Choroid plexus epithelial cells express the adhesion protein P-cadherin at cell-cell contacts and syntaxin-4 in the luminal membrane domain

Abstract

The choroid plexus epithelial cells (CPECs) belong to a small group of polarized cells, where the Na⁺-K⁺-ATPase is expressed in the luminal membrane. The basic polarity of the cells is, therefore, still debated. We investigated the subcellular distribution of an array of proteins known to play fundamental roles either in establishing and maintaining basic cell polarity or in the polarized delivery and recycling of plasma membrane proteins. Immunofluorescence histochemical analysis was applied to determine the subcellular localization of apical and basolateral membrane determinants. Mass spectrometry analysis of CPECs isolated by fluorescence-activated cell sorting was applied to determine the expression of specific forms of the proteins. CPECs mainly express the cell-adhesive P-cadherin, which is localized to the lateral membranes. Proteins belonging to the Crumbs and partitioning defective (Par) protein complexes were all localized to the luminal membrane domain. Par-1 and the Scribble complex were localized to the basolateral membrane domain. Lethal(2) giant larvae homolog 2 (Lgl2) labeling was preferentially observed in the luminal membrane domain. Phosphatidylinositol 3,4,5-trisphosphate (PIP₃) was immunolocalized to the basolateral membrane domain, while phosphatidylinositol 4,5-bisphosphate (PIP₂) staining was most prominent in the luminal membrane domain along with the PIP₃ phosphatase, Pten. The apical target-SNARE syntaxin-3 and the basolateral target-SNARE syntaxin-4 were both localized to the apical membrane domain in CPECs, which lack cellular expression of the clathrin adaptor protein AP-1B for basolateral protein recycling. In conclusion, the CPECs are conventionally polarized, but express P-cadherin at cell-cell contacts, and Lgl2 and syntaxin-4 in the luminal plasma membrane domain.

INTRODUCTION

The choroid plexus (CP) resides inside the ventricle system of the brain, where it secretes the majority of the cerebrospinal fluid (9, 50). A very high transport rate is orchestrated by an asymmetric cellular distribution of specific solute transport proteins in the plasma membranes of the epithelial cells. The net transport is driven by the ubiquitously expressed Na⁺-K⁺-ATPase, which is expressed in the luminal (i.e., cerebrospinal fluid facing) membrane of choroid plexus epithelial cells (CPECs) as opposed to its typical basolateral localization (12, 38, 46). Thereby, CPECs belong to a small group of neuroepithelia where several key basolateral transporter proteins are expressed in the luminal/apical plasma membrane domains (18, 38, 45). Although the ultrastructural analyses of the CP epithelium unequivocally point to a classical polarization of the epithelial cells, the unusual localization of the Na⁺-K⁺-ATPase and the cytoskeletal arrangement still spur debate on the basic polarity of CPECs.

Epithelial cell polarity is established and maintained by an array of specific polarity proteins or apical and basolateral determinants (48). The apical-basolateral polarity of epithelial cells is initiated on the basis of interactions with the basement membrane and neighboring cells, and established as polarity proteins inside the cells translate these interactions/cues into a specific orientation (63). Thereby, the polarity proteins become organized and determine the organization of intracellular and plasma membrane components (5, 14, 49). The expression of cadherin types in the CPECs has been reported previously; however, uncertainty exists regarding which cadherin types are expressed because of species differences and antibody specificity issues (8, 24, 27, 36). This basic epithelial trait might impact the membrane accumulation properties of basic polarity proteins and other plasma membrane proteins such as solute transport proteins.

In mammalian cells, the apical plasma membrane domain is established and maintained by the coordinated actions of the Crumbs complex [protein crumbs homolog, protein associated with Lin-7 1 (Pals1) and InaD-like protein (Patj)] and the partitioning defective (Par) protein complexes [Par-3, Par-6, protein kinase C-ζ (PKCζ) and (Cdc42)], while the Scribble complex (protein scribble homolog), Lethal (2) giant larvae protein homolog (Lgl), and Disks large homolog 1 (Dlg1) establishes and maintains the basolateral membrane domain alongside Par-1 (16, 19, 21, 33). The cellular organization of two phosphatidylinositol phosphate lipids, phosphatidylinositol 4,5-bisphosphate (PIP₂), phosphatidylinositol 3,4,5-trisphosphate (PIP₃), and their converting enzymes also contribute to the establishment and maintenance of apical-basolateral polarity (11, 53). The phosphatidylinositol 4,5-bisphosphate 3-kinase (PI3K) keeps the basolateral membrane domain free of PIP₂ by converting it to PIP₃, while the phosphatase and tensin homolog (Pten) catalyzes the opposite reaction in the apical membrane domain. Both lipids and their inter-converting enzymes are important for identity and establishment of each of their membrane domains (11, 15, 37, 53).

In polarized epithelial cells, a given ion solute transport protein accumulates in a specific plasma membrane domain based on an interplay between direct selective membrane targeting, selective membrane stabilization, and transcytosis. Motifs in the amino acid chain and/or in the glycans in the plasma membrane proteins are recognized by the cellular sorting machinery in the Golgi apparatus, which forms targeted vesicles for trafficking. Vesicles targeting a given plasma membrane domain undergo site-selective tethering and insertion of membrane proteins into the plasma membrane. These processes are governed by a complex system of transport vesicles and recycling compartments and also involve a multitude of Rab proteins, the exocyst complex, and the soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) proteins (17, 55, 59, 61). SNARE proteins contribute significantly to the distinction between the different membrane domains in the cell, by their membrane-specific distribution of complementary sets of vesicle SNARE proteins (vSNARE) in the vesicle membranes and target SNARE proteins (tSNARE) in the target membranes. In polarized epithelia, two such tSNARE proteins are syntaxin-3 in the apical plasma membrane and syntaxin-4 in the basolateral plasma membrane (32, 52, 58). Furthermore, the basolateral accumulation of some plasma membrane proteins depends on basolateral recycling, where continued basolateral exocytosis of once internalized proteins depends on the epithelial specific clathrin adaptor AP-1B (13, 41).

