Microtubule-Dependent Apical Polarization of Basement Membrane Matrix mRNAs in Mouse Epithelial Cells

Abstract

The basement membrane (BM) demarcating epithelial tissues undergoes rapid expansion to accommodate tissue growth and morphogenesis during embryonic development. To facilitate the secretion of bulky BM proteins, their mRNAs are polarized basally in the follicle epithelial cells of the Drosophila egg chamber to position their sites of production close to their deposition. In contrast, we observed the apical rather than basal polarization of all major BM mRNAs in the outer epithelial cells adjacent to the BM of mouse embryonic salivary glands using single-molecule RNA fluorescence in situ hybridization (smFISH). Moreover, electron microscopy and immunofluorescence revealed apical polarization of both the endoplasmic reticulum (ER) and Golgi apparatus, indicating that the site of BM component production was opposite to the site of deposition. At the apical side, BM mRNAs colocalized with ER, suggesting they may be co-translationally tethered. After microtubule inhibition, the BM mRNAs and ER became uniformly distributed rather than apically polarized, but they remained unchanged after inhibiting myosin II, ROCK, or F-actin, or after enzymatic disruption of the BM. Because Rab6 is generally required for Golgi-to-plasma membrane trafficking of BM components, we used lentivirus to express an mScarlet-tagged Rab6a in salivary gland epithelial cultures to visualize vesicle trafficking dynamics. We observed extensive bidirectional vesicle movements between Golgi at the apical side and the basal plasma membrane adjacent to the BM. Moreover, we showed that these vesicle movements depend on the microtubule motor kinesin-1 because very few vesicles remained motile after treatment with kinesore to compete for cargo-binding sites on kinesin-1. Overall, our work highlights the diverse strategies that different organisms use to secrete bulky matrix proteins: while Drosophila follicle epithelial cells strategically place their sites of BM protein production close to their deposition, mouse embryonic epithelial cells place their sites of production at the opposite end. Instead of spatial proximity, they use the microtubule cytoskeleton to mediate this organization as well as for the apical-to-basal transport of BM proteins.

Introduction

The basement membrane (BM) is a dense thin sheet of extracellular matrix that delineates the borders of many tissues, including epithelia, muscle, and adipose tissues (Jayadev and Sherwood, 2017). The BM sheet is intricately woven from two primary networks: laminin and type IV collagen. These molecular components are interconnected and stabilized by various crosslinking molecules, including nidogen and perlecan (Yurchenco, 2011). The importance of BM is highlighted by the embryonic lethality that results from genetic ablation of core BM components, especially collagen IV and laminin (Wu et al., 2023). The BM and its active remodeling are both required for the branching morphogenesis of mammalian organs, such as mouse embryonic salivary glands (Harunaga et al., 2014; Wang et al., 2021).

During embryonic development, the BM must undergo rapid expansion to accommodate tissue growth and morphogenesis. To permit tissue expansion, numerous micro-perforations of the BM are generated at the growing tips of mouse embryonic organs as well as the posterior side of gastrulating mouse embryos (Harunaga et al., 2014; Kyprianou et al., 2020). These perforations depend on the activities of matrix remodeling enzymes, such as matrix metalloproteinases (MMPs) and a disintegrin and metalloproteases (ADAMs) (Harunaga et al., 2014; Kyprianou et al., 2020).

In addition to matrix remodeling, BM expansion also requires the production and secretion of new matrix components. Many BM components are large molecules and thus pose challenges for cells to secrete. For example, the building block of type IV collagen networks is a heterotrimer approximately 400 nm in length (Yurchenco, 2011). This and other matrix components can be too large to be packaged in regular COPII vesicles, which range from 60 to 90 nm in diameter and mediate cargo transport from the endoplasmic reticulum (ER) to the Golgi apparatus. To accommodate bulky cargos, cells rely on specialized molecular machineries to generate large COPII vesicles. For example, collagen secretion generally requires MIA3/TANGO1, which is localized to ER exit sites and is thought to delay vesicle biogenesis to allow large vesicle formation (Saito et al., 2009; Wilson et al., 2011).

One strategy for facilitating the secretion and local assembly of bulky cargo molecules is to place their sites of production close to their deposition. For example, in the follicle epithelial cells of the Drosophila egg chamber, the mRNAs of major BM components, including collagen IV and laminin, are all polarized basally (Lerner et al., 2013). While the transitional ER (tER) and Golgi stacks are distributed throughout these cells, BM proteins and enzymes dedicated to their biogenesis are enriched in the basal compartments of the ER, which are closest to the targeted basal surface (Lerner et al., 2013). It remains unclear whether this strategy is widely used in other systems, such as the epithelial cells of developing mammalian embryos, which are confronted by the same challenge of secreting large amounts of bulky BM components.

