Synaptotagmin-Like Proteins Control the Formation of a Single Apical Membrane Domain in Epithelial Cells

ARTICLES

Synaptotagmin-like proteins control the formation of a single apical membrane domain in epithelial cells

Manuel Gálvez-Santisteban1,6, Alejo E. Rodriguez-Fraticelli 1,6, David M. Bryant 2, Silvia Vergarajauregui 1,7, Takao Yasuda3, Inmaculada Bañón-Rodríguez 1, Ilenia Bernascone 1, Anirban Datta 2, Natalie Spivak 2,4, Kitty Young2, Christiaan L. Slim 2,7, Paul R. Brakeman 2,4, Mitsunori Fukuda 3, Keith E. Mostov 2,5 and Fernando Martín-Belmonte1,8

The formation of epithelial tissues requires both the generation of apical–basal polarity and the coordination of this polarity between neighbouring cells to form a central lumen. During de novo lumen formation, vectorial membrane transport contributes to the formation of a singular apical membrane, resulting in the contribution of each cell to only a single lumen. Here, from a functional screen for genes required for three-dimensional epithelial architecture, we identify key roles for synaptotagmin-like proteins 2-a and 4-a (Slp2-a/4-a) in the generation of a single apical surface per cell. Slp2-a localizes to the luminal membrane in a PtdIns(4,5)P 2-dependent manner, where it targets Rab27-loaded vesicles to initiate a single lumen. Vesicle tethering and fusion is controlled by Slp4-a, in conjunction with Rab27/Rab3/Rab8 and the SNARE syntaxin-3. Together, Slp2-a/4-a coordinate the spatiotemporal organization of vectorial apical transport to ensure that only a single apical surface, and thus the formation of a single lumen, occurs per cell.

Epithelia represent the most fundamental tissue in metazoa, forming cell has a single apical initiation site, and thus the tube has a single complex layers of cells such as the skin or kidney tubules. The lumen, however, is largely unclear. Similarly, how such machineries epithelial plasma membrane is divided into two domains: apical are controlled at the transcriptional level during morphogenesis of and basolateral, separated by cellular junctions, dependent on the epithelial tissues is poorly understood 3. asymmetric delivery and segregation of membrane proteins and Here, we report a functional screen for regulators of three- lipids1,2. Such plasma membrane asymmetry allows the formation of dimensional (3D) epithelial polarity using MDCK cyst cultures, a central lumen, and hence the evolution of specialized functions for based on transcriptional, RNA-mediated interference (RNAi) and different metazoan tissues3. Epithelial cells create lumens through an morphogenetic analysis. array of morphogenetic mechanisms. Despite this diversity, a series of Synaptotagmin-like proteins (Slps) 1-5 are a family of Rab effectors common molecular events creates biological tubes: vectorial transport involved in regulated exocytosis 10 . Slps harbour an amino-terminal to a nascent apical domain, de novo apical plasma membrane biogenesis, Rab-binding domain (also called the Slp homology domain SHD) and secretion and expansion of the luminal space 4,5. Transport of apical and tandem carboxy-terminal C2 domains involved in Ca 2+ and proteins to the initial site for apical-membrane formation, at which the phospholipid binding, and function in tethering secretory vesicles Par3–aPKC–Cdc42 polarity complex is established, is controlled by a to the plasma membrane 11 . In Drosophila melanogaster, a single Rab11a/8a GTPase cascade and its effectors, the exocyst and Myo5B divergent Slp paralogue, bitesize (Btsz), functions in epithelial (refs 6,7). At the lumen, phosphoinositide asymmetry is concomitantly polarization12 , although whether mammalian Slps function in polarity established with PtdIns(4,5)P2 and PtdIns(3,4,5)P3 localizing to, and generation is unknown. specifying, the apical and basolateral domains, respectively8,9. How We demonstrate that Slp2-a and related family member Slp4-a vectorial exocytic transport is coordinated and directed so that each function in distinct, but complementary, steps of apical transport

1Centro de Biología Molecular Severo Ochoa, Consejo Superior de Investigaciones Cientíﬁcas (CSIC), C/Nicolás Cabrera 1, Madrid 28049, Spain. 2Department of Anatomy, University of California San Francisco, California 94143-2140, USA. 3Department of Developmental Biology and Neurosciences, Graduate School of Life Sciences, Tohoku University, Aobayama, Aoba-ku, Sendai, Miyagi 980-8578, Japan. 4Department of Pediatrics, University of California San Francisco, California 94143-2140, USA. 5Department of Biochemistry and Biophysics, University of California San Francisco, California 94143-2140, USA. 6These authors contributed equally to this work. 7Present addresses: National Heart, Lung and Blood Institute (NHLBI), National Institutes of Health (NIH) Bethesda, Maryland 20824-0105, USA (S.V.); Department of Cell Biology, University Medical Center Groningen, Antonius Deusinglaan 1, Groningen 9713 AV, Holland (C.L.S.). 8Correspondence should be addressed to F.M-B. (e-mail: [email protected])

Received 21 September 2011; accepted 14 June 2012; published online 22 July 2012; DOI: 10.1038/ncb2541

a Canine Genome 2.0 Bioinformatic analysis siRNA screening of Affymetrix array Bibliographic research selected genes Stealth RNAi library: 3 x RNAi sequences per gene qPCR validation of Read-out: percentage of lumen 99 selected genes formation in Matrigel after 72 h 2D-MDCK 3D-MDCK Data set 1 Data set 2 1,597 upregulated genes 47 candidate 1,304 downregulated genes genes 16 regulators (FDR < 0.05) (2 previously identified)

b 140 120 100 ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ 80 ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ 60 ∗ ∗ 40 lumen formation

Percentage of single 20 0 IHH 7B2 FUZ JIP1 DAK PKIG BST2 DYSF SNX2 SNX4 SNX5 GAL4 FRZB NGEF GRB7 GNB5 RHOU SYT13 SYTL2 FXYD2 IFT172 FABP1 FARP1 FGFR4 CHE11 FYCO1 CDH16 CLDN2 Control PI3KR1 PI3KR2 UNC5C IGFBP6 FBLIM1 S100A4 S100A5 VPS37D TGFBR7 NHERF1 NHERF2 NHERF3 SMTNL2 SYNGR1 TB2/DP1 SEC14L3 STARD10 RANBPL3 ARHGAP24

c Podxl β-cat DNA ZO-1

siRNA: Control Claudin 2 (CLDN2) Smoothelin-like 2 Wrch-1 (RHOU) Fuzzy ( FUZ) (SMTNL2)

d Tissue distribution in all data sets e Breast cancer versus normal (Oncomine) 80 Kidney Breast Kidney cancer versus normal Skin 60 Colorectal Prostate Lung 40 Gastric Bladder Ovaries 20 Pancreatic Seminiferous genes (Oncomine)

Head and neck Percentage of microarray 0

Oesophagus data sets with downregulated Brain FUZ SNX5 FRZB RHOU SYTL2 SYT13 IFT172 CHE11 CDH16 CLDN2

05 10 15 PI3KR1 NHERF1 SMTNL2 TB2/DP1 Number of positive-hit genes downregulated STARD10 in cancer versus normal tissue ARHGAP24

Figure 1 A screen for regulators of 3D epithelial polarization. line, the threshold considered for the definition of a positive hit (lumen (a) Experimental design for the function screen of regulators of 3D formation < 75% of control; P < 0.05); green stars, positive hits; red epithelial polarity. MDCK cells were cultured for 36 h in 2D or 3D ( n = 3), dots, knockdowns where the efficiency was below 60%. ∗P < 0.05; n = 3; and control (2D) and experimental (3D) RNA samples were analysed error bars represent s.d. (c) Examples of phenotypes induced by RNAi in using the Affymetrix Canine Genome 2.0 platform. The significance the screen. Seventy-two hour MDCK cysts transfected with siRNA from of the data was determined by LiMMA analysis (false discovery rate four positive candidates (SMTNL2, CLDN2, RHOU and FUZ ) and stained (FDR) < 0.05). A set of significantly upregulated ( >2-fold) genes was for the apical marker Podxl (red), the basolateral marker β-catenin (β-cat; pooled with other genes of interest and then gene overexpression was green), the tight junction marker ZO-1 (white) and nuclei (blue). Scale bars, validated by RT–qPCR. Bioinformatic pathway analyses revealed that 5 µm. ( d) Epithelial cancers with a downregulated 3D polarity gene set. some genes were connected in common functional pathways. A final The expression levels of the candidate gene set in all cancer versus normal set of 47 candidates was selected for stealth siRNA design. MDCK cells expression data sets were analysed using Oncomine (www.oncomine.org). were transfected with siRNAs individually or in pools and cultured to grow The graph shows the number of downregulated genes per type of indicated cysts. The silencing efficiency of the siRNA was determined by RT–qPCR. epithelial cancers (P <0.05 ,n varies in each tissue). ( e) Frequency of gene Then, cells were fixed and stained for Podxl, β-catenin and nuclei to downregulation in breast and kidney cancer data sets. The graph indicates quantify normal lumen formation. The RNAi screening finally resulted in 16 the percentage of data sets with downregulated candidate genes in breast positive hits (see Methods). (b) RNAi screening for polarization regulation. and kidney cancer versus normal tissue microarray data sets (Oncomine, Lumen formation efficiency was quantified for each of the listed 47 siRNA P < 0.05). The yellow arrowheads denote genes downregulated in both treatments. Green dotted line, normal levels as found in control; red dotted breast and kidney cancer.

Table 1 Localization and function of synaptotagmin-like proteins, and their mutants, in MDCK cyst formation. Protein Construct Cell localization Mutant effect Phenotype rescue Early aggregate Lumen initiation Open lumen

Slp2-a Full length Unpolarized PM Junctions PM Apical PM Yes SHD Vesicle/cytosolic Vesicle/cytosolic Vesicle/cytosolic - 1C2AB Vesicle/endosome Vesicle/endosome Subapical - Linker Cytosol Cytosol Cytosol - 1SHD Unpolarized PM Junctions PM Apical PM No C2AB Unpolarized PM Junctions PM Apical PM - mut E11A/R32A Unpolarized PM Junctions PM Apical PM No Rab binding No mut V18A Unpolarized PM Junctions PM Apical PM No Rab3/8 binding Yes

Slp4-a Full-length Apical PM Vesicles Apical PM Yes SHD nuc/cytosol nuc/cytosol nuc/cytosol - 1C2AB Cytosol Cytosol Cytosol - Linker Cytosol Cytosol Cytosol - 1SHD Unpolarized PM Unpolarized PM Unpolarized PM - C2AB Unpolarized PM Unpolarized PM Unpolarized PM - mut I18A Unpolarized PM Vesicles/junctions PM Subapical/junctions PM No Rab3/8 binding No mut V21A Vesicles Vesicles Subapical No Rab3 binding No mut W118S Cytosol Cytosol Cytosol No Rab binding No mut K > Q Vesicles Vesicles Subapical No lipid binding Partial Slp454a Unpolarized PM Vesicles/PM Subapical/junctions PM No SNARE binding No 1305–354 Unpolarized PM Vesicles/PM Junctions PM No SNARE binding No

This table represents a qualitative summary of results regarding WT or mutant Slp protein localization (at different time points) and its ability to rescue the silencing of endogenous protein expression. PM, plasma membrane; nuc, nucleus; mut, mutant. to form a lumen de novo . Slp2-a controls the positioning of Slp4- To analyse a possible relevance of these genes in vivo , we analysed a/Rab27-positive vesicles to target exocytosis to a single PtdIns(4,5)P 2- whether this gene set was downregulated in human cancers, suggesting enriched lumen. Slp4-a regulates the tethering of these vesicles, an importance in the maintenance of a differentiated epithelial through the association with the apical SNARE syntaxin-3 (Stx3), to phenotype in vivo (Fig. 1d). Notably, renal, breast and skin cancers mediate vesicle delivery to the lumen. Thus, through a functional, presented with the strongest downregulation of this gene set. We multi-step screen, we have identified a previously uncharacterized selected one of these genes, SYTL2 (encoding the protein Slp2-a), which mechanism for coordinated vectorial transport, crucial to form a was significantly downregulated in several epithelial cancer data sets single apical domain. (Fig. 1e), to characterize its role in lumen formation.

