letters

Direct interaction of two polarity complexes implicated in epithelial tight junction assembly Toby W. Hurd*||, Lin Gao†||, Michael H. Roh‡, Ian G. Macara†# and Ben Margolis*‡§ *Howard Hughes Medical Institute, ‡Departments of Biological Chemistry and §Internal Medicine, University of Michigan Medical School, Ann Arbor, MI 48109-0650, USA †Center for Cell Signalling, University of Virginia Health Sciences Center, Charlottesville, VA 22908-0577, USA ||These authors contributed equally to this work e-mail: ¶[email protected] or #[email protected]

Published online: 27 January 2003; DOI: 10.1038/ncb923

Tight junctions help establish polarity in mammalian epithelia by forming a physical barrier that separates apical and basolateral membranes. Two evolutionarily conserved multiprotein complexes, Crumbs (Crb)–PALS1 (Stardust)–PATJ (DiscsLost)1–4 and Cdc42–Par6–Par3–atypical protein kinase C (aPKC), have been implicated in the assembly of tight junctions and in polarization of Drosophila melanogaster epithelia5–8. Here we identify a biochemical and functional link between these two complexes that is mediated by Par6 and PALS1 (proteins associated with Lin7). The interaction between Par6 and PALS1 is direct, requires the amino terminus of PALS1 and the PDZ domain of Par6, and is regulated by Cdc42-GTP. The transmembrane protein Crb can recruit wild-type Par6, but not Par6 with a mutated PDZ domain, to the cell surface. Expression of dominant-negative PALS1-associated tight junction protein (PATJ) in MDCK cells results in mis-localization of PALS1, members of the Par3–Par6–aPKC complex and the tight junction marker, ZO-1. Similarly, overexpression of Par6 in MDCK cells inhibits localization of PALS1 to the tight junction. Our data highlight a previously unrecognized link between protein complexes that are essential for epithelial polarity and formation of tight junctions.

or Sdt cause polarity defects in D. melanogaster epithelia12. The mammalian homologue of Sdt is PALS14. Similar to Sdt, PALS1 interacts with the tail of mammalian Crb isoforms through its PDZ domain. PALS1 utilizes an L27 domain to interact with a multi-PDZ domain protein called PATJ4,13. Similarly, Discs Lost1, a D. melanogaster orthologue of PATJ, can interact with an L27 domain from Sdt4. Thus, analogous complexes containing Crb–PALS1–PATJ

a L

L

IP

L

IP

L

IP HA–Par3 HA–Par6 Myc–PALS1

HA–Par3

–

HA–Par6

–

–

+

Myc–PALS1 –

–

+

–

–

–

+

+

+

+

+

–

–

+

+

+

+

+

+

+

b

c L L

pithelial cell lines provide an excellent model system for exploring the molecular basis of how polarity is established and maintained. Polarized epithelial cells contain distinct apical and basolateral membranes, each with their own complement of cell-surface proteins9 and a tight junctional seal at the superior aspect of the lateral surface to prevent admixture of membranes. Genetic and biochemical evidence has implicated multiple protein complexes in the assembly and maintenance of tight junctions. Many components of these complexes possess post-synaptic density 95/discs large/zona occludens 1 (PDZ) domains. PDZ domains commonly, but not exclusively, bind to the extreme carboxyl terminus of proteins and are often found in membrane-associated guanylate kinase (MAGUK) proteins, where they are paired with Src-homology 3 (SH3) and guanylate kinase domains10. MAGUK proteins are crucial for several aspects of cell polarity11. One MAGUK protein required for cell polarity in D. melanogaster epithelia is Stardust (Sdt)2,3. Sdt interacts with the transmembrane protein Crb through its PDZ domain. The two proteins colocalize in a subapical complex at the top of the lateral membrane, corresponding to the site in the epithelia of higher organisms where the tight junction forms. Mutations in either Crb

