The endocytic fission protein EHD1 interacts with tubulin and regulates microtubule function

Abstract

The Eps15 Homology Domain protein-1 (EHD1) is an ATPase and key endocytic regulatory protein required for optimal receptor recycling, and primary ciliogenesis. Over the past decade, a central role for EHD1 has been identified in the fission of endosomes. Despite these findings, additional evidence has also pointed at a potential function of EHD1 in the regulation of microtubules. Herein, we demonstrate that EHD1 regulates the distribution of endosomes, and conversely, centrosome depletion alters EHD1 localization in cells. We show that endogenous EHD1 is found in a complex with various endogenous tubulins including TUBB3, TUBB1, α-tubulin and γ-tubulin, interactions that are independent of intact microtubules, and appear to be indirect. Depletion of key individual EHD1 interaction partners that are known to bind tubulin fail to impede EHD1-tubulin interactions, suggesting that either several proteins are capable of mediating EHD1’s connection with microtubules, or that the bridging interaction partner remains to be identified. Functionally, EHD1 depletion leads to impaired microtubule regrowth and decreased end-binding protein displacement, suggesting a role for EHD1 in modulating microtubule plus-end dynamics. Finally, EHD1’s role in microtubule regulation appears to be evolutionarily conserved, as single-cell stage C. elegans embryos with a dysfunctional EHD1/RME-1 protein displayed enhanced tubulin accumulation at metaphase spindle poles. Our findings strongly support a previously unaddressed role for EHD1 in microtubule regulation.

1. Introduction

Endocytic trafficking occurs when receptors on the cell surface are internalized into vesicles that fuse with a compartment known as the early or sorting endosome (SE) [1]. At the SE, cargo receptors are sorted and delivered to membrane domains for packaging into vesicles and/or tubules that undergo fission and are subsequently transported to the degradative pathways or recycled to the plasma membrane [2]. Among the best characterized regulators of these pathways is the RAB family of small GTP-binding proteins [3]. Through their interactions with multiple effectors, the RABs control key trafficking events including fusion of vesicles with target organelles [4]. Another key family of proteins that regulates endocytic events are the C-terminal EHD proteins [5,6]. The best characterized EHD protein is EHD1 [7–9], which shares common effectors with RABs including RAB11-FIP2 [10], Rabenosyn-5 [11,12] and Rabankyrin-5 [13], binding to these proteins through EH-domain/asparagine-proline-phenylalanine motif interactions [14–20]. EHD1 has been implicated primarily in the fission of vesicles and tubules that promotes receptor recycling [21–24].

Many proteins that regulate endocytic trafficking have also been linked to control of primary cilia generation. For example, both RAB8 and RAB11 have been implicated in primary ciliogenesis [25–27], along with EHD1 [28,29]. Studies have demonstrated that EHD1 is required at an early stage of primary ciliogenesis to remove the centriolar capping protein, CP110, from the mother centriole which then facilitates formation of the ciliary vesicle [28–30]. Our lab has recently demonstrated that one mechanism by which EHD1 promotes CP110 removal from the mother centriole is through its control of centriolar satellite movement [30,31]. Centriolar satellites are non-membrane-bound protein aggregates that move along microtubule tracks and cycle in and out of the centrosomal region, potentially serving as reservoirs that store key centriolar proteins [32]. Our study demonstrated that centriolar satellites deliver the E3 ligase, HERC2, to the mother centriole to ubiquitinate CP110 and facilitate its proteasomal degradation [30]. However, the absence of EHD1 leads to impaired centriolar satellite delivery to the mother centriole, suggesting a potential role for EHD1 in microtubule regulation [30].

Several lines of evidence support the notion that EHD1 might regulate microtubules, in addition to its effect on centriolar satellite movement. First, we have previously documented a role for EHD1 in mitosis, with increased numbers of binucleated cells observed upon its depletion [33]. Second, knock-down of EHD1 in human and worm cells leads to a coalescence of endosomes in the perinuclear region of the cell, suggesting potentially altered microtubule function [7,9]. EHD1 also interacts with the centrosomal protein CDK5RAP2 (also known as Cep215) and affects centriole disengagement and duplication, with EHD1-depleted cells exhibiting defective centriole disengagement [34]. In addition, both EHD1- and Rab10-containing tubular endosomes underwent dispersion upon microtubule depolymerization using nocodazole [9,35]. Moreover, a Rab10 effector, MICAL-L1, which also interacts with and recruits EHD1 to endosomes, directly interacts with various tubulin proteins which comprise microtubules [29]. Third, the close relationship between EHD1 and RAB proteins, that have defined roles in mediating motor protein and microtubule function, suggests a potential connection between EHD1 and microtubules [10–13]. Finally, a recent study suggests that transgenic EHD1 introduced into cancer cells can be co-immunoprecipitated with TUBB3 [36]. When coupled with the potential function of EHD1 on centriolar satellites that may be related to microtubule regulation, we hypothesized that EHD1 may be a microtubule interacting protein that regulates microtubule function.

2. Results

To assess the impact of EHD1 depletion on microtubule-based endosome distribution, we first reduced EHD1 expression in HeLa cells using siRNA oligonucleotides. After validation of effective depletion by immunoblotting (Fig. 1C), we immunostained the cells on coverslips with antibodies to the endosomal marker EEA1 and used DAPI to mark nuclei. As demonstrated in Fig. 1A and the accompanying inset, mock-treated cells displayed EEA1-containing endosomes in a pattern reflecting a combination of perinuclear and cytoplasmic localization. However, upon EHD1 knock-down, EEA1 endosomes showed a greater propensity to localize to the perinuclear region, and fewer endosomes were scattered throughout the periphery of the cells (Fig. 1B and inset). Quantification determined that the percent of endosomes in the perinuclear region was increased by about two-fold in cells depleted of EHD1 (Fig. 1D), and that at increasing distances from the perinuclear region, fewer EEA1-containing endosomes were consistently observed in the knock-down cells (Fig. 1E). Since the transport of EEA1-containing endosomes away from the perinuclear area into the cytoplasm is microtubule dependent, overall, these data support the notion that EHD1 depletion impacts microtubule function, either directly or potentially via regulation of motor proteins.

Fig. 1.

Impact of siRNA-mediated EHD1 knock-down on the distribution of early endosomes. HeLa cells were treated with oligonucleotides that target EHD1 for 48 h, fixed, immunostained and imaged. A,B Representative fields of cells taken from mock-treated (A) and EHD1-depleted (B) HeLa cells. White denotes EEA1-stained early endosomes, and blue marks DAPI-stained nuclei. White dashed lines mark cell borders, and yellow hexagons mark sample perinuclear regions of interest for quantification. Insets depict the cells in which the regions of interest have been marked. C, Immunoblot validating decreased EHD1 protein levels upon siRNA treatment. D, Graph shows data from three independent experiments depicting the percentage of EEA1-endosomes in the perinuclear region. Imaris software was used to quantify the number of endosomes per cell, both within and outside a 4 μm-sized region of interest (ROI) manually placed near the nucleus. E, Distances from the ROI for all endosomes were calculated, binned and plotted for mock and EHD1-depleted cells. Statistical significance was calculated with an unpaired two-tailed t test for normally distributed samples and with the Mann–Whitney two-tailed t test for comparisons which did not meet the assumption of normality. Scale Bar,10 μm.