Here, we show that all major BM mRNAs are polarized apically instead of basally in the outer epithelial cells of mouse embryonic salivary glands. Both the ER and Golgi apparatus also have apical polarization, suggesting BM component production occurs opposite to its deposition site in these epithelial cells. BM mRNAs at the apical side appear to be co-translationally tethered to the ER. The apical polarization of BM mRNAs and ER depends on microtubules, while the Golgi-to-plasma membrane vesicle trafficking depends on the microtubule motor kinesin-1. In summary, our research reveals the diverse strategies that various organisms can employ to secrete bulky BM components. While Drosophila follicle epithelial cells conveniently locate their BM protein production areas near their deposition sites, mouse embryonic epithelial cells position these sites at the opposite end of the cell, relying on the microtubule cytoskeleton for maintaining this organization and mediating the cargo transport.

Results

Basement membrane mRNAs are polarized apically in mouse embryonic salivary gland outer epithelial cells

To characterize the subcellular localization of BM mRNAs, we performed single-molecule mRNA fluorescence in situ hybridization (smFISH) on whole-mount mouse embryonic salivary glands (Wang, 2019). In smFISH, each mRNA molecule is hybridized by 30–50 singly labeled short probes that collectively generates a single fluorescence spot of a near-diffraction-limit size (Fig. 1A) (Raj et al., 2008). We first examined mRNAs of Col4a1, which encode the α1 subunit of collagen IV, a major BM component. Our results revealed a 3-fold enrichment of Col4a1 mRNAs on the apical side compared to the basal side in outer epithelial cells (Fig. 1B, D, E; Video S1). In addition, the expression of Col4a1 mRNAs was higher in outer epithelial cells adjacent to the BM than interior epithelial cells (Fig. 1B). Apical polarization was observed for all examined mRNAs encoding components of the BM or other extracellular matrix, including collagen IV α2 (Col4a2), laminin α5 (Lama5), laminin β1 (Lamb1), laminin γ1 (Lamc1), perlecan (Hspg2), and fibronectin (Fn1) (Fig. 1C–E).

Figure 1. Various subcellular localization patterns of mRNAs in mouse embryonic salivary gland outer epithelial cells.

(A) Schematics of the mouse submandibular salivary gland at embryonic day 13 (E13) and the single-molecule mRNA FISH method. (B) Confocal fluorescence image showing an E13 salivary gland co-stained for Col4a1 mRNA, collagen IV protein, E-cadherin protein, and nuclear DNA (with DAPI). Col4a1 mRNA encodes a subunit of collagen IV, a major component of the basement membrane. Dashed box in the left image indicates the zoomed region shown in the right two images. (C) Confocal fluorescence images showing expression patterns of various mRNAs in the E13 salivary gland. Yellow dashed line indicates the position of basement membrane, which marks the basal side of the outer epithelial cells. (D) Schematic depicting quantification of the polarity index. Note that a polarity index of 0.5 indicates a 3-fold increased mRNA density in the apical side compared to the basal side. (E) Plot of the polarity index of various mRNAs in the outer epithelial cells adjacent to basement membrane of mouse embryonic salivary glands. Each dot corresponds to one gland. Error bars: standard deviation. Numbers of samples: Pax9, 9; Sox9, 6; Sox10, 9; Emg1, 6; Gapdh, 6; Col4a1, 61; Col4a2, 5; Lama5, 7; Lamb1, 6; Lamc1, 17; Hspg2, 11; Fn1, 8; Net1, 4; Cyb5r3, 3. Statistics: One-Way ANOVA (p=3.5e-44) with Tukey’s post-hoc test. All comparisons were made with Pax9. n.s., not significant. **, p<0.01. ***, p<0.001. Scale bars, 10 μm.

To determine whether mRNAs are globally enriched on the apical side of these cells, we examined mRNAs encoding proteins that do not enter the secretary pathway. We found that mRNAs of transcriptional factors and metabolic enzymes (Pax9, Sox9, Sox10, Emg1, Gapdh) were uniformly distributed (Fig. 1C–E; Video S2). We next examined mRNAs of Net1 and Cyb5r3, which were previously reported to be polarized basally in the mouse intestinal epithelium (Moor et al., 2017). We found that Net1 and Cyb5r3 mRNAs were also polarized basally in the outer epithelial cells of mouse embryonic salivary glands (Fig. 1C–E). Thus, while mRNAs encoding extracellular matrix components are polarized apically, other mRNAs can adopt diverse subcellular localization patterns in mouse embryonic epithelial cells.

