Biomedical Research (Tokyo) 37 (2) 127–139, 2016

Comprehensive proteome analysis of brush border membrane fraction of ileum of knockdown mice

1 2 2 2 1 1, 3 Saori YOSHIDA , Toshiyuki FUKUTOMI , Toru KIMURA , Hiroyuki SAKURAI , Ryo HATANO , Hiroto YAMAMOTO , 3 3 3 1 Ken-ichi MUKAISHO , Takanori HATTORI , Hiroyuki SUGIHARA , and Shinji ASANO 1 Department of Molecular Physiology, College of Pharmaceutical Sciences, Ritsumeikan University, 1-1-1 Noji-Higashi, Kusatsu-City, Shiga 525-8577, Japan; 2 Department of Pharmacology and Toxicology, Kyorin University School of Medicine, 6-20-2 Shinkawa, Mitaka, Tokyo, 181-8611, Japan; and 3 Department of Pathology, Shiga University of Medical Sciences, Seta Tsukinowa-cho, Otsu-City, Shiga 520-2192, Japan (Received 21 January 2016; and accepted 12 February 2016)

ABSTRACT Ezrin is an binding which cross-links membrane with directly or indirectly via PDZ domain-containing scaffold proteins. It is mainly expressed at the brush bor- der membrane (BBM) of gastrointestinal tracts, and is involved in the construction of microvilli structure and the functional expression of complexes at the cell surface. To pre- cisely study the roles of ezrin on the expression of membrane proteins at the cell surface, here we prepared the BBM fractions of ileums from the wild-type and ezrin-knockdown (Vil2kd/kd) mice, analyzed them by mass spectrometry, and compared their proteomic patterns. Totally 313 proteins were identified in the BBM fractions. Several transport proteins, cytoskeleton-associated proteins, and trafficking proteins were up- or down-regulated in the BBM fraction of the ileum in the Vil2kd/kd mice. Among them, the expressions of i) Na+/H+ exchanger regulatory factor 1 (a PDZ do- main-containing scaffold protein), ii) sodium monocarboxylate transporter 1, which contains a PDZ domain-binding motif at their carboxy-terminal, and iii) chloride intracellular channel protein 5 were down-regulated at the BBM fraction of the ileum in the Vil2kd/kd mice, suggesting that ez- rin is involved in their expression in the BBM.