Given the impact of EHD1 knock-down on microtubule function and endosome distribution in cells, we tested whether endogenous EHD1 interacts with various tubulins, the major components of microtubules. While a previous study identified such a potential interaction [36], to our knowledge no studies have been published showing interactions between endogenous proteins. Using antibodies directed against endogenous EHD1, co-immunoprecipitations showed that in HeLa cells levels of EHD1 were detected in an immune complex with endogenous α-tubulin (Fig. 2A), TUBB3 (Fig. 2B), TUBB1 (Fig. 2C), and γ-tubulin (Fig. 2D). For TUBB3 pulldowns, a clear reciprocal band of EHD1 could also be detected (Fig. 2B; TUBB3 IP). The precise electrophoretic mobility of bands can vary modestly between lysate and IP eluate lanes, a common effect attributable to the distinct composition of immunoprecipitated samples (including high immunoglobulin content and elution buffer constituents) that can subtly influence migration. Notably, the co-IP signals for each tubulin fall within the mobility range defined by the corresponding reference lanes run on the same gel. In addition, similar interactions could be observed from RPE-1 cell lysates demonstrating an interaction between EHD1 and TUBB3 (Supplemental Fig. 2A), γ-tubulin (Supplemental Fig. 2B), and α-tubulin (Supplemental Fig. 2C). Overall, these data support the idea that EHD1 and various tubulin isoforms form either direct or indirect interactions.

Fig. 2.

Endogenous EHD1 interacts with α-tubulin, β1-tubulin, β3-tubulin and γ-tubulin in HeLa cells. Lysates collected from HeLa cells were subjected to immunoprecipitation by anti-EHD1, an anti-tubulin antibody or immunoprecipitated with beads only (control). The bead-bound fractions were separated using SDS-PAGE, transferred to nitrocellulose and then immunoblotted with antibodies against α-tubulin (A), β3-tubulin (TUBB3) (B), β1-tubulin (TUBB1) (C) or γ-tubulin (D) for the lysate, EHD1 coIP, anti-tubulin IP, and beads only lanes. The blots were then stripped with 3 M guanidine thiocyanate, washed, and blotted using anti-EHD1 antibody to confirm EHD1 pull down/immunoprecipitation efficacy and reciprocal interactions. Lower and higher exposure times are marked (upper panels). Depicted blots are representatives taken from three independent experiments.

We next asked whether EHD1 can interact with individual tubulin proteins, or whether the interaction requires intact microtubules. To test this, we took either untreated HeLa cells (Fig. 3A and B), or HeLa cells with depolymerized microtubules following treatment with nocodazole (Fig. 3C and D), and subjected them to co-immunoprecipitations with EHD1. As demonstrated, in both untreated (Fig. 3E) and treated HeLa cells (Fig. 3F), we observed TUBB3 interactions with EHD1. These data suggest that intact microtubules are not required for tubulin-EHD1 interactions.

Fig. 3.

Depolymerized tubulin interacts with EHD1. HeLa cells were treated with 20 μM nocodazole for 35 minutes at 37°C to depolymerize microtubules, lysed and immunoprecipitated with anti-EHD1 antibody or beads only. Representative images and insets from untreated (A,B) and nocodazole-treated (C,D) HeLa cells stained using anti-α-tubulin antibody to label intact microtubules (A,B) and depolymerized microtubules (C,D). Lysates collected from both untreated (E) and nocodazole-treated (F) HeLa cells were subjected to immunoprecipitation by anti-EHD1 antibody or immunoprecipitation with beads only (control). The bead-bound fractions were separated using SDS-PAGE, transferred to nitrocellulose and then immunoblotted with anti-TUBB3 antibody (E,F) for EHD1 coIP and lysate lanes, and anti-EHD1 antibody to confirm immunoprecipitation (EHD1 IP lane). Gels denoting tubulin in the coIP lane are taken from higher exposures during enhanced chemiluminescence. Depicted blots are representative of three independent experiments. Scale Bar, 10 μm.

The interaction between EHD1 and the various tubulins could be either direct or indirect. To evaluate whether the interactions were likely direct, we purified the EH domain of EHD1 (EH-1) as a GST-fusion protein. We selected this domain because to date, it is the region that interacts with almost all known EHD1 interaction partners [6], and it is both relatively easy to purify and stable, unlike the full-length EHD1 protein. We then subjected purified GST-EH-1 to pull-down assays with purified α/β tubulin. As demonstrated, we were unable to observe a direct interaction between EH-1 and α/β tubulin (Fig. 4A). While this does not entirely rule out the possibility that the proteins might bind directly, a direct interaction is unlikely.

Fig. 4.

EHD1-tubulin interactions are unlikely direct and are not mediated by MICAL-L1. (A) In-vitro binding was conducted by incubating 2 μg of purified porcine αβ-tubulin heterodimers with either 8 μg of GST alone (control) or GST-EH-1 purified on glutathione resin for 3 h at 4°C. Bead bound fractions were subjected to SDS-PAGE and then transferred to nitrocellulose and blotted with anti-α-tubulin antibody (A, top) and anti-GST antibody (A, bottom). B-D, HeLa cells were treated with siRNA oligonucleotides targeting MICAL-L1 for 72 h. (B) Immunoblotting was done to validate depletion of MICAL-L1. Lysates from mock and MICAL-L1 siRNA treated cells were subjected to coimmunoprecipitation by anti-EHD1 antibody or immunoprecipitation with beads only (control). The bead-bound fractions were separated using SDS-PAGE, transferred to nitrocellulose and then immunoblotted with anti-TUBB3 antibody (C,D) for EHD1 coIP and lysate lanes, and with anti-EHD1 antibody to confirm immunoprecipitation (EHD1 IP lane). Gels denoting TUBB3 in the coIP lane are taken from higher exposures. Depicted blots are representatives taken from three independent experiments.

MICAL-L1 is a key EHD1 interaction partner, binding EH-1 through the second of its two asparagine-proline -phenylalanine (NPF) motifs [37–39]. Moreover, it directly interacts with all major tubulin isoforms [29]. Accordingly, we assessed whether knock-down of MICAL-L1 (validated in Fig. 4B) would render a loss of the interaction between EHD1 and TUBB3. However, as demonstrated, knock-down of MICAL-L1 did not interfere with the interaction between EHD1 and TUBB3 (Fig. 4C and D). Additional experiments knocking-down RAB8, RAB10 and CDK5RAP2, all implicated in tubulin binding and/or microtubule function [40–45], did not prevent EHD1 from interacting with tubulin (Supplementary Table 1). These experiments collectively suggest that while EHD1 likely binds indirectly to tubulins through a mediating protein(s), the identity of the bridging protein(s) remains unknown.