Basement membrane mRNAs colocalize with the endoplasmic reticulum

Because BM components need to be synthesized in the ER and routed through the Golgi apparatus to be secreted, we examined the subcellular localizations of these organelles by immunostaining. We found that both the ER and Golgi are strongly polarized to the apical side of the outer epithelial cells in developing salivary glands (Fig. 2A–B, E–F). In contrast, mitochondria were uniformly distributed, while ribosomes were basally enriched in these cells (Fig. 2C–D, G–H). The apical enrichment of ER and Golgi, as well as the relatively uniform distribution of mitochondria, were further confirmed by our transmission electron microscopy data (Fig. 2I) and serial block face scanning electron microscopy analysis (Video S3). To determine the spatial relationships between BM mRNAs and secretary organelles, we performed smFISH for Col4a1 mRNA with co-immunostaining of an ER or Golgi marker, which showed that the Col4a1 mRNA colocalized with the ER but not the Golgi (Fig. 2J–K). Therefore, the primary secretory pathway organelles, along with mRNAs encoding secreted BM proteins, are all apically polarized, indicating that these cells may have specialized their apical compartment for producing BM components, which are eventually secreted at the basal side.

Figure 2. Basement matrix mRNAs colocalize with the endoplasmic reticulum (ER).

(A-H) Confocal fluorescence images showing the expression patterns of ER (anti-PDI), Golgi (anti-RCAS1), mitochondria (anti-AIF), and ribosomes (18S rRNA smFISH) in E13 salivary glands. Images in (E-H) are magnified views of outer epithelial cells. (I) Transmission electron microscopy images of an E13 salivary gland (left three panels). The right panel shows annotations of several organelles. (J-K) Confocal fluorescence images showing the expression patterns of Col4a1 mRNAs compared to ER (anti-PDI) or Golgi (anti-RCAS1) in the outer epithelial cells of E13 salivary glands. Cyan arrowheads indicate colocalized mRNA and ER signals at the apical side of outer epithelial cells. Yellow dashed lines indicate the positions of basement membrane. Scale bars: 100 μm (A-C), 2 μm (I), 10 μm (D-H, J-K).

Secreted proteins commonly have a signal peptide at their N-terminus, which is recognized by the signal recognition particle (SRP) that targets the ribosome-mRNA-nascent peptide chain complex to the ER (Akopian et al., 2013). To test whether a signal peptide alone is sufficient to mediate apical polarization of mRNAs in these cells, we used lentivirus to express an exogenous reporter transgene with or without a signal peptide in salivary gland epithelial explants (Fig. 3A–B) (Sekiguchi et al., 2023). Both the mRNA and protein products of the reporter lacking a signal peptide were distributed uniformly in the outer epithelial cells (Fig. 3C–F). In contrast, use of the signal peptide sequence from Col4a1 (collagen IV α1), Col4a2 (collagen IV α2), or Ins2 (insulin) all resulted in apical enrichment of the reporter mRNA and protein (Fig. 3C–F). Thus, the observed apical polarization of BM mRNAs appeared to be mediated by co-translational tethering to the ER.

Figure 3. Signal peptide is sufficient to mediate the apical polarization of an exogenous reporter mRNA.

(A) Schematics of the lentivirus constructs expressing an exogenous sfGFP reporter (scFv-GCN4-sfGFP-GB1) with or without a signal peptide (sigP). (B) Schematics of the experiment. (C) Maximum intensity projections of confocal fluorescence images showing the expression patterns of the sfGFP reporter protein (green) and mRNA (magenta). (D) An example showing raw, filtered, and segmented images of the sfGFP reporter smFISH. (E) Schematic depicting quantification of the polarity index. (F) Plot of polarity index of reporter mRNAs without or with the indicated signal peptides. Each dot is the averaged value from 3–8 cells of one explant culture. Numbers of samples: 6 explants per experimental group. Error bars: standard deviation. Statistics: One-Way ANOVA (p=5.8e-8) with Tukey’s post-hoc test. All comparisons were with the No sigP group. ***, p<0.001. Scale bars: 5 μm.

The apical polarization of basement membrane mRNAs depends on microtubules

To determine how the apical polarization of BM mRNAs and secretary organelles is maintained, we examined the localization patterns of Col4a1 mRNAs in salivary glands treated with various pharmaceutical inhibitors or the enzyme collagenase. We found that the apical polarization of Col4a1 mRNAs was ablated upon microtubule inhibition by nocodazole or colchicine, but it remained unaffected after BM disruption by collagenase, F-actin inhibition by latrunculin A or cytochalasin D, myosin II inhibition by blebbistatin, or ROCK inhibition by Y27632 (Fig. 4A–B). Similarly, the ER and Golgi also became uniformly distributed upon microtubule inhibition by nocodazole treatment (Fig. 4C–D). Under the collagenase treatment and the acute F-actin inhibition conditions we used, the BM was completely ablated, and disruption of F-actin organization was also evident (Fig. 4E). Therefore, the apical polarization of BM mRNAs and secretary organelles in mouse salivary gland epithelial cells depends on microtubules, but not on the presence of the BM, F-actin, or myosin contractility.