RESULTS Slp2-a associates with, and regulates the formation of, the Identiﬁcation of 14 previously uncharacterized regulators of luminal membrane epithelial morphogenesis The mammalian Slp family has been shown to regulate primarily We performed a multi-step, functional screen for regulators specifically Rab27-dependent membrane trafficking and secretion15–17, but the of 3D epithelial architecture and morphogenesis. We first conducted function of these proteins in mammalian epithelial morphogenesis is a microarray-based differential expression analysis comparing unknown. We confirmed that Slp2-a protein levels were upregulated the transcriptome of MDCK cells undergoing apical –basolateral in 3D when compared with 2D cultures by western blot (14-fold polarization either in the traditional monolayer culture (2D), or enrichment at 72 h; Fig. 2a), validating the qPCR data (Supplementary as 3D cysts grown in basement membrane extract (3D), wherein Fig. S2). Next, we characterized Slp2-a localization in MDCK cysts. MDCK cells self-assemble to form a 3D monolayer ( Fig. 1a). Notable On plating MDCK into 3D, the apical podocalyxin (Podxl) localized transcriptional differences were observed during 3D morphogenesis to the peripheral surface of early aggregates, before it is internalized with 1,597 upregulated, and 1,304 downregulated probes detected into vesicles and delivered to the contact between two cells, where (Supplementary Fig. S1). To prioritize functional analyses, upregulated lumen is formed de novo 7,9,18. In early aggregates, Slp2-a localized genes were subjected to bioinformatic analysis to reconstruct potential to the plasma membrane, enriched at cellular junctions ( Fig. 2b molecular pathways and known components of epithelial polarization. and Table 1). A pool of Slp2-a became apparent on internalized Using this approach, a set of 99 upregulated genes was selected for Podxl-positive transcytosing vesicles near the cell–cell contact (Fig. 2b, secondary validation by quantitative PCR (qPCR; Supplementary 16 –20 h). On lumen formation (24 –48 h), Slp2-a localized to the apical Fig. S2 and Table S1). Finally, 47 candidate genes were targeted membrane (Fig. 2b, 48 h). through short interfering RNA (siRNA), including the known polarity Slp2-a depletion perturbed 3D lumen formation in a dose-dependent regulators claudin-2 (CLDN2) and Wrch-1 ( RHOU ) as internal manner (Fig. 2c,e), without affecting the cell polarity in cells growing controls13,14(Fig. 1b and Supplementary Table S2). We found a set in monolayers (Supplementary Fig. S3A). Slp2-a depletion results of 14 genes previously uncharacterized to be required for this process in abnormal morphology with multiple small lumens, and the (Fig. 1b,c, green stars and Supplementary Table S3). These included accumulation of Podxl ( Fig. 2d), that we identify as transcytotic tight and adherens junctions, Rho GTPases, lipid signalling and vesicles (Supplementary Fig. S5C). Notably, stable expression of membrane trafficking proteins. RNAi-resistant human Slp2-a (GFP–Slp2-a; Fig. 2f) completely

NATURE CELL BIOLOGY ADVANCE ONLINE PUBLICATION 3 © 201 2 Macmillan Publishers Limited. All rights reserved. ARTICLES a b 12 h 16 h 20 h 24 h 48 h 24 h 48 h 72 h GFP–Slp2-a Podxl 2D 3D 2D 3D 2D 3D Mr (K) WB: 150 β-cat Slp2-a WB: Tub 50 L

20 ∗ 2D 15 3D GFP–Slp2-a Podxl ∗ 10 ∗ 5

(arbitrary units) β-cat 0 3D Slp2-a induction

normalized to 2D 24 h 24 h 48 h 72 h c d e 120 l -1 -2 -3 siRNA: Control Slp2-a-3 Magnification siRNA: β-cat 100 Contro Slp2-a Slp2-a Slp2-a Pool ∗ ∗ Mr (K) Podxl 80 ∗∗ Nuclei 150 WB: 60 100 Slp2-a L 40 20 WB: lumen formation

50 Percentage of single Tub 0 siRNA: -1 -2 -3 Pool Control Slp2-a Slp2-a Slp2-a

f MDCK GFP–Slp2-a g hsiRNA control i siRNA: Control Slp2-a-3 120 siRNA Slp2-a Slp2-a localization -3 -3 GFP–Slp2-a 100 Podxl 12 h siRNA: ∗ Control Slp2-a Control Slp2-a β-cat 80 Mr (K) WB: 150 L L 60 100 Slp2-a 20 h 40 WB:

50 Tub lumen formation 20 Percentage of single 0 24 h MDCK GFP–Slp2-a

Figure 2 Slp2-a is required for epithelial morphogenesis. ( a) Top, western in cells transfected with control siRNA or Slp2-a siRNA. Values are blot (WB) showing the induction of Slp2-a in MDCK cells growing in 2D mean ± s.d. (n = 5; ≥100 cysts per experiments). (f) Knockdown of and 3D at different time points (24–72 h). Bottom, expression quantified Slp2-a by siRNA in cells stably expressing GFP–Slp2-a. MDCK cells by densitometry ( n = 4). (b) Localization of GFP–Slp2-a during lumen stably expressing GFP–Slp2-a were transfected with Slp2-a or control formation. MDCK cells stably expressing GFP–Slp2-a were grown in 3D siRNAs. Total lysates were blotted for Slp2-a using α-tubulin as a loading and fixed at different time points. Cysts were stained with Podxl (red) control. (g) Rescue effect of GFP–Slp2-a in cells silenced for Slp2-a on and β-catenin (β-cat; blue). The arrows indicate localization to apical lumen formation. Cells were stained for Podxl (red) and β-catenin (blue). plasma membrane; the arrowheads indicate localization to cell–cell Scale bars, 10 µm. ( h) Quantification of cysts with normal lumens in cells junctions; scale bar, 5 µm. ( c) Downregulation of Slp2-a by siRNA. MDCK expressing GFP–Slp2-a and transfected with control or Slp2-a siRNAs, cells were transfected with three different siRNA duplexes targeting n = 3. ( i) Slp2-a localization during lumen initiation. In early aggregates canine Slp2-a, and siRNA efficiency was analysed by western blotting. (12 h), Slp2-a localizes to cell–cell junctions at sites of apical vesicle (d) Effect of Slp2-a siRNA-mediated silencing on lumen formation. fusion. After the lumen is initiated (24 h), Slp2-a remains polarized Cells were transfected with a pool of siRNA to knockdown Slp2-a or at the apical membrane. Green lines, Slp2-a; blue ovals, nuclei; black siRNA control and plated to form cysts for 72 h. Markers are Podxl (red), lines, basolateral membrane. In all panels error bars represent s.d.; β-catenin (green) and nuclei (blue). The arrows indicate apical membrane ∗P < 0.05; ∗∗ P < 0.005; L, lumen; areas outlined in micrographs are localization; the arrowheads indicate localization to intracellular apical magnified in the associated images. Uncropped images of blots are shown vesicles. Scale bars, 5 µm. ( e) Quantification of cysts with normal lumens in Supplementary Fig. S8. rescued lumen formation and morphogenesis of cysts with endogenous during lumen formation7–9. To elucidate the control of Slp2-a Slp2-a knockdown (Fig. 2g,h). Moreover, these results indicate that localization, we analysed Slp2-a domains during cyst formation. In Slp2-a is the predominant variant required for epithelial morphogenesis early aggregates, the C2 domains localized to the plasma membrane in 3D-MDCK (ref. 19 ). and to cell –cell junctions, but not to Podxl vesicles ( Fig. 3a and Supplementary Fig. S4b, 12 –20 h). Once lumens formed, the C2A/B The SHD and C2 domains play non-redundant roles in targeting fragment localized exclusively to the apical membrane ( Fig. 3a and Slp2-a to membranes Supplementary Fig. S4b, 24 –48 h). In contrast, the SHD fragment Slp-family proteins share an N-terminal Rab27-binding domain (the was predominantly cytoplasmic, and partially localized to Podxl SHD), a linker region and two c-terminal tandem C2 domains vesicles in early aggregates and subapically in mature cysts ( Fig. 3a, (phospholipid and/or protein interaction sites)20,21. Slp2-a could bottom panels). Deletion of the C2 domains (GFP –Slp2-a1C2A/B) potentially therefore connect Rab GTPases and phosphoinositides resulted in a similar localization to the SHD, whereas the linker region

a 12 h 20 h 24 h 48 h C2A/B–GFP C2A/B–GFP Podxl C2A/B–GFP C2A/B–GFP C2A/B–GFP Podxl Podxl Podxl Podxl β-cat β-cat β-cat β-cat

β-cat Merge

C2A/B L

SHD–GFP SHD–GFP Podxl SHD–GFP SHD–GFP SHD–GFP Podxl Podxl Podxl Podxl β-cat β-cat β-cat β-cat

β-cat MergeMerge SHD

L Magnification Magnification Magnification Magnification

b d 12 h 20 h 36 h Slp2-a NSHD C2A C2B C PHD–GFP PHD–GFP Slp2-a C2A/B C2A C2B Cherry–Slp2-a Podxl ∆SHD C2A C2B SHD SHD Podxl Merge ∆C2A/B SHD Linker Magnification Magnification Magnification c e f C2A/B ∆ Ionomycin added GST–Slp2-a t = 15 s t = 30 s t = 40 s t = 50 s M FL ∆C2A/B GST SHD r (K) GST–FL-Slp2-aGST-Slp2-a PHD–GFP 250 LPA S1P Linker 150 LPC PtdIns(3,4)P2 PtdIns(3,5)P P 100 PtdIns 2 P C2A P P PtdIns(3)P PtdIns(4,5)P2 75 P Cherry–Slp2-a P C2B P PtdIns(4)P PtdIns(3,4,5)P3 P 50 PtdIns(5)P PA Podxl PE PS Apical 37 vesicles PC Blank membrane

Figure 3 Slp2-a requires SHD and C2A/B domains for correct localization. acid. PS, phosphatidylserine. (d) Co-localization of Cherry–Slp2-a and

(a) Localization of GFP–Slp2-a C2A/B, and SHD during lumen morphogenesis. PtdIns(4,5)P2 during early cyst formation. MDCK cells stably expressing MDCK cells stably expressing different GFP–Slp2-a constructs were grown Cherry–Slp2-a were transfected with the PtdIns(4,5)P2 probe (PHD–GFP) in 3D to form cysts. Cysts were fixed at different time points (12, 20, 24 and grown in cysts. Cysts were fixed at different time points (12, 20, and 48 h) and co-stained to detect Podxl (red) and β-catenin (β-cat; blue). 36 h) and co-stained to detect Podxl (blue). The arrowheads indicate The arrowheads indicate apical membrane; the arrows indicate localization apical membranes; the arrows indicate cell–cell junction membrane to cell–cell junctions. Scale bars, 5 µm. ( b) Scheme of the Slp2-a constructs localization. Scale bars, 5 µm. ( e) Apical Slp2-a localization depends on used. Different domains and truncated forms of Slp2-a were cloned for PtdIns(4,5)P2. Cysts expressing PHD–GFP (top panels) and Cherry–Slp2-a characterizing Slp2-a function. (c) PIP-binding ability of Slp2-a. GST-tagged (bottom panels), were treated with ionomycin, which stimulates endogenous 1C2A/B full-length Slp2-a (GST–FL-Slp2-a) and 1C2A/B (GST–Slp2-a ), which PLC activity to deplete membrane PtdIns(4,5)P 2, and were analysed by should be unable to bind phospholipids, were expressed and purified video-microscopy (0.1 s exposure every 1 s). Still images at different time in bacteria. PIP-strip membranes were incubated with 1 µg ml −1 of GST points after ionomycin addition are presented. The arrowheads indicate (control), GST–FL-Slp2-a or GST–Slp2-a 1C2A/B and then membranes were apical membrane localization. Scale bars, 10 µm. ( f) Schematic of Slp2-a blotted with anti-GST. A scheme of the PIP-strip membrane is shown. The association with the apical plasma membrane. Slp2-a C2A/B domains arrowheads indicate specific PIP2 binding. The red lines highlight the PIP2 bind PIP2 and localize Slp2-a to the lumen initiation site and the apical species. LPA, lysophosphatidic acid. LPC, lysophosphatydilcholine. membrane. The SHD domain binds apical vesicles. For all panels, areas PI, phosphatidylinositol. PE, phosphatidylethanolamine. PC, outlined in micrographs are magnified in the associated images. Uncropped phosphatidylcholine, S1P, sphingosine-1-phosphate. PA, phosphatidic images of blots are shown in Supplementary Fig. S8. was cytoplasmic (Supplementary Fig. S4). These results suggest that becoming progressively enriched to the lumen in morphogenesis whereas the SHD binds to apical vesicles, the C2 domains target (Fig. 3d). Both PtdIns(4,5)P2 and Cherry–Slp2-a disappeared rapidly Slp2-a to membranes. from the apical membranes on ionomycin treatment, which causes 23 Notably, the distribution of the C2 domains resembles PtdIns(4,5)P 2 PIP depletion at the membrane (Fig. 3e and Supplementary Video localization during cyst formation9. Furthermore, Slp2-a, and par- S1). Taken together, these results confirm that Slp2-a requires alogues, bind selectively to PtdIns(4,5)P 2 (refs 12 ,17 ), although it could the C2 domains for PtdIns(4,5)P 2 binding and apical membrane bind also to phosphatidylserine 16 . We found that C2 domains bound localization, whereas the SHD region targets Slp2-a to apically specifically to PIP 2 species, but not to phosphatidylserine ( Fig. 3c). destined vesicles (Fig. 3f). Lact-C2–GFP, a probe for phosphatidylserine, presents non-polarized membrane localization in cysts (Supplementary Fig. S4e). Given the Slp2-a targets Rab27 vesicles to the lumen initiation site 9 established role of PtdIns(4,5)P 2 in apical membrane specification , to form the lumen 22 and higher cellular abundance , we reasoned that PtdIns(4,5)P 2 Nearly all described functions of mammalian Slps required the may target Slp2-a to plasma membranes. During cyst formation, SHD domain21 . In contrast, Btsz, the sole Slp paralogue in