A recent study has shown that a key EHD1 interaction partner and RAB5 effector, Rabankyrin-5 [13], interacts with dynein, a minus-end directed motor protein [46]. Indeed, a previous study from our lab demonstrated that Rabankyrin-5 knock-down led to perinuclear accumulation of retromer-marked endosomes, phenocopying the effect of EHD1 knockdown on endosome distribution (see Fig. 1) [13]. Accordingly, we hypothesized that Rabankyrin-5 might also interact with tubulin and thus mediate the interaction between EHD1 and microtubules. To test this hypothesis, we transfected cells with GFP-Rabankyrin-5 and then immunoprecipitated the lysates with either anti-GFP antibodies or antibodies directed against EHD1 (Fig. 5A). The whole cell lysates showed expression of GFP-Rabankyrin-5, TUBB3 and EHD1 (Fig. 5A; left panels). Anti-GFP and anti-EHD1 pull-down Rabankyrin-5 and EHD1, respectively, as anticipated (Fig. 5A; top and bottom panels). Rabankyrin-5 also precipitated EHD1 (Fig. 5A; bottom panel), as expected, although due to low GFP-Rabankyrin-5 expression levels, in this experiment the reciprocal pull-down was not observed. Significantly, both EHD1 and GFP-Rabankyrin-5 precipitated TUBB3, suggesting that Rabankyrin-5 might be a potential linker between EHD1 and microtubules (Fig. 5A; middle panel).

Fig. 5.

Rabankyrin-5 interacts with tubulin and EHD1, but does not mediate EHD1-tubulin interactions. (A) HeLa cells were transfected with GFP-Rabankyrin-5 DNA for 48 h and then collected lysates were subjected to coimmunoprecipitation using anti-EHD1 antibody (EHD1 coIP lane) or anti-GFP antibody coated beads (GFP coIP lane). The bead bound fractions were then separated using SDS-PAGE, transferred to nitrocellulose and blotted with anti-GFP and anti-EHD1 antibodies to confirm immunoprecipitation (GFP-Rabankyrin-5 and EHD1 coIP lanes, respectively) and with TUBB3 to confirm coimmunoprecipitation in both the lanes. (B) HeLa cells were treated with two different siRNA oligonucleotides (Oligos A and B) targeting Rabankyrin-5 for 48 h. Immunoblotting was done to validate depletion of Rabankyrin-5. (C) Lysates from mock and Rabankyrin-5 siRNA treated cells were subjected to coimmunoprecipitation by anti-EHD1 antibody or immunoprecipitation with beads only (control). The bead-bound fractions were separated using SDS-PAGE, transferred to nitrocellulose and then immunoblotted with anti-TUBB3 antibody for EHD1 coIP and lysate lanes. The blot was then stripped with 3 M guanidine thiocyanate, washed, and blotted using anti-EHD1 antibody to confirm immunoprecipitation (lower panel). Depicted blots are representatives taken from three independent experiments.

To assess whether Rabankyrin-5 mediates the connection between EHD1 and tubulin, we first depleted Rabankyrin-5 from HeLa cells using siRNA. As demonstrated, both Oligo A and Oligo B led to significant depletion of Rabankyrin-5 levels (Fig. 5B). We then compared the amount of TUBB3 pulled down by mock-treated cells, and cells lacking Rabankyrin-5 following depletion with either Oligo A or Oligo B (Fig. 5C). As shown, immunoprecipitation with anti-EHD1 led to pulldown levels of EHD1 and TUBB3 that were similar (Fig. 5C). These experiments suggest that despite the potential for Rabankyrin-5 to link between EHD1 and TUBB3, EHD1 still retains an association with TUBB3 in the absence of Rabankyrin-5.

Since EHD1 interacts with various tubulin isoforms and its knock-down affects organelles trafficked along microtubules, we elected to evaluate whether ablating the centrosome—the organelle from which most microtubules are generated, has an impact on EHD1. To do this, we used the inhibitor centrinone to deplete centrosomes over 3 days [47]. Untreated cells displayed a concentration of EHD1-containing vesicles/endosomes in the perinuclear area, in part overlapping with γ-tubulin which marks the centrioles (Fig. 6A–C). In the centrinone-treated cells, we observed a loss of γ-tubulin-marked centrioles (red channel), and a significant decrease in EHD1-containing structures localized to the peri-nuclear region (Fig. 6D–F). Minimal differences were observed in EHD1 expression upon centrinone treatment, suggesting that the effects observed are primarily due to altered protein localization (Fig. 6G). To quantitatively assess this difference, we measured the total area of EHD1-containing structures in the peri-nuclear region, and demonstrated a significant decrease in the centrinone-treated cells (Fig. 6H). These data suggest a mutually functional relationship between EHD1 and microtubules.

Fig. 6.

Impact of centrinone-mediated centrosomal ablation on EHD1 distribution. HeLa cells were either left untreated, or treated with 150 nM centrinone for 72 h. Cells were methanol fixed and immunostained with anti-γ-tubulin and anti-EHD1 antibodies. (A-F) Representative images from untreated (A-C) and centrinone-treated (D-F) cells. EHD1 is labeled in white, DAP1 is labeled in blue, and γ-tubulin is labeled in red. Insets focus on representative cells depicting EHD1 distribution and the centrioles indicated by γ-tubulin. (G) Lysates from untreated and centrinone-treated HeLa cells were immunoblotted for EHD1 to determine effect on protein levels. (H) Graph representing area of EHD1 structures within 3 μm of the nucleus indicated by DAPI, quantified using Imaris. Values were taken from three independent experiments. Statistical significance was calculated with an unpaired two-tailed t test for normally distributed samples and with the Mann–Whitney two-tailed t test for comparisons which did not meet the assumption of normality. Scale Bar, 10 μm.

To address whether EHD1 depletion impacts microtubule function, we performed a nocodazole washout assay in mock-treated and EHD1 knock-down cells (validated in Fig. 7K, upper gel). At steady-state, we did not observe detectable differences in the microtubule distribution pattern between mock-treated and EHD1-depleted cells (Fig. 7A and F). Upon nocodazole treatment, both mock and depleted cells displayed depolymerized microtubules (Fig. 7B and G). 8 minutes after washout of the nocodazole, mock-treated cells showed the start of a rapid repolymerization of microtubules, which was less evident in the EHD1 knock-down cells (Fig. 7; compare C with H). This difference was further exacerbated at time points of 10 and 13 min. washout (Fig. 7; compared D with I, and E with J). The data quantified are shown in Fig. 7L. Given the impaired microtubule regrowth, we hypothesized that these defects might potentially be mediated by the end-binding protein, EB1. EB1 regulates microtubule dynamics through its binding to the growing end of the microtubule to promote polymerization and stabilization [48,49]. To this aim, we transfected EB1-GFP into HeLa cells and gene-edited CRISPR/Cas9 EHD1 knock-out HeLa cells (validated in Fig. 7K, lower gel), and used live image analysis to assess the mean track length and track displacement of EB1-GFP at the growing end of microtubules. Mean track represents the total distance traveled by EB1-GFP along its entire trajectory, whereas track displacement indicates the straight line distance between the start and end points of the track. As demonstrated in Fig. 7M and N, both track length and displacement of EB1-GFP at the plus ends of microtubules were shortened in the absence of EHD1 (see Supplemental Fig. 1. for representative examples), consistent with the slower microtubule regrowth observed in the nocodazole washout experiments. Overall, these data suggest that EHD1 plays a role in facilitating microtubule assembly, possibly though modulation of plus-end microtubule dynamics.