Thus, multiple cellular components can influence the luminal membrane accumulation of the Na⁺-K⁺-ATPase in CPECs. We previously extended earlier studies on a preferential subluminal localization of ankyrin and the spectrin cytoskeleton in CPECs (8). However, a more exhaustive investigation of the basic cellular traits in CPECs is warranted to define the reason for the unusual luminal membrane accumulation of several key transport proteins. Here, we investigated the expression and cellular localization of such apical and basolateral determinants in the native choroid plexus tissue by a combination of mass spectrometry analysis of isolated CPECs and immunohistochemistry. We find that the epithelial cells are normally polarized with respect to all major polarity protein complexes and the membrane phospholipids PIP₂ and PIP₃. However, the canonical basolateral membrane tSNARE, syntaxin-4, is mainly localized to the luminal membrane domain along with an interacting vSNARE synaptosomal-associated protein 23 (Snap23). Finally, the CPECs lack the basolateral recycling endosome-specific clathrin adaptor protein AP-1B. Thus, we identified new targets for future investigations of the atypical membrane accumulation of plasma membrane proteins in CPECs, which could probably be relevant for studies of other neuroepithelia as well as in basic cellular polarity biology.

MATERIALS AND METHODS

Animals.

A total of 52 male c57/bl6jbom mice were used at age 9–14 wk (Taconic). All procedures conformed to Danish animal welfare regulations and were approved by the Danish Animal Experiments Inspectorate, Ministry of Food, Agriculture and Fisheries (j.n. 2012-15-2935-00004).

Electron microscopy.

Mice were perfusion fixed via the heart with 4% paraformaldehyde in a phosphate-buffered salt solution (PBS, in mM: 167.2 Na⁺, 150.0 Cl⁻, 7.2 $HP{O}_4^{2-}$, 2.8 H₂$P{O}_4^{-}$, with pH 7.4). Subsequently, the CP tissue was postfixed in 2% glutaraldehyde in 0.1 M cacodylate buffer. The tissue was incubated in 1% OsO₄ for 1 h. After washing in 0.1 M cacodylate and 0.05 M maleate buffer (pH 5.2), the tissue was incubated in 0.5% uranyl acetate in 0.05 M maleate buffer (pH 5.2) for 1 h, before dehydration in graded ethanol and propylene oxide. The tissue was infiltrated 1 h with epoxy (Epon 812, TAAB) and propylenoxide 1:1 and in pure epoxy for 2 days. Ultrathin sections were prepared using a Reichert ultramicrotome, mounted on 200 mesh nickel grids, and stained with uranyl acetate and lead citrate. Images were recorded in a FEI-Morgagni microscope operating at 80 kV by using a CCD camera (MegaViewIII, SIS).

Fluorescence-activated cell sorting of CPECs.

For each of four replicates, eight mice were euthanized under isoflurane anesthesia, and CP from all four ventricles of six mice were dissected and collected in 4°C HEPES-buffered salt solution (HBS, in mM: 145 Na⁺, 3.6 K⁺, 1.8 Ca²⁺, 0.8 Mg²⁺, 138.6 Cl⁻, 0.8 $\mathrm{S}{\mathrm{O}}_4^{2-}$, 2 $HP{O}_4^{2-}$, 10 HEPES, 5.5 glucose, pH 7.4). The pooled CP tissues were incubated in 50 µg/ml concanavalin A fluorescein (Vector) in HBS for 10 min at 37°C, digested in 2 μg/ml dispase (Invitrogen) and 2 μg/ml collagenase B (Roche) in calcium-free HBS 30 min at 37°C, and incubated in 1:1 mixture of TrypLE Select Enzyme (Thermo-Fisher) and cell culture trypsin/EGTA (Thermo-Fisher) supplemented with 1 mg/ml DNase (Sigma) for 10 min at 37°C. The cell preparation was inspected by microscopy, passed through a 50-µm filter, and propidium iodide was added before FACS for exclusion of dead cells. Cells were sorted into fluorescein-positive and fluorescein-negative samples by four-way purity sorting on a FACSAria III (BD Biosciences). Undigested CP tissue served as control in further analysis.

Protein mass spectrometry.

FACS-purified CPECs were homogenized in 8 M urea buffer with 2 M thiourea. The sample was reduced, alkylated, digested (trypsin, Promega), and desalted on C18 columns (Waters), as described before (57). The peptides were separated by high-pH RPLC using a Dionex Ultimate 3000 LC system (Thermo Scientific) with a ZORBAX Extended-C18 LC column (2.1 × 150 mm, 5 μm, Agilent). Buffer A (25 mM NH₄FA in 100% H₂O, pH 10) and buffer B (25 mM NH₄FA in 90% ACN, pH 10) were used for gradient separation. The gradient was 0–10% B (0–10 min), 10–35% B (10–50 min), and 35–80% B (50–64 min), with 32 fractions collected every 2 min. The 32 fractions were further pooled into eight by mixing equal-time-interval fractions, for example, fraction 1, 9, 17, and 25 were mixed together. The resulting eight fractions were lyophilized in a SpeedVac and then resuspended in 0.1% formic acid (FA) for LC-MS/MS analysis.

Analysis was performed by nano liquid-chromatography (nLC) (EASY-nLC 1000, Thermo Scientific) coupled to a mass spectrometer (Q Exactive, Thermo Fisher Scientific) through an EASY-Spray nano-electrospray ion source (Thermo Scientific). A pre-column (Acclaim PepMap 100, 75 µm × 2 cm, nanoviper fitting, C18, 3 µm, 100 Å, Thermo Scientific) and analytical column (EASY-Spray Column, PepMap, 75 µm × 15 cm, nanoviper fitting, C18, 3 µm, 100 Å, Thermo Scientific) were used to trap and separate peptides, respectively. For nLC separation, buffer A was 0.1% FA and buffer B was 95% ACN-0.1% FA. A 30-min gradient of 1% to 35% buffer B was used for peptide separation. Mass spectrometry constituted of full scans (m/z 300–1800) at a resolution of 70,000 (at m/z 200) followed by up to 10 data-dependent MS/MS scans at a resolution of 17,500. HCD collision energy was 28%. Dynamic exclusion of 30 s as well as rejection of precursor ions with charge state +1 and above +8 were employed.

MS data analysis.

MS raw files were searched against a mouse protein database (RefSeq database downloaded 13 Oct. 2014, containing 58,513 sequences) using both the SEQUEST algorithm (embedded in Proteome Discoverer 1.4, Thermo Scientific) and the MASCOT (version 2.5, linked to Proteome Discoverer 1.4 as well). Precursor and fragment mass tolerances were 10 ppm and 0.02 Da, respectively; tryptic peptides with at most two missed cleavage sites were considered.

NH₂-terminal acetylation and methionine oxidation were set as dynamic modifications, while cysteine carbamidomethylation was set as static modification. False discovery rate (FDR) was calculated using Percolator, and only rank 1 and high-confidence peptides (with a target FDR q-value below 0.01) were included in the final results.

