white). (A) Cooverexpression of ens and fragments that include amino acids 568–777 of Nin alters MT organization, forming a robust juxtanuclear MTOC (pink arrowheads) associated with late pupal lethality. (B) LamC staining labels the inner nuclear membrane and is reduced where Nin-ens ectopic MTOCs are formed (orange arrowheads). Additionally, nuclear DNA spills into the cytoplasm (yellow arrowheads).

Using a similar approach, we mapped the domain that binds dynein. In agreement with previous work that identified NIN as an activator of dynein (Redwine et al., 2017; Reck-Peterson et al., 2018) that binds directly to DLIC1 and DLIC2 via its N-terminal 87 amino acid EF hand domains (Celestino et al., 2019; Lee et al., 2020), the N-terminus of Drosophila Nin (Nin^1-100 and Nin^1-451) coimmunoprecipitated with Dlic-GFP (Figure 6C, green arrows) while central (Nin^454-571) and C-terminal Nin fragments (Nin^454-1091 and Nin^568-1091) did not (Figure 6C, red arrows). A longer N-terminal fragment, Nin^1-571, was unexpectedly not detected in the co-IP with Dlic-GFP, possibly due to lower expression levels. We were unable to detect an association between Dynein intermediate chain (Dic) and Nin (Supplemental Figure 4B). A summary of the mapping of ens and Dlic binding domains to Nin is shown in Figure 8.

FIGURE 8:

Summary of Nin domains and model depicting its cooperative role with ensconsin at the MTOC. (A) Summary of Nin domains mapped from this work that includes the Dlic and ens binding domains and the nuclear localization domain. (B) Location of Nin domains on the three-dimensional structure of Nin-PB predicted by AlphaFold (Jumper et al., 2021; Varadi et al., 2022). Color coding matches that in (A). (C) Model for the possible roles of Nin and ens at the fat body perinuclear MTOC, building on the model proposed recently (Zheng et al., 2020). KASH = Klarsicht, ANC-1, Syne Homology domain. CTD = C-terminal domain. CKK = CAMSAP1, KIAA1078, KIAA1543 domain. MT BD = MT binding domain. KBD = Kinesin binding domain.

Nin synergizes with ens to promote MT assembly

Our next goal was to evaluate whether Nin’s interaction with dynein or ens was sufficient for MTOC formation. Ens and Dlic colocalize with MTs at the MTOC (Supplemental Figure 5, A and B). Dlic-GFP was predominantly nuclear when overexpressed and appeared to drive full-length Nin into the nucleus (Supplemental Figure 5C). Remarkably, cooverexpression of ens-GFP with full-length Nin or constructs containing a region near Nin’s C-terminus (Nin^454-1091, Nin^568-1091, and Nin^550-924) produced robust juxtanuclear MTOCs that superseded the perinuclear MTOC (Figure 7A). Overexpression of Nin or ens alone did not induce an ectopic MTOC, whereas together, they synergize to organize MTs. These robust, ectopic MTOCs disrupted nuclear integrity as evidenced by alterations to the pattern of LaminC (LamC), a nuclear lamin nucleoskeleton component (Figure 7B, orange arrowheads). LamC staining was diminished at the nucleus adjacent to the position of the ectopic MTOC (Figure 7B, orange arrowheads), and chromosomal material appears to have spilt into the cytoplasm (Figure 7B, yellow arrowheads).

We infer from these results that amino acids 568-777 of Nin are sufficient for the formation of an MTOC in cooperation with ens. This is consistent and overlaps with the MT-independent ens-interacting domain identified from the co-IP experiments (see Figure 6B). When MTs are disrupted, Nin and ens still coalesce, but fail to organize into a single focus, indicating that MTs contribute to the organization of the single ectopic MTOC (Supplemental Figure 5D). The synergy between Nin and ens in forming an MTOC is unique to that partnership as overexpression of Nin^1-1091 together with either Dlic or Dhc did not overtly alter MT organization (Supplemental Figure 5C).

DISCUSSION

In this study, we generated a set of transgenic Nin deletion constructs in Drosophila that enabled us to map Nin’s localization, partner-binding, and MT-regulating domains in vivo in the larval fat body. From these data, we identify domains of Nin that contribute to its localization at the MTOC on the nuclear surface and inside the nucleus, we confirm the orthologous association of the N-terminus with Dlic that was previously mapped on human Nin, and we map the domain that associates with ens (Figure 8). From coexpression assays, we found a synergistic interaction between ens and Nin’s ens-binding domain that is sufficient to generate a robust ectopic MTOC.