The ERM proteins (Ezrin/Radixin/Moesin) are actin small intestine and renal proximal tubules, and regu- binding proteins which cross-link membrane pro- lates the functional expression of membrane trans- teins with cytoskeleton directly or indirectly via the porters at the epithelial cell surface and construction PSD-95/Discs-large/ZO-1 (PDZ) domain-containing of microvilli structure (16). scaffold proteins such as Na+/H+ exchanger regula- Functional roles of ezrin in vivo had been studied tory factors (NHERF) 1 or 2 (6, 38). They regulate using knockout (Vil2−/−) mice. The Vil2−/− mice did scaffolding, forming, and fixing functional protein not survive past weaning, however, the small intes- complexes on the plasma membrane (6, 14). Among tine of neonatal Vil2−/− mouse retained microvilli the ERM proteins, ezrin is predominantly expressed structure and epithelial polarity, suggesting that ez- on the brush border membrane (BBM) of stomach, rin is not required for the formation of brush border microvilli nor for the establishment or maintenance of epithelial polarity (33). On the other hand, the Address correspondence to: Shinji Asano, Ph.D. Department of Molecular Physiology, College of Phar- same group reported that ezrin is important for in- maceutical Sciences, Ritsumeikan University, 1-1-1 Noji- testinal development and homeostasis especially in Higashi, Kusatsu-City, Shiga 525-8577, Japan villus morphogenesis and maintenance by use of Tel: +81-77-561-5908, Fax: +81-77-561-5908 conditional knockout mice of Vil2 (7). Inde- E-mail: [email protected] pendently, Tamura et al. (37) prepared ezrin knock- 128 S. Yoshida et al. down (Vil2kd/kd) mice in which expression level of MATERIALS AND METHODS ezrin was decreased to <5% compared with the wild-type mice. The Vil2kd/kd mice exhibited growth Generation of Vil2kd/kd mice. Vil2kd/kd mice were kind retardation and high mortality up to their weaning gifts from Professor Tsukita of the Graduate School period. The grown-up Vil2kd/kd mice suffered from of Frontier Biosciences, Osaka University. Up to the achlorhydria because of impairment of secretagogue- weaning period, mortality of the mice was very high, stimulated membrane fusion of tubulovesicles with as reported elsewhere (37). However, technical im- apical plasma membranes in parietal cells, with their provement of our handling and feeding methods structure of gastric glands being changed (37, 43). successfully reduced mortality at 50 days after birth On the other hand, the mice contained apparently from 93% to 80%, which enabled us to study the normal microvilli structure in the small intestine un- phenotypes of adult mutant mice (43). Genotyping der the electron microscopy (37). Recently, we re- of mice was performed by PCR of mouse tail ge- ported that apical cell surface expression of a sodium nomic DNA, by using a combination of primers phosphate co-transporter, Npt2a was impaired in the specific for the wild-type allele and for the targeted renal proximal tubules of Vil2kd/kd mice although the allele. The forward and reverse primers for the wild- tubule retained normal microvilli structure under the type allele were 5’-GTGTGGCACTCTGCCTTC electron microscopy. The impairment of cell surface AAG-3’ and 5’-CATGGTGCCACACAGGACTC-3’, expression of Npt2a resulted in hypophosphatemia respectively. The reverse primer for the targeted al- and loss of phosphate in urine (16). We also newly lele was 5’-AGCGGATCTCAAACTCTCCTC-3’. reported that apical cell surface expression and pro- The mRNA expression levels of ezrin in the ileum tein kinase A-mediated activation of cystic fibrosis segment of Vil2kd/kd mice was 2% compared with transmembrane conductance regulator (CFTR) to- that of wild-type mice (data not shown). gether with an anion exchanger 2 and 1 were impaired in the bile ductal cholangiocytes of Antibodies. Anti-ezrin (#3145) (polyclonal), anti- Vil2kd/kd mice, which resulted in reduced bile flow ERM (#3142), and anti-glyceraldehyde 3-phosphate and intrahepatic cholestasis (17). These results sug- dehydrogenase (GAPDH) (14C10) antibodies were gest that ezrin plays important roles on cell surface purchased from Cell Signaling Technology (Danvers, expression of functional membrane transporter com- MA, USA). Anti-ezrin (3C12) antibody (monoclonal) plexes in vivo. However, comprehensive analysis of was purchased from Acris Antibody (Hiddenhausen, proteins which are down- or up-regulated in the pres- Germany). Anti-radixin (ab52495) and anti-NHERF1 ence or absence of ezrin has not been reported yet. (ab3452) antibodies were purchased from Abcam Recently, comprehensive analysis of proteins ex- (Cambridge, UK). Anti-Na+, K+-ATPase α-subunit pressed on the isolated plasma membrane has been (C464.6), anti-villin (C-19), and anti-β-actin (AC-15) performed by proteomic studies by use of mass spec- antibodies were purchased from Santa Cruz Biotech- trometry. Such proteomic analysis was carried out nology (Santa Cruz, CA, USA). Anti-sodium mono- for isolated apical and basolateral membrane vesi- carboxylate transporter 1 (SMCT1) (SLC5A8) cles from renal cortex and collecting ducts (9, 44). antibody was purchased from Biorbyt (Cambridge, Donowitz et al. reported the proteome of the mouse UK). Anti-protein disulfide-isomerase (PDI) (1D3) jejunal brush border membrane vesicles (BBMVs) antibody was purchased from Enzo Life Sciences (10). Quantitative proteomic studies were also per- (New York, USA). Anti-CLIC5 antibody was pur- formed with the BBMVs purified from the small in- chased from Alomone Labs (Jerusalem, Israel). An- testines of NHERF1 and NHERF2 knockout mice ti-moesin antibody was a kind gift from Professor (11, 12). In these studies, NHERF1 and 2 were Tsukita of the Graduate School of Frontier Biosci- found to regulate the expression of transport pro- ences, Osaka University. teins, signaling proteins, trafficking proteins, and proteins involved in proliferation, cell division and Histochemical studies. Small intestines of 8-weeks cell adhesion. On the other hand, there has been no old female mice were flushed with cold saline, and report for quantitative proteomic analysis of the BBM the small intestines were dissected into three seg- prepared from Vil2kd/kd mouse intestines. In this study, ments: duodenum, jejunum, and ileum. Duodenum we prepared the BBM fraction from the ileum seg- was collected as a segment 5 cm under the stomach. ments of wild-type and Vil2kd/kd mice, and compared Jejunum was collected as a segment 10 cm in length their quantitative proteomic patterns. following the duodenum. Ileum was collected as a segment 10 cm above the cecum. They were opened Brush border membrane fraction of ezrin knockdown mouse 129 lengthwise, and were fixed in a paraformaldehyde- for 15 min at 4°C, and the supernatant was collected based fixing solution overnight at 4°C, embedded in and stocked as a total tissue lysate at −80°C. The paraffin and cut into 1–2 μm-thick sections. Depar- total lysate was centrifuged at 8,000 × g for 15 min affined and rehydrated slices were subjected tohe- at 4°C, and the supernatant was centrifuged at matoxylin and eosin (HE) staining. HE slides were 100,000 × g for 90 min at 4°C. The pellet (as a crude scanned using the NanoZoomer digital scanner membrane fraction) was re-suspended in a suspend- (Hamamatsu Photonics, Shizuoka, Japan). Numbers ing buffer solution containing 150 mM D-mannitol, of villi per 1 mm length were counted from HE digi- 2.5 mM EGTA, 6 mM Tris-HCl (pH 7.1), and sus- tal images. The average numbers of villi were deter- pended through a 25 G needle. CaCl2 solution (1 M) mined by counting 3–5 fields from three 8-weeks old was added to the suspension to give a final concen- female mice of each genotype (littermates). Length tration of 10 mM, and the suspension was kept on ice of villi and depth of crypt were computed from HE for 15 min, followed by centrifugation at 3,000 × g digital images using NDP.view software (Hamamatsu for 15 min at 4°C. The pellet was re-suspended in Photonics). The average length of villi and depth of the suspending buffer solution and stored at −80°C crypts were determined by measuring at least 10 vil- as the basolateral membrane-rich fraction. The su- li and 10 crypts from three mice of each genotype pernatant was centrifuged at 100,000 × g for 30 min (littermates). at 4°C, and the pellet was re-suspended in the sus- pending buffer solution as the BBM fraction. Immunofluorescence microscopy. Tissue samples of 8-weeks old female mice were prepared, fixed, em- Characterization of BBM fraction by enrichment of bedded, sliced and deparaffined as reported in the marker enzymes. BBM purification was assessed by section of histochemical studies. Deparaffined and abundance of marker proteins of each subcellular rehydrated slices were subjected to antigen retrieval compartments; villin, Na+, K+-ATPase α-subunit, by boiling for 45 min in the Immunosaver (Nisshin PDI, and GAPDH as markers for BBM, basolateral EM, Tokyo, Japan), followed by treatment with 10% membrane (BLM), ER, and , respectively. goat serum for 30 min at room temperature. Slides were washed in PBS, and then sections were incu- In solution digestion and TMT labeling for MS anal- bated with primary antibodies overnight at 4°C. ysis. The BBM fraction solubilized in a 100 mM Slides were washed with PBS containing 0.03% ammonium bicarbonate solution containing 12 mM Tween 20 (PBS-T), and then incubated with Alexa sodium deoxycholate and 12 mM sodium N-lauroyl Fluor 488-labeled goat anti-mouse IgG (H+L), Al- sarcosinate was digested with trypsin (Proteomics exa Fluor 594-labeled goat anti-rabbit IgG (H+L), grade; Roche Life Science, Switzerland). Tryptic di- and 1 μg/mL 4’,6-diamidino-2-phenylindole (DAPI) gests were treated according to the Phase-transfer for at room temperature for 1 h. After washed with surfactants (PTS) protocol (26) and desalted using PBS-T, the sections were mounted with fluorescent C18-StageTips (31). Briefly, membrane proteins mounting medium (VECTASHIELD, Vector Labora- were reduced with 10 mM dithiothreitol for 30 min tories Inc., CA) and examined using a confocal laser at 55°C, alkylated with 60 mM iodoacetamide for scanning microscope (FV-1000D IX-81; Olympus, 30 min at room temperature, and digested with 1 : 20 Tokyo, Japan). (w/w) trypsin for 16 h at 37°C. An equal volume of an organic solvent, ethyl acetate, was added to di- Preparation of BBM fractions. The BBM fractions gested samples, and the mixtures were acidified by were prepared as described previously with some 1% trifluoroacetic acid, and vortexed to transfer de- modifications (36). Small intestines of 8-weeks old tergents to the organic phase. After centrifugation, female mice were flushed with cold PBS, and the the aqueous phase containing peptides was collected small intestines were dissected into three segments: and desalted using C18-StageTips. The tryptic pep- duodenum, jejunum, and ileum. The small intestinal tides were labeled with Tandem Mass TagTM (TMT) mucosa was gently scraped off with a coverslip and Reagents (Thermo Fischer Scientific, Bremen, Germa- homogenized in a homogenizing buffer solution ny) following the manufacturer’s instruction. Labeled containing 300 mM D-mannitol, 5 mM ethylene gly- peptide sample was fractionated by SDB-StageTips col tetraacetic acid (EGTA), 12 mM Tris-HCl (pH 7.1) with reversed-phase in basic pH (31), and fraction- with protease inhibitors with a Polytron homogeniz- ates were cleaned-up and concentrated with C18- er (Kinematica AG) at the maximum speed for StageTips. 1 min. The homogenate was centrifuged at 3,000 × g 130 S. Yoshida et al.