Tissue fixation and immunohistochemistry.

Mice were perfusion fixed via the heart with 3% paraformaldehyde in PBS, as described previously (8). In brief: after fixation, the brain was removed, postfixed for 2 h, dehydrated in ethanol and xylene, and embedded in paraffin wax, enabling 2-μm sectioning using a rotary microtome (Leica). The sections were dewaxed and stepwise rehydrated, before epitopes were retrieved by boiling the sections in TEG buffer: 10 mM Tris buffer with 0.5 mM EGTA (pH 9), or in 10 mM citrate-buffer (pH 6). The epitopes were quenched with 50 mM NH₄Cl in PBS, and unspecific binding was blocked by washing with 1% BSA in PBS with 0.2% gelatin and 0.05% saponin. Sections were incubated overnight at 4°C with primary antibody diluted in 0.1% BSA in PBS with 0.3% Triton X-100. Primary antibodies are listed in Table 1, and positive control tissues included kidneys and small intestines (not shown). For fluorescence visualization of the primary antibodies, AlexaFluor 488- or 555-coupled donkey anti-rat, -goat, -rabbit, or -mouse secondary antibodies (Invitrogen) were used, and cell nuclei were visualized using Topro-3 counterstaining (Invitrogen). Sections were mounted with a coverslip in Glycergel anti-fade medium (DAKO) and analyzed using a Leica DMIRE2 inverted microscope with a TC5 SPZ confocal unit using × 63/1.32 numerical aperture HCX PI Apo oil objectives. Sections from at least three mice were analyzed for each antibody labeling.

Table 1.

Primary antibodies used in this study

Target	Antibody No.	Technique	Dilution	Host	Source
Acetylated tubulin	T6793	Paraffin–TEG	1:1,000	Mouse	Sigma-Aldrich
Cdc42	sc-34314	Floating	1:50	Goat	Santa Cruz Biotech.
Collagen type IV	0868124	Paraffin–TEG	1:50	Rabbit	MP Biomedicals
Dlg1	PA1-741	Paraffin–TEG/floating	1:25	Rabbit	Thermo Fisher
Patj	Anti-PATJ	Paraffin–TEG	1:500	Rabbit	Massey-Harroche, Le Bivic
Laminin (all forms)	ab7463	Paraffin–TEG	1:50	Rabbit	Abcam
Lgl1	1735	Paraffin–TEG	1:25	Mouse	Patrick Humbert
Lgl2	ab55423	Paraffin–citrate/floating	1:1,000	Mouse	Abcam
Na+-K+-ATPase α1	3B-0/56-0	Paraffin–TEG	1:5,000	Mouse	(25)
N-/N2-cadherin	363 003	Paraffin–TEG	1:1,000	Rabbit	SYnaptic SYstems
Par-1	Orb266676	Paraffin–TEG	1:100	Rabbit	Biorbyt
Par-3	07-330	Paraffin–TEG	1:100	Rabbit	Merck Millipore
Par-6	sc-166405	Floating	1:50	Mouse	Santa Cruz Biotech.
P-cadherin	PAB013Mu01	Paraffin–TEG	1:200	Mouse	Cloud Clone
Pten	138G6	Paraffin–TEG	1:50	Rabbit	Cell Signaling Tech.
PIP3	Z-P345b	Floating	1:100	Mouse	Echelon Biosciences
PIP2	Z-B045	Floating	1:100	Mouse	Echelon Biosciences
Pals1	17710-1-AP	Paraffin–TEG	1:100	Rabbit	Proteintech
Crumbs 3	crumbs 3A	Paraffin–TEG	1:200	Rat	Massey-Harroche, Le Bivic
PKCζ	sc216	Paraffin–TEG	1:500	Rabbit	Santa Cruz Biotech.
Scribble	sc-11048	Paraffin–TEG	1:50	Goat	Santa Cruz Biotech.
Snap23	111 202	Paraffin–TEG	1:500	Rabbit	SYnaptic SYstems
Syntaxin-3	sc-47437	Paraffin–TEG	1:100	Goat	Santa Cruz Biotech.
Syntaxin-4	ab101879	Paraffin–TEG/floating	1:50	Rabbit	Abcam
Syntaxin-4	110 043	Paraffin–TEG	1:500	Rabbit	SYnaptic SYstems
βIII-Spectrin	SPTBN2/1584	Paraffin–TEG	1:500	Mouse	Novus Biologicals

Several immunohistochemistry techniques were applied. Most antibodies were applied to sections of paraffin wax-embedded tissue, and epitopes were retrieved by microwave boiling in either TEG buffer or citrate buffer. Other antibodies were used for direct labeling of the uncut floating tissue.

Free-floating tissue immunohistochemistry.

Mice were perfusion fixed and tissues postfixed as described above. Choroid plexus epithelial tissues were dissected from the brain, while control tissues kidney and small intestines were sectioned on a Vibratome 1500 (Technical Products International) before staining. The performance of the antibodies was dependent on target-retrieval in either 10 mM citrate buffer (pH 6) or TEG buffer for 2 h at 60°C. Cells were permeabilized in 50 mM NH₄Cl-PBS with 0.2% saponin, and incubated overnight at 4°C with primary antibody diluted in 50 mM NH₄Cl-PBS with 0.05% saponin. Primary antibodies are listed in Table 1. Tissues were postfixed 30 min in 4% paraformaldehyde before fluorescence visualization of the primary antibodies as described above. Tissues were mounted on Cell-Tak (BD Biosciences) covered glass slides with a coverslip and Glycergel anti-fade medium (DAKO) and analyzed as described above.

RESULTS

Basic ultrastructural orientation of the CPECs and cell adhesion proteins.