Slp2-a and PtdIns(4,5)P2 co-localized at the plasma membrane, Drosophila, does not require a Rab-binding domain for epithelial

a D D bsiRNA control c siRNA: Control Slp2-a d SH SH ∆ ∆ 120 siRNA Slp2-a GFP– ∆SHD GFP– ∆SHD

SHD Podxl Podxl 1.5 Rab27a

∆

β-cat β-cat 1.5 MDCK GFP– GFP– 100 Rab27b

Slp2-a Slp2-a (a.u.) l

L (a.u.) ol siRNA: C KD C KD C KD 80 WB: 150 WB: ∗∗ ∗∗ 1.0 1.0 Slp2-a 100 GFP 60 endog. WB: GFP–Slp2-a 50 Tub 40 GFP–Btsz2 GFP–Btsz2 0.5 0.5

Podxl Podxl mRNA quantification mRNA

MDCK Btsz2–GFP 20 β-cat β-cat quantification mRNA normalized to contro to normalized M (K) C-KD C- KD contr to normalized r 0.0 0.0

WB: Percentage of single lumen 0 formation (relative to control) siRNA: 250 GFP siRNA: SHD M (K) ∆ Rab3b r Btsz2–GFP Control Rab27a Rab27b Control WB: Rab27a/b 150 Slp2-a MDCK 100 endog. Btsz2–GFP WB: 50 Tub GFP–Slp2-a f g GST–Slp2-a h siRNA control e siRNA Slp2-a Rab27a/b Rab3b 120 siRNA: Control 120 100 Input WT V18A E11A/R32AGST Mr (K) DNA / GFP– 100

80 WB: GFP Rab27a 50 ∗∗ 80 ∗∗ Podxl / 60 ∗∗ GFP– Rab3b 50 60 ∗∗ -cat

β 40 GFP– 20 Rab8a 50 40 Coomassie 20 Percentage of single lumen 0 formation (relative to control) siRNA: 150 25

Percentage of single lumen 0 Rab3b formation (relative to control) Magnification L ControlRab27aRab27b MDCK WT V18A E11A Rab27a/b R32A GFP–Slp2-a i siRNA: Control Slp2-a j Magnification k GFP–V18A GFP–V18A GFP–Rab27a β-cat GFP–Rab27a GFP–Rab27a GFP–Rab27a Podxl Podxl Ch–Slp2-a -cat -cat Podxl Podxl V18A β β β-cat β-cat L 16 h L Ch–Slp2-a Podxl L Podxl β-cat GFP–Slp2-a siRNA control GFP–E11A/R32A GFP–E11A/R32A GFP–Rab27a β-cat GFP–Rab27a Podxl Podxl Ch–Slp2-a β-cat β-cat Podxl β-cat GFP–Rab27a

E11A/R32A GFP–Rab27a L 20 h Ch–Slp2-a Podxl Podxl β-cat Slp2-a

GFP–Slp2-a Podxl β-cat GFP–Rab27a β-cat GFP–Rab27a Ch–Slp2-a

Podxl siRNA β-cat

24 h Ch–Slp2-a Podxl

Figure 4 Slp2-a binds Rab27 to form the apical membrane. ( a) Knockdown E11A/R32A. GST (control) or GST–Slp2-a (WT, V18A and E11A/R32A) of Slp2-a in cells stably expressing GFP–Slp2-a 1SHD or Btsz2–GFP beads were used to pulldown fluorescent protein-tagged Rab3b, Rab8a (Btsz2) at 72 h after siRNA transfection. C, control; KD, knockdown; or Rab27a from total cell lysates. Bottom lane, Coomassie staining of endog, endogenous; WB, western blot. (b) Quantification of cysts with an independent polyacrylamide gel loaded with GST–Slp2-a constructs normal lumens in cells expressing GFP–Slp2-a 1SHD or Btsz2–GFP and and a representative input. ( h) Quantification of normal cysts in cells transfected with siRNA to Slp2-a or control ( n = 3). (c) Rescue effect expressing GFP–Slp2-a WT, V18A and E11A/R32A mutants transfected of GFP–Slp2-a 1SHD and Btsz2–GFP in cells knocked down for Slp2-a on with siRNA against Slp2-a or control ( n = 3). (i) Images of GFP–Slp2-a lumen formation at 72 h post siRNA transfection. Poxdl, red; β-catenin V18A and E11A/R32A cysts after Slp2-a knockdown. Cysts were fixed (β-cat), blue. Note, Btsz2–GFP localization is not polarized on the 48 h after transfection. Poxdl, red; β-catenin, blue. The arrows indicate plasma membrane of cysts. ( d) Knockdown of Rab27a/b and Rab3b apical membrane. (j) Co-localization of Rab27 and Slp2-a during cyst by siRNA. MDCK cells were transfected with different siRNA duplexes morphogenesis. MDCK cells stably expressing Cherry (Ch)–Slp2-a and targeting canine Rab27a, Rab27b or Rab3b. After 72 h RNA extracts GFP–Rab27a were grown as cysts and fixed after 16, 20, or 24 h. Podxl, were quantified by RT–qPCR; n = 3. ( e) Effect of Rab27a/b or Rab3b blue; β-catenin, white. (k) Effect of the downregulation of Slp2-a on knockdown on cyst formation. Cysts were fixed 48 h after transfection. GFP–Rab27a cells 36 h after siRNA transfection. The arrowheads indicate Silencing of Rab27a/b or Rab3b was sufficient to disrupt cyst formation Rab27a subapical localization; the arrows indicate co-localization of Podxl and accumulate Podxl in vesicles (arrowheads). Podxl, red; β-catenin, and Rab27a. In all panels values are means ± s.d. of n independent green; nuclei, blue. (f) Quantification of cysts with normal lumens in experiments; ∗P < 0.05; ∗∗ P < 0.005; L, lumen; scale bars, 10 µm; areas cells transfected with siRNA targeting Rab27a, Rab27b, Rab27a/b or outlined in micrographs are magnified in the associated images. Uncropped Rab3b (n = 3). (g) Rab–GTPase interaction with Slp2-a mutants V18A and images of blots are shown in Supplementary Figs S8 and S9. morphogenesis12 (Supplementary Fig. S5b), suggesting Rab–Slp and an SHD-deleted Slp2-a mutant (GFP –Slp2-a1SHD). Importantly, interactions may be dispensable for epithelial polarity. To address neither GFP–Slp2-a1SHD nor Btsz2–GFP was able to rescue the this possibility, we expressed the epithelial Btsz protein (Btsz2 –GFP) defects caused by endogenous Slp2-a knockdown (Fig. 4a–c and

Supplementary Fig. S4b) . These results reveal that the SHD region is either alone or in tandem, localized to apical and basolateral plasma required for epithelial morphogenesis in MDCK cysts. membranes (Fig. 5f, left panels, Supplementary Fig. S6e and Table 1), To determine which Rab GTPase interactions are required for Slp2-a suggesting that they confer nonspecific plasma membrane localization. function, we analysed Rabs that interact with other Slps (ref. 10 ). In support of this, GST –Slp4-a C2 domains bound promiscuously

Slp2-a bound to Rab3b, Rab8a, Rab27a and to a lower extent Rab3a to PtdIns(4,5)P 2 and PtdIns(3,4,5)P3 (Fig. 5g,h and Supplementary (Supplementary Fig. S5A). Although Rab8a/b are required for cyst Fig. S6c). In contrast, the C2A domain of Slp2-a bound specifically 7,18 formation , only knockdown of both Rab27a/b strongly perturbed to PtdIns(4,5)P 2 and localized apically (Supplementary Fig. S4d), lumen formation (Fig. 4d–g), confirming partial isoform redundancy suggesting that only Slp2-a is able to tether Rab27 vesicles to 24 noted from knockout mice . In contrast, although Rab3a-d isoforms apical PtdIns(4,5)P2-enriched plasma membrane. Together, these data are expressed in MDCK (data not shown), silencing of Rab3b was highlight a requisite of the SHD-linker region for apical targeting of sufficient to disrupt cyst formation ( Fig. 4d–g), suggesting Rab3b may Slp4-a, representing a major difference from Slp2-a, and suggesting have subtle non-redundant roles in apical transport 25,26, and epithelial non-redundant roles for Slp2-a and Slp4-a in lumen formation. polarity. These results suggested that Slp2-a could mediate the targeting of apical vesicles loaded with Rab27a/b, Rab3b and/or Rab8a/b. Slp4-a apical localization and function depend on Rab and Next, we generated Slp2-a SHD mutants to disrupt the interaction syntaxin interaction with specific Rabs based on the structure of the Slp2-a/Rab27 Slp4-a functions in docking of secretory granules with the plasma interaction20 . The introduction of a V18A mutation in Slp2-a membrane, a function modulated by the SHD region, which interacts completely abolished the interaction with Rab3b and Rab8a, while with Rab3, Rab8 and Rab27 family members10 . Therefore, we examined preserving Rab27a binding; E11A/R32A mutations also disrupted the contribution of Rab binding to Slp4-a function. the binding to Rab27 ( Fig. 4g). Although both mutants retain To elucidate the role of Rab binding, we examined the ability of apical localization (Fig. 4i and Table 1), GFP –Slp2-aV18A, but not SHD mutations to bind to Rab27/3/8 and to rescue the phenotype GFP–Slp2-aE11A/R32A, completely rescued the Slp2-a knockdown of endogenous Slp4-a knockdown (Fig. 6a,b)27 . In contrast to apical phenotype (Fig. 4h,i and Supplementary Fig. S5f). Together, these data wild-type (WT)-Slp4-a, removal of the Rab-interacting region of indicate that although Slp2-a can bind multiple Rabs, Rab27a/b is Slp4-a1SHD resulted in both cytoplasmic and non-polarized membrane necessary and sufficient for Slp2-a function in lumen morphogenesis. localization of Slp4-a ( Fig. 6c, Supplementary Fig. S6e and Table 1), as Next, we analysed Rab27 localization. Before lumen formation, for Slp4-aW118S (no Rab binding) expression (Fig. 6a–c, Supplementary GFP–Rab27a co-localized with Podxl and Slp2-a in vesicles tran- Fig. S6f and Table 1). Uncoupling of Slp4-a from Rab3b (V21A) scytosing to the lumen ( Fig. 4j, top and middle panels and Supple- resulted in subapical localization ( Fig. 6c and Supplementary Fig. S6f). mentary Fig. S5c). Once lumen initiation was completed, Rab27a In contrast, co-uncoupling of Rab8 and Rab3b (I18A) resulted in localized to a subapical compartment, whereas Slp2-a localized the targeting of a pool of Slp4-a to the basolateral membrane, in apically (Fig. 4j, bottom panels and Supplementary Fig. S5c,e). Finally, addition to subapical vesicles ( Fig. 6a–c, Supplementary Fig. S6f and Slp2-a knockdown caused the scattering of Rab27a vesicles close Table 1). Whereas WT-Slp4-a was able to rescue endogenous Slp4-a to the plasma membrane ( Fig. 4k, bottom panels). Thus, Slp2-a is knockdown, none of the Rab-binding mutants was able to restore required to localize Rab27a. Taken together, these results indicate lumen formation (Fig. 6b,c and Table 1). These results indicate that that Slp2-a binds to Rab27a-loaded apical vesicles and targets them Rab27, Rab8 and Rab3 binding are required for Slp4-a localization to initiate the lumen. and function. The different localization of the mutants also suggests that Rab27 is required for Slp4-a targeting to vesicles, Rab8 would Slp4-a also functions in lumen biogenesis be necessary to exclude Slp4-a from the basolateral membrane, and In addition to Slp2-a, mammalian cells express four other Slp-family Rab3 may be important for subsequent Slp4-a transport to the proteins (Slp1-5), and four closely related Slac2s (Slp homologue lumen initiation membrane. lacking C2 domains) 10 . To determine whether other Slp proteins Next we analysed the role of phospholipid binding for Slp4-a, function in epithelial polarization, we analysed their expression during using a phospholipid-uncoupled mutant (K > Q; ref. 16 ). Slp4 -aK>Q lumen formation. Notably, whereas Slp2-a was the sole Slp upregulated localized to subapical puncta of apical membranes (Supplementary Fig. in 3D at early times ( Fig. 5a, 3D-14 h), Slp1 and Slp4-a were upregulated S6g and Table 1), and conferred a partial rescue to morphogenesis at later times ( Fig. 5a, 3D-36 h). (Fig. 6b), suggesting that although the C2 domains can confer In contrast to Slp1, Slp4-a silenced cysts presented acute defects with membrane localization, interaction of Slp4-a with other factors (Rabs the formation of multiple lumens and internal vesicles ( Fig. 5b,c). In and SNARE complexes) may partially compensate for the lack of addition, Slp4-a knockdown did not affect polarity or ciliogenesis in C2 domain function. monolayers (Supplementary Fig. S3a,b). Moreover, we observed that Slp4, in contrast to other Slps, binds to SNARE proteins by its linker Slp2-a is specifically induced before Slp4-a in 3D cultures ( Fig. 5d), domain28,29. To analyse SNARE binding, we generated a chimaeric suggesting that Slp2-a is required earlier than Slp4 in lumenogenesis. construct of Slp4-a bearing a Slp5 linker domain ( Slp454) that was We next examined the localization of Slp4-a. Endogenous Slp4a/b unable to bind to Stx3 ( Fig. 6d). Slp454 showed a subapical and and GFP–Slp4-a associated with apical membranes at all stages of basolateral localization, and failed to rescue lumen formation on polarization (Fig. 5e and Supplementary Fig. S6d), thus presenting endogenous Slp4-a knockdown (Fig. 6e,f and Supplementary Fig. S6i a different localization pattern from Slp2-a ( Fig. 3 and Table 1). Next, and Table 1), suggesting an important role for Stx3 binding to we examined Slp4-a domains in cysts. The C2 domains (C2A/B), Slp4-a. Next, we mapped the Stx3-binding domain of Slp4-a, and

a 4 ∗∗ b shRNA: Control Slp1 Slp4-a c 150 β-cat ∗ ∗ 2D 36 h ∗ 3D 14 h Podxl 3 3D 36 h DNA 100