Figure 4. The apical polarization of matrix mRNAs depends on microtubules.

(A) Confocal fluorescence images of Col4a1 mRNA expression patterns in outer epithelial cells of E13 salivary glands under the indicated treatment conditions. Yellow dashed lines indicate the epithelial-mesenchymal boundary, which marks the basal side of the outer epithelial cells. (B) Plot of the polarity index of Col4a1 mRNAs under various treatment conditions. See Fig. 1D for the quantification method. Note that a polarity index of 0.5 represents 3-fold greater mRNA density in the apical side compared to the basal side. Each dot represents one gland. Numbers of samples: Control, 18; Nocodazole, 14; Colchicine, 3; Collagenase, 7; Latrunculin A, 4; Cytochalasin D, 4; Blebbistatin, 4; Y27632, 5. Error bars: standard deviation. Statistics: One-Way ANOVA (p=3.7e-13) with Tukey’s post-hoc test. All comparisons were with Control. n.s., not significant, p>0.05. ***, p<0.001. (C-D) Confocal fluorescence images of the ER (C) or Golgi (D) localization under the indicated treatments. (E) Confocal fluorescence images of the basement membrane and F-actin under the indicated treatments. Yellow arrowheads indicate the epithelial-mesenchymal boundary. Scale bars, 10 μm.

The apical-to-basal cargo transport depends on kinesin-1

We next asked how newly synthesized BM components could be delivered from the apical side to the basal membrane for deposition. To visualize the vesicles mediating cargo transport from the Golgi to the plasma membrane, we used the small GTPase Rab6, which is generally required for secretion of cargos without a transmembrane domain, including BM components (Homma et al., 2019). Consistent with its established role, immunostaining of Rab6 revealed its prominent enrichment in the Golgi, as well as its presence on intracellular vesicles, in the outer epithelial cells of embryonic salivary glands (Fig. 5A). To visualize vesicle dynamics, we used lentivirus to express mScarlet-tagged Rab6a in the epithelial explants (Fig. 3B) (Bindels et al., 2017). We observed abundant Rab6a-marked vesicle dynamics in most outer epithelial cells, but relatively few moving vesicles in some other cells, suggesting that secretion was tightly regulated in these epithelial cells (Fig. 5B). Importantly, vesicle dynamics was strongly inhibited after treatment by kinesore, which competes for cargo-binding sites on kinesin-1 motors (Fig. 5B–C, Video S4) (Randall et al., 2017). Thus, the apical-to-basal cargo transport in mouse embryonic salivary gland epithelial cells depends on the microtubule motor kinesin-1.

Figure 5. Vesicle dynamics in outer epithelial cells depends on the microtubule motor kinesin-1.

(A) Confocal fluorescence images of the outer epithelial cells in E13 salivary glands. Cyan arrowheads indicate Rab6-marked vesicles. (B) Temporal-coded Airy Scan time-lapse images of mScarlet-I-Rab6a in outer epithelial cells treated with solvent control (DMSO) or kinesore. The color code was chosen so that neighboring time frames have different colors to highlight vesicle dynamics. (C) Plot of normalized dynamic vesicle signals in control or kinesore-treated explants. Statistics: Student’s t-test. **, p<0.01. Error bars: standard deviation. Scale bars: 20 μm (A), 5 μm (B).

Discussion

Our research reveals the diverse cellular strategies that different organisms may use to accommodate the effective secretion of bulky basement membrane (BM) components. We showed that major BM mRNAs and the primary secretary organelles are all polarized apically in the outer epithelial cells of mouse embryonic salivary glands, contrary to the basal polarization observed in the Drosophila egg chamber (Lerner et al., 2013). While Drosophila epithelial cells place their sites of BM protein production close to their deposition, mouse embryonic epithelial cells place their sites of production at the distal, apical end. Instead of spatial proximity, they use the microtubule cytoskeleton to mediate this organization as well as for the apical-to-basal transport of BM proteins.

A common feature of Drosophila follicle epithelial cells and mouse embryonic salivary gland epithelial cells is the compartmentalization of BM production, although it is at two opposite sides of these two types of cells. Restricting BM production within a smaller region may increase the efficiency of BM protein biogenesis and secretion, since some BM components require specialized molecular machinery. For example, the export of bulky components from the ER commonly requires MIA3/TANGO1 (Wilson et al., 2011). In mammalian cells, the biosynthesis of collagens depends on the collagen-specific chaperone HSP47 for triple helix assembly (Nagata, 2003). Compartmentalization may be a widely used strategy to facilitate the biosynthesis of bulky matrix components. For example, the production site of aggrecan, a different type of extracellular matrix component, is localized to a distinct ER region in avian chondrocytes (Vertel et al., 1989).