Transmission electron micrographs of sections from Epon-embedded tissue show the basic features of the CPECs (Fig. 1A): a basement membrane facing the interstitial connective tissue, basolateral labyrinth, and junctional complexes close to the luminal membrane. Occasionally, cilia were also observed at the luminal membrane (not shown). Immunohistochemical analysis of thin paraffin sections confirmed the expression of the basement membrane protein collagen IV opposite the choroid plexus epithelial luminal membrane marker Na⁺-K⁺-ATPase α1 (Fig. 1B). The collagen IV staining seemed more pronounced in the capillary basement membrane than in the epithelial basement membrane. Laminin staining was equally strong in the endothelial and epithelial basement membranes (Fig. 1C). In the same micrograph, the luminal cilia are marked by acetylated tubulin. MS analysis of FACS-isolated CPECs confirmed the previously reported expression of ankyrin 3 and several spectrin forms, with the addition of βIII-spectrin, P-cadherin, and N2-cadherin (8) [Table 2; full MS analysis and database link are provided in a recent publication (10)]. FACS-isolated cell populations were negative for specific endothelial, fibroblast, glial, and neuronal markers by mass spectrometry (http://interpret.au.dk/). RT-PCR of FACS-isolated CPECs was applied to verify the molecular expression of βIII-spectrin, P-cadherin, and N2-cadherin and also indicated expression of N-cadherin mRNA (not shown). Figure 1D exemplifies the immunolocalization of βIII-spectrin to the luminal plasma membrane domain of mouse CPECs. P- and N-cadherin were both localized to the basolateral plasma membrane domain (Fig. 1, E and F). According to the manufacturer, the N-cadherin antibody might react with N2-cadherin as well.

Fig. 1.

CPEC ultrastructure and basic cell polarity. A: transmission electron micrograph showing the ultrastructure of an Epon embedded CPE cell. B: immunofluorescence double-labeling of mouse brain sections for collagen IV (red) in the basement membrane and Na⁺-K⁺-ATPase α1 (green) in the brush border. C: immunofluorescence double-labeling for laminin (red) in the basement membrane and acetylated tubulin (green, asterisk) in the motile cilia. D: immunofluorescence labeling for βIII-spectrin (green) mainly in the luminal membrane domain. E: similar labeling for the placental cell adhesion protein P-cadherin (green) especially corresponding to the lateral membranes and basal labyrinths. F: immunolabeling for the neuronal cell adhesion protein N-cadherin (green) also corresponding to the lateral membranes and basal labyrinths. Fluorescence micrographs were overlaid on the corresponding differential interference contrast (DIC) images. Nuclei were visualized by Topro-3 (blue). Arrows indicate the luminal plasma membrane, whereas arrowheads mark the basal plasma membrane. The images are representative for micrographs from at least 3 mice. VL, ventricle lumen; MV, microvilli; N, nucleus; JC, junctional complex; M, mitochondria; IS, lateral intercellular space; BL, basal labyrinth.

Table 2.

Cadherins, ankyrins, and spectrins identified by MS analysis of FACS-isolated CPECs

Accession No.	Description	Coverage, %	Unique Peptides	PSMs
56606025	Cadherin-12 (N2-cadherin)	1.13	1	1
45496816	Cadherin-3 (P-cadherin)	7.80	4	11
116256491	Ankyrin-3 (ankyrin G)	3.19	4	8
295054266	Spectrin α chain, nonerythrocytic 1 (αII)	76.46	1	1,025
568913212	Spectrin α chain, nonerythrocytic 1 (αII)	76.21	0	1,038
568913216	Spectrin α chain, nonerythrocytic 1 (αII)	76.58	1	1,035
55926127	Spectrin β chain, brain 2 (βIII)	14.41	14	76
84490394	Spectrin β chain, erythrocytic (βI)	11.59	16	64
117938332	Spectrin β chain, nonerythrocytic 1 (βII)	70.25	135	809

The numbers represent pooled data from 4 experiments The accession numbers of identified transporters are shown along with the protein names (description and common abbreviation), the mean sequence coverages, the number of unique peptides identified, and the total number of identified peptide sequences. PSMs, peptide spectrum matches.

Crumbs complex is localized to the luminal membrane domain in CPECs.

The Crumbs complex consists of the transmembrane protein Crumbs and the cytosolic associated proteins Pals1 and Patj (3). These proteins are considered to be ubiquitously expressed in polarized cells, and Crumbs 2, Crumbs 3, Pals1, and Patj were all detected by MS analysis of FACS-isolated mouse CPECs (Table 3). Immunohistochemical analysis expectedly localized Crumbs 3 (Fig. 2A), Pals1 (Fig. 2B), as well as Patj (Fig. 2C) to the luminal plasma membrane domain of CPECs. The Patj labeling was especially pronounced at the tight junction area. Similar staining patterns were obtained in renal medullary collecting ducts from the same mice (not shown).

Table 3.

Basic polarity proteins identified by MS analysis of FACS-isolated CPECs

Accession No.	Description	Coverage, %	Unique Peptides	PSMs
254675207	Protein crumbs homolog 2 (Crumbs 2)	4.13	4	24
29244038	Protein crumbs homolog 3 (Crumbs 3)	15.04	1	4
9625023	MAGUK p55 subfamily member 5 (Pals1)	37.48	20	66
568928300	InaD-like protein (Patj)	20.49	22	46
122937357	Protein kinase MARK2 (Par-1b)	23.15	9	50
61888842	Partitioning defective 3 (Par-3)	1.98	1	2
253314520	Partitioning defective 6 β (Par-6)	9.70	3	9
568962235	PIP(4,5)3-kinase β (PI3K)	0.76	1	1
568990942	Protein scribble (Scrib)	8.44	10	35
226874865	Lethal (2) giant larvae protein 1 (Lgl1)	11.00	6	15
568973450	Lethal (2) giant larvae protein 2 (Lgl2)	28.39	18	43
6753364	Cell division control protein 42 (Cdc42)	43.98	7	86
356995919	Disks large homolog 1 (Dlg1)	37.27	24	91
295293129	Disks large homolog 3 (Dlg3)	42.62	12	37

The numbers represent pooled data from 4 experiments. The accession numbers of identified transporters are shown along with the protein names (description and common abbreviation), the mean sequence coverages, the number of unique peptides identified, and the total number of identified peptide sequences. PSMs, peptide spectrum matches.

Fig. 2.

Localization of the Crumbs complex in CPECs. Immunofluorescence labeling of mouse brain sections for Crumbs-3 (A), Pals1 (B), and Patj (C). The left panels are lower-magnification overviews, while the right panels are higher-magnification micrographs of the same or similar tissue sections. All three proteins were localized to the luminal membrane domain of the CPECs, but Patj seemed most abundant in the tight junction area. Fluorescence micrographs were overlaid on the corresponding DIC images. Nuclei were visualized by Topro-3 (blue). Arrows indicate the luminal plasma membrane, whereas arrowheads mark the basal plasma membrane. The images are representative for micrographs from at least 3 mice.

Par complex is localized to the luminal membrane domain in CPECs.