2 L L 50 ∗∗ ∗∗ lumen formation

1 (relative to control) Percentage of single Amount of mRNA

(relative to control) 0 shRNA: 0 Slp1 Control Slp4-1Slp4-2 Slp1 Slp5 Slp2-a Slp4-a Slac2A Slac2B 3 d g 2 2 2 )P

6.0 3.0 Slp2-a 3D Slp4-a 5.0 PtdIns PtdIns(3)PPtdIns(4)P PtdIns(5)P PtdIns(3,4)PPtdIns(3,5)PPtdIns(4,5)PPtdIns(3,4,5PS Slp2-a 2D 2.5 Slp2-a D 14 h) 14 D 4.0 Slp4-a 3D C2A Slp4-a 2D GST– Slp2-a

3.0 2.0 C2B tive to 2 to tive 2.0 3D/2D (a.u.) 1.5 RNA quantification quantification RNA 1.0 C2A

m GST–

(a.u. rela (a.u. Slp4-a 0.0 mRNA relative expression 1.0 C2B 14 h 36 h 72 h 5 d 7 d 14 h 36 h 72 h 5 d 7 d e 12 h 20 h f C2AB SHD h Slp4-a Merge Slp4-a Merge Slp4-a Podxl DNA Slp4-a Podxl DNA Podxl Podxl Relative PIP2/PIP3 binding β-cat β-cat 3

1 Slp4-a Podxl Slp4-a Podxl Merge Merge L L 0 Relative binding (a.u.) C2A C2B C2A C2B GST–Slp2-a GST–Slp4-a

24 h 48 h Slp4-a Slp4-a Relative PIP2/PS binding Slp4-a Merge Slp4-a Merge Podxl Podxl 60 β-cat DNA 40

L Magnification 20 Podxl Podxl 0 Slp4-a Podxl Slp4-a Podxl Relative binding (a.u.) C2A C2B C2A C2B GST–Slp2-a GST–Slp4-a

Figure 5 Slp4-a is required for epithelial morphogenesis. ( a) Analysis of (n = 3). (e) Localization of GFP–Slp4-a in stably expressing cells during Slps and Slac2s expression in 2D versus 3D. Slp1, Slp2-a, Slp4-a, Slp5, lumen formation. Cysts were stained for Podxl (red) and β-catenin (blue). Slac2A and Slac2B expression was evaluated at different time points (14 The arrowheads indicate Slp4-a co-localization with Podxl. (f) Localization and 36 h) by RT–qPCR in MDCK cells grown in 2D and 3D. Slp3, Slac2C of GFP–Slp4-a C2A/B and SHD during lumen morphogenesis. Cysts were and the related Rab27 effectors Noc2 and rabphilin3 were not expressed fixed at 96 h and co-stained to detect Podxl (red) and β-catenin (blue). The in MDCK ( n = 4). (b) Effect of Slp4-a silencing on lumen formation. Cells arrows indicate the basolateral plasma membrane. (g) Phospholipid-binding stably expressing Slp4, Slp1 or scramble shRNA were plated to form cysts ability of Slp2-a and Slp4-a C2 domains. Purified GST-tagged Slp2-a and for 72 h. Cells were stained to detect Podxl (red), β-catenin (β-cat; green) Slp4-a C2A and C2B domains were incubated with beads covered with and nuclei (blue). The arrowheads indicate intracellular Podxl vesicles. phosphoinositides or phosphatidylserine (PS). ( h) Quantification of relative (c) Quantification of cysts with normal lumens in cells expressing scramble, binding of phosphoinositides and PS to Slp2-a and Slp4-a C2 domains. The

Slp1 or Slp4-a shRNA (n = 3). (d) Quantification of Slp2-a and Slp4-a panels show the PIP2/PIP3 (top) and PS /PIP2 binding ratios (bottom). Note, mRNA in cells grown on filters or in Matrigel. MDCK cells were grown the Slp2-a C2A domain binds mainly to PIP 2, whereas other C2 domains on filters to confluence (2D) or in Matrigel (3D). mRNA expression was show similar binding abilities to PIP 2, PIP 3 or PS. In all panels values are evaluated at different times by RT–qPCR. Data were normalized to 2D levels means ± s.d. of n independent experiments; ∗P < 0.05; ∗∗ P < 0.005; L, at 14 h post-plating. Left panel, Slp2-a (blue lines) and Slp4-a (green lines) lumen; scale bars, 10 µm; areas outlined in micrographs are magnified in the mRNA expression patterns in 2D (solid lines) or 3D (dashed lines). Right associated images. Uncropped images of blots are shown in Supplementary panel, mRNA expression as the 3D/2D coefficient at different time points Fig. S9.

a c WT-Slp4-aSlp4-aI18A Slp4-aV21A Slp4-aW118S Slp4-a∆SHD GST–Slp4-a GFP–Slp4-a Podxl Input SHD β-cat WT I18A V21A W118S ∆ GST Mr (K) GFP– Rab27a 50 WB:

GFP– anti-GFP rescue Rab3b 50 WB: GFP–Slp4-a Podxl Rab8a 25 Slp4-a anti-Rab8a M (K) r Coomassie

150 cation

100 25 shRNA β-cat Merge

agnifi M b shRNA Slp4-a f shRNA Slp4-a shRNA control 120 Input IP: anti-T7 GFP–Slp454a d e Podxl 100 β-cat Control ∗ Slp454 120 shRNA Slp4-a 80 ∗∗ ∗∗ –Slp4 ∗∗ T7 T7–Slp454Mock T7–Slp4T7– Mock 100 ∗∗ 60 FLAG– 80 ∗∗ Munc18a WB: ∗∗ 40 60 ∗∗ anti-FLAG Slp454a Podxl FLAG– 40 20 Stx3 20

WB: lumen formation

0 (relative to control)

Percentage of single 0 Percentage of single lumen ∆SHD anti-T7 β-cat Merge formation (relative to control) MDCK WTI18A V21A W118S K>Q Slp454a Slp4∆305-354 GFP–Slp4-a MDCK g 12h 24h 36h h i Control shRNA Stx3 j shRNA: ControlStx3 GFP–Slp4-a GFP–Slp4-a GFP–Slp4-a Stx3 β-cat β-cat Slp4-a Cherry–Stx3 Cherry–Stx3 Cherry–Stx3 β-cat Podxl Podxl Podxl WB: β-catenin β-catenin β-catenin DNA DNA DNA DNA Stx3 WB: GAPDH KD (%): 100 4 Control k 150 GFP–Slp4-a Cherry–Stx3 GFP–Slp4-a Cherry–Stx3 GFP–Slp4-a Cherry–Stx3 β-cat Podxl β-cat Podxl Slp4-a Podxl 100 ∗∗ β-catenin β-catenin β-catenin DNA DNA DNA 50 lumen formation Slp4-a–KD

(relative to control) 0 Magnification Magnification Percentage of single shRNA: Control Stx3

Figure 6 Slp4-a binding to the plasma membrane, Rabs and Stx3 of the Stx3–Slp4 interaction using cells expressing GFP–Slp454 or is required for apical membrane formation. ( a) Rab–GTPase binding GFP–Slp4-a1305–354 and knocked down for Slp4-a (n = 3). (f) Localization to Slp4-a mutants. Purified GST-tagged Slp4-a (WT, I18A, V21A, of GFP–Slp454 in Slp4-a silenced cells. Podxl, red; β-catenin, W118S and 1SHD) or GST (control) proteins were used to pulldown blue. Arrowheads show vesicular localization; arrows show basolateral fluorescent-protein-tagged Rab3b, Rab8a or Rab27a from total cell membrane. (g) Localization of Cherry–Stx3 and GFP–Slp4-a during lysates. Membranes were blotted with anti-GFP or anti-Rab8a. Bottom cyst development. Cherry–Stx3 co-localized with GFP–Slp4-a at the lane, Coomassie staining of an independent polyacrylamide gel periphery of early aggregates (arrowheads), intracellular vesicles and loaded with different GST–Slp4-a constructs and inputs. WB, western cell–cell contacts. As lumens formed, Cherry–Stx3 concentrated at the blot. (b) Quantification of cysts with normal lumens in GFP–Slp4-a nascent luminal membrane with GFP–Slp4-a (arrowheads). Nuclei (blue). (WT, 1SHD, I18A, V21A, W118S and K > Q) cells expressing (h) Intracellular localization of GFP–Stx3 (arrows) in Slp4-a knockdown scramble or Slp4-a shRNA (n = 3). (c) Rescue effect of GFP–Slp4-a cysts (72 h). β-catenin, red; nuclei, blue. (i) GFP–Slp4-a localization WT, Rab-binding defective mutants or membrane-binding defective in 48 h cysts knocked down for Stx3. Podxl (red), β catenin (green) Slp4-aK>Q. MDCK cysts (72 h) expressing GFP–Slp4-a WT, 1SHD, and nuclei (blue). Note the Slp4-a basolateral mis-localization in the I18A, V21A, W118S or K > Q were knocked down (KD) for Slp4-a, Stx3 knockdown cysts (yellow arrows). The arrowheads indicate Podxl and stained for Podxl (red) and β-catenin (β-cat; blue). Arrowheads vesicles. (j) Downregulation of Stx3 in MDCK cells stably expressing Stx3 indicate co-localization with Podxl; arrows indicate co-localization with shRNA. Total lysates were blotted for Stx3 and GAPDH (loading control). β-catenin. (d) Co-immunoprecipitation assay of Slp4a binding to Stx3. (k) Quantification of the effect of Stx3 silencing in cyst formation ( n = 3). T7–Slp4-a or T7–Slp454 beads were incubated with FLAG–Munc18-2 In all panels values are means ± s.d of n independent experiments; and FLAG–Stx3 lysates. FLAG-tagged proteins were detected with ∗P < 0.05; ∗∗ P < 0.005; scale bars, 5 µm; areas outlined in micrographs HRP-conjugated anti-FLAG. Input 1:10 of immunoprecipitate (IP) are magnified in the associated images. Uncropped images of blots are volume. (e) Quantification of cysts with normal lumens in the absence shown in Supplementary Fig. S10. identified amino acids 305–354 to be essential for Stx3 binding Stx3 functions as a critical apical 30 SNARE. In cysts, GFP –Stx3 (Supplementary Fig. S6h). Consistently, expression of a construct co-localized with Slp4-a at the nascent luminal membrane ( Fig. 6g). No- lacking amino acids 305–354 of Slp4-a (GFP –Slp4-a1305–354) also tably, Stx3 knockdown resulted in disruption of lumen formation and failed to localize to the apical plasma membrane and to rescue the redistribution of a pool of GFP –Slp4-a to the basolateral membrane lumen formation on endogenous Slp4-a knockdown (Fig. 6e and (Supplementary Fig. S6i–k), and a similar localization observed on the Supplementary Fig. S6j and Table 1). removal of the Stx-binding region of Slp4-a (Supplementary Fig. S6i,j).