Full-length tagged Nin (Nin^1-1091) localizes primarily to the nuclear surface, while lower levels are detected inside the nucleus. We further determined that a 114-amino acid domain at positions 454–567 is necessary and sufficient to target Nin into the nucleus (Figure 8, A and B, green). All subfragments of Nin containing this domain showed significant localization within the nucleus; however, the full-length Nin protein had relatively low levels inside the nucleus, possibly due to the presence of competing domains in the N- and C-terminal regions that target Nin to the nuclear surface. A construct lacking this domain (Nin^∆451-567) localized less within the nucleus than full-length Nin, reinforcing the significance of this domain in targeting Nin to the nucleus.

Localization of Nin to the nucleus was shown in human cultured cells overexpressing hNinein (Cheng et al., 2006) and in Drosophila muscle cells overexpressing Drosophila Nin (Rosen et al., 2019). In both cases, nuclear localization was linked to Nin SUMOylation (Cheng et al., 2006; Rosen et al., 2019). If this mechanism holds in fat body cells, it suggests that the nuclear targeting domain identified at 454-567 in Drosophila Nin is a candidate domain for SUMO modification.

Expression of the N-terminal 100 amino acids of Nin was also localized within the nucleus, but larger fragments that included this domain did not localize predominantly within the nucleus. Therefore, unlike the nuclear localization domain at 454-567, this domain was not sufficient for targeting, and we suggest that it passively localizes into the nucleus due to its small size and may not play a role in the nuclear localization of full-length Nin. The expected molecular weight of this domain with the engineered tag is ∼44 kDa, which is within the limits (30–60 kDa) of passive diffusion between nucleus and cytoplasm (Keminer and Peters, 1999; Wang and Brattain, 2007; Timney et al., 2016). This could explain why longer fragments such as Nin^1-353 and Nin^1-451 do not prominently localize inside the nucleus but are instead localized primarily to the perinuclear MTOC.

N- and C-terminal Nin-TagRFP-Myc constructs localize to the MTOC at the nuclear surface. The contribution of multiple domains of Nin conferring localization to the MTOC may point to its interactions with several distinct partners, including MTs, ens, etc. that anchor Nin at the MTOC. The fat body nuclear surface has prominent circumferential MTs (Zheng et al., 2020), and recruitment of full-length Nin is largely, but not completely, dependent on them. Depolymerizing MTs may release a MT-interacting pool of Nin, freeing up Nin to aggregate, a phenotype similar to higher levels of Nin overexpression. We propose that Nin localizes to the nuclear surface through a combination of MT-dependent and -independent modes, with the MT-independent mode being dependent on Msp300, Patronin, and/or ens. Nin is a MT-associated protein, and so the MT-dependent mode is likely attributed to the high density of circumferential MTs associated with the fat body nucleus. Both pools of Nin likely contribute to MTOC function.

Human DLIC1 was shown to bind directly to NIN via a pair of Ca⁺⁺-independent EF hand domains at the N-terminal 87 amino acids of NIN (Lee et al., 2020), a region that is highly conserved between human and Drosophila nineins (Zheng et al., 2016). We confirm that Drosophila Nin is also associated with Dlic and have mapped this interaction to the N-terminal end of Nin (Figure 8, A and B, pink).

When Dlic-GFP was overexpressed in fat body cells, it localized in the nucleus. This was unexpected as Dlic’s role as a subunit of the dynein motor complex places its function with MTs in the cytoplasm. Other studies have used this and other tagged UAS-Dlic transgenes in tissues other than the fat body; however, their localization was not detected or reported to be nuclear (Pandey et al., 2007; Emre et al., 2011; Wainman et al., 2012; Baumbach et al., 2015; Inaba et al., 2015). Endogenous Dlic localized at the plasma membrane and at the perinuclear MTOC. Although Dlic-GFP overexpression drove Nin into the nucleus, this nuclear localization dynamic did not overtly impact MT organization. Furthermore, loss of either Dhc or Dlic or inactivation of dynactin does not affect Nin localization to the perinuclear MTOC or MT organization in fat body cells. Although the interaction between Nin and Dlic may be important in other cellular processes, this study did not reveal a role for it in MT organization in the fat body.