IP

Pre

Pre

IP

IP

Par3

PALS1

E

PATJ Par3 Myc–PALS1 IP: Par6 IP: Myc–PALS1

Figure 1 Par6 binds PALS1. a, Myc–PALS1 immunoprecipitates HA–Par6 , but not HA–Par3, in the absence of HA–Par6 . Myc–PALS1 was immunoprecipitated from HEK293 cells transiently transfected with Myc–PALS1, HA–Par6wt and/or HA–Par3, and subsequently immunoblotted for HA–Par3, HA–Par6 and Myc–PALS1. Protein expression levels in cell lysates (L) are shown adjacent to the Myc–PALS1 immunoprecipitations (IP). In the control, anti-Myc immunoprecipitations were performed on untransfected cells. b, Immunoprecipitation of endogenous Par6 coprecipitates endogenous Par3 and PALS1. Par6 was immunoprecipitated from cell lysates of wild-type MDCK I cells and immunoblotted for endogenous Par3 and PALS1. c, Myc–PALS1 immunoprecipitates endogenous Par3 from MDCK cell extracts. Myc–PALS1 was immunoprecipitated from cell lysates of MDCK cells stably expressing Myc–PALS1and immunoblotted for endogenous Par3 and PATJ. In the control lane (Pre), rabbit pre-immune serum was used to perform the immunoprecipitations. Protein expression levels in cell lysates are also shown.

NATURE CELL BIOLOGY VOL 5 FEBRUARY 2003 www.nature.com/naturecellbiology

©2003 Nature Publishing Group

137

L

L

Myc–PALS1(∆GUK)

Myc–PALS1(∆SH3)

Myc–PALS1(∆PDZ)

Myc–PALS1wt

HA–Par6

HA–Par6

–

1–271

mut.PDZ

Myc–PALS1

102–371

GST–PALS1(1–181)

L

GST

c

GST–PALS1(1–181)

371

GST

271

GST–PALS1(1–181)

102

Par6

Myc–PALS1

wt

PALS1 675

Myc–PALS1

+

+

+

+

+

d

IP: HA–Par6

GUK

Cell lysate IP: HA–Par6

PDZ

181 CRIB PDZ

1

4.1B SH3

GST

U1 1

L27

b L27

a

Myc–PALS1(∆U1 + L27N)

letters

HA–Par3

Myc–PALS1 HA–Par6

HA–Par6

–

+

+

–

–

–

–

–

–

+

+

Myc–PATJ

+

–

–

+

+

–

–

–

–

HA–Par6(mut.PDZ)

Cell lysate

f Bound to glutathione beads

e

HA–Par6

GST–PALS1(1–181) Blot: anti-HA Cell lysate

50% input

–

+

GST–PALS1(1–181)

–

–

Myc–PALS1

GST

+

–

HA–Par6(102–371)

–

HA–Par6

HA–Par6(1–371)

HA–Par3

HA–Par6(1–271)

Myc–PATJ

S–Par6(1–271)

g Myc–PALS1 HA–Par6

IP: HA–Par6

Myc–Cdc42 Myc–PALS1 Cell lysate

HA–Par6 Myc–Cdc42

Myc–PALS1

+

+

+

+

HA–Par6

–

+

+

+

–

–

Q61L

T17N

Myc–Cdc42

Figure 2 Characterization of Par6–PALS1 interaction. a, The domain structure of PALS1 and Par6. b, Deletion of the N-terminal 181 residues of PALS1 prevents binding of Par6. HA–Par6 was immunoprecipitated from HEK293 cells transiently transfected with HA–Par6 Myc–PALS1wt or PALS1 lacking the U1 and L27N, PDZ, SH3 or GUK domains. Subsequent immunoprecipitates were blotted for Myc–PALS1and HA–Par6 . Myc–PALS1 expression levels were determined by immunoblotting cell lysates (bottom). c, Par6 binds to the N terminus of PALS1. A recombinant GST fusion protein consisting of the U1 and L27N domains of PALS1 (GST–PALS1(1–181)) or GST alone were immobilized on agarose beads and incubated with cell extract from HEK293 cells transiently transfected with HA–Par6, HA–Par3 or Myc–PATJ. Precipitating proteins were detected by immunoblotting for HA–Par6 , HA–Par3 or Myc–PATJ. Cell lysates were also immunoblotted to monitor protein expression levels. d, Coprecipitation of Myc–PALS1and HA–Par6 . Cos7 cells were cotransfected with Myc–PALS1 and different HA-tagged Par6 constructs.