The Par complex consists of the cytosolic proteins Par-3, Par-6, PKCζ, and Cdc42 (2). The full complex normally localizes in close proximity to the junctional complexes, while Par-6 and PKCζ are localized to the entire luminal membrane domain. All four proteins were identified by MS analysis of FACS-purified CPECs (Table 3). The immunohistochemical analysis localized Par-3 predominantly to the area of the junctional complexes (Fig. 3A), whereas PKCζ was distributed across the entire luminal membrane domain in CPECs (Fig. 3B). Immunostaining for Cdc42 and Par-6 was only positive when the free-floating tissue protocol was applied. With this technique, Cdc42 was also immunolocalized to the luminal membrane domain (Fig. 3C), and similar staining was less convincingly obtained for Par-6 (Fig. 3D). Similar staining patterns for Par-6, Par-3, and PKCζ were obtained in renal medullary collecting ducts from the same mice (not shown).

Fig. 3.

Localization of the components of the Par complex in CPECs. A: immunofluorescence labeling of mouse brain sections for Par-3. The left panel is a lower-magnification overview, while the right panel is a higher-magnification micrograph from another mouse brain section. Immunolabeling was confined to the tight junction area and to a lesser degree to the luminal membrane domain. B: similar immunolabeling for PKCζ in CPECs. The left panel is a lower-magnification overview, and the right panel is a higher-magnification micrograph. Immunolabeling was pronounced corresponding to the luminal membrane domain. Fluorescence micrographs are overlaid on the corresponding DIC images. C: immunolabeling of free-floating CP for Cdc42. The left panel is a lower-magnification overview, while the right panel a higher-magnification micrograph. Fluorescence signal was observed in the luminal membrane domain. D: immunostaining of free-floating CP for Par-6 (left) and the corresponding DIC image (right). Labeling was observed corresponding to the luminal membrane domain in some cells. Nuclei were visualized by Topro-3 (blue). Arrows indicate the luminal plasma membrane, whereas arrowheads mark the basal plasma membrane. The images are representative for micrographs from at least 3 mice.

Scribble complex and Par-1 are localized to the basolateral membrane domain in CPECs.

The Scribble complex is involved in maintenance of the basolateral membrane and consists of the cytosolic proteins Scribble, Dlg1, and Lgl isoforms 1 and/or 2. All of these proteins were detected by MS analysis of CPECs along with another basolateral determinant, Par-1 (Table 3). Scribble was immunolocalized to the basolateral membrane domain in CPECs (Fig. 4A) along with Par-1 (Fig. 4B), and Dlg1 (Fig. 4C). Dlg1 staining seemed to be distributed more diffusely in the basolateral domain in paraffin sections (left), but was more distinct basolaterally when stained in free-floating tissue. Both scribble and Dlg1 were also localized to the basolateral membrane domain in renal medullary collecting ducts (not shown). Lgl1 was stably immunolocalized to the basolateral membrane domain in CPECs among the tested mice (Fig. 5A), while Lgl2 staining was more variable. The most frequent observation was a luminal membrane domain labeling (Fig. 5B, left). However, even littermates processed in parallel showed either unpolarized distribution beneath the entire plasma membrane or predominantly basolateral membrane labeling (not shown). Lgl2 labeling was observed exclusively at the luminal membrane domain by free-floating staining (Fig. 5B, right).

Fig. 4.

Localization of basolateral polarity determinants in CPECs. A: immunofluorescence labeling of mouse brain sections for Scribble. The left panel is a lower-magnification overview, while the right panel is a higher-magnification micrograph. Scribble immunolabeling was confined to the basolateral membrane domain of CPECs. B: similar labeling for Par-1 in CPECs. The left panel is a lower-magnification overview, and the right panel is a higher-magnification micrograph. Par-1 immunolabeling was also mainly observed in the basolateral membrane domain. C: immunolabeling of Choroid plexus epithelial tissue for Dlg1. The left panel shows labeling pattern in a mouse brain section, whereas the right panel represents immunolabeling of free-floating CP. The Dlg1 labeling was pronounced in the basolateral membrane domain, especially by the free-floating CP staining. Except for images of free-floating tissue staining, the fluorescence micrographs are overlaid on the corresponding DIC images. Nuclei are visualized by Topro-3 (blue). Arrows indicate the luminal plasma membrane, whereas arrowheads mark the basal plasma membrane. The images are representative for micrographs from at least 3 mice.

Fig. 5.

Localization of Lgl forms in CPECs. A: immunofluorescence labeling of mouse brain sections for Lgl1. The left panel is a lower-magnification overview, while the right panel represents a higher-magnification micrograph. Lgl1 immunolabeling was most prominent corresponding to the basolateral membrane domain of CPECs. B: examples of immunolabeling of CPE tissue for Lgl2. The left panel shows labeling pattern in a mouse brain section, whereas the right panel represents immunolabeling of free-floating CP tissue. The Lgl2 labeling was pronounced in the luminal membrane domain in these example micrographs. Except for images from free-floating tissue staining, the fluorescence micrographs are overlaid on the corresponding DIC images. Nuclei are visualized by Topro-3 (blue). Arrows indicate the luminal plasma membrane, whereas arrowheads mark the basal plasma membrane. The images are representative for micrographs from at least 3 mice.

PIP₂ and PIP₃ are conventionally polarized in CPECs.

The cellular distribution of the plasma membrane lipids PIP₃ and PIP₂, and their converting enzymes, also contribute to the establishment and maintenance of epithelial apical-basolateral polarity (11, 53). The basolateral exclusion of PIP₂ and accumulation of PIP₃ are governed by PI3K. Conversely, luminal exclusion of PIP₃ and PIP₂ accumulation is accomplished by the action of Pten (53). PIP₂ immunolabeling was most predominant in the luminal membrane domain by free-floating staining (Fig. 6A), whereas PIP₃ was immunolocalized to the basolateral membrane domain by the same technique (Fig. 6B). PI3K was detected in CPECs by MS analysis (Table 3), but not by immunolabeling. Pten seemed to localize primarily to the luminal membrane domain of CPECs by immunohistochemistry on paraffin sections (Fig. 6C).

Fig. 6.