a Slp4-a Slp2-a β-cat b Merge Slp4-a Podxl β-cat h

16 h Control Rab27 Slp2-a

20 h Slp2-a Podxl Rab3

siRNA Magnification c 24 h Merge Slp2-a Podxl β-cat Rab8 Stx3

Podxl Slp4-a Control 16 h Apical PIP2-enriched Slp4-a vesicles membrane

20 h shRNA Magnification d Slp2-a overexpression -cat Slp4-a Slp2-a β PIP2

24 h Stx3 Slp2-a Slp4 Rab27a Magnification Control Single lumen

PIP2 e β-cat Podxl Merge + DNA Slp2-a Magnification siRNA f∗∗ g ∗∗ Stx3 KD Slp2-a 120 ∗∗ 30 NS Slp4 KD Impaired fusion Vesicle retention ∗∗ 100 ∗∗ PIP NS 25 Stx3 2 80 shRNA NS 20 Slp4-a 60 15 Slp4-a 40 10 Delayed and Slp2-a KD mislocalized fusion >1 lumen per cell siRNA 20 5 Plasma membrane 2 apical surfaces/cell ≥

Slp2-a Percentage of cysts with Endosomes Percentage of single lumen formation (relative to control) 0 0 shRNA / / Apical membranes

Slp4-a Slp2-a Slp4-a Control Slp2-a Slp2-a Slp4-a Slp4-a Control Slp2-aSlp4-a Magnification siRNA siRNA shRNAsiRNA shRNAsiRNA shRNA shRNA

Figure 7 Slp2-a regulates Slp4-a targeting to determine single apical plasma membranes. (f) Quantification of cysts with normal lumens in control, membrane formation. (a) Slp2-a and Slp4-a localization during lumen Slp2-a knockdown, Slp4-a knockdown or Slp2-a/4-a double knockdown initiation. Cysts stably expressing GFP–Slp4-a and Cherry–Slp2-a were (n =3). (g) Quantification of cysts presenting two or more apical surfaces per fixed after 16, 20 and 24 h. Podxl (blue, bottom panels) and β-catenin cell in control, Slp2-a knockdown, Slp4-a knockdown or Slp2-a/4-a double (β-cat; blue, top panels). The arrows indicate vesicular Slp4-a; the knockdowns (n =3). In all panels values are mean ± s.d. from n independent arrowheads indicate Slp2-a/Slp4-a co-localization at the nascent luminal experiments; NS, not significant; ∗∗ P < 0.005; scale bars, 5 µm; areas membrane. (b) Effect of Slp2-a knockdown (KD) on Slp4-a localization. outlined in micrographs are magnified in the associated images. ( h) Model Slp4-a localization becomes basolateral after Slp2-a knockdown and of Slp2-a/4-a function in epithelial polarization. Top panel, Slp2-a targets co-localizes partially with Podxl (red) in vesicles (arrows). The arrowheads Rab27-positive endosomes to the PIP 2-enriched membrane. Slp4-a binds to indicate scattered Podxl vesicles. (c) Effect of Slp4-a knockdown on Rab3 and Stx3 to be delivered to the lumen initiation site in Rab27-positive Slp2-a localization. After Slp4-a knockdown, Slp2-a localization at cellular vesicles. As Slp4-a is delivered in Rab27-positive vesicles, its targeting junctions is unaffected. Note the accumulation of Podxl (red) in vesicles depends on Slp2-a function. Therefore, Slp2-a directs localization of the

(arrowheads). (d) Effect of Slp2-a overexpression on GFP–Slp4-a and Slp4-a/Stx3-inﬂuenced vesicle tethering activity to the single PIP 2-enriched GFP–Rab27 in 24 h cysts. Slp4-a or Rab27a co-localized with Slp2-a initiation site, and thus a single lumen per cell. Bottom panel, when and β-catenin (blue) at cellular junctions (arrows). ( e) Effect of double Slp4-a or Stx3 are perturbed, vesicles cannot be correctly tethered and Slp2-a/Slp4-a knockdown on lumen formation. Cells knocked down for apical vesicles accumulate. When Slp2-a is disrupted, vesicles are tethered Slp4-a, Slp2-a or Slp4-a/Slp2-a for 48 h were ﬁxed and stained for nuclei ectopically to different positions of the plasma membrane, resulting in (blue), Podxl (red) and β-catenin (green). The arrowheads indicate apical multiple apical domains in the same cell.

In contrast, Slp4-a knockdown did not disrupt the apical localization lumen. We thus examined their localization during lumen formation. of Stx3 ( Fig. 6h). These data suggest that Stx3 association with Slp4-a, In early aggregates, Slp2-a localized mainly to cell –cell junctions, by interaction with the linker domain, directs recruitment of Slp4-a to whereas Slp4-a co-localized with Podxl at the cell –extracellular matrix apically destined vesicles, and initiation of de novo lumen formation. interface (Fig. 7a, 16 h, top panels and Supplementary Fig. S7c). As internalized Podxl transcytosed to the cell –cell contact, Slp4-a Slp2-a regulates Slp4-a function to produce a single apical co-localized to these vesicles, whereas Slp2-a was mainly at the cell –cell surface per cell junctions (Fig. 7a, 20 h and Supplementary Fig. S7c). Finally, Slp2-a/4-a Our data thus far indicate that Slp2-a and Slp4-a function in co-localization was evident once lumens formed (Fig. 7a, 24 h and distinct, non-redundant steps in Rab-dependent transport to form the Supplementary Fig. S7c).

Next we examined whether their activities were mutually dependent. controls the clustering of apically destined vesicles through the

Knockdown of Slp2-a caused the scattered distribution of small interaction with Rab27a/b and the association with PtdIns(4,5)P2 at Podxl vesicles and the redistribution of GFP –Slp4-a to the lateral the nascent apical membrane. Subsequently, Slp4-a acts as an effector plasma membrane, partially co-localizing with Podxl vesicles (Fig. 7b, of Rab27a/b, Rab8a/b and Rab3b on vesicles, to couple the vesicles to arrows). In contrast, Slp4-a knockdown cysts presented clusters Stx3-mediated fusion events at the plasma membrane to create the of Podxl vesicles ( Fig. 7c), but GFP–Slp2-a localized normally to lumen. Inhibition of this Slp2-a/4-a pathway perturbs this transport cell–cell contacts, suggesting that Slp2-a localization and function are and ultimately causes the formation of multiple apical membranes. independent, or upstream, of Slp4-a. Furthermore, overexpression of These data support our model that Slp2-a and Slp4-a function in a GFP–Slp2-a forced endogenous Rab27a and Slp4-a mis-recruitment spatiotemporal cascade to control vectorial apical transport ( Fig. 7h), to cell –cell contacts at early time points ( Fig. 7d), whereas the a fact supported by their sequential transcriptional upregulation converse effect of GFP –Rab27a on Slp2-a was not observed (not during cyst formation. shown). These results suggest that Slp2-a functions to regulate the An interesting question is how Slp2-a/4-a may coordinate this positioning of Rab27 vesicles upstream of Slp4-a-mediated vesicle vectorial transport to form a single lumen. In non-polarized cells, both docking, thus controlling the position of the apical membrane Slp2-a and Slp4-a are considered as negative regulators of secretion, and subsequent lumen. on the basis of the fact that their overexpression attenuates secretory To test this hypothesis, we silenced Slp2-a and Slp4-a alone or granule release15–17,27,32. Indeed, Slp4-a can interact with the closed together (Fig. 7e–g). Strikingly, we observed that whereas Slp4-a (non-fusion-forming) conformation of SNARE complexes 29 . To this knockdown induced accumulation of Podxl vesicles close to the end, transient overexpression of Slp2-a or Slp4-a in the presence of membrane (Fig. 7e, middle panels), Slp2-a knockdown cysts possessed endogenous protein consistently reduced single-lumen-formation rates some cells simultaneously developing more than one apical membrane (Supplementary Fig. S7A,B), a trend that could be strongly reversed by (Fig. 7e, top panels, quantification Fig. 7g). Moreover, although expression of SHD-deleted Slp2-a/4-a, suggesting that they may act as dual knockdown cysts presented a mixture of both phenotypes negative regulators of vesicle trafficking. However, Slp2-a/4-a are also (Fig. 7e, bottom panels), they perturbed single lumen formation clearly required for single lumen formation, thus suggesting a scenario to a level resembling that of Slp4-a knockdown alone, supporting where, rather than being considered as negative or positive regulators the notion that Slp4-a functions downstream of Slp2-a ( Fig. 7f). of exocytosis, Slp2-a/4-a act as molecular traffic wardens 32 , controlling Although cysts with multiple lumens have been observed previously vectorial exocytosis through ensuring vesicles dock and fuse only at on knockdown of trafficking proteins, the knockdown of Slp2- singular membrane domains to form a single, coordinated luminal a is unique in that this is the first time, to our knowledge, space between neighbouring cells. that a cell can participate simultaneously in the generation of We identified, for the first time, a transcriptionally regulated multiple luminal surfaces. molecular pathway that controls the formation of a single apical surface These results indicate that Slp2-a and Slp4-a form part of a core per cell, addressing a major, long-term unanswered question in cell apical transport pathway that controls the positioning of Rab27a/b biology. The study of the transcriptional machinery responsible for vesicles, and their subsequent Rab8/Rab3b/Stx3-dependent fusion lumenogenesis in vivo presents a major future challenge to both cell with the apical plasma membrane, respectively, to form a single and developmental biology.

PtdIns(4,5)P2-enriched apical membrane and lumen during de novo apical domain biogenesis (Fig. 7h). METHODS Methods and any associated references are available in the online DISCUSSION version of the paper. One of the central, unsolved questions in epithelial biology concerns Note: Supplementary Information is available in the online version of the paper how apical–basolateral polarity is coordinated between neighbouring cells to form a common, single luminal region. Making use of the fact ACKNOWLEDGEMENTS that MDCK can undergo polarization into either in 2D or 3D, we have We thank C. M. Ruiz-Jarabo for comments on the manuscript, and members of the uncovered a gene set that specifically facilitates the transition to the Martin-Belmonte laboratory for discussion. We thank M. ter Beest, J. Peränen, and 3D architecture. Interestingly, most of this gene set is downregulated K. Simons, Max Planck Institute of Molecular Cell Biology and Genetics (MPI-CBG), Dresden, Germany, for generous gifts of reagents, and the Mostov laboratory for in some glandular epithelial cancers, indicating a potential clinical kind assistance. This work was supported by grants from the Human Frontiers relevance in maintaining a polarized phenotype ( Fig. 1). Science Program (HFSP-CDA 00011/2009), Marie Curie (IRG-209382), MICINN During the development of 3D polarity, apical membrane (BFU2008-01916), (BFU2011-22622) and CONSOLIDER (CSD2009-00016) to F.M-B.; by NIH R01DK074398, R01AI25144 and R01DK91530 to K.M., and The components are delivered to the site for lumen initiation, between March of Dimes Basil O’Connor Starter Research Award to P.R.B. A.E.R-F. is the neighbouring cells, to initiate a luminal space de novo 7,31. What had recipient of a JAE fellowship, from CSIC; M.G-S. is the recipient of a FPI fellowship, remained unclear, though, was how this trafficking pathway was from MICINN; and I.B-R. is the recipient of an AECC fellowship. An institutional Grant from the Fundación Ramón Areces to CBMSO is also acknowledged. organized to direct vectorial transport to a singular position, thus allowing formation of a single apical membrane per cell. Here, we AUTHOR CONTRIBUTIONS have identified a molecular pathway specifically upregulated during M.G-S., A.E.R-F., D.M.B., S.V. and F.M-B. designed the experiments. M.G-S., A.E.R- 3D morphogenesis that controls the formation and positioning of F., D.M.B., S.V., T.S., I.B.R., I.B., A.D., N.S., K.Y. and C.L.S. did the experimental work. M.G-S., A.E.R-F., D.M.B., K.E.M. and F.M-B. analysed the experiments. a single apical membrane per cell, through the complementary P.R.B. and M.F. provided reagents. F.M-B., D.M.B. M.G-S. and A.E.R-F. wrote the functions of Slp2-a and Slp4-a ( Fig. 7h). We demonstrate that Slp2-a manuscript.