Protein identification and quantification by NanoLC- satory up-regulation of radixin and moesin in the il- MS/MS with reverse phase. Each fractionated peptide eum of Vil2kd/kd mouse. prepared above was injected into a trap column (C18, 0.3 × 5 mm; L-column, Chemicals Evaluation and Effects of ezrin knockdown on mouse intestinal Research Institute, Tokyo, Japan), and an analytical structure column (C18, 0.075 × 120 mm; Nikkyo Technos, Next, we studied the effects of ezrin knockdown on Tokyo, Japan), which was attached to NanoLC-MS/ the structure of small intestine because villus mor- MS system. NanoLC-MS/MS analysis was conduct- phogenesis and maintenance were impaired in the ed by an LTQ Orbitrap Velos mass spectrometer ezrin knockout (Vil2−/−) mice (7). The whole length (Thermo Fisher Scientific) equipped with a nanoLC of small intestine of Vil2kd/kd mice was not signifi- interface (KYA, Tokyo, Japan) and a nanoHPLC cantly different from that of wild-type mice (29.3 ± system DiNa (KYA). Purified peptides were intro- 0.9 cm and 31.5 ± 0.9 cm for wild-type and Vil2kd/kd duced from NanoLC to the LTQ Orbitrap Velos, a 8 weeks-old female mice, respectively). Among the hybrid ion-trap Fourier transform mass spectrometer. small intestinal segments, the structure of Vil2kd/kd Full MS and MS/MS scans were followed by higher mouse ileum was comparable with that of wild-type energy collisionally activated dissociation (HCD). mouse ileum (Fig. 2). We measured the number and The database search engine; Proteome Discoverer height of villus, and crypt depth of three 8 weeks-old 1.4 (Thermo Scientific) and MASCOT 2.4 (Matrix female littermate mice as described in the MATERI- Science) were used to identify and quantify proteins ALS AND METHODS. There was no significant from the MS, MS/MS, and reporter ion spectra of difference in the number of villus (9.8 ± 1.0 and peptides. Peptide mass data were matched by search- 9.4 ± 0.7 for wild-type and Vil2kd/kd mice, respective- ing UniprotKB/Swiss-prot release 2015_07 (24-Jun- ly), height of villus (152 ± 14 μm vs. 188 ± 6 μm), 15) database. False discovery rate (FDR) (40) was and crypt depth (59 ± 5 μm vs. 66 ± 1 μm) between calculated by enabling peptide sequence analysis us- the wild-type and Vil2kd/kd mice. Therefore, we pre- ing Percolator (19). High confidence peptide identifi- pared the BBM fraction from the ileum segments of cation was obtained by setting a target FDR threshold Vil2kd/kd and wild-type littermate mice, and compared of ≤1.0% at the peptide level. their quantitative proteomic patterns.

Validation of BBM fraction RESULTS Validation of the BBM preparation was performed Expression of ERM proteins in the mouse small in- by western blotting with antibodies which specifi- testine cally detect proteins in each organelle and specific First, we compared the expression of each ERM membrane domains (Fig. 3). The total tissue lysate protein in the ileum of the wild-type and Vil2kd/kd and BBM fractions were prepared from the Vil2kd/kd mice by immunofluorescence and western blotting. and wild-type littermate mouse ileums, and enrich- All three ERM proteins were expressed in the ileum, ment of specific proteins was estimated. Villin, a however, their location was quite different between marker of BBM, was abundant in the BBM fraction them. Ezrin was highly enriched in the brush border compared with the total lysate. Conversely, PDI and of villus epithelial cells. On the other hand, moesin GAPDH, which are markers for ER and cytosol, re- and radixin were not found in the brush border of spectively, were not observed in the BBM fraction. the epithelial cells, rather they were mainly found in Na+, K+-ATPase α-subunit, a marker of BLM, was endothelial cells (Fig. 1A, B). Immunofluorescence not detectable in the BBM fraction of the wild-type performed in the absence of primary antibody mouse ileum whereas a faint band of Na+, K+-ATPase showed very low level of background fluorescence α-subunit appeared in the BBM fraction of the (data not shown), which suggested that the patterns Vil2kd/kd mouse ileum. Na+, K+-ATPase α-subunit was shown in Fig. 1A and B were specific. Fig. 1C shows more enriched in the different BLM-rich fraction the western blotting pattern of total tissue lysate of rather than the BBM fraction both in the wild-type the ileum from the wild-type and Vil2kd/kd mice with and Vil2kd/kd mouse ileums (data not shown). an anti-ERM antibody which can detect all three ERM proteins. The results indicate that the expres- MS analysis sion of ezrin is much higher than that of radixin and Totally 313 proteins in three of three independent moesin in the ileum in the wild-type. Ezrin was not preparations from the ileum BBM fractions of the detected in the Vil2kd/kd mice. There was no compen- wild-type mice were identified, which included cy- Brush border membrane fraction of ezrin knockdown mouse 131