Localization of polarized plasma membrane phospholipids in CPECs. A: immunofluorescence labeling of free-floating choroid plexus for the membrane lipid PIP₂. The left panel is a lower-magnification overview, while the right panel is a higher-magnification micrograph. Immunostaining mainly accumulated corresponding to the luminal membrane domain. B: immunofluorescence labeling of free-floating CP for the membrane lipid PIP₃. The left panel is a lower-magnification overview, and the right panel is a high-magnification micrograph. The immunoreactivity was confined to the basolateral membrane domain. C: immunofluorescence labeling of mouse brain sections for Pten. The left panel is a lower-magnification overview, while the right panel exemplifies a higher-magnification micrograph. Pten immunoreactivity was most pronounced in the luminal membrane domain. The fluorescence micrographs of brain sections were overlaid on the corresponding DIC images. Nuclei are visualized by Topro-3 (blue). Arrows indicate the luminal plasma membrane, whereas arrowheads mark the basal plasma membrane. The images are representative for micrographs from at least 3 mice.

Basolateral target SNARE syntaxin-4 is localized luminally in CPECs.

As mentioned in the introduction, polarized insertion of new plasma membrane proteins relies on controlled vesicle sorting, trafficking, tethering, and plasma membrane fusion governed by vesicle and target SNARE proteins, the exocyst, and Rab proteins. The exocyst complex components 1–8, the apical target SNARE syntaxin-3, the basolateral target SNARE syntaxin-4, as well as the interacting vesicle SNARE Snap23 were all identified in MS analysis of CPECs (Table 4). Several well-characterized vesicle ras-related proteins (Rab proteins) relating to plasma membrane turnover were also identified by MS analysis (Table 5 and Refs. 14, 22). Among the clathrin adaptor proteins, we did not detect the epithelia-specific basolateral recycling clathrin adaptor protein AP-1B subunit AP1µ2 (or µ1B) by MS analysis (Table 6). RT-PCR analysis of FACS-isolated mouse CPECs and human choroid plexus mRNA confirmed the lack of molecular expression of AP1µ2 (not shown). As previously observed, syntaxin-3 was localized to the luminal membrane domain in CPECs (Fig. 7A). Multiple efforts were made to immunolabel for syntaxin-4. In all three cases of positive staining, the applied antibodies produced either exclusive or predominant luminal membrane domain labeling in CPECs (Fig. 7, B and C). In renal medullary collecting ducts, syntaxin-3 was localized to the luminal membrane, while syntaxin-4 was basolateral (not shown). Snap23 immunoreactivity was also observed consistently in the luminal membrane domain of CPECs (Fig. 7D).

Table 4.

Exocyst components and SNAREs identified by MS analysis of FACS-isolated CPECs

Accession No.	Description	Coverage, %	Unique Peptides	PSMs
576796187	Exocyst complex component 1	12.97	7	13
21313438	Exocyst complex component 2	7.90	7	16
568983232	Exocyst complex component 3	15.98	7	11
83921574	Exocyst complex component 4	18.56	14	45
46402177	Exocyst complex component 5	7.77	4	8
62526126	Exocyst complex component 6	4.48	2	9
568943120	Exocyst complex component 6B	7.05	3	8
247269443	Exocyst complex component 7	17.15	7	19
37674218	Exocyst complex component 8	12.01	1	8
2670778802	Syntaxin-3	18.12	5	22
6678177	Syntaxin-4	36.91	7	23
295317330	Synaptosomal-associated protein 23 (Snap23)	57.14	9	37

The numbers represent pooled data from 4 experiments. The accession numbers of identified transporters are shown along with the protein names (description and common abbreviation), the mean sequence coverages, the number of unique peptides identified, and the total number of identified peptide sequences. PSMs, peptide spectrum matches.

Table 5.

Vesicle transport Rab proteins identified by MS analysis of FACS-isolated CPECs

Accession No.	Description	Coverage, %	Unique Peptides	PSMs
6679593	Ras-related protein Rab-3A	28.64	4	40
171184402	Ras-related protein Rab-4A	26.15	2	19
568945132	Ras-related protein Rab-4B	22.58	2	19
13385374	Ras-related protein Rab-5A	42.33	5	44
28916687	Ras-related protein Rab-5B	45.12	5	52
113866024	Ras-related protein Rab-5C	62.04	8	56
38372905	Ras-related protein Rab-8A	57.97	4	77
27734154	Ras-related protein Rab-8B	47.34	3	61
7710086	Ras-related protein Rab-10	37.50	7	85
6679583	Ras-related protein Rab-11B	62.39	14	103
106507168	Ras-related protein Rab-12	16.46	2	18
21311975	Ras-related protein Rab-13	14.85	1	34
18390323	Ras-related protein Rab-14	80.00	17	166
45593126	Ras-related protein Rab-20	20.60	3	6
33859751	Ras-related protein Rab-21	58.11	14	78
148747177	Ras-related protein Rab-22A	30.41	4	11
37718983	Ras-related protein Rab-35	31.84	3	55

The numbers represent pooled data from 4 experiments. The accession numbers of identified transporters are shown along with the protein names (description and common abbreviation), the mean sequence coverages, the number of unique peptides identified, and the total number of identified peptide sequences. PSMs, peptide spectrum matches.

Table 6.

Clathrin adaptor proteins identified by MS analysis of FACS-isolated CPECs

Accession No.	Description	Coverage, %	Unique Peptides	PSMs
88853578	AP-1 complex subunit β-1	64.05	31	397
667751602	AP-1 complex subunit γ-1	42.46	27	110
160707961	AP-1 complex subunit γ-like 2	9.99	6	16
6671557	AP-1 complex subunit μ-1	64.07	24	107
40254484	AP-1 complex subunit σ-2	32.50	1	20
569011644	AP-1 complex subunit σ-2	34.64	1	20
116256510	AP-2 complex subunit α-1	50.16	33	226
163644277	AP-2 complex subunit α-2	46.91	28	152
21313640	AP-2 complex subunit β	52.72	26	338
6753074	AP-2 complex subunit μ	59.77	24	125
161086984	AP-2 complex subunit σ	26.06	4	16
163310776	AP-3 complex subunit β-1	24.52	21	66
6671565	AP-3 complex subunit δ-1	18.02	17	57
568988254	AP-3 complex subunit μ-1	40.43	9	42
170763483	AP-3 complex subunit μ-2	19.14	4	27
6753078	AP-3 complex subunit σ-1	19.69	3	9
163644298	AP-4 complex subunit μ-1	3.34	1	1
82546849	AP-5 complex subunit ζ-1	4.09	2	2

The numbers represent pooled data from 4 experiments. The accession numbers of identified transporters are shown along with the protein names (description and common abbreviation), the mean sequence coverages, the number of unique peptides identified, and the total number of identified peptide sequences. PSMs, peptide spectrum matches.

Fig. 7.