Fig. 7.

EHD1 knock-down impairs microtubule regrowth. HeLa cells were either mock-treated or treated with siRNA oligonucleotides directed at EHD1 for 48 h. Both mock and siRNA-treated cells were stained using anti-α-tubulin antibody to mark the microtubule network in untreated cells (A,F). The remaining cells were treated with 20 μM nocodazole for 35 min. at 37°C to depolymerize microtubules and then immediately fixed to validate microtubule depolymerization (B,G: depolymerization panels), or subjected to microtubule regrowth by replacing nocodazole-laden media with complete media (washout) and fixed after 8 minutes (C,H: 8 min. washout panels), 10 minutes (D,I: 10 min. washout panels) and 13 minutes (E,J: 13 min. washout panels). Insets focus on representative cells displaying the phenotype at corresponding time points. (K) Immunoblot validating EHD1 depletion upon siRNA treatment (upper gel), and immunoblot validation of EHD1 knock-out cells (lower gel). (L) Graph depicting area of microtubule regrowth measured using ZEN software by manually drawing a contour around the area of radial microtubule extension at each time point. (M,N) EB1-GFP was transfected into HeLa cells or CRISPR knock-out cells lacking EHD1, and cells were subjected to live imaging to monitor EB1-GFP. Graphs display EB1-GFP microtubule track length (M) and EB1-GFP displacement length (N). Values were taken from three independent experiments. Statistical significance was calculated with an unpaired two-tailed t test for normally distributed samples and with the Mann–Whitney two-tailed t test for comparisons which did not meet the assumption of normality. Scale Bar, 10 μm.

EHD proteins are highly conserved from all multicellular organisms, including plants, flies, zebrafish and worms [5]. RME-1, the single C. elegans EHD1 homolog, has been studied extensively and plays a similar role in trafficking to that of the mammalian EHD1 [5,7,8]. However, whether EHD1 affects microtubule function in worms has not been studied to date. To address this, we compared the intensity of mCherry-tubulin at the metaphase spindle poles of 1-cell stage C. elegans embryos from worms with either wild-type RME-1, or embryos derived from the rme-1(b1045) mutant allele with dysfunctional RME-1 [7] (Fig. 8). As demonstrated, a significant enhancement in tubulin intensity at the spindle poles was observed in cells with the mutant rme-1 (Fig. 8; compare A to B: quantified in C). In contrast control measurements of cytoplasmic mCherry-tubulin displayed no major difference in wild-type or mutant rme-1 1-cell stage embryos, indicating that total tubulin levels were likely unchanged in the rme-1 mutants (Fig. 8D). These data support the notion that EHD1 plays a role in microtubule regulation that is evolutionarily conserved.

Fig.8.

rme-1(b1045) mutants exhibit increased tubulin intensity at metaphase spindle poles in 1-cell stage C. elegans embryos. Representative image of a 1-cell C. elegans embryo expressing mCherry–tubulin and wild-type rme-1, showing normal tubulin intensity at metaphase spindle poles (A). Representative image of a 1-cell C. elegans embryo expressing mCherry–tubulin and the rme-1(b1045) mutant allele, showing elevated tubulin intensity at metaphase spindle poles (B). Quantification of the mean tubulin fluorescence intensity (arbitrary units, AU) at the metaphase spindle poles in embryos expressing wild-type rme-1 (control) versus rme-1(b1045) (C), and quantification of mean cytoplasmic mCherry–tubulin fluorescence in 1-cell stage embryos immediately prior to metaphase onset (arbitrary units, AU) (D). Each point represents the mean tubulin intensity measured at a single metaphase spindle pole or within the cytoplasm of an individual embryo. Error bars indicate standard deviation; center lines denote the mean. n = number of spindle poles (C) or embryos (D) analyzed. Unpaired two-tailed t-test; p < 0.0001 (spindle pole tubulin intensity), p = 0.3100 (cytoplasmic tubulin intensity). Scale Bar, 10 μm.

3. Discussion

Over a quarter of a century ago it was observed that EHD1 depletion or mutations in key EHD1 domains lead to an altered distribution of endosomes [7–9]. Since that time, EHD1’s role has been strongly implicated as a central protein in endocytic membrane trafficking, and many recent studies collectively support its function in the process of endosome fission [21–24,50–53].

Despite the growing number of studies highlighting EHD1’s role in endosome fission and subsequently, the recycling of internalized receptors, particularly through a network of tubular recycling endosomes, certain aspects of EHD1 function remain unclear. In particular, it is hard to reconcile a role in fission with EHD1’s impact on endosome distribution (as noted above), cytokinetic defects [33], central spindle formation [33], kinetochore-microtubule interactions [33], centriole disengagement and duplication [34], as well as centriolar satellite motility [30].

Accordingly, we sought to better understand the relationship between EHD1 and microtubules. Based on the earlier studies that qualitatively showed that EHD1 function affects endosome distribution in cells, we now provide quantitative measurements of the impact of EHD1 depletion on endosome retention in the perinuclear region, as well as the dispersion distances of endosomes from the peri-nuclear region. As expected, major differences in distribution were observed.

A potential interaction between tubulins and EHD1 might help provide an explanation for how EHD1 regulates microtubule function. A recent study using transgenic proteins showed that EHD1 could co-immunoprecipitate with the tubulin TUBB3 [36]. Our study expands on these findings and is the first to demonstrate that endogenous EHD1 co-immunoprecipitates with TUBB3, as well as TUBB1, α-tubulin and γ-tubulin, in HeLa and RPE-1 cells. While we were unable to identify a linker between EHD1 and tubulin, based on our data it is likely that the interactions are indirect. Indeed, we were unable to detect a direct tubulin interaction with the EH domain of EHD1, the domain which is responsible for almost every EHD1 interaction identified to date. Knock-down of either MICAL-L1 or Rabankyrin-5, each of which binds to tubulin and to EHD1 [29,46], did not disrupt EHD1-tubulin interactions. However, it remains possible that both proteins are capable of linking EHD1 to microtubules, and that depletion of either one individually is insufficient to impair the binding of EHD1 to tubulin. Future studies could investigate simultaneous knock-down of both proteins, although technical solutions will be required since such treatment appears toxic to the cells.