138

HA–Par6 was immunoprecipitated and bound Myc–PALS1 was detected by immunoblotting. The Par6 PDZ mutant (mut.PDZ) harbours KPLG167–170AAAA mutations. e, The Par6 PDZ domain is required for efficient interaction with the PALS1 N terminus. GST–PALS1(1–181) was immobilized on agarose beads and incubated with cell lysates from HEK293 cells transiently transfected with either full length HA–Par6(1–371), HA–Par6(102–371), HA–Par6(1–271) or HA–Par6 (mut.PDZ). Expression levels of the HA–Par6 mutants are also shown (bottom). f, Direct interaction of PALS1 with Par6. Purified GST or GST–PALS1(1–181) was bound to glutathione beads and incubated with purified S–Par6(1–271). GST proteins were detected using anti-GST antibody and S–Par6(1–181) was detected using HRP-tagged S protein. g, The PALS1–Par6 interaction is regulated by Cdc42. Cos7 cells were cotransfected with Myc–PALS1and HA–Par6 in the absence or presence Cdc42Q61L or Cdc42T17N. HA–Par6 was immunoprecipitated and bound Myc–PALS1 was detected by immunoblotting.

NATURE CELL BIOLOGY VOL 5 FEBRUARY 2003 www.nature.com/naturecellbiology

©2003 Nature Publishing Group

letters and Crb–Sdt–DiscsLost (Dlt) are found in mammalian and D. melanogaster epithelia, respectively14. A different complex, which colocalizes with PALS1 and Sdt in D. melanogaster and mammalian epithelia, is composed of Par3, Par6, aPKC and Cdc42-GTP5–8. The Par3–Par6–aPKC complex is important for determining polarity in many cell types, including D. melanogaster neuroblasts, the Caenorhabditis elegans zygote and mammalian epithelial cells11,15,16. Par6 contains a PDZ domain and an adjacent Cdc42/Rac-interactive binding domain (CRIB)-like motif, both of which are required for binding to the small GTPases Cdc42 and Rac7,8,17. The N terminus of Par6 binds aPKC-λ and ζ7,8,16. In turn, both aPKCs and Par6 can interact with a multi-PDZ domain protein called Par3, also known as atypical PKC isotypespecific interacting protein (ASIP) and Bazooka18,19. The spatial and functional similarities of Crb–PALS1–PATJ and Par3–Par6–aPKC led us to examine whether these complexes interact. First, we transiently expressed Myc–PALS1 with haemagglutinin (HA)–Par6 or HA–Par3 in HEK 293 cells and found that Par6, but not Par3, co-immunoprecipitated with PALS1 (Fig. 1a). However, when all three proteins were expressed together, we detected a complex of PALS1 with both Par6 and Par3. Importantly, we could detect endogenous PALS1 in an immunoprecipitation of endogenous Par6 from MDCK cells (Fig. 1b). Although Fig. 1b demonstrates that the antibodies we raised detect PALS1 in MDCK cells by immunoblotting, they are not useful for immunoprecipitation. Accordingly, we used a MDCK cell line expressing Myc–PALS1 and found that endogenous Par3 and PATJ were both co-immunoprecipitated with Myc–PALS1 (Fig. 1c). Collectively, these data indicate that the Crb–PALS1–PATJ and Par3–Par6–aPKC complexes can associate with one another through a PALS1–Par6 interaction. PALS1 is a multi-domain protein (Fig. 2a). To identify the region of PALS1 that interacts with Par6, we tested the ability of various Myc–PALS1 deletion mutants to co-immunoprecipitate with HA–Par6. We found that removal of the N terminus of PALS1, which contains the U1 and L27N domains, reduced binding to Par6 (Fig. 2b). Furthermore, a glutathione S-transferase (GST) fusion protein containing the U1 and L27N domains (PALS1(1–181)) could bind to overexpressed Par6, but not to Par3 (Fig. 2c). However, GST–L27N alone could not bind Par6 (data not shown), suggesting that the binding site for Par6 exists in the PALS1 U1 domain. Neither the N-terminal aPKC-binding site nor the C terminus of Par6 were required for PALS1 association, but mutation of the Par6 PDZ domain reproducibly reduced binding to PALS1 (Fig. 2d, e). Importantly, bacterially expressed GST–PALS1(1–181) bound specifically to S-tagged Par6(1–271), demonstrating that the interaction is direct (Fig. 2f). Next, we co-expressed PALS1 and Par6 with the constitutively active Cdc42 mutant, Cdc42Q61L, which binds to Par6, or with inactive Cdc42 mutant, Cdc42T17N, which does not. Interestingly, Cdc42Q61L (which is constitutively GTPbound) markedly enhanced the association of PALS1 with Par6, suggesting the interaction might be controlled by cell signalling events (Fig. 2g). We then examined the functional implications of interactions between Par6 and the Crb–PALS1–PATJ complex. We demonstrated that Crb3 is endogenously expressed in MDCK cells and exists in a complex with PATJ and PALS1 (ref. 20). When co-expressed with Myc–Crb3 in HeLa cells, Par6 was recruited to the plasma membrane (Fig. 3a). However, Par6 containing the mutant PDZ domain was not recruited (Fig. 3b). Conversely, Crb3-∆ERLI, a Crb3 mutant that lacks four amino acid residues at its C terminus and cannot bind PALS1, was unable to recruit Par6 to the membrane (Fig. 3c). Therefore, we suggest that Crb3 promotes recruitment of Par6 to the membrane through PALS1. Finally, we examined if components of the Crb–PALS1–PATJ or Par3–Par6–aPKC complexes were mislocalized in MDCK cells expressing dominant-negative mutants that disrupt these polarity