Localization of SNARE proteins in CPECs. A: immunofluorescence labeling of mouse brain sections for the luminal target SNARE syntaxin-3. The left panel is a lower-magnification overview, while the right panel is a higher-magnification micrograph. Syntaxin-3 immunolabeling was confined to the luminal membrane domain. B: immunofluorescence labeling for the basolateral target SNARE syntaxin-4 (Abcam). The left panel shows the labeling pattern in a mouse brain section, whereas the right panel represents immunolabeling of free-floating CP tissue. Syntaxin-4 immunolabeling was mainly confined to the luminal membrane domain. C: immunofluorescence labeling using a different anti-syntaxin-4 antibody (SySy). The left panel is a lower-magnification overview, while the right panel exemplifies a higher-magnification micrograph. The immunoreactivity was most pronounced corresponding to the luminal plasma membrane domain, although the expected basolateral staining was also observed. D: immunolabeling of mouse brain sections for the vesicle SNARE Snap23. The left panel is a lower-magnification overview, while the right panel is a higher-magnification micrograph. Snap23 immunoreactivity was confined to the luminal plasma membrane domain. Except for the image from free-floating tissue staining, the fluorescence micrographs are overlaid on the corresponding DIC images. Nuclei are visualized by Topro-3 (blue). Arrows indicate the luminal plasma membrane, whereas arrowheads mark the basal plasma membrane. The images are representative for micrographs from at least 3 mice.

DISCUSSION

In this study, we describe the expression and subcellular localization of the basic cell polarity machinery in native CPECs. This is to define the target molecules for future investigations of the atypical distribution of certain plasma membrane proteins in this tissue as well as other neuroepithelia. An immense body of knowledge exists on the fundamental polarity proteins and the interactions that generate and maintain epithelial cell polarization. The vast majority of information arises from studies of Caenorhabditis elegans, Drosophila melanogaster, and mammalian cell cultures (3, 54). Fewer studies exist on basic polarity proteins and membrane lipids in native mammalian tissues. The renal proximal tubules and the retinal pigment epithelium are prominent examples of well-studied native epithelia (10a, 28, 40, 43). Here, we report the successful transfer of knowledge on basic cell polarity to the choroid plexus epithelium, a tissue studied mainly as a secretory model cell-type but also for its unusual membrane accumulation of key ion transport proteins (1, 36).

From the ultrastructural analysis presented here and in the literature, there is little doubt that CPECs reside on a basement membrane facing the interstitium (56). Our combination of ultrastructural and immunofluorescence analyses validated the localization of the basement membrane and the junctional complex at the lateral membrane. Microvilli and cilia were observed at the luminal surface as described previously. We have previously immunolocalized ankyrin and several spectrin proteins in CPECs (8). In our MS analysis, it seemed that an additional spectrin type was expressed in CPECs, βIII-spectrin. Thus, it seems that the main proportion of CPEC spectrin cytoskeleton consists of αII-spectrin in combination with βI-, βII-, and βIII-spectrin beneath the luminal membranes of these cells. The antibody previously applied to localize E-cadherin to the CPECs turns out to cross-react with P-cadherin, according to the manufacturer (8). Therefore, we revisited the cadherin expression in these cells and find molecular evidence for P-cadherin and less convincingly N- and/or N2-cadherin expression. Here, P- and N-cadherin were both localized to the lateral membrane domain including the basal labyrinth where adjacent cells develop contacts. Consistent with some previous reports (24, 27, 36), expression of E-cadherin in CPECs was neither detected by MS analysis nor RT-PCR, which leaves the basolateral P-cadherin as the most probable cell-cell adhesion protein in this tissue. Interestingly, B-cadherin is the bird homologue of mammalian P-cadherin and was identified in an early study of choroid plexus cell polarity in chicken (36). When overexpressed in fibroblasts, E-cadherin recruited both non-erythrocyte spectrin (αII) and the Na⁺-K⁺-ATPase to the cell-cell contacts, whereas B-cadherin overexpression failed to recruit these proteins to cell-cell contacts. Thus, the identification of P-cadherin as the major cadherin in CPECs might constitute a basic trait needed for atypical Na⁺-K⁺-ATPase localization. Indeed, this notion is supported by loss in the polarized Na⁺-K⁺-ATPase distribution when the E-/P-cadherin ratio was elevated in bovine retinal pigment epithelium (6). Thus, the CPECs display typical epithelial specializations of the basal, lateral, and luminal surfaces and adhere to their neighboring cells, but the type of cell-cell contacts may be one permissive factor for atypical membrane protein distribution in these cells.

As earlier introduced, the Crumbs complex is important for the establishment and maintenance of the apical membrane domain (3). We consistently found the three components of the complex at the luminal membrane domain of the CPECs, and here it probably serves similar functions as in other epithelia. Interestingly, the Crumbs complex has a synergistic collaboration with the luminal Par complex in apical-basolateral polarity (20, 29). Therefore, it was not surprising that Par-3, PKCζ, Cdc42, and less convincingly Par-6, were localized at the same site as the Crumbs complex in CPECs. The performance of antibodies usually varies with the applied protocols. For the Cdc42 and Par-6 antibodies, the paraffin embedding, de/rehydration, and/or boiling seemed to prevent immunoreactivity in the tissue, which was better preserved in the more native state of the free-floating staining protocol. The Scribble complex is important for establishment and maintenance of the basolateral membrane domain, and we found Scribble, Dlg1, Lgl1, as well as Par-1 expressed in the basolateral membrane domain of CPECs. Lgl2 was surprisingly immunolocalized mainly to the luminal membrane domain in the cells. The variable staining pattern could of course be caused by variations among mice in the PKCζ-mediated phosphorylation status of Lgl2, which governs the proteins dissociation from the apical pole and subsequent accumulation at the basal pole of the cell (44). However, a more simple explanation is a variable performance of the antibodies depending on technical issues including the efficiency of the tissue fixation. The MS analysis indicates phosphorylation of the relevant amino acid residues of Lgl2 in favor of the first option (not shown), but the identified phospho-peptides are shared between Lgl1 and Lgl2, leaving the observation inconclusive. Thus, the Lgl2 localization revealed by the free-floating staining technique may therefore be the strongest indication of luminal membrane domain accumulation of Lgl2 in CPECs. This localization can have some consequences for the cellular fate of syntaxin-4 as detailed below.