While the mode of EHD1 regulation of microtubules remains to be fully elucidated, our data support the idea that EHD1 is required for microtubule regrowth (Fig. 7), suggesting a potential role in the modulation of microtubule regulators. Such regulators might include microtubule nucleation regulators, plus-end stabilizers (i.e, end-binding proteins) as well as microtubule severing and remodeling proteins. Indeed, EB1 tracks showed modestly reduced distance and displacement in the EHD1 knock-out cells. Finally, it is noteworthy that EHD1 regulation of microtubules appears to be evolutionarily conserved, as rme-1 mutant worms display retention of tubulin at metaphase spindle poles. Indeed, the rme-1(b1045) mutant has previously been shown to exhibit a reduced brood size and developmental defects [7], and microtubule dynamics are essential for germline proliferation, meiotic spindle function, and early embryonic divisions in C. elegans. Although we observe elevated tubulin signal at spindle poles in rme-1 mutants, this does not necessarily indicate enhanced microtubule polymerization. Increased tubulin accumulation at spindle poles can also reflect impaired microtubule turnover, defective depolymerization, or altered release and redistribution of microtubules from spindle poles. Thus, elevated tubulin intensity is compatible with disrupted microtubule dynamics rather than increased polymerization per se.

Importantly, our microtubule regrowth assays reveal abnormalities in microtubule behavior following depolymerization, supporting the conclusion that microtubule dynamics—not simply polymer mass—are perturbed in rme-1 mutants. Such defects would be expected to compromise spindle function and cell division, providing a plausible explanation for the developmental defects observed in vivo.

Overall, our study highlights a key role for EHD1 in binding tubulin and modulating microtubule function, expanding the known role of this protein beyond endosome fission and shedding new light on microtubule-dependent processes including centrosome duplication, cytokinesis and primary ciliogenesis.

4. Material and methods

4.1. Cell Lines and Treatments

The HeLa cell line (ATCC-CCL-2) was obtained from ATCC and cultured in complete DMEM (high glucose) (Thermo Fisher Scientific, Carlsbad, CA), with 10% FBS (Atlas Biologicals), 1x penicillin-streptomycin (Gibco) and 2 mM L-glutamine (Gibco). CRISPR/Cas9 gene-edited EHD1 knock-out cell line was obtained from GenScript (Piscataway, NJ). The hTERT RPE-1 human epithelial cell line (ATCC-CRL4000) was obtained from ATCC and grown in DMEM/F12 (Thermo Fisher Scientific, Carlsbad, CA) with 10% FBS, 1x penicillin-streptomycin, 2 mM L-glutamine and 1x non-essential amino acids (Thermo Fisher Scientific, Waltham, MA). To prevent Mycoplasma and other contamination, 100 μg/ml Normocin (Thermo Fisher Scientific) was added to culture media, and cells were routinely tested for Mycoplasma contamination. For EHD1 knock-down, small-interfering RNA (siRNA) oligonucleotides targeting EHD1 (Sigma, #71015) were used to transfect HeLa cells using Lipofectamine RNAi/MAX (Invitrogen, Carlsbad, CA) for 48 h in the absence of antibiotics as per the manufacturer’s protocol. siRNA oligonucleotides targeting MICAL-L1 (Sigma, #361239), Rabankyrin-5 (Sigma, Oligo A #34936; Oligo B #23276), Rab8a (Sigma, #339466), Rab10 (Dharmacon, #L-010823-00-0010) and CDK5RAP2 (Dharmacon, #L-019154-00-0020) were each diluted in DharmaFECT reagent (Dharmacon, Lafayette, CO) to transfect HeLa cells for 72 h (48 h for Rabankyrin-5 siRNA) in complete media as per the manufacturer’s protocol. To test the efficiency of the siRNA-mediated knock-down, immunoblotting was performed.

4.2. Antibodies and Reagents

The antibodies used in this study were: mouse anti-α-tubulin (Sigma, #T9026 DM1a, at 1:500 for immunoblotting and 1:100 for immunofluorescence), mouse anti-TUBB3 (ProteinTech, #66375, at 1:500 for immunoblotting), rabbit anti-TUBB3 (Biomatik, #CAU-22722, at 1:3000 for immunoblotting), mouse anti-TUBB1 (Santa Cruz, #sc58884, at 1:1200 for immunoblotting), mouse anti-γ-tubulin (Sigma, #T5326, at 1:100 for immunofluorescence and 1:500 for immunoblotting), mouse anti-MICAL-L1 (Novus/Abnova, #H00085377-B01P, at 1:700 for immunoblotting), rabbit anti-GST (Millipore), rabbit anti-EHD1 (LifeTein, #1795, at 1:200 for immunoblotting; Abcam, #AB109311, at 1:100 for immunofluorescence), rabbit anti-Rabankyrin-5 (Invitrogen #PA5-24640, at 1:200 for immunoblotting), mouse anti-HA (Biolegend, #901513 at 1:300 for immunoblotting), mouse anti-GFP (Roche #1184460001 at 1:200 for immunoblotting), mouse anti-myc (Serotec #MCA2200B at 1:2000 for immunoblotting), mouse anti-CDK5RAP2 (Sigma/Atlas, #HPA046529, at 1:200 for immunoblotting), mouse anti-GAPDH-Horseradish peroxidase (HRP) (ProteinTech, #60004, at 1:5000 for immunoblotting), rabbit anti-EEA1 (Cell Signaling, #3288, at 1:30 for immunofluorescence), rabbit anti-Rab8a (Sigma, #R5530, at 1:250 for immunoblotting), rabbit anti-Rab10 (Abcam, #237703, at 1:800 for immunoblotting). Secondary antibodies used included: Alexa 568-conjugated goat anti-mouse (Life Technologies, #A11031), Alexa 488-conjugated goat anti-rabbit (Life Technologies, #A11034), Alexa 488-conjugated goat anti-mouse (Life Technologies, #A11029), all used at 1:500 for immunofluorescence; donkey anti-rabbit-HRP (Amersham, #NA934), goat anti-mouse-HRP (Jackson, #115-035-008), goat anti-mouse light-chain-HRP (Jackson Immunol, #115-035-174), mouse anti-rabbit light chain-HRP (Jackson, #211-032-171, at 1:3000 for immunoblotting), goat anti-rabbit light chain-HRP (Boster Biologicals, #ST14A23C54) used at 1:5000 for immunoblotting. DAPI (#D1306) was obtained from Thermo Fisher Scientific. The inhibitors used in the study were: Centrinone (TargetMol, #T14927), Nocodazole (LKT Laboratories). Both the inhibitors were reconstituted in DMSO to make the stock and then diluted in complete media for use in the experiments at the indicated concentrations.

4.3. Quantifying dispersal of EHD1-containing structures upon centrinone treatment

HeLa cells plated on coverslips were incubated for 72 h in 150 nM centrinone (TargetMol, #T14927) diluted in complete media. Both untreated and centrinone-treated coverslips were then immediately fixed using chilled 100% methanol for 5 min at −20°C, and then stained using anti-EHD1 and anti-γ tubulin antibodies diluted in PBS buffer containing 0.1% Triton-X-100 and 0.5% BSA for 1 h at room temperature. The coverslips were subsequently washed 3x with PBS to remove unconjugated primary antibodies and then stained with appropriate fluorochrome-conjugated secondary antibodies. Following a 1 h incubation with secondary antibodies, the coverslips were washed 3X and then mounted in Fluoromount G Mounting Medium (Southern Biotech). Z-stack images were taken using a Zeiss LSM 800 confocal microscope with a 63x/1.4 NA oil objective, and maximal intensity orthogonal projections were obtained in the Zen Blue Software. Quantification was done using the Imaris 9.9.1 (Oxford Instruments) software to first render all the nuclei (in blue) and EHD1 structures (in green) in an image. The area of EHD1-containing structures was computed for three cells chosen at random in each image. This was done using the filter tab, to select all EHD1 structures within 3 μm of each of the chosen nuclei by setting the “shortest distance to surface” at 3.00 as the upper threshold. The areas of all EHD1 structures within 3 μm of the nucleus was then summed and represented as “Area Sum.”