From the co-IP data, we identified a Nin-ens interaction domain between amino acids 568–777 of Nin (Figure 8, A and B, yellow), which was consistent with the colocalization of Nin and ens from IF staining. We further demonstrated that, when overexpressed, Nin and ens function synergistically to form an ectopic ncMTOC. When Nin constructs that include amino acids 568–777 are overexpressed with ens, robust ectopic MTOCs are formed adjacent to the nucleus. Interestingly, Nin’s interaction with overexpressed ens is sufficient to overcome Nin’s nuclear localization (compare Nin^454-1091 Figure 7, A and B to Supplemental Figure 1A). In contrast to the developing muscle, where multiple foci formed from Nin-ens overexpression (Rosen et al., 2019), there is only one Nin-ens MTOC per fat body cell. Overexpression of Nin-TagRFP-Myc fragments or ens-GFP alone in the fat body produced no overt effects, and flies were viable, whereas cooverexpression of fragments containing Nin’s ens-binding domain together with ens-GFP was late pupal lethal. In contrast with these synergistic effects of Nin and ens expression in the fat body, similar overexpression experiments in the developing muscle showed that Nin overexpression was lethal and was suppressed, rather than enhanced, by cooverexpression of ens (Rosen et al., 2019). Higher expression of Nin in the fat body above the levels achieved with Nin-TagRFP-Myc (expression of 2X Nin-TagRFP-Myc, Nin-GFP, or Eos-Nin) was lethal, but the cause of lethality is unclear. Whether the MTOC that Nin-ens generates involves other factors remains to be determined. It likely does not require γ-tubulin because γ-tubulin is expressed at very low levels in the fat body and was not required for MT assembly at the primary fat body MTOC (Zheng et al., 2020).

In the fat body, Nin-ens ectopic MTOCs were positioned near the nucleus, disrupted the nuclear envelope, altered nuclear morphology, and resulted in the leakage of chromosomal material into the cytoplasm. One function of the fat body MTOC under normal conditions may be to ensure proper nuclear morphology. Normally, MTs in the fat body are organized both in circumferential bundles surrounding and with their minus ends anchored at the nuclear surface (Zheng et al., 2020). With the generation of Nin-ens ectopic MTOCs, the nuclear morphology changes appear to be the result of the forces generated by imposing MTs emanating from the newly established ectopic MTOC as depletion of MTs under these conditions attenuated disruption of nuclear morphology. The disruption of LamC patterning at the nuclear periphery is consistent with the fragility of these nuclei, as mutations in lamins can produce a similar phenotype (Davidson and Lammerding, 2014). Moreover, impinging MTs can impact nuclear morphology, a phenomenon also correlated with the reduction of lamin signal (Biedzinski et al., 2020; Heffler et al., 2020). LamC signal was not disrupted by Nin-ens cooverexpression when MTs were also disrupted, pointing to the role of Nin-ens-organized MTs in nuclear morphology changes and the loss of structural integrity of the nucleus. Presumably, the lethality associated with Nin and ens cooverexpression is due to these nuclear disruptions.

Altogether, our findings reveal novel features of domains required for Nin’s localization to the fat body MTOC on the nuclear surface, localization inside the nucleus, its interactions with dynein and ens, and how those interactions impact Nin localization and MT organization in fat body cells. Furthermore, we identify a synergistic effect between Nin and ens capable of coordinating an MTOC, implying a mechanistic connection between this partnership normally at the fat body perinuclear and possibly at other MTOCs. It will be interesting to discover whether human NIN has a cooperative role with MAP7 (ens orthologue) in organizing MTs and whether this connection has disease relevance.

MATERIALS AND METHODS

Request a protocol through Bio-protocol.

Generation of UAS-Nin-TagRFP-Myc transgenes

The Nin coding sequence (Nin-RB isoform) was amplified by polymerase chain reaction (PCR) from the LD21844 cDNA clone (RRID:DGRC_5314) using the primers 1.FWD and 1.REV. The PCR product, which included the entire ORF, was cloned into the pENTR/D-TOPO vector (Thermo Fisher Scientific, Catalogue #K240020). Using Gateway LR cloning, the coding sequence was inserted into pBID-UASC-GRM (RRID:Addgene_35203, [Wang et al., 2012]), creating a C-terminal TagRFP-Myc-tagged Nin construct (pBID-UASC-Nin-GRM). A version of pBID-UASC-Nin-GRM was also generated to be RNAi resistant to the Nin^HMS23837 RNAi line (RRID:BDSC_62414). The primers used to create this plasmid introduced silent mutations (CAAGAGAT­TTCAAGTCT­CCAG, mutations underlined) in the RNAi recognition motif (CAGG­AAATCAGTTCA­CTGCAA). Either of these plasmids was used as a template for the construction of Nin constructs shown in Figure 1. PCR fragments were inserted into EcoR1- and Bsu36I-digested pBID-UASC-GRM plasmid using NEBuilder HiFi DNA Assembly (New England BioLabs, Catalogue #E5520S). Primers were designed using SnapGene and purchased from Integrated DNA Technologies. Primers and sequences are shown below. Constructs were verified by sequencing and transgenes were then inserted at VK40(3R) attP docking site by ΦC31 integration by GenetiVision, and screened in our lab. Transgenic lines were selected by expression of mini-white and confirmed by Western blotting and IF staining.