Fig. 1 (A, B) Immunofluoresence ob- servation of the ileum sections of wild- type and Vil2kd/kd mice with the anti-ezrin (red) and anti-moesin antibodies (green) (A), and with the anti-ezrin (green) and anti-radixin antibodies (red) (B). Merge patterns were also presented. Scale bars, 100 μm. The expression pattern of ezrin was different from that of radixin and moesin in the ileum. In addition, neither moesin nor radixin compensated the loss of ezrin in the Vil2kd/kd mouse ileum. (C) Western blots of 10 μg tissue lysate of the wild-type and Vil2kd/kd mouse ileum with the an- ti-ERM antibody. The ileum expresses all three ERM proteins with the ex- pression of ezrin being highest. toskeleton-associated proteins (54 proteins: 17.3%), matched BBM preparations, with the variability of transporters (32 proteins: 10.2%), signaling proteins the ratio in the multiple data acquisition being less (23 proteins: 7.3%), nuclear proteins (18 proteins: than 30%. We identified changes (increase or de- 5.8%), trafficking proteins (22 proteins: 7.0%), en- crease) in abundance of 19 proteins in three of three zymes (87 proteins: 27.8%) and others or unknown matched BBM preparations between the wild-type (77 proteins: 24.6%). The lists of transport proteins, and Vil2kd/kd mouse ileum as shown in Table 2 with cytoskeleton-associated proteins, and trafficking pro- the ratios in Vil2kd/kd/wild-type mice. These 19 pro- teins were shown in Table 1. The total number of teins were characterized as 1) transport proteins (4 proteins found in the ileum BBM fraction was com- proteins), 2) cytoskeleton-associated proteins (6 pro- parable with the number of proteins previously re- teins), 3) trafficking proteins (2 proteins), 4) en- ported in the jejunal BBMV (10). zymes (3 proteins), and 5) others (4 proteins). Proteins identified in both wild-type and Vil2kd/kd Three transport proteins down-regulated in the mouse ileum BBM fraction were to be compared. Vil2kd/kd mouse were SMCT1 (SLC5A8), transmem- Three independent matched BBM preparations were brane channel-like protein 4 (TMC4), and chloride analyzed. Proteins were considered to be up-regulat- intracellular channel protein 5 (CLIC5). SMCT1 is a ed in the Vil2kd/kd mouse when the ratio of their pro- transporter of lactate, pyruvate and short-chain fatty teins in Vil2kd/kd/wild-type mice was > 2.0, and down- acids (8, 27). TMC4, a member of transmembrane regulated when the ratio was < 0.6, in three of three channel-like proteins (TMCs), is a putative ion chan- 132 S. Yoshida et al.

Fig. 2 Sections of wild-type and Vil2kd/kd mouse ileum were stained by HE. Scale bar, 100 μm. The villus structure was maintained in the Vil2kd/kd mouse with the number of villi being unchanged between wild-type and Vil2kd/kd mice. nel or modifier with its physiological function being unknown at present (21). CLIC5 is a member of pu- tative intracellular chloride channels (CLICs), which is expressed in the apical domains of epithelial cells of small and large intestines, and interacts with the cortical actin cytoskeleton via ezrin (5). On the oth- er hand, one transport protein up-regulated in the Vil2kd/kd mouse was a member of ZnT transporter, transporter 10 (SLC30A10) (18). Recently, its primary function has been shown to be manganese transport (24). Two cytoskeleton-associated proteins down-regu- lated in the BBM fraction of Vil2kd/kd mouse ileum were a PDZ domain-containing scaffold protein, NHERF1, and ezrin itself. On the other hand, four cytoskeleton-associated proteins were up-regulated in the BBM fraction of Vil2kd/kd mouse ileum. Epitheli- al cell adhesion molecule (EpCAM) is an actin-bind- ing protein involved in homotypic cell-cell adhesions (3). INAD-like (INADl) protein is a scaffold protein containing 10 PDZ domains which is expressed at Fig. 3 Validation of the BBM fractions. The total tissue ly- sate (1 μg) and BBM fractions (1 μg) prepared from the wild- the apical membrane and tight junction of entero- kd/kd 2+ type and Vil2 mouse ileums were examined by western cytes (2, 23). Adseverin is a Ca -dependent actin blotting with antibodies against ezrin, villin, GAPDH, PDI, filament-severing protein, which is highly expressed Na+, K+-ATPase α-subunit, and β-actin, respectively. in kidney and small intestine, and regulates the actin cytoskeleton architecture by severing and capping F- actin strands (25, 34). Claudin-3 is a tight junction cell surface (32). VAMP8, a member of vesicular integral membrane protein (29). SNARE proteins, is involved in regulated secretion Two trafficking proteins up-regulated in the BBM of the entire exocrine system (39). fraction of Vil2kd/kd mouse ileum were Protein S100- Three enzymes up-regulated in the BBM fraction A10 and vesicle-associated membrane protein 8 of Vil2kd/kd mouse ileum were putative adenosylho- (VAMP8). Protein S100-A10 is targeted to membranes mocysteinase 2, intestinal-type alkaline phosphatase, via a complex formation with annexin A2, and regu- and glutathione S-transferase A1. lates the intracellular trafficking of target proteins to Brush border membrane fraction of ezrin knockdown mouse 133