4.4. Quantification of EEA1 distribution upon EHD1 knock-down

Mock- and EHD1-siRNA treated HeLa cells plated on coverslips were fixed with 4% paraformaldehyde (in PBS) for 10 min at room temperature. Once fixed, coverslips were stained with anti-EEA1 antibody diluted in PBS buffer containing 0.2% saponin and 0.5% BSA for 1 h at room temperature. The coverslips were subsequently washed 3X with PBS to remove unconjugated primary antibodies and then stained with appropriate fluorochrome-conjugated secondary antibodies for 30 min. at room temperature. The coverslips were then mounted in Fluoromount and imaged using a Zeiss LSM 800 confocal microscope (Carl Zeiss) with a 63×/1.4 NA oil objective. Quantification was done using Imaris 9.9.1 (Oxford Instruments) software to first render all the nuclei (blue) and EEA1 endosomes (white). Three cells were chosen at random from each image for quantification. A 10-vertex surface with a radius of 4 μm was created and placed as a region of interest (ROI) adjacent to the nucleus, capturing the most endosome-dense peri-nuclear region. Then all the EEA1-marked endosomes were selected, and their distances from the perimeter of the ROI were calculated using the “shortest distance to surface” command. Distance values for all EEA1-marked endosomes within or contacting the ROI were denoted as zero and considered “peri-nuclear.” Counts representing the distance values that were zeroed and in the periphery were then used to calculate the ‘percentage of endosomes in the peri-nuclear region’ as: [Total number of endosomes in the peri-nuclear region (all zeroed distances) ÷ Total number of endosomes in the cell]×100. While “counts” were used to calculate percentage-wise distribution, individual distance measurements were binned as increasing distances from the ROI into the cell periphery.

4.5. Microtubule Regrowth Assay and Measuring Microtubule Track and Displacement Length

Microtubule regrowth: Mock and EHD1-siRNA treated HeLa cells were plated on coverslips and a subset was designated as “untreated.” These were fixed using chilled 100% methanol for 5 min at −20°C, and then stained using anti-α tubulin antibody (to depict the microtubule network) diluted in PBS buffer containing 0.1% Triton-X-100 and 0.5% BSA for 1 h at room temperature. The coverslips were subsequently washed 3X with PBS to remove unconjugated primary antibodies and then stained with appropriate fluorochrome-conjugated secondary antibodies. The coverslips were then mounted using Fluoromount. Then the remaining coverslips were treated with 20 μM nocodazole (LKT Laboratories) diluted in complete media for 35 min. at 37°C to induce microtubule depolymerization. One subset of coverslips was immediately fixed, stained and mounted as described, to confirm depolymerization. Microtubule regrowth was allowed in the rest of the coverslips by replacing the nocodazole-containing media with complete media (washout) and incubation at 37°C. To capture the dynamics of microtubule regrowth, coverslips were immediately rinsed with PBS and fixed (as described earlier) at different time points of washout: 8 min., 10 min., and 13 min. The coverslips were then stained and mounted. Imaging for the microtubule network, depolymerization and recovery was done using the Zeiss LSM 800 confocal microscope (Carl Zeiss) with a 63×/1.4 NA oil objective. Using the “spline” command in ZEN software, a contour was manually drawn around the microtubule aster to measure the area (μm²) displaying a radial extension from the microtubule organizing center. The area of microtubule extension was used as a metric for microtubule regrowth, with 0 μm² being the microtubule area immediately after treatment due to lack of any detectable microtubules as a consequence of nocodazole-induced depolymerization.

EB1-GFP Tracking: Control and EHD1 knock-out HeLa cell lines were used. The EHD1 knock-out HeLa cell line was generated by GenScript (Piscataway, NJ) using CRISPR/Cas9 gene-editing. The knock-out was validated by immunoblotting of lysates collected from control and knock-out cell lines. Both control and EHD1 knock-out HeLa cells plated on glass-bottom dishes were transiently transfected with EB1-GFP (Addgene plasmid #39399) using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) for 16 h in the absence of antibiotics as per the manufacturer’s protocol. Imaging was then performed using the Zeiss LSM 800 Confocal microscope (Carl Zeiss) with 63×/1.4 NA oil objective, whilst keeping the cells in pre-warmed Opti-MEM medium (Gibco) containing 2 mM L-Glutamine and 30% FCS. The master gain and laser power were manually adjusted to optimize the acquisition of EB1-GFP movement, taking an image every 0.5 s for 90 s spanning 180 frames. To quantify the EB1 tracks, Imaris 9.9.1 (Oxford Instruments) software’s particle tracking analysis feature was used. For each movie, all EB1-GFP structures detected across all 180 frames (90 s) were rendered as “surfaces” using the appropriate threshold adjusted to minimize background noise. Then, autoregressive motion algorithm with appropriate threshold was used to generate “tracks” from the trajectory of every detected EB1 structure across all 180 frames. 10 tracks in each movie were manually selected (while ensuring EB1 movement along the track). For each of these tracks, Imaris was used to compute the statistics for “Track Length” which indicates actual path length (in μm) spanned by the EB1 structure; and “Track displacement length” which indicates the shortest distance (in μm) between the initial point (when the EB1 structure for the selected track was first detected) and final point (when the EB1 structure for the selected track is last detected before it disappears). For each replicate, a total of 100 tracks from 10 images were computed for both mock and EHD1 knock-out, and then statistical analysis was performed using GraphPad Prism 10.2.3 (GraphPad, San Diego, CA) on data obtained from three independent experiments and plotted.

4.6. In vitro binding protein-protein interactions

8 μg of purified, glutathione-resin bound bait proteins: GST and GST-EH1 were each washed 3X with TGMC buffer (20 mM Tris-HCl pH 7.9, 20% glycerol, 5 mM MgCl₂, 0.5% NP-40, 1 mM DTT, 0.2 mM PMSF, 0.1 M NaCl. The bait proteins were then diluted in 20 μL of TGMC buffer and incubated with 3 U of micrococcal nuclease for 10 min at 30°C. Purified porcine αβ-tubulin heterodimers (Cytoskeleton, Inc. #T240A) reconstituted in general tubulin buffer (80 mM PIPES pH 6.9, 2 Mm MgCl₂, 0.5 mM EGTA) were used as target proteins. 2 μg of the tubulin heterodimers were diluted in 30 μL of TGMC buffer. The in vitro binding reaction was then set up by adding the target protein to the bait proteins on ice and rocking at 4°C overnight. The bead bound-protein complexes were precipitated by centrifugation, washed 6X with TGMC buffer, and protein complexes were analyzed by immunoblotting.