COMPETING FINANCIAL INTERESTS 16. Kuroda, T. S. & Fukuda, M. Rab27A-binding protein Slp2-a is required for peripheral The authors declare no competing financial interests. melanosome distribution and elongated cell shape in melanocytes. Nat. Cell Biol. 6, 1195–1203 (2004). 17. Yu, M. et al . Exophilin4/Slp2-a targets glucagon granules to the plasma membrane Published online at www.nature.com/doifinder/10.1038/ncb2541 through unique Ca2+-inhibitory phospholipid-binding activity of the C2A domain. Reprints and permissions information is available online at www.nature.com/reprints Mol. Biol. Cell 18, 688–696 (2007). 18. Sato, T. et al . The Rab8 GTPase regulates apical protein localization in intestinal 1. Mostov, K., Su, T. & ter Beest, M. Polarized epithelial membrane traffic: conservation cells. Nature 448, 366–369 (2007). and plasticity. Nat. Cell Biol. 5, 287–293 (2003). 19. Fukuda, M., Saegusa, C. & Mikoshiba, K. Novel splicing isoforms of synaptotagmin- 2. Rodriguez-Boulan, E., Musch, A. & Le Bivic, A. Epithelial trafficking: new routes to like proteins 2 and 3: identification of the Slp homology domain. Biochem. Biophys. familiar places. Curr. Opin. Cell Biol. 16, 436–442 (2004). Res. Commun. 283, 513–519 (2001). 3. Bryant, D. M. & Mostov, K. E. From cells to organs: building polarized tissue. Nat. 20. Chavas, L. M. et al . Elucidation of Rab27 recruitment by its effectors: structure of Rev. Mol. Cell Biol. 9, 887–901 (2008). Rab27a bound to Exophilin4/Slp2-a. Structure 16, 1468–1477 (2008). 4. Datta, A., Bryant, D. M. & Mostov, K. E. Molecular regulation of lumen 21. Fukuda, M. Versatile role of Rab27 in membrane trafficking: focus on the Rab27 morphogenesis. Curr. Biol. 21, R126–R136 (2011). effector families. J. Biochem. 137, 9–16 (2005). 5. Lubarsky, B. & Krasnow, M. A. Tube morphogenesis: making and shaping biological 22. Di Paolo, G. & De Camilli, P. Phosphoinositides in cell regulation and membrane tubes. Cell 112, 19–28 (2003). dynamics. Nature 7112, 651–657 (2006). 6. Roland, J. T. et al . Rab GTPase-Myo5B complexes control membrane recycling and 23. Zoncu, R. et al . Loss of endocytic clathrin-coated pits upon acute depletion epithelial polarization. Proc. Natl Acad. Sci. USA 108, 2789–2794 (2011). of phosphatidylinositol 4,5-bisphosphate. Proc. Natl Acad. Sci. USA 104, 7. Bryant, D. M. et al . A molecular network for de novo generation of the apical surface 3793–3798 (2007). and lumen. Nat. Cell Biol. 12, 1035–1045 (2010). 24. Bolasco, G. et al . Loss of Rab27 function results in abnormal lung epithelium 8. Gassama-Diagne, A. et al . Phosphatidylinositol-3,4,5-trisphosphate regulates the structure in mice. Am. J. Physiol. Cell Physiol. 300, 466–476 (2011). formation of the basolateral plasma membrane in epithelial cells. Nat. Cell Biol. 25. van IJzendoorn, S. C., Tuvim, M. J., Weimbs, T., Dickey, B. F. & Mostov, K. E. 8, 963–970 (2006). Direct interaction between Rab3b and the polymeric immunoglobulin receptor 9. Martin-Belmonte, F. et al . PTEN-mediated apical segregation of phosphoinositides controls ligand-stimulated transcytosis in epithelial cells. Dev. Cell 2, controls epithelial morphogenesis through Cdc42. Cell 128, 383–397 (2007). 219–228 (2002). 10. Kuroda, T. S., Fukuda, M., Ariga, H. & Mikoshiba, K. The Slp homology domain of 26. Schlülter, O. M., Schmitz, F., Jahn, R., Rosenmund, C. & Südhof, T. C. A complete synaptotagmin-like proteins 1-4 and Slac2 functions as a novel Rab27A binding genetic analysis of neuronal Rab3 function. J. Neurosci. 24, 6629–6637 (2004). domain. J. Biol. Chem. 277, 9212–9218 (2002). 27. Fukuda, M., Kanno, E., Saegusa, C., Ogata, Y. & Kuroda, T. S. Slp4-a/granuphilin-a 11. Ishii, N. et al . A case of recurrent gall bladder cancer responding to chemotherapy regulates dense-core vesicle exocytosis in PC12 cells. J. Biol. Chem. 277, with gemcitabine after endoscopic metallic biliary stent implantation. Gan To Kagaku 39673–39678 (2002). Ryoho 35, 1403–1405 (2008). 28. Tsuboi, T. & Fukuda, M. The Slp4-a linker domain controls exocytosis 12. Pilot, F., Philippe, J. M., Lemmers, C. & Lecuit, T. Spatial control of actin through interaction with Munc18-1.syntaxin-1a complex. Mol. Biol. Cell 17, organization at adherens junctions by a synaptotagmin-like protein. Nature 442, 2101–2112 (2006). 580–584 (2006). 29. Fukuda, M., Imai, A., Nashida, T. & Shimomura, H. Slp4-a/granuphilin-a interacts 13. Bagnat, M., Cheung, I. D., Mostov, K. E. & Stainier, D. Y. Genetic control of single with syntaxin-2/3 in a Munc18-2-dependent manner. J. Biol. Chem. 280, lumen formation in the zebrafish gut. Nat. Cell Biol. 9, 954–960 (2007). 39175–39184 (2005). 14. Brady, D. C., Alan, J. K., Madigan, J. P., Fanning, A. S. & Cox, A. D. The transforming 30. Low, S. H. et al . The SNARE machinery is involved in apical plasma membrane ρ family GTPase Wrch-1 disrupts epithelial cell tight junctions and epithelial trafficking in MDCK cells. J. Cell Biol. 141, 1503–1513 (1998). morphogenesis. Mol. Cell Biol. 29, 1035–1049 (2009). 31. Schluter, M. A. et al . Trafficking of Crumbs3 during cytokinesis is crucial for lumen 15. Holt, O. et al . Slp1 and Slp2-a localize to the plasma membrane of formation. Mol. Biol. Cell 20, 4652–4663 (2009). CTL and contribute to secretion from the immunological synapse. Traffic 9, 32. Gomi, H., Mizutani, S., Kasai, K., Itohara, S. & Izumi, T. Granuphilin molecularly 446–457 (2008). docks insulin granules to the fusion machinery. J. Cell Biol. 171, 99–109 (2005).

METHODS linker domain (1305–354) was cloned into pEGFP-C1. All constructs were verified Two-step functional screening (microarrays, RT –qPCR, RNAi, Oncomine). by sequencing. A microarray-based differential expression analysis was conducted using the Affymetrix Canine Genome 2.0 platform. MDCK type II cells were grown in P100 Cells and 3D culture. T23-MDCKII and MDCKII cells were grown as described dishes to form 2D monolayers or 3D cysts in Matrigel (at 10 5 cells ml −1). Total previously9. MDCK cells stably expressing GFP –Slp2-a (full length and mutants), RNA was isolated at 36 h and purified using RNeasy (Qiagen) and 5 µg of RNA Cherry–Slp2-a, GFP–Rab27a, GFP–Rab8a and PLCδ-PH–GFP (PHD–GFP) were was submitted for microarray analysis (n = 3) using the Affymetrix platform at made by co-transfection with the blasticidin-resistant gene and selected for ten days Parque Científico de Madrid (Cantoblanco). The raw microarray data are deposited with 0.5 µg ml −1 blasticidin. MDCK cells stably expressing GFP–Slp1, GFP–Stx3 in NCBI GEO (Gene Expression Omnibus), accessible online using the number or GFP –Slp4-a (full length and mutants) were selected for ten days using G418 GSE32495. A LiMMA (false discovery rate < 0.05) analysis revealed a set of 1,597 (0 .5 mg ml −1). Cysts and Transwell cultures were prepared as described before 34 . upregulated genes. The resulting data set was analysed to select a maximum of 100 upregulated genes for qPCR validation. To further examine this selection, we Microscopy. Immunofluorescence microscopy of cysts was performed as previ- used bioinformatic/bibliographic searches to analyse all 1,597 upregulated genes ously described9,34. Per condition, >100 cysts per experiment were analysed. For for previously published references into function, and Gene Ontology terms from early time points, cysts were grown up to 24 h and two/three-cell stage cysts human or mouse orthologues in NCBI databases. We selected genes with Gene were classified on the basis of Podxl and β-catenin localization either as: formed Ontology terms, or references into function related to processes or mechanisms preapical-patch or presence of internal vesicles. involved in changes in cell signalling, cell architecture and organ morphogenesis. The list included a comprehensive list of Gene Ontology terms related to cell RNAi. The RNAi sequences and qPCR primers are listed in Supplementary Table S3. polarity, membrane trafficking, cell-to-cell junction assembly and remodelling, Briefly, 25 nucleotide stealth siRNA duplexes targeting messenger RNA sequences cell cycle regulation, cytoskeleton regulation and cell division, among others of canine Slp2-a were purchased from Invitrogen. Sequences were submitted to (complete list on demand). The second selection approach used STRING software BLAST search to ensure targeting specificity. For siRNA transfection, MDCK cells (http://string-db.org/) to select genes interacting with pathways previously known were trypsinized and then nucleofected (Lonza) with siRNA duplexes or scrambled to have a role in epithelial architecture or morphogenesis. From the resulting list, siRNA. After 24-h incubation, cells were resuspended and plated in 12-well plates we selected 99 genes on the basis of bibliographic research and designed specific and in coverglass chambers coated with Matrigel to grow cysts. Total cell lysates primers to perform qPCR analysis validation of their overexpression pattern in 3D from 3D cultures were analysed by western blotting or RT-qPCR to confirm the cyst formation. siRNA efficiency. After qPCR validation, a stealth siRNA library was custom designed to target 47 Stable RNAi was achieved by viral short hairpin RNA (shRNA), essentially as validated candidate genes (Invitrogen). To perform the siRNA screening, MDCK previously described7. In all instances, knockdown was verified by western blot or cells were transfected with siRNA using Nucleofector-II (Lonza). Transfected cells RT-qPCR procedures (Brilliant-II SYBR Green Kit, Agilent), and normalizing to were cultured for two days in 3D conditions and RNA extracts were analysed by RT- GAPDH expression. RNAi and RT-qPCR primers are presented in Supplementary qPCR to check the silencing efficiency (Supplementary Table S2). Gene expression Table S3. Stx3 shRNA is as described previously 35 . Slp1 and Slp4-a shRNA silencing was verified by RT –qPCR procedures (SYBR RT-qPCR premix, Applied lentiviruses were constructed in pLKO.1-puro according to the Addgene pLKO.1 Biosystems), and normalizing to GAPDH or HPRT expression. For functional protocol (www.addgene.org) using iRNAi (www.mekentosj.com), and target analyses, transfected MDCK cells were grown for three days in Matrigel to form sequences were based on an (AA)N19 algorithm. RNAi sequences were submitted cysts, and lumen formation efficiency was quantified by confocal microscopy using to BLAST (NCBI) to verify target specificity, with SYTL4 sequences targeting the following markers to assess lumen formation: localization of the apical protein common regions to Slp4-a and Slp4-b transcripts. GFP-tagged human Slp4-a, which Podxl, integrity of the actin cytoskeleton (F-actin; phalloidin), adherens junctions is not targeted by anti-canine shRNA, was used for Slp4-a knockdown and rescue (β-catenin), tight junctions (ZO-1), nuclei (DNA; DAPI). experiments.