In essence, our findings highlight the complexities and distinctive adaptations of the cellular mechanisms governing BM secretion, suggesting the need for further investigations into other mammalian systems. It paves the way for future research into the regulatory networks controlling these processes, which could have broader implications for tissue engineering and regenerative medicine.

Materials and Methods

Mouse embryos

Mouse experiments were conducted under animal study protocols 14-745, 17-845, and 20-1040 with approval from the NIDCR Animal Care and Use Committee (ACUC). Mouse embryos at the desired gestational stages were isolated from timed pregnant ICR (CD-1) outbred mice obtained from Envigo. All mouse embryos were used without sex identification (mixed sexes).

Salivary gland isolation and culture

Mouse submandibular salivary glands were isolated from 13-day embryos as previously described (Sequeira et al., 2013). Briefly, the mouse embryo was decapitated using a scalpel (Fine Science Tools, 10011–00 and 10003–12). The removed head was then held sideways with forceps (Fine Science Tools, 11251–20), and another cut was made across the mouth to separate the mandible with the tongue from the top part of the head. Under a dissecting microscope, the mandible was positioned with the tongue side down on a glass plate. Using forceps, a cut was made at the mandible’s midline to expose the tongue and the two attached submandibular glands. Once surrounding tissues were cleared away with forceps, the glands were removed from the tongue and placed in a dish with base medium, which was DMEM/F-12 medium (Thermo Fisher, 11039047) supplemented with 1X PenStrep (Thermo Fisher, 15140163). The isolated glands were then cultured on polycarbonate filters (MilliporeSigma, WHA110405) floating on base medium supplemented with 150 μg/mL vitamin C (MilliporeSigma, A7506) and 50 μg/mL transferrin (MilliporeSigma, T8158). Up to 3 filters were placed on 1 mL culture medium in a 35 mm tissue culture dish and cultured at 37°C with 5% CO₂.

Single epithelial bud isolation and culture

Single bud isolation was performed as previously described (Sekiguchi et al., 2023). Briefly, the adjacent sublingual gland was removed from the submandibular gland during dissection to ensure all epithelial buds were from submandibular glands. The glands were then treated with 150 μL 2 units/mL dispase (Thermo Fisher, 17105041; diluted in base medium) in a well of a Pyrex spot plate (Fisher Scientific 13–748B) for 15 min at 37°C. Up to 10 glands were treated in each well. After dispase treatment, glands were washed twice with 5% BSA (w/v; MilliporeSigma, A8577; diluted in base medium) in the same well to quench dispase. Under a dissecting microscope, glands were subjected to repetitive pipetting using a 200 μL pipettor set at 100 μL with a low-retention tip (Rainin, 30389187), until the mesenchyme was dissociated into single cells while epithelial buds remained intact. Salivary epithelial buds were subsequently rinsed 3 times by being transferred to new wells of the spot plate prefilled with 150 μL 5% BSA (w/v) using a 20 μL pipettor with a low-retention pipette tip (Rainin, 30389190). Care was taken during transfer to minimize carryover of mesenchymal cells. To mitigate evaporation, the wells of the spot plate were covered with coverslips during preparations of next steps, such as lentivirus transduction and explant culture. Epithelial buds were further washed 2–3 times in base medium without BSA before the next steps.

Epithelial buds were cultured in ultra-low attachment 96-well V-bottom plates (S-bio, MS-9096VZ) as previously described (Wang et al., 2021a). Briefly, each bud was cultured separately in each well in 100 μL explant culture medium, which was base medium supplemented with 0.5 mg/mL growth factor-reduced Matrigel (Corning, 356231; stock 9–10 mg/mL), 200 ng/mL FGF7 (R&D Systems, 5028-KG-025), and 1X ITS supplement (Thermo Fisher, 41400045). In practice, an aliquot of Matrigel was thawed on ice or at 4°C early before salivary gland dissections. The explant culture medium was prepared at 2X concentration with a total volume of (n + 2) × 50 μL, where n is the sample number. The wells for explant culture were pre-filled with 45 μL base medium (DMEM/F-12 with 1X PenStrep), and one bud was transferred into each well in precisely 5 μL medium using a low-retention pipette tip. 50 μL 2X explant culture media was then added to each well. The explants were cultured at 37°C with 5% CO₂.