Very early events in establishing cell polarity are the separate accumulation of two phosphatidylinositol-phosphates in the plasma membrane lipid bilayer, with enrichment of PIP₂ in the luminal membrane and PIP₃ in the basolateral membrane (15, 26, 37). This polarization into two domains can be observed even in nonepithelial cells and persists even under cell division and migration (7, 42). As mentioned above, the exclusion of PIP₂ from the basolateral membrane is mediated by the action of the soluble PI3K, which transforms PIP₂ to PIP₃ and resides in the basolateral domain of epithelial cells (15, 53). Conversely, the exclusion of PIP₃ from the luminal membrane is governed by the phosphatase Pten, which converts PIP₃ to PIP₂ (62). Like the basic apical-basolateral polarity proteins, the system of asymmetrical distribution of the phosphatidylinositol derivatives PIP₂ and PIP₃ seems to be conserved in CPECs. Thus, the Crumbs, Par, and Scribble complexes as well as PIP₂ and PIP₃ all seem to be localized to the expected cellular domains in CPECs, and it is reasonable to conclude that CPECs are equipped with the same conventional machinery for basic apical-basolateral polarity as other epithelia (Fig. 8). Only Lgl2 seems to deviate from this general pattern.

Fig. 8.

Schematic representation of the localization of basic luminal and basolateral membrane domain markers in CPECs. The cell-cell adhesive cadherins, P-cadherin, and N- and/or N2-cadherin (P-cad and N-cad, respectively) localize to the lateral membrane domains including the basal labyrinths. The Par complex (including Par-3, PKCζ, Cdc42, and Par-6) and the Crumbs complex (including Crb3, Patj, and Pals1) are both localized in the luminal membrane domain and/or tight junction area. Par-1 is like the Scribble complex (Scrib, Lgl1, and Dlg1) localized towards the basolateral pole of the cells, whereas Lgl2 is mainly observed in the luminal membrane domain. The polarized plasma membrane phospholipids PIP₂ and PIP₃ are expressed in the luminal and basolateral membrane domains, respectively, and the phosphatase involved in converting PIP₃ to PIP₂, Pten, is localized luminally as PIP₂. The plasma membrane target SNAREs syntaxin-3 (Stx3) and syntaxin-4 (Stx4) are both mainly localized to the luminal membrane domain, along with the vesicle SNARE Snap23. The luminal expression of Lgl2 and syntaxin-4 with the interacting Snap23 is an unanticipated result.

The polarized delivery of a membrane protein from intracellular cargo vesicles involves the recruitment of the exocyst complex to vesicles and the fusion of vesicles with specific membrane domains involving a host of Rab proteins (22, 61). We detected all the exocyst complex components and several Rab proteins involved in vesicle trafficking to and from the plasma membrane by MS analysis of CPECs. Nevertheless, we were unsuccessful in immunolabeling CPECs for the selected proteins in these groups, and we did not pursue the subcellular localization further and conclude that these components in cellular vesicle trafficking and membrane fusion are expressed in CPECs as expected. Syntaxins comprise a large group of SNARE proteins, which provide specificity to the fusion of cytosolic transport vesicles to organelles or plasma membranes (23). Syntaxin-3 is described as a luminal membrane target SNARE, whereas syntaxin-4 is the only known basolateral target SNARE in polarized epithelial cells (32). We previously localized the target SNARE protein syntaxin-3 to the luminal membrane of CPECs (8). We have here validated this finding by applying other syntaxin-3-specific antibodies. Syntaxin-4 was also immunolocalized in CPECs using different antibodies. Two of these produced only luminal membrane domain labeling, whereas the third stained the basal labyrinth area in addition to the luminal membrane domain. Validity to these findings was also provided by the conventional distribution of syntaxin-3 and -4 in renal inner medullary collecting ducts (IMCDs) of the same tissue slide. The IMCDs were chosen despite discrepancy regarding collecting duct syntaxin-3 and -4 exists in the literature. Syntaxin-4 was first described as an apical protein in IMCD (principal cells), while syntaxin-3 was restricted to the basolateral membrane of intercalated cells (34, 35). SDS target retrieval expanded the renal syntaxin-3 expression pattern to include virtually all basolateral membrane domains along the nephron and collecting ducts (4). However, the same year Weimbs and coworkers reported what we regard as the most convincing study of renal expression of these syntaxins (31). In this study, syntaxin-3 was localized to the apical membranes and syntaxin-4 to the basolateral membranes, similar to what was already reported for MDCK cells (31, 32). It is tempting to speculate that the lack of E-cadherin and the clathrin adaptor protein AP-1B involved in basolateral recycling is implicated in the luminal accumulation of Lgl2 and syntaxin-4. Indeed, Lgl has been shown to interact with syntaxin-4 (39) and AP-1B deficiency has been linked directly to mislocation of syntaxin-4 in MDCK cells (47). This simple scheme for mislocalization of the basolateral tSNARE and the potential following accumulation of certain basolateral membrane proteins in the luminal plasma membrane is, however, contradicted by observations from studies on renal proximal tubules. Like CPECs and retinal pigment epithelial cells, renal proximal tubular cells are devoid of AP-1B but express syntaxin-4 as well as the Na⁺-K⁺-ATPase in the basolateral membrane domain (10a, 60). Interestingly, the renal proximal tubule cells express E-cadherin as well as N-cadherin (30), but not P-cadherin.

In conclusion, CPECs exhibit many of the conventional characteristics regarding basic cell polarity. The novel observations deviating from other epithelia are 1) the expression of P-cadherin and probably N- and/or N2-cadherin instead of E-cadherin, 2) the localization of Lgl2 in the luminal membrane domain, 3) the lack of AP-1B expression, and 4) the luminal localization of syntaxin-4. Future studies are warranted to explore which of these deviations are involved in the atypical localization of certain basolateral plasma membrane proteins such as the Na⁺-K⁺-ATPase in the luminal membrane of CPECs and other neuroepithelia.

GRANTS

This study was supported by Aarhus University Research Foundation (AUFF) AU-Ideas, the InterPrET (Interactions of Proteins in Epithelial Transport) Center, and the Danish Council for Independent Research–Medical Sciences.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

I.B.C., H.H.D., and J.P. conceived and designed research; I.B.C., E.N.M., and J.P. performed experiments; I.B.C., E.N.M., H.H.D., and J.P. analyzed data; I.B.C., E.N.M., H.H.D., and J.P. interpreted results of experiments; I.B.C., E.N.M., and J.P. prepared figures; I.B.C. and J.P. drafted manuscript; I.B.C., E.N.M., H.H.D., and J.P. edited and revised manuscript; I.B.C., E.N.M., H.H.D., and J.P. approved final version of manuscript.