Antibodies. Antibodies against α-tubulin (1:5,000; T9026, Sigma-Aldrich), Rab27a Virus production and transduction. Lentivirus production was performed essen- (1:200; R4655, Sigma-Aldrich), GFP (1:500; a5455, Invitrogen), Rab8a (1:1,000, tially as previously described 7. For lentivirus transductions, subconfluent MDCK 610845, BD Biosciences), mRFP/Cherry (1:250; PM005, MBL), Stx3 (1:200; Ab4113, cultures, 1–4 h after plating, were infected with virus-containing supernatants for Abcam), GST (1:5,000; sc138), β-catenin (1:1,000, sc7199) and Slp4-a (1:100; 34448) 12 –16 h at 37 ◦C. Viral supernatants were then diluted 1:1 with growth medium, from Santa Cruz Biotechnology were commercial primary antibodies. The Slp2-a and cultured for a further 48 h. Transduced cells were selected by passage into antibody was raised as a polyclonal serum against the Slp2-a SHD region and used as appropriate antibiotic-containing medium. Puromycin (5 µg ml −1), and blasticidin previously described33 . Podxl antibody was a gift from the Ojakian laboratory (State (12.5 µg ml −1) were used. University of New York Downstate Medical Center, USA). ZO-1 (1:500; R4076) was from DSHB. Peroxidase-conjugated donkey anti-mouse IgG and anti-rabbit Statistics. Single lumen formation was quantified as previously described 7. The IgG were used as secondary antibodies for western blots (Jackson ImmunoResearch percentage of cysts with a single lumen was determined, and normalized to control Laboratories). Alexa Fluor-conjugated secondary antibodies (Alexa Fluor 405, 488, cysts as 100%. Values are mean ± s.d. from ≥3 replicate experiments, with n ≥ 100 555 or 647; Invitrogen) and TOPRO-3 or DAPI (for nuclear/DNA staining) were cysts per replicate. For RT-qPCR experiments, the percentage of remaining mRNA used in the microscopy protocols. in each knockdown condition was normalized to the HPRT level, and represented as a percentage of the control (scramble shRNA) mRNA levels. The significance was Plasmids. Slp2-a and Slp4-a (full length and mutants) were cloned into calculated using a paired, two-tailed Student’s t-test. ∗P < 0.05, ∗∗ P < 0.001. either pEGFP-C1/C2 or pmCherry-C1 vector (Takara Bio). The human Slp4- a complementary DNA template was from Open Biosystems (Thermo Fisher). Rab–GTPase pulldown. Rab GTPase–Slp protein pulldowns were performed Plasmids provided were: pEGFP-Slp1 (J. Peränen, University of Helsinki, Finland), using HEK293T cells overexpressing GFP-tagged Rab proteins, and GST-tagged pEGFP-STX3a (M. ter Beest, University of Chicago, Illinois, USA), pENTR- Slp proteins. HEK293T cells expressing GFP-tagged Rab proteins were lysed in Rab3a/b/c/d (B. Goud, Institut Curie, Paris, France), pEGFP-Rab8a (M. Montoya, 0.1% SDS, 1% Triton X-100, 0.5 mM dithiothreitol and ×1 TBS buffer with a CNIC, Madrid, Spain) and pEGFP-Rab27a/b (WT and dominant negative) protease inhibitor cocktail and sodium orthovanadate. Cell debris and nuclei were (J. Hammer, NIH, USA). For bacterial expression of GST-tagged full-length and removed by centrifugation at 14,000 g for 2 min at 4 ◦C, and lysates were precleared mutant proteins, Slp2-a and Slp4-a were cloned into pGEX-4T1 vector (Promega) and incubated in rotation with 100 ng of the relevant GST protein-loaded beads or pDEST15 (Invitrogen). Slp2-a (V18A, E11A/R32A; ref. 16), and Slp4-a (V21A, (GE Amersham) for 30 min, using GST alone as the control, in the presence of W118S, I18A, K > Q; refs 17,27) mutants were generated using Quickchange XL a non-hydrolysable GTP analogue (Sigma). Beads were centrifuged and washed (Stratagene). To disrupt C2 domain –lipid interactions in Slp4-a (C2AB K > Q; five times, dried using aspiration, and resuspended in 40 µl Laemmli buffer before K410Q, K412Q, K416Q, K564Q, K566Q, K571Q), three of four lysine residues of western blot analysis. the PIP-binding consensus motifs present in synaptotagmin and synaptotagmin-like family C2 domains [K(K/R)KTXXK(K/R)] were mutated to glutamine in both C2 Co-immunoprecipitation assays. Co-immunoprecipitation assays in COS-7 domains, as reported for Slp2-a (ref. 16). To disrupt Stx interactions in Slp4-a, the cells were performed essentially as described previously 28,29. In brief, pEF-FLAG- linker domain of Slp4-a was substituted for the linker domain of Slp5 (GFP –Slp454a) Stx3, pEF-FLAG-Munc18-2, pEF-T7-Slp454 (a linker domain-swapping construct by subcloning the chimaeric SHD Slp4-linkerSlp5 fragment from a previously reported between Slp4-a and Slp5 (ref. 28)) or pEF-T7-Slp4a linker deletion constructs Slp454b construct28 into a GFP –Slp4-aC2AB plasmid. T7-tagged Slp4-a, Slp454a (that is, linker, amino acids 144–354; F1, amino acids 144–240; F2, amino acids and Slp4-a linkers were used for in vitro Stx3 binding experiments as previously 215–304; and F3, amino acids 272–354) were transfected into COS7 cells by using reported28,29. The Slp4-a construct lacking the Stx3-interacting amino acids of the Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions.

Cells were collected 48 h after transfection and homogenized in a homogenization GST–C2B from Slp2-a or Slp4-a, washed five times, dried and resuspended in buffer28,29. After removal of insoluble materials by centrifugation, cell lysates 100 µl sample buffer before western blot analysis. Enhanced chemiluminescence were obtained. The associations between T7-tagged proteins and FLAG-tagged blotting of membranes was developed by immunostaining with an anti-GST Stx3/Munc18-2 in the cell lysates were evaluated by immunoprecipitation using antibody (Sigma-Aldrich) and HRP-conjugated donkey anti-mouse IgG (Jackson anti-T7 tag antibody-conjugated agarose beads (Merck Biosciences) as described Immunoresearch). previously36 . Immunoreactive bands were visualized with horseradish peroxidase (HRP)-conjugated anti-T7 tag antibody (1:10,000 dilution; Novagen, 69522-4) and HRP-conjugated anti-FLAG tag M2 antibody (1:10,000 dilution; Sigma- 33. Imai, A., Yoshie, S., Nashida, T., Shimomura, H. & Fukuda, M. The small GTPase Aldrich, A8592) and detected by enhanced chemiluminescence (GE Healthcare). Rab27B regulates amylase release from rat parotid acinar cells. J. Cell Sci. 117, The Rab–GTPase co-immunoprecipitation assay was performed as previously 1945–1953 (2004). 34. Rodriguez-Fraticelli, A. E. et al . The Cdc42 GEF Intersectin 2 controls mitotic described7. spindle orientation to form the lumen during epithelial morphogenesis. J. Cell Biol. 189, 725–738 (2010). −1 PIP-strip and lipid bead protein binding assays. A solution of 1 µg ml 35. Schuck, S., Manninen, A., Honsho, M., Fullekrug, J. & Simons, K. Generation of of purified protein (GST, GST –Slp2-a, GST–Slp4-a) was incubated with PIP- single and double knockdowns in polarized epithelial cells by retrovirus-mediated strip membranes according to the instructions of the manufacturer (Echelon RNA interference. Proc. Natl Acad. Sci. USA 101, 4912–4917 (2004). Bioscience). Lipid beads (Echelon) prepared with different phosphoinositides 36. Fukuda, M. & Kanno, E. Analysis of the role of Rab27 effector Slp4-a/Granuphilin-a or phosphatidylserine were incubated with 2 µg of purified GST –C2A or in dense-core vesicle exocytosis. Methods Enzymol. 403, 445–457 (2005).

DOI: 10.1038/ncb2536

A Upregulated Downregulated 15

5 Log Experiment (3D-MDCK) 5 10 15 B Log Control (2D-MDCK) 6 4 2 0 -2

Log Ratio (3D/2D) -4 -6 5 10 15 Log Signal

Figure S1 Transcriptional profiling of 3D vs. 2D MDCK cultures. (A) downregulated (red) genes are plotted. Significance of upregulated MDCK cells were grown in 2D and 3D conditions and purified RNA and downregulated genes was calculated using FDR-LiMMA analysis extracts were submitted for Canine Genome 2.0 Affymetrix cDNA (FIESTA software). (B) Signal-ratio (3D/2D) M-A plot of microarray results chip microarray analysis. Mean results of overexpressed (green) and presented in A.

Amount of mRNA relave to 2D 36h Amount of mRNA relave to 2D 36h (arbitrary units) 0,5 1,5 2,5 3,5

(arbitrary units) 10 0 1 2 3 0 1 2 3 4 5 6 7 8 9

7B2 ARHGAP24 ADRBK2 ANKRD28 AXIN2 AQP5 ARL3 CAM16 AXUD1 BREVIN BST2 FARP1 CADM1 CAMK1D FUZ CASC4 CHE11 GAL4 CiB1 CLD2 GBR7 DAK DIRC2 IGFBP6 DLG2 DP1 IHH DYNLC2B DYSF EFCAB1 PIK3R1 2D 36h EFCAB10 FBLIM1 RANBP3L FGFR4 FRZB RHOU FXYD2 FYCO1

S100A5 3D 14h GAL3 GNB5 SEC14L3 HECW2 HHAT HINT2 SLP2A ICK IFT172 STARD10 IFT74 IP6K2 TGFBR3 JIP KIF5B MUC20 Amount of mRNA relave to 2D 36h MYADM MYO18A (arbitrary units) NHERF1 10 20 30 40 50 60 0 NHERF2 OCIN PCM1 PIK3R2 FABP1 PKIG PLLP2 PTPLAD2 RAB10 S100A13 NGEF S100A4

S100P 36h2D D3h3 4 3D 36h 3D 14h 2D 36h SEC10 SEC62 SMTNL2 NHERF3 SNX17 SNX2 SNX4 3D 14h3D SNX5 STX1B SYNGR1 SYT13 SYNJ2BP1 TMEM205 TSPAN8

VAMP1 36h3D VAPB UNC5C VLA6 VPS37D WBSCR16

Figure S2 RT-qPCR validation of selected genes. Relative mRNA gene-specific exon-spanning reactions and quantitative data was expression of a selected group of genes upregulated during cyst formation. analyzed using SDS software (Applied Biosystems), and normalized to MDCK cells were plated either in 2D, or 3D to grow cysts and lysates were 2D expression levels. Different charts are used to distribute results in collected for 2D at 36h, and at 14 and 36h in 3D. After mRNA extraction different scales using relative arbitrary units (top chart, ≤2-fold; bottom- and cDNA polymerization, samples were analyzed by qPCR using left, ≤10-fold; bottom-right >10-fold). N=4.

A B siRNA Control siRNA Slp2a shRNA Control shRNA Slp4 Cell Packing 0.8 *

0.6 Nuclei

0.4 Podxl

0.2 Cell Height/Width Ratio Height/Width Cell -cat β

0.0

x-z siRNA Slp2a shRNA Slp4 siRNA Control shRNA Control C D siRNA Control siRNA Slp2a 6 100%

4 m) μ

50%

2 % CIliated cells CIliated % Nuclei

/ Cilia Length ( Length Cilia

0% 0 shRNA Control shRNA Slp4

siRNA Slp2a shRNA Slp4 siRNA Slp2 shRNA Slp4 siRNA Control shRNA Control siRNA Control shRNA Control Acetilated tubulin

E siRNA Control siRNA Slp2-a F Vesicles APM 14h 36h 36h Non-internalized 120% * 100%

-cat

/ 80%

GFP GFP

/ 60% **

(% of cyst) of (% 40%

GFP-Podxl Merge GFP αGFP Merge GFP αGFP Merge GFP αGFP 20% Podxl Podxl Podxl

Antibody localization after uptake after localization Antibody 0% Control Slp2-a KD

Figure S3 Slp-family proteins are not essential for 2D cell polarity or at 4ºC, after washing, they were grown for 24 hours more, allowing lumen ciliogenesis. (A) Confocal microscopy images of 6d culture of MDCK cells in development to occur. Then, they were fixed and staining for `-catenin (blue) transwells. Cells transfected with Slp2-a siRNA or Slp4-a shRNA or controls and a secondary antibody to detect the uptaken _-GFP. Inmediately after were grown in a 2D-monolayer in transwells for 6d, fixed and stained to binding (14h), control and Slp2-a silenced cysts showed peripheral GFP-PCX detect Podxl (red), `-catenin (green) and DNA (blue). Three confocal sections (green) and _-GFP (red). 24 hours after treatment, control cysts presented are shown for each treatment. Top panels show x-y apical section. Middle podocalyxin, mainly in the apical plasma membrane (arrows), revealing PCX- panels show x-y mid-height section. Bottom panels show x-z section. Bars, GFP is transcytosed from peripheral membranes to the new developed apical 20 +m. (B) Quantification of cell packing in monolayers shown in A. Cell membrane. However, Slp2-a knock-down cysts showed accumulations of packing measurements are cell height-width ratios of control (black bars) vesicles containing GFP-PCX (green) and _-GFP (red), suggesting a defect or Slp-KD cells (white bars). Values represent the average of three different in trafficking of transcytosed PCX. Yellow, colocalization of podocalyxin and experiments ± S.D. N=3 (> 150 cells/experiment, * P<0.05). (C) Confocal bound antibodies, blue indicates `-catenin. Bar, 10 +m. (F) Quantification microscopy images of 6d culture of MDCK cells in transwells. Cells treated of antibody localization after uptake in control or Slp2-a KD cysts. Cysts as in A were fixed and stained to detect primary cilia using acetyl-tubulin as in Figure A were analyzed by confocal microscopy. Cysts were classified a marker (green) and DNA (blue). Bars, 10 +m. (D) Quantification of ciliated as “vesicles”, “apical plasma membrane” (APM) or “not-internalized” cells and cilia length in monolayers shown in C. Values represent the average depending on the localization presented by the _-GFP (Red) antibody was of three different experiments ± S.D. N=3 (> 100 cells/experiment). (E) found in internal vesicles, apical membrane or peripherally, respectively. Localization of apical and basolateral proteins in control cysts or silenced Values are mean ± SD from three different experiments. N=3 ( >100 cysts/ for Slp2-a. 14 hours after plating, cysts were treated with _-GFP for 30’ experiment; **, P < 0.005; *, P<0.05).