Primer name	Sequence	Length
1.FWD	CACCATGGAGGTATCCGCCGAT	22-mer
2.FWD	GACCCAGCTTTCTTGTACAAAGTGGTTGAT	30-mer
3.FWD	CTTTGAATCACAAGACGCATACCAAACGATGGAGGTATCCGCCGATCCGTAC	52-mer
4.FWD	CTTTGAATCACAAGACGCATACCAAACGATGGAACTGGCCCAAACGTCGAGCAGCATT	58-mer
5.FWD	CGCAAGTGCACCGAAGGAGAG	21-mer
6.FWD	CTTTGAATCACAAGACGCATACCAAACGATGCGCAAGTGCACCGAAGGA	49-mer
7.FWD	CAAGAGATTTCAAGTCTCCAGTCAGAGATCGAGGATTTGCGACAG	45-mer
8.FWD	CTTTGAATCACAAGACGCATACCAAACGATGGATTCGCCGAGCAAAACACAT	52-mer
9.FWD	CTTTGAATCACAAGACGCATACCAAACGATGGGAAAAAGTCCAGCCAGCTCA	52-mer
1.REV	AGGCATGCCAGGCAGTCCA	19-mer
2.REV	GACTCTCTCCCATGTGAAGCCCTC	24-mer
3.REV	ATCAACCACTTTGTACAAGAAAGCTGGGTCGTATGTGTTGTTTAGCGGCTCATC	54-mer
4.REV	ATCAACCACTTTGTACAAGAAAGCTGGGTCGGTGTCCAGTGATTCCACG	49-mer
5.REV	ATCAACCACTTTGTACAAGAAAGCTGGGTCCACGGCAAGCAAAGCCA	47-mer
6.REV	ATCAACCACTTTGTACAAGAAAGCTGGGTCGAGCTTAATGTTCTGTTCGAGAAG	54-mer
7.REV	ATCAACCACTTTGTACAAGAAAGCTGGGTCGGTGCACTTGCGCTGTTT	48-mer
8.REV	ATCGCTCTGCTCTCCTTCGGTGCACTTGCGCTTAATGTTCTGTTCGAGAAG	51-mer
9.REV	CTGGAGACTTGAAATCTCTTGAGCCTTCTCCTCGAGCATTTGAAT	45-mer
10.REV	ATCAACCACTTTGTACAAGAAAGCTGGGTCCAGTGCCGACGGACTGCT	48-mer

Fly stocks

Flies were maintained on standard food. Crosses were conducted at 29°C unless otherwise stated. For crosses involving tub-GAL80^ts-7, crosses were started at 25° and then moved to 29° on the third day. w¹¹¹⁸ was used as wild-type.

Identifiers and sources for the fly strains used in this study can be found below. Bloomington Drosophila Stock Center = BDSC, Vienna Drosophila Resource Center = VDRC.

Mutant, RNAi, transgenic, or driver line used	Source
w[1118]	RRID:BDSC_3605
Nin[1]	(Zheng et al., 2016)
αTub84B[JF01373]	RRID:BDSC_31389
Dhc64C[GL00543]	RRID:BDSC_36583
Dlic[GD9681]	RRID:BDSC_41686
ens[HMS00933]	RRID:BDSC_40825
klar[HMS01612]	RRID:BDSC_36721
klar[GD9271]	VDRC_32836
Luc[JF01355]	RRID:BDSC_31603
Msp300[KK112156]	VDRC_107183
msps[HMS01906]	RRID:BDSC_38990
Patronin[GD11946]	VDRC_27654
UAS-DCTN2-p50 (Dynamitin)	RRID:BDSC_8784
UAS-HA-Dhc	(Silvanovich et al., 2003)
UAS-Dlic-GFP	(Zheng et al., 2008)
UAS-ens-GFP	(Rosen et al., 2019)
UAS-GFP	RRID:BDSC_5430
UAS-mCherry	RRID:BDSC_35787
UAS-Eos-Nin (III)	(Rosen et al., 2019)
UAS-Nin-GFP	(Zheng et al., 2016)
UAS-Nin-6 × Myc	(Zheng et al., 2016)
UAS-CFP-spastin	(Du et al., 2010)
Mef2-GAL4	RRID:BDSC_27390
SPARC-GAL4	RRID:BDSC_77473
SPARC-GAL4, GFP-Rab5	(Zheng et al., 2020)
tub-GAL4[LL7]	RRID:BDSC_5138

The following lines were generated for this study:

Line
y[1], w[*];; UAS-Nin[1-1091]-TagRFP-Myc
y[1], w[*];; UAS-Nin[1-100]-TagRFP-Myc
y[1], w[*];; UAS-Nin[1-242]-TagRFP-Myc
y[1], w[*];; UAS-Nin[1-353]-TagRFP-Myc
y[1], w[*];; UAS-Nin[1-451]-TagRFP-Myc
y[1], w[*];; UAS-Nin[Δ451-567]-TagRFP-Myc
y[1], w[*];; UAS-Nin[1-571]-TagRFP-Myc
y[1], w[*];; UAS-Nin[454-571]-TagRFP-Myc
y[1], w[*];; UAS-Nin[454-1091]-TagRFP-Myc
y[1], w[*];; UAS-Nin[568-1091]-TagRFP-Myc
y[1], w[*];; UAS-Nin[550-924]-TagRFP-Myc
y[1], w[*];; UAS-Nin[778-1091]-TagRFP-Myc
w[*]; Nin[1]; SPARC-GAL4/TM6B, Hu[1], Tb[1]
w[*]; Nin[1]; UAS-Nin[1-1091]-TagRFP-Myc
w[*]; Nin[1]; UAS-Nin[1-100]-TagRFP-Myc
w[*]; Nin[1]; UAS-Nin[1-242]-TagRFP-Myc
w[*]; Nin[1]; UAS-Nin[1-353]-TagRFP-Myc
w[*]; Nin[1]; UAS-Nin[1-451]-TagRFP-Myc
w[*]; Nin[1]; UAS-Nin[Δ451-567]-TagRFP-Myc
w[*]; Nin[1]; UAS-Nin[1-571]-TagRFP-Myc
w[*]; Nin[1]; UAS-Nin[454-571]-TagRFP-Myc
w[*]; Nin[1]; UAS-Nin[454-1091]-TagRFP-Myc
w[*]; Nin[1]; UAS-Nin[568-1091]-TagRFP-Myc
Nin[1]; SPARC-GAL4, mCherry.nls[3]/TM6B, Hu[1], Tb[1]
w[*]; SPARC-GAL4, UAS-ens-GFP/TM6B, Hu[1], Tb[1]
SPARC-GAL4, UAS-Dlic-GFP/TM6B, Hu[1], Tb[1]
w[*]; UAS-Nin-GFP; tub-GAL4[LL7], tub-GAL80[ts-7]/TM6B, Hu[1], Tb[1]
αTub84B[JF01373], UAS-Nin[550-924]-TagRFP-Myc (II)
w[*];; SPARC-GAL4, UAS-Nin[1-1091]-TagRFP-Myc
w[*];; SPARC-GAL4, UAS-Nin[1-100]-TagRFP-Myc
w[*];; SPARC-GAL4, UAS-Nin[1-242]-TagRFP-Myc
w[*];; SPARC-GAL4, UAS-Nin[1-353]-TagRFP-Myc
w[*];; SPARC-GAL4, UAS-Nin[1-451]-TagRFP-Myc
w[*];; SPARC-GAL4, UAS-Nin[454-571]-TagRFP-Myc
w[*];; SPARC-GAL4, UAS-Nin[1-571]-TagRFP-Myc
w[*];; SPARC-GAL4, UAS-Nin[Δ451-567]-TagRFP-Myc
w[*];; SPARC-GAL4, UAS-Nin[568-1091]-TagRFP-Myc
w[*];; SPARC-GAL4, UAS-Nin[550-924]-TagRFP-Myc
w[*];; SPARC-GAL4, UAS-Nin[778-1091]-TagRFP-Myc
w[*];; SPARC-GAL4, UAS-Nin[454-1091]-TagRFP-Myc

Survey of Nin lethality

GAL4 driver virgin females were crossed to the indicated transgenes. UAS-mCherry and UAS-GFP/TM6B were used as controls. Progeny were screened for viability (i.e., eclosion from pupal case). Pupal lethality was also assessed and quantified. Crosses were repeated three times, and the results were averaged. RStudio “Ghost Orchid” Release (8b9ced18, 2021-11-08) 2021.09.1 Build 372 was used to graph the results compiled in a separate spreadsheet. The following code was used:

library(“ggplot2”)

Nin$Tag ← factor(Nin$Tag, levels = c(“Nin-TagRFP-Myc”, “Nin-GFP”, “Eos-Nin”, “mCherry”, “GFP”))

Nin$GAL4 ← factor(Nin$GAL4, levels = c(“tub-GAL4”, “SPARC-GAL4”, “GAL4-Mef2”))

ggplot(Nin,

 aes(x = Tag,

  y = Percent,

  fill = Key)) +

 geom_bar(stat = “identity”,

   #color = “black”,

   position = “stack”) +

 scale_fill_grey(start = 0.8, end = 0.1) + theme_classic() +

 facet_grid(∼ GAL4) +

 #labs(title = “Nin OE Lethality”) +

 theme(axis.text.x = element_text(angle = 45, hjust = 1))

NLS database search

The sequences of Nin PB from amino acids 1–100 and 454–567 were separately analyzed with the following programs to check for nuclear localization sequences: NLStradamus (Nguyen Ba et al., 2009) and NLS Predictor (NovoPro). Default settings were used for each program.