a Myc–CRB3

HA–Par6 Anti-Myc Anti–HA

Myc–CRB3/ HA–Par6

Myc–CRB3/ HA–Par6

Merge

b HA–Par6(mut.PDZ) Anti-Myc Anti–HA

Myc–CRB3/ HA–Par6(mut.PDZ)

Myc–CRB3/ HA–Par6(mut.PDZ)

Merge

c Myc–CRB3-∆ERLI Anti-Myc Anti–HA

Myc–CRB3-∆ERLI/ HA–Par6

Myc–CRB3-∆ERLI/ HA–Par6

Merge

Figure 3 CRB3 recruits Par6 to the plasma membrane. a, Myc–Crb3 recruits HA–Par6 to the plasma membrane. HeLa cells were transiently transfected with Myc–Crb3 and/or HA–Par6wt. Cells were fixed 48 h after transfection and stained with mouse monoclonal anti-Myc (green) rabbit polyclonal anti-HA (red) antibodies. Nuclei were stained with DAPI (blue). b, Deletion of the four C-terminal residues of Myc–Crb3 prevents recruitment of HA–Par6 to the plasma membrane. HeLa cells were transiently transfected with HA–Par6wt and/or Crb3 lacking the C-terminal E, R, L and I residues (Myc–Crb3-∆ERLI). Cells were stained as in a. c, Mutation of the Par6 PDZ domain prevents its recruitment to the plasma membrane by Myc–Crb3. HeLa cells were transiently transfected with Myc–Crb3 and/or a Par6 containing the KPLG167–170AAAA mutation in the PDZ-encoding region (HA–Par6 mut.PDZ). Cells were stained as in a.

NATURE CELL BIOLOGY VOL 5 FEBRUARY 2003 www.nature.com/naturecellbiology

©2003 Nature Publishing Group

139

letters

a

b

Myc–PALS1

Myc–PALS1

ZO-1

ZO-1

PALS1

aPKCζ

Myc–PATJ1(1–238) ZO-1 PALS1

Myc–PALS1(1–238) ZO-1 aPKCζ

Myc–PATJ1(1–238) ZO-1 PATJ

c

d

Myc–PATJ1(1–238)

wt

Par6(wt) PALS1

CRB3 E-cadherrin

Par6(102–371)

Par6(mut.PDZ) PALS1

Figure 4 PATJ(1–238) disrupts aPKCζζ localization and tight junction biogenesis. a, An MDCK cell line stably expressing Myc–PATJ(1–238) was grown in lowcalcium (5 µM) media and subsequently incubated in normal-calcium (1.8 mM) media for 24 h. Cells were then fixed, permeabilized, and immunostained with antibodies against the indicated proteins. b, MDCK cells stably expressing Myc–PALS1 or Myc–PATJ(1–238) were grown as in a. MDCK cells expressing Myc–PALS1 were immunostained with mouse monoclonal anti-Myc (red) and rabbit polyclonal antiaPKCζ (green). Cells expressing Myc–PATJ(1–238) were stained with mouse monoclonal anti-ZO-1 (red) and rabbit polyclonal anti-aPKCζ (green). Square panels represent x-y sections, whereas rectangular panels represent x-z sections (apical to the

top, basolateral to the bottom). c, MDCK cells expressing Myc–PATJ(1–238) were treated as in a. Subsequently, cells were immunostained with anti-E-cadherin and anti-Crb3 antibodies (green and red, respectively). Crb3 localizes to the apical surface, whereas E-cadherin is present at the lateral membrane. d, Overexpression of Par6 inhibits relocalization of PALS1 to cell–cell contacts during tight junction reassembly. MDCK cell lines stably expressing different Myc-tagged Par6 constructs were subjected to calcium withdrawal after reaching confluence (as described previously). Cells were fixed 6 h after re-addition of calcium to the medium and stained for endogenous PALS1. Untransfected cells are shown as a control.