A Linker-Slp2a Magnification BΔSHD-Slp2a Magnification C ΔC2AB-Slp2a Magnification Linker-GFP-Slp2a Linker-GFP ΔSHD-GFP-Slp2a ΔSHD-GFP ΔC2AB-GFP-Slp2a ΔC2AB-GFP Podxl Podxl Podxl β-cat β-cat β-cat

20h 20h 12h Podxl β-cat Podxl β-cat β-cat

Podxl

Linker-GFP-Slp2a Linker-GFP ΔSHD-GFP-Slp2a ΔSHD-GFP ΔC2AB-GFP-Slp2a ΔC2AB-GFP Podxl Podxl Podxl β-cat β-cat β-cat

24h 24h Podxl β-cat Podxl β-cat 20h Podxl β-cat

Linker-GFP-Slp2a Linker-GFP ΔSHD-GFP-Slp2a ΔSHD-GFP ΔC2AB-GFP-Slp2a ΔC2AB-GFP Podxl Podxl Podxl β-cat β-cat β-cat

48h 24h Podxl β-cat 48h Podxl β-cat Podxl β-cat

D E Lact-C2-GFP Lact-C2-GFP gp135 C2A-GFP C2A-GFP C2B-GFP C2B-GFP gp135 Podxl L Podxl F-actin β-cat β-cat L

Podxl β-cat Podxl β-cat F-actin Merge

Figure S4 Slp2-a- 6C2AB, Slp2-a- 6SHD, Slp2-a-linker and PS-probe form cysts for 12-24h, fixed and stained to detect Podxl (red) and `< catenin localization. (A) Confocal microscopy images of cysts stably expressing GFP- (blue). Bar, 10 +m. (D) Confocal microscopy images of cysts stably expressing Slp2-a linker domain. Cells were plated to form cysts for 20-24-48h, fixed GFP- Slp2-a C2A (left panels) and C2B (right panels). Cells were plated to and stained to detect Podxl (red) and `< catenin (blue). Bar, 10 +m. (B) form cysts for 12-24h, fixed and stained to detect Podxl (red) and `< catenin Confocal microscopy images of cysts stably expressing GFP- Slp2-a 6SHD. (blue). Bar, 10 +m. (E) Confocal microscopy images of cysts stably expressing Cells were plated to form cysts for 20-24-48h, fixed and stained to detect a phosphatidylserine (PS) fluorescent probe (Lact-C2-GFP). Cells were plated Podxl (red) and `< catenin (blue). Bar, 10 +m. (C) Confocal microscopy to form cysts for 72h, fixed and stained to detect Podxl (red) and F-actin images of cysts stably expressing GFP- Slp2-a 6C2AB. Cells were plated to (blue). Note the lack of asymmetric polarization of PS at the PM. Bar, 10 +m.

A B GFPGFP GFP Relative Slp2-a Btsz2 GFP Input Slp2aGSTGST binding: GFP GFP

Rab27a +++ Input α Input α

150 KDa 250 KDa αGFP Rab8a + 25 KDa αRab27a Rab3a + Merge Magnification Rab3b ++ C Rab27aGFP Rab27aGFP Podxl Rab3c - β− cat 16h Rab3d +/- Podxl β− cat

Rab11a -

WB:GFP Rab27aGFP D Merge Magnification 20h Rab27 Rab27 Podxl β− cat Podxl F-Actin

Podxl F-Actin Rab27aGFP

24h Podxl β− cat E Merge Magnification Rab27aGFP Rab27aGFP Slp2aCherry Podxl Rab27aGFP

Slp2aCherry Podxl 48h Podxl β− cat

Rab27aGFP F Slp2a V18A E11A/R32A MDCK GFP GFP MDCK GFP siRNA: C-KD C- KD C- KD C- KD C- KD WB:Slp2a 150 Slp2aGFP 72h Slp2a 100 Podxl β− cat

50 WB:Tub kDa

Figure S5 Slp2-a controls apical membrane trafficking through interaction correspond to five different stages of lumen initiation (AP vesicle aggregation with Rab27a. (A) Slp2-a-interacting Rab GTPases. GST-Slp2-a or GST [16h], vesicle fusion [20h], preapical-patch formation [24h], lumen expansion (control) beads were used to pull-down fluorescent protein-tagged Rab3a/b/ [48h], and mature cyst [72h]). Arrowheads, Podxl-positive vesicles; arrows, c/d, Rab8a, Rab11a or Rab27a from total cell lysates. The table indicates lumen. Bar, 5 +m (left panels). (D) Confocal microscopy images of endogenous relative binding results from 3 different experiments. (B) Immunoprecipitation Rab27a/b. MDCK cells were plated to form cysts for 72h and stained to detect of Rab27-GTP in GFP-Slp2-a or GFP-Btsz2 MDCK cysts. GFP-Slp2-a or GFP- Rab27 (green), Podxl (red) and F-actin (blue). Bar, 10 +m. (E) Confocal Btsz2 were immunoprecipiated using anti-GFP (or control) beads and analyzed microscopy images of Slp2-a and Rab27 in cysts at 72h. MDCK cells stably to detect binding of endogenous Rab27. (C) Localization of GFP-Rab27a expressing Cherry-Slp2-a and GFP-Rab27a were grown as cysts and fixed at during cyst morphogenesis. MDCK cells stably expressing GFP-Rab27a were 4d and stained to detect Podxl (blue). Bar, 10 +m. (F) Western-blot of Slp2-a grown to form cysts and fixed after 16, 20, 24, 48 and 72h. Samples were downregulation by siRNA in WT, V18A or E11A/R32A-Slp2-a-GFP stably stained for Podxl (red), and `-catenin (blue). Confocal microscopy images expressing MDCK cells. Tubulin was used as loading control.

GST-Slp4a qPCR: Slp4 A B C WTC2A C2B C2AB D 125% LPA S1P SLP4a N SHD ZnF SHD C2A C2B C 6 d 100% LPC PI(3,4)P2

WT PI PI(3,5)P2 75%

SHD nuc PI(3)P PI(4,5)P2

ΔSHD / 50% ** ** Linker PI(4)P PI(3,4,5)P3 SLP4a C2A 25% C2B PI(5)P PA

C2AB PE PS Slp4a/b 0% PC Blank shRNA: Slp4-1 Slp4-2 Control E GFP-Slp4a / Podxl / nuc F GFP-Slp4a / Podxl / β-cat ΔSHD Linker C2A C2B V21A W118S I18A

L L L L L

T7-Slp4-a Control Slp4-a KD G H I Slp454a J Δ305-354 GFP-Slp4a / Podxl / nuc LinkerF1 F2 F3 Mock C2AB K>Q Flag-Munc18-2

WB:α-Flag -cat β

Flag-Stx3 (Input) /

Flag-Munc18-2 Podxl / / WB:α-Flag Flag-Stx3 IP: α-T7

Slp454a Podxl Δ305-354 Podxl

L GFP-Slp4a ∗ WB:α-T7 L IP: α-T7 L L β-cat Merge β-cat Merge Galvez-Santisteban et al. Figure S6 1 2 3 4 5

Figure S6 Localization of other mutants and constructs of Slp4-a. (A) 10 +m. (G) Confocal microscopy images of cells stably transfected with GFP- Quantification of the silencing of Slp4-a by RT-qPCR. Values represent the Slp4-a K>Q mutant which uncouples phospholipid binding. Cells were plated average of ≥ three different experiments ± S.D., normalized to control levels to form cysts for 48 h, fixed and stained to detect Podxl (red) and nuclei (**, P < 0.005). (B) Scheme of Slp4-a/b, and Slp4-a domain fragments (blue). Arrowheads, subapical vesicular localization of Slp4-a. Bar, 10 +m. (H) utilized. (C) PIP-binding assay. GST-tagged Slp4-a WT and C2A, C2B or Mapping of Slp4-a syntaxin binding site. Beads coupled with each T7-Slp4a C2AB were expressed and purified in bacteria. PIP-strip membranes were linker deletion mutant (i.e., AA144-240; AA215-304; and, AA272-354) were incubated with 1 +g/ml concentration of either GST-Slp4-a WT, C2A, C2B incubated with FLAG-Munc18-2 and FLAG-Stx3, and co-immunoprecipitated or C2AB fusion proteins and then membranes were blotted with an anti-GST FLAG-tagged proteins were detected with HRP-conjugated anti-FLAG antibody antibody. A scheme of the PIP-strip membrane is shown (left panel). (D) (Blot: anti-FLAG, middle panel). The asterisk indicates non-specific bands of Confocal microscopy images of 72h cysts stained with anti-Slp4-a antibody immunoglobulin light chain used for the IP. Note that only the F3 construct (green) and nuclei (blue). Note apical localization of Slp4-a. Bar, 20 +m. exhibited Munc18-2-dependent Stx3 binding activity (lane 4), indicating (E) Confocal microscopy images of cells stably transfected with GFP-Slp4-a that amino acids 305-354 of the Slp4a linker domain are required for Stx3 fragments. Cells stably expressing GFP-Slp4-a linker, C2A, C2B or 6SHD binding. (I) Confocal microscopy images of scramble shRNA expressing cells fragments were plated to form cysts for 72 h, fixed and stained to detect stably expressing GFP-Slp4-5-4-a chimera. Cells were plated to form cysts Podxl (red) and nuclei (blue). Arrowheads, subapical vesicular localization for 48 h, fixed and stained to detect Podxl (red) and `catenin (blue). Arrows, of Slp4-a. Arrows, BL PM localization of Slp4-a. Bar, 10 +m. (F) Confocal lateral plasma membrane localization of Slp4-5-4-a. Bar, 10 +m. (J) Confocal microscopy images of cells stably transfected with GFP-Slp4-a Rab GTPase microscopy images of Slp4-a or scramble shRNA expressing cells stably mutants. Cells stably expressing GFP-Slp4-a V21A, W118S or I18A mutants transfected with GFP-Slp4-a 6305-354. Cells were plated to form cysts for 48 were plated to form cysts for 48 h, fixed and stained to detect Podxl (red) and h, fixed and stained to detect Podxl (red) and `catenin(blue). Arrows, lateral `-catenin (blue). Arrowheads, subapical vesicular localization of Slp4-a. Bar, plasma membrane localization of GFP-Slp4-a 6305-354. Bar, 10 +m.

B A 150% 200% **

100% 150%

100% ** ** * 50% 50% (relative to control) (relative to control) % Single lumen formation lumen % Single % Single lumen formation lumen % Single 0% 0%

WT SHD SHD WT Δ SHD SHD C2A C2B MDCK MDCK Δ Linker C2AB GFP-Slp2a GFP-Slp4a C 1.0

Podxl β-cat

0.5 Colocalization correlation value

0.0

Slp2-a Slp4-a Slp2-a Slp4-a Slp2-a Slp4-a

Early Preapical Lumen Galvez-Santisteban et al. Figure S7

Figure S7 Overexpression and colocalization analysis of Slp2-a and (C) Correlation analysis for Slp2-a and Slp4-a quantitative-colocalization Slp4-a during cyst morphogenesis. (A) Quantification of cysts with normal using Podxl (black bars, apical membrane marker) and `-catenin lumens in cells transiently overexpressing GFP (control), or GFP-Slp2-a (white bars, basolateral marker) in three different stages during cyst constructs (WT, 6SHD, or SHD). Values are mean ± SD from three different morphogenesis. GFP-Slp2-a or GFP-Slp4-a expressing cells were grown to experiments. N=3(>100 cysts/experiment; *, P < 0.05; **, P < 0.005). (B) form cysts for 12 (early), 24 (preapical) or 72h (lumen), fixed and stained Quantification of cysts with normal lumens in cells stably overexpressing to analyze Podxl or `-catenin localization. Quantitative colocalization GFP (control), or GFP-Slp4-a constructs (WT, 6SHD, SHD, Linker, analysis between GFP signal and each marker was performed using ImageJ. C2A, C2B and C2AB-GFP). Values are mean ± SD from three different Values are mean ± SD from three independent cyst cultures. N=3 (>10 experiments. N=3 (>100 cysts/experiment; *, P < 0.05; **, P < 0.005). cysts/experiment).

Figure 2A Figure 2C Figure 2F

KDa 250 KDa 150 100 250 150 100

Figure 3C

Figure 4A KDa 250 150 100

KDa 250 150 100

Figure S8 Scans of uncropped blots.

Figure 4G

GST-Slp2-a

Input WT V18A E11A/R32A KDa 250 150

100 75

Figure 5G

Figure S8 continued

Figure 6A

GST-Slp4-a Input kDa WT I18A V21A W118S SHD GST 250 6 150 100 75

50 37

25 20 15

Figure 6D

Figure 6J

kDa kDa 250 50 150 100 75 37 50

25 37

shRNA Stx3 shRNA shRNA scramble shRNA

Figure S8 continued

Figure S5A

Figure S5B

Figure S5F

kDa kDa 250 150 250 100 150 100 75 75

50 50

Figure S8 continued

Figure S6C Figure S6H

Figure S8 continued

Supplementary Table 1 RTqPCR validation of gene overexpression (2D, 14h, 36h) Supplementary Table 2 RNAi and RNA expression analysis in silencing experiments Supplementary Table 3 Quantification of lumen morphogenesis (RNAi screening)

Supplementary Video 1 Slp2-a apical localization depends on PIPs. MDCK cells stably expressing PHD-GFP and Slp2-Cherry were grown in cysts and set up for videomicroscopy. Cysts were treated with Ionomycin for 5 minutes during recording.