Immunostaining and imaging

Fat bodies were dissected from two wandering third instar larvae in 1X DPBS (Life Technologies, Ref#14080-055) and mounted on poly-Lysine-treated slides (Kao and Megraw, 2004; VWR, Catalogue#89085-339) in 12 μL of 4% paraformaldehyde (PFA, Spectrum, Catalogue#P1010). Two to four slides were prepared for each condition. After ∼8 min in PFA, a siliconized 22 × 22 coverslip (VWR, Catalogue#48366-227) was applied, and the tissues were allowed to flatten for ∼1 min before the slide was snap frozen in liquid nitrogen. Once frozen, the slide was removed from the liquid nitrogen, and the coverslip was quickly pried off using a razor blade. The slide was then immediately placed into a Coplin jar filled with phosphate-buffered saline (PBS), pH 7.3 (137 mM NaCl (OmniPur, Catalogue#7760), 2.7 mM KCl (OmniPur, Catalogue#7300), 8 mM Na₂HPO₄ (VWR, Catalogue#0404-500G), 1.4 mM KH₂PO₄ (Millipore, Catalogue#529568-250GM) before application of primary antibodies.

Once all samples had been processed, slides were dried, and a hydrophobic ring was drawn around the tissue sample using a Liquid Blocker Super PAP Pen (Electron Microscopy Sciences, Catalogue#71312). The sample was incubated overnight at 4°C with a blocking solution (PBS, 0.5% bovine serum albumin [Boehringer Mannheim Corp., Catalogue#100-021], 0.1% saponin [Sigma, Catalogue#S-2149]) in a dark humid chamber. The next day, slides were incubated with primary antibody either overnight at 4°C or for 4–6 h at room temperature. Following incubation with primary antibodies, samples were washed 3 × 10 min with PBS. Samples were then incubated with secondary antibodies for 1.25 h at room temperature. After secondary antibody incubation, samples were washed 3 × 5 min with PBS and then mounted in 15 μL of mounting media (80% Glycerol (Alfa Aesar, Catalogue#36646); 0.1 M Tris•HCl (Calbiochem, Catalogue#648311), pH 8.8; 0.05% p-phenylenediamine [Sigma Aldrich, Catalogue#P-1519]). Slides were imaged and analyzed on a Nikon A1R or AX confocal microscope with an Apo TIRF 60X/1.49 oil objective using NIS-Elements version 4.6 or 5.42 software. At least six larvae were imaged for each experiment, and the experiment was repeated at least three times.

Antibodies and stains

The following antibodies were used in this work: rat anti-α-tubulin (YL1/2; 1:1000 for immunofluorescence (IF); 1:3333 for immunoblotting (WB); Invitrogen, RRID:AB_2210201), mouse anti-Dlic monoclonal (P5F5; 1:5000 IF; gift from Tom Hays [Mische et al., 2008]), mouse anti-Dic (MAB1618; 1:1000 WB; EMD Millipore, RRID:AB_1674698), rabbit anti-ensconsin (1:20 IF; 1:1000 WB; gift from Vladimir Gelfand, [Barlan et al., 2013]), rabbit anti-HA tag (C29F4; 1:1000 IF; Cell Signaling Technology, Catalogue#3724, RRID:AB_1549585), chicken anti-GFP (1:5000 WB; Aves Labs, RRID:AB_2307313), rabbit anti-GFP (1:10000 WB; Invitrogen, RRID:AB_221570), mouse anti-LamC (LC28.26; 1:100 IF; DSHB deposited by P. A. Fisher, RRID:AB_528339, [Riemer et al., 1995]), mouse anti-Myc tag (9B11; 1:5000 WB; Cell Signaling Technology, RRID:AB_331783), and guinea pig anti-Nin (1:1000 WB; gift from Eric Lécuyer). The following stains were used for IF staining: DAPI (DNA; 1 μg/mL; Sigma), Phalloidin-iFluor 488 (F-actin; 1:1000; AAT Bioquest, Catalogue#23115), Phalloidin-iFluor 568 Conjugate (F-actin; 1:1000; Invitrogen). Nin transgenes, ens-GFP, and Dlic-GFP did not require additional staining.

Quantification and analysis

To visualize a fluorescence intensity profile, images were opened in NIS-Elements AR 4.6 software > Measure > Intensity Profile. A single z-slice near the center of the stack was selected for measurement, and a ∼30 μm line was drawn bisecting the nucleus. The fluorescence intensity profile was calculated automatically by the software.