Table 1 Transport proteins Cytoskeleton-associated proteins Protein Protein 1. Anoctamin-6 1. Actin, alpha cardiac muscle 1 2. 2. Actin, cytoplasmic 1 3. ATP-binding cassette sub-family G member 2 3. Actin-related protein 2/3 complex subunit 1B 4. ATP-binding cassette sub-family G member 5 4. Actin-related protein 2/3 complex subunit 4 5. ATP-binding cassette sub-family G member 8 5. Actin-related protein 3 6. Canalicular multispecific organic anion transporter 1 7. Chloride intracellular channel protein 1 6. Adseverin 8. Chloride intracellular channel protein 5 7. Alpha-actinin-4 9. -like protein 4 (SLC44A4) 8. Band 4.1-like protein 3 10. Cysteine-rich protein 1 9. Brain-specific angiogenesis inhibitor 1-associated protein 11. Ileal sodium / bile acid 2-like protein 1 12. MFS-type transporter (SLC18B1) 10. Brain-specific angiogenesis inhibitor 1-associated protein 13. Multidrug resistance protein 1 (P-glycoprotein) 2-like protein 2 14. Neutral and basic amino acid transport protein rBAT 11. Cadherin-related family member 5 (SLC3A1) 12. Claudin-3 15. Niemann-Pick C1-like protein 1 13. Cofilin 1 16. Potential phospholipid-transporting ATPase IC 14. Coronin-1B 17. Protein tweety homolog 3 15. Coronin-1C 18. Sodium/glucose cotransporter 1 (SLC5A1) 19. Sodium/myo-inositol cotransporter 2 (SLC5A11) 16. Coronin-2A 20. Na+, K+-ATPase alpha-1 subunit 17. Destrin 21. Na+, K+-ATPase beta-1 subunit 18. Epithelial cell adhesion molecule 22. Sodium monocarboxylate transporter 1 (SMCT1) (SLC5A8) 19. Ezrin 23. Sodium-dependent neutral amino acid transporter B(0)AT1 20. F-actin-capping protein alpha subunit (SLC6A19) 21. Filamin-B 24. Sodium-dependent phosphate transport protein 2B (Na/Pi-2b) 22. Harmonin (SLC34A2) 25. 13 member 2 (SLC13A2) 23. INAD-like protein 26. Solute carrier family 15 member 1 (PEPT1) (SLC15A1) 24. Keratin, type I cytoskeletal 19 27. Solute carrier family 26 member 6 (SLC26A6) 25. Keratin, type II cytoskeletal 8 28. Solute carrier family 52 (riboflavin transporter) member 3 26. LIM and SH3 domain protein 1 (SLC52A3) 27. Mucin-13 29. Transmembrane channel-like protein 4 (TMC4) 28. Myosin light polypeptide 6 30. Transmembrane channel-like protein 5 (TMC5) 29. Myosin-14 31. Zinc transporter 10 (SLC30A10) 30. Na+/H+ exchange regulatory factor1 (NHERF1) 32. Zinc transporter ZIP4 (SLC39A4) 31. Na+/H+ exchange regulatory factor3 (NHERF3) 32. Plastin-1 Trafficking proteins 33. Profilin 1 Protein 34. Protein cordon-bleu 1. ADP-ribosylation factor 3 35. Protein crumbs homolog 3 2. ADP-ribosylation factor 5 36. Ras GTPase-activating-like protein IQGAP 1 3. Annexin A2 37. Ras-related C3 botulinum toxin substrate 1 4. Annexin A4 38. Syntenin-1 5. Annexin A5 39. Transforming protein RhoA 6. Annexin A6 40. Tubulin alpha-1C chain 7. Annexin A7 41. Tubulin beta-5 chain 8. Annexin A11 9. Annexin A13 42. Myosin Ia 10. Copine-3 43. Myosin Id 11. Protein S100-A10 44. Myosin-5B 12. Ras-related protein Rab-1A 45. Myosin-6 13. Ras-related protein Rab-2A 46. Myosin-7A 14. Ras-related protein Rab-6A 47. Myosin-7B 15. Ras-related protein Rab-7A 48. Villin-1 16. Ras-related protein Rab-8A 49. 14-3-3 protein beta/alpha 17. Ras-related protein Rab-11A 50. 14-3-3 protein epsilon 18. Ras-related protein Rab-5C 19. Ras-related protein Rab-10 51. 14-3-3 protein eta 20. Ras-related protein Rab-35 52. 14-3-3 protein sigma 21. Syntaxin-binding protein 2 53. 14-3-3 protein theta 22. Vesicle-associated membrane protein 8 (VAMP8) 54. 14-3-3 protein zeta/delta 134 S. Yoshida et al.

Table 2 Up- and down-regulated proteins Protein Avg. Ratio (Vil2kd/kd: WT) 1) Transport protein (4 proteins) Up-regulated Zinc transporter 10 (SLC30A10) 2.71 ± 0.52 Down-regulated Sodium monocarboxylate transporter 1 (SMCT1) (SLC5A8) 0.57 ± 0.06 Transmembrane channel-like protein 4 (TMC4) 0.46 ± 0.11 Chloride intracellular channel protein 5 (CLIC5) 0.35 ± 0.05

2) Cytoskeleton-associated proteins (6 proteins) Up-regulated Epithelial cell adhesion molecule (EpCAM) 3.99 ± 1.25 INAD-like protein (INADl) 2.52 ± 0.42 Adseverin 2.24 ± 0.28 Claudin-3 2.02 ± 0.41 Down-regulated Na+/H+ exchange regulatory factor 1 (NHERF1) 0.53 ± 0.05 Ezrin 0.17 ± 0.03

3) Trafficking proteins (2 proteins) Up-regulated Protein S100-A10 3.08 ± 0.53 Vesicle-associated membrane protein 8 (VAMP8) 2.21 ± 0.38

4) Enzyme (3 proteins) Up-regulated Putative adenosylhomocysteinase 2 3.73 ± 1.18 Glutathione S-transferase A1 2.40 ± 0.39 Intestinal-type alkaline phosphatase 2.24 ± 0.35

5) Others (4 proteins) Up-regulated Guanine nucleotide-binding protein G(κ) subunit alpha 2.39 ± 0.28 Sushi domain-containing protein 2 2.23 ± 0.60 Down-regulated Calcium and integrin-binding protein 1 0.55 ± 0.02 Alpha-defensin 20 0.40 ± 0.14

Validation of NHERF1, SMCT1 and CLIC5 down- lower than those of the wild-type. regulation in the BBM fraction of Vil2kd/kd mice The immunofluorescence study of the ileum showed The MS results showed the down-regulation of that NHERF1 was co-localized with ezrin at the api- NHERF1, SMCT1, TMC4 and CLIC5 in the BBM cal surface of enterocytes in the wild-type mouse fraction from the ileum of Vil2kd/kd mice. Among whereas the expression of NHERF1 did not gather them the decreased expression levels of NHERF1, on the apical surface of enterocytes in the Vil2kd/kd SMCT1 and CLIC5 were confirmed by western blot- mouse (Fig. 5A). This result suggests that the ex- ting (Fig. 4). Western blot studies for TMC4 could pression of NHERF1 in the BBM was impaired in not be performed because there were no good anti- the Vil2kd/kd mice, which is in a good agreement with bodies available. The expression levels of NHERF1, the previous finding reported in ezrin knockout mice SMCT1 and CLIC5A (an alternative splicing variant (7). CLIC5 was also co-localized with ezrin at the of CLIC5 which is expressed in small and large in- apical surface of enterocytes in the wild-type mouse testines) in the BBM fraction and the total tissue ly- whereas the expression of CLIC5 in the enterocytes sate of the Vil2kd/kd mouse ileum were significantly was impaired in the Vil2kd/kd mouse (Fig. 5B). Simi- Brush border membrane fraction of ezrin knockdown mouse 135