Single-molecule mRNA fluorescence in situ hybridization (smFISH)

Probes for smFISH were designed using the Stellaris probe designer. Sequences of all smFISH probes used probe sequences can be found in Table S1. Some probes were synthesized by LGC Biosearch Technologies with TAMRA-C9, CAL Fluor Red 590 or Quasar 670 dyes. Other probes were ordered as unlabeled DNA oligos and enzymatically labeled with ddCTP-TMR (Jena Biosciences, NU-850-TAM) using terminal transferase (New England Biolabs, M0315L) as previously described (Gaspar et al., 2017).

smFISH was applied to wholemount E13 salivary glands or epithelial explant cultures as detailed previously (Wang, 2019). Briefly, samples were fixed using 4% paraformaldehyde in PBS either at room temperature (RT) for an hour or overnight at 4°C. After rinsing in PBSTx (PBS + 0.2% Triton-X-100), samples were dehydrated stepwise in cold 30%, 50%, 70% and 100% methanol in PBSTx on ice. After dehydration, samples can be preserved in 100% methanol at −20°C for up to 3 months, but most samples were processed within 1 week. For smFISH, samples were rehydrated stepwise in cold 70%, 50%, and 30% methanol in PBSTx on ice. Following a 10-minute PBSTx rinse at RT, they were permeabilized in 0.5% SDS in PBS, and then equilibrated in smFISH Wash Solution (2X SSC and 10% formamide in RNase-free water). For hybridization, samples were incubated in smFISH Hybridization Solution (2X SSC, 10% formamide, 10% dextran sulfate and 50 μg/mL yeast tRNAs in DEPC-treated water) containing 50 nM probes (1–2 nM each probe) at 37°C for 12 to 16 hours. After hybridization, samples were washed in smFISH Wash Solution for 30 min at RT, stained with 0.5 μg/mL DAPI in smFISH Wash Solution for 2 hours at RT, washed 2 more times for 30 min at RT, rinsed in 2X SSC (K D Medical, RGF-3240) and mounted in ProLong Diamond Anti-fade Mountant (Thermo Fisher, P36961) for imaging.

Small-molecule inhibitors and collagenase treatment

For treatment with small-molecule inhibitors or collagenase, the desired concentration was supplemented to the culture media under the filter (for intact glands) or in the 96-well plate (for epithelial explants). The same dilution factor of solvent was used as control. For nocodazole (e.g., MilliporeSigma, M1404), stock was 1 mg/mL (~3 mM) in DMSO; 1 μg/mL (~3 μM) was used for a 2-hour or 4-hour treatment. For colchicine (MilliporeSigma, C9754), stock was 40 mg/mL (~100 mM) in DMSO, and 40 μg/mL (~100 μM) was used for a 2-hour treatment. For collagenase (Elastin Products Company, CL103), stock was 2 mg/mL in water; 20 μg/mL was used for a 24-hour treatment. For latrunculin A (MilliporeSigma, 428026), stock was 1 mM in DMSO; 200 nM was used for a 4-hour treatment. For cytochalasin D (MilliporeSigma, C2618), stock was 5 mg/mL (~10 mM) in DMSO; 1 μg/mL (~2 μM) was used for a 4-hour treatment. For blebbistatin (MilliporeSigma, 203391), stock was 25 mM; 50 μM was used for a 17-hour treatment. For Y27632 (MilliporeSigma, Y0503), stock was 10 mM; 20 μM was used for a 17-hour treatment. For kinesore (MilliporeSigma, SML2361), stock was 2 mg/mL (3.73 mM) in DMSO; 50 μM was used and live imaging of epithelial explant cultures was started after 1 hour incubation.

Plasmids

Lentiviral plasmids for expressing the sfGFP reporter (scFv-GCN4-sfGFP-GB1) with or without a signal peptides, or mScarlet-I-tagged Rab6a, were constructed by Gibson Assembly (Gibson et al., 2009). The scFv-GCN4-sfGFP-GB1 fragment was amplified from pHR-scFv-GCN4-sfGFP-GB1-NLS-dWPRE (Addgene, # 60906).

Lentivirus packaging

Lentivirus packaging was performed as previously described (Wang et al., 2021). Briefly, lentiviral plasmids for expressing various reporters were co-transfected with psPAX2 (Addgene, 12260) and pMD2.G (Addgene, 12259) into HEK293T cells by calcium co-precipitation. The supernatant containing lentiviruses were collected at 36- and 60-hours post transfection and pooled together. The supernatant was filtered through a 0.45 μm filter (MilliporeSigma, SE1M003M00), and concentrated using PEG (System Biosciences, LV825A-1) following the manufacturer’s instructions. Viral titer was estimated using Lenti-X GoStix Plus (Takara, 631281) after 100X dilution following the manufacturer’s instructions. Concentrated lentiviruses were stored at −80°C.