Nuclear positioning was quantified using NIS-Elements AR 4.6 software, as previously described (Zheng et al., 2020). The distance between the centroid of the autothresholded nuclear signal (DAPI staining) and the centroid of the auto- or manually thresholded cell boundary (Phalloidin staining) was measured for each cell. Over 100 cells were measured across at least three slides (experiments). Outliers (as defined by Tukey’s fences) were removed from the data set.

Fiji (Fiji is just ImageJ, ImageJ2 version 2.14.0/1.54f build c89e8500e4) was used to quantify nuclear circularity, MT density, and Nin localization at the nuclear surface. Six larvae across three slides (experiments) were used to measure around 100 cells for each experimental condition. The average background signal for five nonperinuclear regions in each image was subtracted from the measurements. Results were analyzed using Student’s t-test on GraphPad Prism 10 version 10.2.0 (355) analysis software.

For nuclear circularity, LamC signal was thresholded with smoothing, the image was converted to binary, and ROIs were created for each in-focus nucleus. Then, ROIs were measured with the circularity tool selected. Scores range from 0–1 with 1 being a true circle.

For MT density measurements, LamC or DAPI signal was thresholded with smoothing, the image was converted to binary, and ROIs were created for each in-focus nucleus. The ROIs were then dilated using the default settings to encompass the MT signal around the nucleus. Finally, ROIs were measured with the area and mean grey values tools selected to get an integrated intensity of the MT signal with this formula:

Integrated Intensity = Area of ROI × (Mean Intensity − Background Intensity)

For Nin localization measurements, a nuclear ROI was created, and the nuclear circularity measurements were measured with area and mean grey value tools selected to get the integrated intensity of Nin signal within the nucleus. Then, a dilated ROI was created for the MT intensity measurements and measured to get the integrated intensity of the Nin signal within and surrounding the nucleus. Nuclear integrated intensity was subtracted from dilated integrated intensity to get the integrated intensity outside the nucleus. The ratio of integrated intensity outside the nucleus compared with the total integrated intensity of the dilated ROI gave the percentage of Nin that localized to the nuclear surface. Three larvae were analyzed and at least 10 cells were measured for each.

GFP-Trap magnetic beads

pBiex-1 GFPNanobody_SNAP plasmid (RRID:Addgene_82711) was expressed in E. coli strain BL21(DE3)pLysE and then purified by Ni²⁺-immobilized metal affinity chromatography. A 1-cm diameter column was packed with 1.5 mL packed bed volume of Chelating Sepharose Fast Flow (Amersham Biosciences, Catalogue#17-0575-01). The column was washed with water before being charged with 5 mL of 10% NiSO₄ (Millipore Sigma, Catalogue#227676) and washed again with water, followed by a five-volume wash of Solution A (20 mM Tris, pH 8.0; 100 mM NaCl). The E. coli lysate was mixed with Ni²⁺-charged sepharose and rocked at 4°C in the column for 15 min for binding. The lysate was then eluted from the column followed by a 10-volume Solution A wash and five to 10 volume wash of Solution B (20 mM Tris, pH 8.0; 100 mM NaCl; 15 mM Imidazole (Millipore Sigma, Catalogue#IX0005). Finally, column flow was stopped, and 2.5 mL of Solution C (20 mM Tris, pH 8.0; 100 mM NaCl; 300 mM Imidazole) was added. The column was then incubated at 4°C for 10 min before collecting the eluent. The purified GFPNanobody was dialyzed against PBS before coupling to NHS Mag Sepharose (Millipore Sigma, Catalogue#GE28-9440-09) following the product protocol.

coimmunoprecipitation

Twenty to forty larvae were collected, washed in water, dried on a paper towel, and then frozen at –80°C until all samples and controls had identical numbers of larvae. Using a micro pestle (VWR International Pestle, Catalogue#47747-366) and mechanical tissue homogenizer (VWR International Pellet Mixer, Ref#47747-370), larvae were homogenized in 500 μL of Lysis Buffer (10 mM Tris•HCl, pH 7.5; 150 mM NaCl (OmniPur, Catalogue#7760); 0.5 mM EDTA (OmniPur, Catalogue#4005); 0.5% Nonidet P40 Substitute (VWR, Catalogue#E109-100ML)) containing protease inhibitors (1 mM 1,10-phenanthroline monohydrate (Sigma-Aldrich, Catalogue#P9375-5G); 0.5 mM PMSF (Sigma, Catalogue#EM-7110); 1 mM benzamidine hydrochloride hydrate (Sigma-Aldrich, Catalogue#B6506-5G); 1X protease inhibitor cocktail (Sigma-Aldrich, Catalogue#P8340-1ML) and 0.4 mM nocodazole (Sigma, Catalogue#M1404-2MG). The lysates were incubated on ice for 30 min with light vortexing every 10 min. Then, lysates were centrifuged at 15,000×g for 6–7 min at 4°C and the cleared lysate was added to a precooled tube containing 300 μL of Dilution Buffer (10 mM Tris•HCl, pH 7.5; 150 mM NaCl; 0.5 mM EDTA) with protease inhibitors. This diluted lysate was then incubated with equilibrated GFP-Trap Magnetic beads (Chromotek, Catalogue#gtmak, RRID:AB_2631358) or beads made in-house (see above) for 45 min-1 h at 4°C with agitation. Following incubation, beads were washed 3 × 15 min with Dilution Buffer with protease inhibitors before being resuspended in 2X SDS–PAGE Loading Buffer (100 mM Tris•HCl, pH 6.8; 4% sodium dodecyl sulfate (J.T. Baker, Catalogue#4095-02); 0.02% Bromophenol Blue (FisherBiotech, Catalogue#BP115-25); 20% Glycerol; 5% BME (OmniPur, Catalogue#6010) for immediate analysis via Western blotting. For co-IP Western blots, input and unbound fractions represent ∼1% of the total lysate and IP fractions represent 45–50% of the total immunoprecipitate.