DISCUSSION In this paper we tried to perform proteome analysis of the BBM fraction purified from small intestines. Among the small intestinal segments, the villus number of Vil2kd/kd mouse jejunum was less than that of wild-type littermate mouse (data no shown). On the other hand, the villus structure and number of Vil2kd/kd mouse ileum were comparable with those of wild-type. Although the Vil2kd/kd mouse suffered from achlorhydria, the pH values of intra-intestinal fluid in the ileum were similar between the wild-type and Vil2kd/kd mice. Therefore, we can rule out the possi- bility that different intra-intestinal pH affected pro- tein expression on the BBM. Here we focused on the results of proteome analysis of the BBM frac- tion to compare the expression patterns between the wild-type and Vil2kd/kd mice. The results were partly confirmed by western blot in combination with im- munofluorescence. Among the down- and up-regulated proteins, we especially focused on the membrane transport pro- teins and scaffold proteins. A PDZ domain-contain- ing scaffold protein, NHERF1, was down-regulated in the ileum BBM fraction of the Vil2kd/kd mice. Cell surface expression of NHERF1 in the enterocytes seems to be impaired in the absence of ezrin as shown in the immunofluorescence (Fig. 5A). This Fig. 4 The total tissue lysate (5 μg for CLIC5, NHERF1, result is in good agreements with the previous find- and β-actin; 10 μg for SMCT1) and BBM fractions (1 μg for ing that the expression of NHERF1 at the cell sur- NHERF1; 5 μg for CLIC5 and β-actin; 10 μg for SMCT1) kd/kd face of small intestine was down-regulated in the prepared from the wild-type and Vil2 mouse ileums −/− were examined by western blotting with antibodies against ezrin knockout (Vil2 ) mice (7). NHERF1 was re- SMCT1, CLIC5, NHERF1 and β-actin, respectively. ported to be required for organization and stabiliza- tion of active phosphorylated ERM proteins at the apical membranes of intestinal and renal epithelial lar studies for SMCT1 could not be performed be- cells (28). It should be noted that ezrin was identi- cause there were no good antibodies available for fied as one of the down-regulated proteins found in immunofluorescence. the jejunum BBMV of NHERF1 knockout mice by The levels of NHERF1, SMCT1, a previous proteomics study (11). These results sug- TMC4 and CLIC5 in the Vil2kd/kd mouse ileums gest that the interaction between ezrin and NHERF1 were 98%, 80%, 86%, and 80% of those in the should be important for the stable expression of wild-type, respectively, from the results by use of both proteins at the BBM. NHERF1 contains two SurePrint G3 Mouse Gene Expression Microarray tandem PDZ domains (PDZ1 and PDZ2), and the (Agilent). On the other hand, the gene expression ERM protein-binding region, and acts as a scaffold levels of ezrin in the Vil2kd/kd mouse ileums were of many proteins. Most of the ligand proteins such 11% of those in the wild-type. These results suggest as CFTR, Npt2a, and Na+/H+ exchanger 3 (NHE3) that these were not significantly down-regulat- bind to the PDZ1 domain. The simultaneous associ- ed by the mRNA expression level. On the other ation of these membrane transport proteins and ez- hand, the gene expression level of SLC30A10 in the rin, as an anchoring protein for protein kinase A Vil2kd/kd mouse ileums was 283% higher than that in (13), on the common platform of NHERF1 is very the wild-type, suggesting that SLC30A10 was up- important for physiological regulation of the trans- regulated by the mRNA expression level. port proteins. In fact, association of ezrin and NHE3 or CFTR on the platform of NHERF1 is essential 136 S. Yoshida et al.

Fig. 5 (A) Immunofluoresence observation of the ileum sections of wild-type and Vil2kd/kd mice with the anti-ezrin (green) and anti-NHERF1 antibodies (red). (B) The observation was performed with the anti-ezrin (green) and anti-CLIC5 antibod- ies (red). Merge patterns were also presented with the staining with DAPI (blue). Scale bar, 100 μm. for the cAMP-dependent phosphorylation and regu- knockout mice (35). Although NHE3 is known to be lation of NHE3 and CFTR; the cAMP-dependent localized at the BBM of small intestines, it was not phosphorylation of NHE3 results in its inhibition identified as a protein in the BBM fraction in the whereas the phosphorylation of CFTR results in its present proteomics analysis as well as the previous activation (30, 42). Some of the PDZ domain-con- analyses of mouse jejunum (10–12). The reason why taining scaffold proteins including NHERF1 have the proteomics analyses failed to identify NHE3 been suggested to regulate transporters via stabiliza- should be studied in the future study. tion of cell surface expression of the transporters (20). In the MS analysis, SMCT1, TMC4, and CLIC5 Actually, apical cell surface expression of Npt2a was were also down-regulated in the ileum BBM fraction impaired in the renal proximal tubule of NHERF1 of the Vil2kd/kd mice. SMCT1 is a Na+-coupled trans- Brush border membrane fraction of ezrin knockdown mouse 137 porter for lactate, pyruvate and short-chain fatty ac- in the Vil2kd/kd mice. The absence of actin-binding ids: acetate, propionate, and butyrate (8, 27). Mouse protein ezrin may promote changes in actin cytoskel- and human SMCT1s are also involved in transport eton architecture. It should be interesting that Protein of a vitamin, nicotinate (15). SMCT1 contains a S100-A10 was up-regulated at the BBM fraction of PDZ-binding motif (a sequence of GTRL) at its C- the ileum in the Vil2kd/kd mice. The up-regulation of terminus, and interacts with a PDZ domain-contain- Protein S100-A10 may compensate the function of ing scaffold protein NHERF3 (1), which can interact ezrin to target proteins to cell surface. Future stud- with NHERF1 (22). Therefore, ezrin may affect cell ies are necessary to clarify the molecular mecha- surface expression of SMCT1 indirectly via scaffold nisms for their up- and down-regulation in the BBM. proteins, NHERF1 and 3. NHERF3 was also identi- In conclusion, we performed the comprehensive fied as one of the down-regulated proteins in our analysis of the BBM of mouse ileum by proteomics present analysis of Vil2kd/kd mice in two of three study, and for the first time, found that NHERF1, matched BBM preparations (data not shown). Growth SMCT1, TMC4, and CLIC5 were down-regulated at retardation found in the Vil2kd/kd mice may be at the BBM of the ileum in the Vil2kd/kd mice. least in part due to impairment of uptake of mono- carboxylates including vitamin nicotinate. Our west- Acknowledgement ern blot analysis further confirmed the down- regulation of SMCT1 in the ileum BBM fraction of We would thank Prof. Tsukita for giving us the the Vil2kd/kd mice (Fig. 4). It should be also pointed Vil2kd/kd mice. We thank Dr. Yosuke Matsumoto, Mr. out that other membrane transport proteins which Kotoku Kawaguchi, and Ms. Karin Ikeda for their contain a PDZ-binding motif at their carboxy termi- help with breeding and genotyping of mice and tech- nus; PEPT1 (SLC15A1) and Npt2b (SLC34A2) were nical support. This research was supported in part down-regulated in the Vil2kd/kd mice in one or two of by Grant-in-Aids for Scientific Research (21590082 three matched BBM preparations (data not shown). and 24590104) from the Japan Society for the Pro- CLIC5, a member of putative intracellular chlo- motion of Science to S.A., and a High-Tech Research ride channels (CLICs), was also down-regulated in Center Project for Private Universities: matching fund the ileum BBM fraction of the Vil2kd/kd mice. CLIC5 subsidy from the Ministry of Education, Culture, interacts with the cortical actin cytoskeleton in po- Sports, Science and Technology of Japan to S.A. larized epithelial cells (4). CLIC-5A which is ex- Ethical approval: All works with animals were per- pressed in small and large intestines as well as formed with approval of the Animal Ethics Commit- stomach, heart, and kidney, is associated with ezrin, tees of Ritsumeikan University. and recruits F-actin and other actin-binding proteins (5). In the renal glomeruli, CLIC-5A is a component CONFLICTS OF INTERESTS of the complex consisting of ezrin and a sialo mem- brane protein, podocalyxin. It should be noted that The authors declare that there is no conflict of inter- the expression of ezrin was down-regulated in the est. glomeruli of CLIC-5A-deficient mice (41). Down- regulation of CLIC5 was also reported in the jejunum REFERNCES BBMV of NHERF1 knockout mice by a previous 1. Anzai N, Kanai Y and Endou H (2007) New insights into re- proteomics study (11). On the other hand, we found, nal transport of urate. Curr Opin Rheumatol 19, 151–157. for the first time, the down-regulation of SMCT1 2. Assémat E, Bazelliéres E, Pallesi-Pocachard E, Le Bivic A and TMC4 in the ileum BBM fraction of the Vil2kd/kd and Massey-Harroche M (2008) Polarity complex proteins. mice. High mortality and growth retardation of Biochim Biophys Acta 1778, 614–630. kd/kd 3. Balzar M, Winter MJ, de Boer CJ and Litvinov SV (1999) Vil2 mice may be explained by cumulative ef- The biology of the 17-1A antigen (Ep-CAM). J Mol Med 77, fects of down-regulation of many kinds of transport- 699–712. ers in the small intestine. 4. Berryman M and Bretscher A (2000) Identification of a novel Besides the membrane transport proteins and scaf- member of the chloride intracellular channel gene family (CLIC5) that associates with the actin cytoskeleton of pla- fold proteins shown above, several cytoskeleton-as- cental microvilli. Mol Biol Cell 11, 1509–1521. sociated proteins and trafficking proteins were up- 5. Berryman M, Bruno J, Price J and Edwards JC (2004) CLIC- or down-regulated in the BBM fraction of the ileum 5A functions as a in vitro and associates in the Vil2kd/kd mice. Among them it should be noted with the cortical actin cytoskeleton in vitro and in vivo. J that an actin filament-severing protein, adseverin, Biol Chem 279, 34794–34801. 6. Bretscher A, Edwards K and Fehon RG (2002) ERM pro- was up-regulated at the BBM fraction of the ileum 138 S. Yoshida et al.