complexes. In initial studies, we found no effects of dominantnegative PATJ on the formation of stable tight junctions4. However, we were able to detect defects in tight junction assembly using a calcium switch protocol. Expression of dominant-negative PATJ(1–238) caused mis-localization of PALS1 and tight junctions

were fragmented 24 h after calcium switch, as detected by ZO-1 staining (Fig. 4a). Importantly, aPKC also mislocalized away from the fragmented tight junctions (Fig. 4b). However, PATJ(1–238) did not affect the global polarity of cells grown on filters, as localization of Crb3 and E-cadherin remained unchanged (Fig. 4c).

140

NATURE CELL BIOLOGY VOL 5 FEBRUARY 2003 www.nature.com/naturecellbiology

©2003 Nature Publishing Group

letters Overexpression of wild-type Par6 or Par6-∆N, both of which disrupt tight junction assembly7,15, resulted in mis-localization of PALS1 (Fig. 4d). In contrast, overexpression of Par6 containing a mutated PDZ domain that binds poorly to PALS1 (Fig. 2) did not result in the mis-localization of PALS1. In summary, our results identify an unexpected direct interaction between PALS1 and Par6 that links two protein complexes important for epithelial cell polarity. Crb is localized to the apical surface of mammalian and fly epithelia and colocalizes at the tight junction with PALS1 and PATJ2–4,21. Through PALS1, Crb may recruit Par6 and associated proteins to the apical side of the tight junction. Par3 also binds to the junctional adhesion molecule (JAM-1), which may contribute to this localization22,23. Furthermore, our results suggest that the localization of these protein complexes is a codependent process, as the disruption of one complex affects the localization of proteins present in the other. Consistently, components of the Par3–Par6–aPKC complex are mislocalized in the epithelia of flies with Sdt mutations2,3. Further studies will be necessary to determine if the absence of PALS1 affects targeting of Par3–Par6–aPKC in mammalian epithelia. Our attempts to reduce levels of PALS1 by inhibitory RNA (siRNA) approaches have been complicated by the fact that MDCK cells are canine in origin and that siRNA strategies in polarized epithelial cells are still at an early stage. The Par3–Par6–aPKC complex is involved in polarity determination in a wide variety of tissues, including epithelia, neuroblasts, migrating astrocytes and in Danio rerio organogenesis11,24–26. In contrast, the Crb–PALS1–PATJ complex seems to be more specific for epithelia, at least in D. melanogaster, suggesting it functions as a specific adaptor assisting in the localization of the Par3–Par6–aPKC complex in this cell type2,3. However, it remains unclear whether all these proteins are simultaneously present in a single complex, or whether the interactions are dynamic. For example, the PDZ domain of Par6 binds to the N terminus of PALS1, but can also bind to Par3 and is required for binding to Cdc42 (refs 7, 8). Nonetheless, the interaction between PALS1 and Par3 is dependent on Par6, indicating that Par3 and PALS1 do not compete for the PDZ domain of Par6. The finding that a small GTPase stimulates the interaction of PALS1 with Par6 suggests the interactions are probably dynamic. Thus, the binding of Cdc42-GTP to the CRIB motif may induce a conformational change in Par6 that increases binding to PALS1. These results may also explain how small G protein signalling can contribute to the localization of the Par3–Par6–aPKC complex, as described in other systems26–28. Whether interaction of PALS1 with Par6 also controls downstream signalling events, such as the modulation of aPKC activity, remains to be determined. Nonetheless, the identification of a direct interaction between these two protein complexes is an important step in understanding the mechanisms that promote epithelial polarity and formation of tight junctions.

Methods Cell culture MDCKI and HeLa cells were grown in DMEM (Invitrogen; Carlsbad, CA) containing 100U penicillin, 100 µg ml−1 streptomycin sulphate, 2 mM L-glutamine and 10% foetal bovine serum. MDCK Myc–Par6B, Myc–PALS1 and Myc–PATJ(1–238; refs 4,7) clones were propagated in media containing 175 µg ml−1 hygromycin B, 300 µg ml−1 G418 and 100 µg ml−1 zeocin, respectively.