To assess Nin multimerization, UAS-Nin-GFP; tub-GAL4[LL7], tub-GAL80[ts-7]/TM6B flies were crossed to UAS-Nin[1-1091]-TagRFP-Myc flies. To circumvent the lethality that arises with high levels of Nin overexpression, crosses were incubated at room temperature for 3 d and then shifted to 29° to derepress GAL80.

To map the Dlic interaction domain, 100 larvae were collected and processed as above in 1250 μL of Lysis Buffer.

co-IP experiments were repeated at least three times.

Western blotting and analysis

To verify expression from the Nin transgenes generated for this study, SPARC-GAL4 was used to drive expression of the transgenes in the fat body. Two larvae were collected, washed, and lysed using a micro pestle and mechanical tissue homogenizer in 40 μl of 2 × SDS–PAGE Loading Buffer. After incubating at 95°C for 5 min, larval lysates were quickly spun down, and 6 μl was loaded for SDS–PAGE gel electrophoresis on an 8% SDS–PAGE gel. The gel was transferred using a wet high-intensity field transfer for 1 h onto nitrocellulose membrane. The membrane was then blocked with 5% nonfat milk (Publix Instant Nonfat Dry Milk) in Tris-buffered saline (TBS; 0.5 M Tris•HCl, pH 7.5; 1.2 M NaCl) for 1 h at room temperature or overnight at 4°C with rocking. This was followed by a brief wash in TBST (Tris-buffered saline with 0.1% Tween 20 [Fisher Bioreagents, Catalogue#BP337-100]) to rinse out the milk.The membrane was then probed with primary antibodies diluted in TBST for 1.25 h at room temperature or overnight at 4°C with rocking. After washing with TBST 3 × 10 min, the membrane was incubated with secondary antibodies conjugated with IRDye-800CW or IRDye-680LT (1:20,000, LI-COR) for 1.25 h at room temperature with rocking. Blots were scanned on an Odyssey CLx-2666 Infrared Imager (LI-COR Biosciences) using Image Studio v 5.2.5 software (LI-COR).

To measure the expression of various Nin transgenes in the fat body, fat bodies were dissected from three wandering third instar larvae. Genital discs remained, but all other tissues and glands were removed. Fat bodies were then homogenized by pipetting in 20 μL of 2 × SDS–PAGE Loading Buffer and analyzed by Western blotting as above on a 10% SDS–PAGE gel. Blots were scanned on an Odyssey CLx-2666 Infrared Imager (LI-COR Biosciences) using Image Studio v 5.2.5 software (LI-COR). The normalized Nin signal was calculated using the Revert 700 Total Protein Stain Strategy (LI-COR, Catalogue#926-11011) published in the Normalization Handbook (rev. September 2019). Values shown are an average of three experiments normalized to wild-type control.

Abbreviations used:

co-IP

co-immunoprecipitation

DCTN2-p50/Dynamitin

Dynactin 2, p50 subunit

Dhc

dynein heavy chain

Dic

dynein intermediate chain

Dlic

dynein light intermediate chain

ens

ensconsin

IF

immunofluorescence

KD

RNAi knock-down

klar

klarsicht

LamC

LaminC

LINC

linker of nucleoskeleton and cytoskelton

MAP7

microtubule associated protein 7

Msp300

muscle-specific protein 300 kDa

msps

mini spindles

MT

microtubule

MTOC

microtubule-organizing center

ncMTOC

noncentrosomal MTOC

Nin

Ninein

Nlp

Ninein-like protein

NLS

nuclear localization signal/sequence

PCM

pericentriolar material

tub

tubulin

WB

immunoblotting/Western blot