teins and merlin: integrators at the cell cortex. Nat Rev Mol 21. Kurima K, Yang Y, Sorber K and Griffith AJ (2003) Charac- Cell Biol 3, 586–599. terization of the transmembrane channel-like (TMC) gene 7. Casaletto JB, Saotome I, Curto M and McClatchey AI (2011) family: functional clues from hearing loss and epidermodys- Ezrin-mediated apical integrity is required for intestinal ho- plasia verruciformis. Genomics 82, 300–308. meostasis. Proc Natl Acad Sci USA 108, 11924–11929. 22. LaLonde DP, Garbett D and Bretscher A (2010) A regulated 8. Coady MJ, Chang M-H, Charron FM, Plata C, Wallendorff B, complex of the scaffolding proteins PZDK1 and EBP50 with Sah JF, Markowitz SD, Romero MF and Japointe J-Y (2004) ezrin contribute to microvillar organization. Mol Biol Cell The human tumor suppressor gene SLC5A8 expresses a Na+- 21, 1519–1529. monocarboxylate cotransporter. J Physiol 557, 719–731. 23. Lemmers C, Médina E, Delgrossi M-H, Michel D, Arsanto 9. Cutillas PR, Biber J, Marks J, Jacob R, Stieger B, Cramer R, J-P and Le Bivic A (2002) hINADl/PATJ, a homolog of discs Waterfield M, Burlingame AL and Unwin RJ (2005) Pro- lost, interacts with crumbs and localizes to tight junctions in teomic analysis of plasma membrane vesicles isolated from human epithelial cells. J Biol Chem 277, 25408–25415. that rat renal cortex. Proteomics 5, 101–112. 24. Leyva-Illades D, Chen P, Zogzas CE, Hutchens S, Mercado 10. Donowitz M, Singh S, Salahuddin FF, Hogema BM, Chen Y, JM, Swaim CD, Morrisett RA, Bowman AB, Aschner M and Gucek M, Cole RN, Ham A, Zachos NC, Kovbasnjuk O, Mukhopadhyay S (2014) SLC30A10 is a cell surface-local- Lapierre LA, Broere N, Goldenring J, deJonge H and Li X ized manganese efflux transporter, and Parkinsonism-causing (2007) Proteome of murine jejunal brush border membrane mutations block its intracellular trafficking and efflux activi- vesicles. J Proteome Res 6, 4068–4079. ty. J Neurosci 34, 14079–14095. 11. Donowitz M, Singh S, Singh P, Salahuddin FF, Chen Y, 25. Lueck A, Brown D and Kwiatkowski (1998) The actin-bind- Chakraborty M, Murtazina R, Gucek M, Cole RN, Zachos ing proteins adseverin and gelsolin are both highly expressed NC, Kovbasnjuk O, Broere N, Smalley-Freed WG, Reynolds but differentially localized in kidney and intestine. J Cell Sci AB, Hubbard AL, Seidler U, Weinman E, de Jonge HR, 111, 3633–3643. Hogema BM and Li X (2010) Alterations in the proteome of 26. Masuda T, Tomita M and Ishihama Y (2008) Phase transfer the NHERF1 knockout mouse jejunal brush border mem- surfactant-aided trypsin digestion for membrane proteome brane vesicles. Physiol Genomics 42A, 200–210. analysis. J Proteome Res 7, 731–740. 12. Donowitz M, Singh S, Singh P, Chakraborty M, Chen Y, 27. Miyauchi S, Gopal E, Fei Y-J and Ganapathy V (2004) Func- Murtazina R, Gucek M, Cole RN, Zachos NC, Salahuddin tional identification of SLC5A8, a tumor suppressor down- FF, Kovbasnjuk O, Broere N, Smalley-Freed WG, Reynolds regulated in colon cancer, as a Na+-coupled transporter for AB, Hubbard AL, Seidler U, Weinman E, de Jonge HR, short-chain fatty acids. J Biol Chem 279, 13293–13296. Hogema BM and Li X (2011) Alterations in the proteome of 28. Molares FC, Takahashi Y, Kreimann EL and Georgescu M-M the NHERF2 knockout mouse jejunal brush border mem- (2004) Ezrin-radixin-moesin (ERM)-binding phosphoprotein brane vesicles. Physiol Genomics 43, 674–684. 50 organizes ERM proteins at the apical membrane of polar- 13. Dransfield DT, Bradford AJ, Smith J, Martin M, Roy C, ized epithelia. Proc Natl Acad Sci USA 101, 17705–17710. Mangeat PH and Goldenring JR (1997) Ezrin is a cyclic AMP- 29. Morita K, Furuse M, Fujimoto K and Tsukita S (1999) Clau- dependent protein kinase anchoring protein. EMBO J 16, 35– din multigene family encoding four-transmembrane domain 43. protein components of tight junction strands. Proc Natl Acad 14. Fehon RG, McClatchey AI and Bretscher A (2010) Organiz- Sci USA 96, 511–516. ing the cell cortex: the role of ERM proteins. Nat Rev Mol 30. Naren AP, Cobb B, Li C, Roy K, Nelson D, Heda GD, Liao J, Cell Biol 11, 276–287. Kirk KL, Sorscher EJ, Hanrahan J and Clancy JP (2003) A 15. Gopal E, Miyauchi S, Martin PM, Ananth S, Roon P, Smith macromolecular complex of β2 adrenergic receptor, CFTR, SB and Ganapathy V (2007) Transport of nicotinate and and ezrin/radixin/moesin-binding phosphoprotein 50 is regu- structurally related compounds by human SMCT1 (SLC5A8) lated by PKA. Proc Natl Acad Sci USA 100, 342–346. and its relevance to drug transport in the mammalian intesti- 31. Rappsilber J, Mann M and Ishikawa Y (2007) Protocol for nal tract. Pharm Res 24, 575–584. micro-purification, enrichment, pre-fractionation and storage 16. Hatano R, Fujii E, Segawa H, Mukaisho K, Matsubara M, of peptides for proteomics using StageTips. Nat Protoc 2, Miyamoto K, Hattori T, Sugihara H and Asano S (2013) Ez- 1896–1906. rin, a membrane cytoskeletal cross-linker, is essential for the 32. Rescher U and Gerke V (2008) S100A10/p11: family, friends regulation of phosphate and calcium homeostasis. Kidney Int and functions. Pflugers Arch 455, 575–582. 83, 41–49. 33. Saotome I, Curto M and McClatchey AI (2004) Ezrin is es- 17. Hatano R, Akiyama K, Tamura A, Hosogi S, Marunaka Y, sential for epithelial organization and villus morphogenesis Caplan MJ, Ueno Y, Tsukita S and Asano S (2015) Knock- in the developing intestine. Dev Cell 6, 855–864. down of ezrin causes intrahepatic cholestasis by the dysregu- 34. Schafer DA and Cooper JA (1995) Control of actin assembly lation of bile fluidity in the bile duct epithelium. Hepatology at filament ends. Annu Rev Dev Biol 11, 497–518. 61, 1660–1671. 35. Shenolikar S, Voltz JW, Minkoff CM, Wade JB and Weinman 18. Huang L and Tepaamorndech S (2013) The SLC30 family of EJ (2002) Targeted disruption of the mouse NHERF-1 gene zinc transporters – A review of current understanding of their promotes internalization of proximal tubule sodium-phos- biological and pathophysiological roles. Mol Aspects Med 34, phate cotransporter type IIa and renal phosphate wasting. 548–560. Proc Natl Acad Sci USA 99, 11470–11475. 19. Käll L, Canterbury JD, Weston J, Noble WS and MacCoss 36. Sugiura T, Kato Y, Wakayama T, Silver DL, Kubo Y, Iseki S MJ (2007) Semi-supervised learning for peptide identification and Tsuji A (2008) PDZK1 regulates two intestinal solute from shotgun proteomics datasets. Nat Methods 4, 923–925. carriers (Slc15a1 and Slc22a5) in mice. Drug Metabol Dis 20. Kato Y, Watanabe C and Tsuji A (2006) Regulation of drug 36, 1181–1188. transporters by PDZ adaptor proteins and nuclear receptors. 37. Tamura A, Kikuchi S, Hata M, Katsuno T, Matsui T, Hayashi Eur J Pharm Sci 27, 487–500. H, Suzuki Y, Noda T, Tsukita S and Tsukita S (2005) Achlor- Brush border membrane fraction of ezrin knockdown mouse 139