4.7. Protein purification

Bacteria with the requisite plasmids (GST or GST-EH-1) were inoculated in 50 ml Luria-Bertani (LB) broth containing 100 μg/ml ampicillin. After overnight incubation at 37°C, the 50 ml culture was used to inoculate a 1000 ml LB culture (with antibiotic), and incubated at 37°C until the optical density reached 0.4–0.6 at 600 nm. To induce protein expression, 1 mM IPTG was added, and the culture then incubated overnight at 18°C. The next day, bacterial pellets were obtained by centrifugation of the 1 L bacterial culture at 5000 rpm for 15 min. at 4°C. The pellets were then resuspended in chilled PBS containing protease inhibitor (1 tablet/10 ml of PBS). The resuspended bacteria were then lysed using sonication on ice at 35% amplitude for 10 min. in pulses of 20 sec on and 10 sec off. The lysates were then cleared to remove debris by centrifugation at 19000 rpm for 45 min at 4°C. Proteins were then purified on glutathione resin (GenScript, #L00206) by incubating/rocking the cleared lysate with PBS-washed glutathione resin for 3 h on ice. This was followed by 4 washes in PBS, and then the resin-bound GST proteins were stored in PBS.

4.8. Co-immunoprecipitation

All co-immunoprecipitation experiments were performed with the same protocol. Cells growing on either 60 mm or 100 mm dishes were washed 3X with chilled PBS. The cells were collected, and pellets were lyzed in cold Brij lysis buffer (25 mM Tris pH 7.4, 125 mM NaCl, 1 mM MgCl₂, 1% Brij) and freshly added protease inhibitor cocktail. The cell lysate was then collected and incubated with or without (negative control-beads only) anti-EHD1 antibody (LifeTein, #1795) and for the corresponding tubulin immunoprecipitations, lysates were incubated with anti-α-tubulin antibody (Sigma, #T9026 DM1a), anti-TUBB1 antibody (Santa Cruz, #sc58884), anti-TUBB3 antibody (Biomatik, #CAU-22722), anti-γ-tubulin (Sigma, #T5326), overnight at 4°C. The next day protein A agarose beads (Cell Signaling Technology, #9863S) or protein G Sepharose beads (Cytiva, #17061801) were added to the lysate-antibody mix and the negative control and left for rocking at 4°C for 2 h. The samples were then washed 3X with the coIP wash buffer containing 50 mM Tris pH 7.4, 150 mM NaCl, 0.1% NP-40 and Iodoacetamide (1.8 mg/ml). Protein complexes were then eluted from the beads by boiling the sample for 10 min. in the presence of 4X loading buffer. Once eluted, the proteins were then detected by immunoblotting.

For Rabankyrin-5-TUBB3 co-immunoprecipitations, HeLa cells were transfected for 48 h with GFP-Rabankyrin-5 DNA or myc-Rabankyrin-5 DNA or HA-Rabankyrin-5 DNA, diluted in GeneExpresso reagent (Excellgen #EG-1031). Cell collection and lysis was done as indicated earlier. The lysate from GFP-Rabankyrin-5 transfected cells was then incubated with anti-GFP antibody coated beads (GFP Selector iba, #2-9131-020) and left rocking overnight at 4°C, whereas the lysates collected from myc-Rabankyrin-5 and HA-Rabankyrin-5 transfected cells were incubated with or without (negative control-beads only) anti-myc and anti-HA antibodies respectively, and left overnight on rocking at 4°C. The next day protein G Sepharose beads (Cytiva, #17061801) were added to the lysate-antibody mix (for HA- and myc-tagged Rabankyrin-5) and the corresponding negative control and left for rocking at 4°C for 2 h. The beads were washed and subsequently proteins were eluted as indicated earlier. The proteins were then detected by immunoblotting.

4.9. Immunoblotting

To validate the knock-down mediated by siRNA, mock and siRNA-treated cells were washed 3X with chilled PBS and collected with a rubber scraper. Cell pellets were then obtained by centrifugation and then lysed with the lysis buffer containing 50 mM Tris pH 7.4, 150 mM NaCl, 1% NP-40 and 0.5% sodium deoxycholate. Lysates were then cleared of debris by centrifugation at 13.2 k rpm for 10 min at 4°C. The cleared lysates were then boiled in loading buffer containing 250 mM Tris-HCl pH 6.8, 8% SDS, 40% glycerol, 5% β-mercaptoethanol and 0.2% (w/v) bromophenol blue for 10 min. The boiled samples were then separated by 10% SDS-PAGE followed by transfer to nitrocellulose membranes (GE Healthcare, Chicago, IL). The membranes were blocked for 30 min. at room temperature using 5% dried milk in PBS containing 0.3% (vol/vol) Tween-20 (PBST). This was followed by incubation of the membranes with the appropriate primary antibodies diluted in PBST for 1–2 h at room temperature. The unconjugated primary antibodies were removed by washing the membrane 3X with PBST. Membranes were then incubated with the appropriate HRP-conjugated secondary antibody diluted in PBST for 1 h at room temperature, followed by 3 PBST washes. The HRP-conjugated antibodies were then visualized using enhanced chemiluminescence (Bio-Rad, Hercules, CA) and the corresponding digital images were obtained through the iBright Imaging Systems (Invitrogen).

For stripping and re-blotting, the priorly blotted nitrocellulose membrane was first washed with PBST and then kept in 3 M guanidine thiocyanate for 3 minutes on rocking to allow stripping of bound antibodies. This was followed by washing 3X with PBST. The membrane was then reblocked for 30 min. at room temperature using 5% dried milk in PBS containing 0.3% (vol/vol) Tween-20 (PBST). The membrane was then incubated with apt primary and secondary antibodies as earlier described.

4.10. C. elegans studies

Growth and maintenance: E. coli OP50 seeded Modified Youngren’s Only Bacto-peptone (MYOB) agar plates (3.49 mM Tris-Cl, 1.98 mM Tris-base, 2% (w/v) bactopeptone, 34.2 mM sodium chloride, 8 μg/ml cholesterol) were used to grow all the C. elegans strains described in this study. All strains were maintained at 20°C.

C. elegans genetic crosses and strain genotypes: The previously described rme-1(b1045) V mutant was [7] was obtained from the Caenorhabditis Genetics Center (CGC). The IYR028 strain (bsIs20[pNP99: unc-119(+) tbb-1p::mCherry::tbb-2::tbb-2 3′-UTR]) was generated as described previously [54]. C. elegans males expressing mCherry-tubulin were obtained from the IYR028 strain by heat shock at 30 °C for 6 hours. These males were crossed with rme-1(b1045) V hermaphrodites, and the resultant progeny were screened for homozygosity of the mCherry-tubulin transgene using confocal fluorescence microscopy (Nikon Instruments, Inc.). mCherry-tubulin homozygotes were further genotyped by PCR to determine homozygosity for either wild-type rme-1 or the rme-1(b1045) mutation.