Lentivirus transduction of salivary epithelial buds

Each epithelial bud was transduced in a separate well of an ultra-low attachment 96-well V-bottom plate (S-bio, MS-9096VZ). Each bud was transferred into a well with precisely 5 μL carryover medium. 15 μL mixture containing 10 μL lentivirus stock, 4 μL base medium, and 1 μL 160 μg/mL polybrene (MilliporeSigma, H9268) was added to each well for lentivirus treatment. The plate with buds and viruses was incubated in a humidified 37°C incubator for 1–2 hours. Each bud was washed 3 times in base media and cultured for 2 days at 37°C with 5% CO₂ before live imaging or fixation and staining.

Antibodies and other staining reagents

The following antibodies and concentrations or dilutions were used. Primary antibodies: anti-E-cadherin (Thermo Fisher, 13-1900), 1 μg/mL; anti-collagen type IV (MilliporeSigma, AB769), 2 μg/mL; anti-PDI (Cell Signaling Technology, 3501), 1:100; anti-RCAS1 (Cell Signaling Technology, 12290), 1:200; anti-AIF (Cell Signaling Technology, 5318), 1:400; anti-RAB6A (Thermo Fisher, PA5–22127), 2 μg/mL. Secondary antibodies were used at 1:200 (1.5–3 μg/mL): Alexa Fluor 647-labeled donkey anti-rat (Jackson ImmunoResearch, 712-546-150), Alexa Fluor 647-labeled donkey anti-rat (Jackson ImmunoResearch, 712-606-150), Cy2-labeled donkey anti-goat (Jackson ImmunoResearch, 705-225-147), Alexa Fluor 647-labeled donkey anti-goat (Jackson ImmunoResearch, 705-606-147), Rhodamine Red-X-labeled donkey anti-rabbit (Jackson ImmunoResearch, 711-296-152). DAPI (Thermo Fisher, 62247) stock was 5 mg/mL in water; 0.5 μg/mL was used. AF594 phalloidin (Thermo Fisher, A12381) stock was 200 units/mL in methanol; 5 units/mL was used.

Immunostaining

Immunostaining of wholemount salivary glands or epithelial explants was performed as previously described (Wang et al., 2021). Briefly, samples were fixed in 4% paraformaldehyde (Electron Microscopy Sciences, 15710) in PBS for 1 hour at room temperature (RT) or overnight at 4°C, permeabilized in PBSTx (PBS with 0.2% Triton-X-100; Thermo Fisher, 28314) for 30 min at RT, blocked in 5% donkey serum (Jackson ImmunoResearch, 017-000-121) in PBSTx for 2 hours at RT, incubated in primary antibodies diluted in 5% donkey serum in PBSTx for 2 days at 4°C, washed 4X 15 min each in PBSTx at RT, incubated in secondary antibodies diluted in PBSTx for 2 days at 4°C, washed 4X 15 min in PBSTx at RT, rinsed once in PBS, and mounted in antifade mountant (Thermo Fisher, P36930) supported by one or two layers of imaging spacers (Grace Bio-labs, 654004). All incubations were performed in sample baskets (Intavis, 12.440) in a 24-well plate.

Co-immunostaining of samples with smFISH was performed as previously described (Wang, 2019). Several changes were made from the regular immunostaining steps described above. First, permeabilization was omitted since samples were permeabilized during smFISH procedures. Second, the base buffer of all incubations was changed to 2X SSC. Third, donkey serum was only included in the blocking step and omitted during long primary antibody incubation to preserve smFISH signals.