hydria by ezrin knockdown: defects in the formation/expan- (2010) CLIC5A, a component of the ezrin-podocalyxin com- sion of apical canaliculi in gastric parietal cells. J Cell Biol plex in glomeruli, is a determinant of podocyte integrity. Am 169, 21–28. J Physiol Renal Physiol 298, F1492–F1503. 38. Tsukita S and Yonemura S (1999) Cortical actin organization: 42. Weinman EJ, Steplock D, Donowitz M and Shenolikar S lessons from ERM (ezrin/radixin/moesin) proteins. J Biol (2000) NHERF associates with sodium-hydrogen exchanger Chem 274, 34507–34510. isoform 3 (NHE3) and ezrin are essential for cAMP-mediat- 39. Wang C-C, Shi H, Guo K, Ng CP, Li J, Gan BQ, Liew HC, ed phosphorylation and inhibition of NHE3. Biochemistry 39, Leinonen J, Rajaniemi H, Zhou ZH, Zeng Q and Hong W 6123–6129. (2007) VAMP8/endobrevin as a general vesicular SNARE for 43. Yoshida S, Yamamoto H, Tetsui T, Kobayakawa Y, Hatano R, regulated exocytosis of the exocrine system. Mol Biol Cell Mukaisho K, Hattori T, Sugihara H and Asano S (2016) Ef- 18, 1056–1063. fects of ezrin knockdown on the structure of gastric glandu- 40. Wang G, Wu WW, Zhang Z, Masilamani S and Shen R-F lar epithelia. J Physiol Sci 66, 53–65. (2009) Decoy methods for assessing false positives and false 44. Yu M-J, Pisitkun T, Wang G, Shen R-F and Knepper MA discovery rates in shotgun proteomics. Anal Chem 81, 146– (2006) LC-MS/MS analysis of apical and basolateral plasma 159. membranes of rat renal collecting duct cells. Mol Cell Pro- 41. Wegner B, AI-Momany A, Kulak SC, Kozlowski K, Obeidat teomics 5.11, 2131–2145. M, Jahroudi N, Paes J, Berryman M and Ballermann BJ