Primers used for PCR screening were:

Forward: 5′-GGAAGCATTCCTGGAAATG-3′

Reverse: 5′-AAAGTCCTATTGTAAATGAAAA-3′

Expected product sizes: wild-type rme-1 = 2459 bp; rme-1(b1045) mutant = 934 bp.

The IYR040 strain (bsIs20[pNP99: unc-119(+) tbb-1p::mCherry::tbb-2::tbb-2 3′-UTR]) was homozygous for both mCherry-tubulin and wild-type rme-1, and was used as a control. The IYR041 strain (rme-1(b1045) V; bsIs20[pNP99: unc-119(+) tbb-1p::mCherry::tbb-2::tbb-2 3′-UTR]) was homozygous for both mCherry-tubulin and the rme-1(b1045) mutation and was used to assess the effect of rme-1 perturbation on tubulin intensity at the spindle poles.

C. elegans embryo imaging: A Nikon Ti2-E inverted microscope (Nikon Instruments, Inc.) equipped with a CREST X-Light V3 spinning-disk confocal system, a motorized XY stage with encoder, and a 600 μm Z piezo stage was used to image C. elegans embryos dissected from the IYR040 and IYR041 strains. C. elegans embryos were dissected and prepared for imaging on an agarose pad and the coverslips were sealed with Vaseline as described previously [55]. Laser excitation was provided by a Celesta laser launch system (Lumencor, Inc.). Timelapse imaging was performed using the NIS-Elements software (Nikon Instruments, Inc.) by taking 1.5 μm Z-stacks every ~45 seconds using a 60× Plan Apo λD oil immersion objective (NA 1.42, refractive index 1.515) and an ORCA-Fusion BT sCMOS camera (Hamamatsu Photonics). To enable quantitative analysis, all images were acquired using the same imaging conditions: 200 ms exposure time and 70% laser intensity on the 561 nm channel. Maximum intensity Z-projections were obtained using the NIS-Elements software. For visual presentation, brightness and contrast adjustments were applied equally across control and rme-1(b1045) images.

Tubulin intensity measurements: 1-cell stage C. elegans embryos expressing mCherry-tubulin (from the IYR040 and IYR041 strains) were analyzed for tubulin intensity at the spindle poles and the cytoplasm. To measure tubulin intensity, the NIS-Elements software (Nikon Instruments, Inc.) was used to obtain maximum intensity Z-projections of the original .nd2 files. Fixed intensity scaling was applied with minimum and maximum display values set to 80 and 800, respectively (gamma = 1) for all images. For spindle pole tubulin intensity measurements, a circular region of interest (ROI) with a constant area of ~15.8 μm² was placed over each spindle pole and in the cytoplasm. Mean spindle pole intensity (in arbitrary units) was calculated by subtracting the cytoplasmic background from the spindle pole ROI intensity using the measurement tools in NIS-Elements. For cytoplasmic tubulin intensity measurements, a circular ROI of 128.97 μm² was placed within the cytoplasm of 1-cell stage pre-metaphase C. elegans embryos, immediately prior to metaphase onset, and positioned away from centrosome-nucleated spindle microtubules. Pre-metaphase embryos were specifically selected to permit placement of a large, representative cytoplasmic ROI without confounding contributions from the spindle. Mean cytoplasmic tubulin fluorescence intensity was quantified for each embryo using NIS-Elements software. Statistical analysis of spindle pole and cytoplasmic tubulin intensities in C. elegans: GraphPad Prism 10.5 software (GraphPad Software, Inc., San Diego, CA) was used to perform statistical analyses of the tubulin intensities at metaphase spindle poles and the cytoplasm of 1-cell C. elegans embryos. A two-tailed unpaired t-test was used to compare tubulin intensities between control and rme-1(b1045) mutant embryos. The error bars denote the standard deviation (s.d.) above and below the mean and the middle lines represent the mean. Differences were considered statistically significant if p<0.05.

4.11. Statistical Analysis

For the analysis shown in each experiment (except 4.10 for C. elegans studies), data obtained from three independent experiments were quantified either via Zen or Imaris software and then exported and processed for statistical analysis using GraphPad Prism 10.2.3 (GraphPad, San Diego, CA) software. The graphs shown depict mean with standard deviation. Normal distribution was tested using the D’Agostino and Pearson normality test. Statistical significance was calculated with an unpaired two-tailed t test for normally distributed samples and with the Mann–Whitney two-tailed t test for comparisons which did not meet the assumption of normality.

Supplementary Material

Supplemental Fig. 1

Supplemental Figure 1 Legend EB1-GFP tracks in mock and EHD1 knock-out cells.

Mock and CRISPR/Cas9-edited EHD1 knock-out HeLa cells were transfected with EB1-GFP for 16 h and then live imaged and quantified for EB1-GFP track length/displacement length (using Imaris software) as described in methods. EB1 structures were rendered as “surfaces” and their corresponding trajectories detected across a full cycle (180 frames for 90 seconds) of live imaging were used to generate “tracks” for each EB1-GFP surface. The wide field image was zoomed in Imaris to focus on one representative track for visualization and representation. Shown is one isolated EB1-GFP structure (highly intense green-colored surface outlined in green at the start of the track) and the associated “track” generated by Imaris following movement. Color coding of the tracks is indicative of the time during which the track is detected. Each track is displayed in “Cylinder” style of 0.04 μm diameter under “Track Style” command of Imaris for better visualization. Track displacement length as computed by Imaris was 5.65 μm for Mock HeLa, A; and 4.69 μm for EHD1 knockout, B. Videos were then exported from Imaris selectively for the key frame range within which the EB1-GFP movement was detected (Frames 46–68 for A, and Frames 1–20 for B) for each sample from a total of 180 frames. Displayed videos were manually set to play at a rate of 3 frames per second and are representatives from three independent experiments. Scale bar, 0.5 μm

Supplemental Fig. 2

Supplemental Figure 2 Legend Endogenous EHD1 interacts with α-tubulin, β3-tubulin and γ-tubulin in RPE-1 cells.

Lysates collected from RPE-1 cells were subjected to immunoprecipitation by anti-EHD1, an anti-tubulin antibody or immunoprecipitated with beads only (control). The bead-bound fractions were separated using SDS-PAGE, transferred to nitrocellulose and then immunoblotted with antibodies against β3-tubulin (TUBB3) (A), γ-tubulin (B), or a-tubulin (C) for the lysate, EHD1 coIP, anti-tubulin IP, and beads only lanes. The blots were then stripped with 3 M guanidine thiocyanate, washed, and blotted using anti-EHD1 antibody to confirm EHD1 pull down/immunoprecipitation efficacy (and reciprocal interactions). Lower and higher exposure times are marked (upper panels). Depicted blots are representatives taken from three independent experiments.

Highlights:

• Endogenous EHD1 interacts with TUBB1, TUBB3, α-tubulin and γ-tubulin.

• EHD1 depletion impacts endosome distribution in mammalian cells.

• EHD1 depletion increases tubulin at 1-cell C. elegans metaphase spindle poles.

• EHD1 is required for normal microtubule regrowth.

• EHD1 affects EB1 track length and displacement at microtubule growing ends.