Electron microscopy

E13 salivary glands were fixed in a fixative of 2.5% glutaraldehyde and 2% formaldehyde in sodium cacodylate buffer with 2 mM calcium chloride for 5 minutes at room temperature (RT), followed by 2–3 hours incubation in the same fixative on ice. The samples were then washed 5X 3 minutes in cold cacodylate buffer with 2 mM calcium chloride, incubated in a reduced osmium solution for 1 hour on ice, and washed 5X 3 minutes in ddH₂O at RT. The reduced osmium solution was prepared right before use by mixing an equal volume of 4% aqueous osmium tetroxide and a solution containing 3% potassium ferrocyanide in 0.3M cacodylate buffer with 4mM calcium chloride. The samples were then incubated in 1% (w/v) freshly prepared thiocarbohydrazide (TCH) solution for 20 minutes at RT, washed 5X 3 minutes in ddH₂O at RT, incubated in 2% osmium tetroxide in ddH2O for 30min at RT, washed 5X 3 minutes in ddH₂O at RT, incubated in 1% uranyl acetate (aqueous) at 4°C overnight, washed 5X 3 minutes in ddH₂O at RT, and incubated in Walton’s lead aspartate staining solution in a 60°C oven for 30 minutes. To prepare Walton’s lead aspartate solution, 0.066 g lead nitrate was dissolved in 10 mL 0.4% (w/v) L-aspartic acid in water and pH adjusted to 5.5 with 1N KOH. The lead aspartate solution was then incubated in a 60°C oven for 30 minutes before usage (no precipitate should form). After staining, the samples were washed 5X 3 minutes in ddH₂O at RT, dehydrated in freshly prepared, ice-cold ethanol solutions at 20, 50%, 70%, 90%, 100%, 100% (5 minutes each), placed in anhydrous ice-cold acetone and incubated for 10 minutes at RT. After dehydration, EPON-Aradite resin was infiltrated into the samples by incubation in 25%, 50%, 75% resin in acetone for 2 hours each at RT, followed by incubation in 100% resin overnight at RT, and another incubation in fresh 100% resin for 2 hours at RT. For polymerization, samples with the resin were incubated in a 60°C for 48 hours. For transmission electron microscopy (TEM), thin sections of the samples (~90 nm thick) were imaged on an FEI Tecnai T12 transmission electron microscope operating at 120 keV. For serial block face scanning electron microscopy (SBF-SEM), the embedded plastic block was imaged using a Sigma Zeiss SEM with fully automated Gatan 3View system operating at 1.5 keV.

Live imaging of epithelial explants

Single-bud salivary epithelial explant cultures (referred to as “explants” hereafter) were mounted in custom-assembled imaging chambers (see below) and imaged using the Fast Airy Scan module on a Zeiss 880 LSM microscopy system. A 40X NA 1.2 water immersion objective was used. To assemble the imaging chamber, a 4-well silicone chamber was removed from an ibidi dish (ibidi, 80466) and attached to a Bioinert dish (ibidi, 81150). 2 mL 2% agarose (Millipore Sigma, A9539) in base medium (dissolved by a microwave) was added outside of the silicone chamber to hold the chamber in place. 30 μL 2% agarose was added inside each chamber to form a thin agarose pad and incubated for 5 min at room temperature to allow the agarose to solidify. A 30 μL glass micropipette (Drummond Scientific Company, 2-000-030) was used to poke 1 or 2 micro wells in each chamber towards the center of the dish bottom. 100 μL DMEM/F12 medium was added to each well, and a pair of forceps was used to remove the agarose insert from each micro well to expose the glass surface. The medium in each well was then replaced by 80 μL DMEM/F12 medium supplemented with 200 ng/mL FGF7 (R&D Systems, 5028-KG-025), 1X ITS (Thermo Fisher, 41400045) and 0.5 mg/mL growth factor-reduced Matrigel (Corning, 356231; stock 9–10 mg/mL). Explants were transferred from the 96-well plate into the imaging chamber using low-retention pipette tips (Rainin, 30389190), which were cut using a pair of scissors (Fine Science Tools, 14090–09) ~2 mm from the tip to provide a larger opening. Each explant was gently pushed into a microwell using forceps with bent tips for immobilization. Prior to imaging, 1 mL base medium was added outside of the chamber to help maintain humidity.

Image analysis and quantification

Image processing, analysis and quantification were mostly performed using Fiji (Schindelin et al., 2012). Customized ImageJ Macro and Python scripts were used for automating or facilitating image analysis and data visualization. All scripts are available at this GitHub repository: https://github.com/snownontrace/scripts-ECM-mRNA-polarization.

smFISH dot counting was performed using a suite of custom-written ImageJ macros as previously described (Wang, 2019). Briefly, smFISH images were smoothened by a Gaussian filter, contrast enhanced by a morphological top-hat filter (Legland et al., 2016), and local maxima points beyond a user-specified threshold level were identified and counted. For polarity index quantification of wholemount glands, the polygon tool in Fiji was used to manually mark the epithelial-mesenchymal boundaries, and the apical and basal regions of interest (ROIs) were computed by shrinking the epithelial ROI by two 7.5 μm steps. For polarity index quantification of single cells in epithelial explant cultures, the polygon tool was used to mark the boundary of the cell, and the apical and basal ROIs were computed by splitting the cell ROI into two parts with the same height.

For quantification of dynamic Rab6a vesicles, frame-by-frame differences were computed and summed up within a user-drawn basal ROI using the polygon tool in Fiji. The total dynamic signal was divided by the average intensity within the bright Rab6a signal at the Golgi apparatus on the apical side for normalization of expression levels.

Highlights.

• Basement membrane mRNAs are polarized apically in mouse embryonic salivary gland outer epithelial cells

• Basement membrane mRNAs seem to be co-translationally tethered with the ER

• The apical polarization of basement membrane mRNAs and ER depends on microtubules

• The apical-to-basal cargo transport depends on the microtubule motor kinesin-1