Immunofluorescence microscopy and imaging For recruitment assays, HeLa cells grown on glass coverslips were transfected with Fugene6 transfection reagent (Roche, Indianapolis, IN) according to the manufacturer’s instructions, with 0.5 µg per well Myc–Crb3/Myc–Crb3-∆ERLI (ref. 20) and/or 0.5 µg per well HA–Par6wt/HA–Par6mut.PDZ (ref. 7) and left for 48 h. For MDCK immunostaining, MDCK cells were seeded at high density onto 12-mm Transwell membrane filters (0.4-µm pore size, Corning Costar, New York, NY). Coverslips and filters were fixed in 4% formaldehyde/PBS for 20 min, permeabilized in 0.1% Triton X-100/PBS for 10 min and blocked in 2% goat serum/PBS for 60 min. Coverslips/filters were incubated with primary antibodies in blocking solution (monoclonal anti-Myc 9E10 (1:1000), polyclonal anti-HA Y11 (1:1000), polyclonal anti-PALS1 (1:200; ref. 4), polyclonal anti-PATJ (1:200; ref. 4), polyclonal anti-Crb3 (1:250), monoclonal anti-E-Cadherin (Sigma, St Louis, MO; 1:1600), polyclonal anti-aPKCζ (Upstate Biotechnology, Lake Placid, NY; 1:500) and monoclonal anti-ZO-1 (Zymed, San Francisco, CA;

1:1000)) overnight in a humidified chamber at 30 °C. After extensive washes with 2% goat serum/PBS, filters were incubated with fluorochrome-conjugated secondary antibody (1:1,000 in 2% goat serum/PBS) for 1 h at 30 °C. Subsequently, filters were washed several times with 2% goat serum/PBS and mounted on glass slides with ProLong antifade reagent (Molecular Probes, Inc, Eugene, OR). Immunofluorescence microscopy was performed at the University of Michigan Morphology and Image Analysis Laboratory with a Zeiss LSM510 Axiovert 100M inverted confocal microscope (Carl Zeiss, Inc., Thornwood, NY).

Immunoprecipitation and immunodetection Plasmids (2 µg) pRK5-Myc-PALS1(wt, ∆U1 + L27N, ∆PDZ, ∆SH3 and ∆GUK), pRK5-Myc-PATJ, pKH3-Par6B(wt, 102–371, 1–271 and mut.PDZ) and pHA3–Par3 were transfected using Fugene6 into HEK293 cells grown to 50% confluency on 10-cm dishes. After 48 h, cells were collected in 0.5 ml lysis buffer (50 mM Tris-HC at pH 7.4, 150 mM sodium chloride, 10% glycerol, 1% Triton X-100, 1.5 mM magnesium chloride, 1 mM EGTA, 10 mM sodium fluoride, 10 mM Na4P2O7, 1 mM Na3VO5, 1 mM phenyl methylsulphonyl fluoride, 10 µg ml−1 leupeptin and 20 µg ml−1 aprotinin). For MDCK immunoprecipitation and immunodetection experiments, cells were grown to confluence on 15-cm dishes and extracted in 1 ml of lysis buffer. Lysates were cleared by centrifugation (12,000g for 15 min at 4 °C), with a fraction kept for immunoblotting. The remainder was used for immunoprecipitation with monoclonal anti-Myc 9E10, polyclonal anti-HA Y11 or anti-Par6 T20 (Santa Cruz Biotechnology, Santa Cruz, CA), as described previously. Precipitated proteins were washed three times with cold PBS supplemented with 10% glycerol, eluted with sample buffer and resolved on 4–12% gradient bis-Tris gels using the NuPAGE electrophoresis system (Invitrogen). Proteins were electrophoretically transferred from the gels onto nitrocellulose membranes. Membranes were blocked in 5% bovine serum albumin/Tris-buffered saline (TBS) for 30 min and incubated with primary antibody in 5% bovine serum albumin/TBS for 2 h at room temperature. After extensive washing with 0.1% Triton X100/TBS, membranes were soaked in TBS supplemented with 5% skimmed milk powder and secondary antibody conjugated to horseradish peroxidase. Incubation with secondary antibody was performed for 1 h at room temperature or for 4 h at 4 °C. Membranes were washed with 0.1% Triton X100/TBS and bands were visualized using ECL reagent (Amersham Biosciences, Piscataway, NJ). For pull-down experiments, HEK293 lysates were incubated with GST or GST–PALS1(1–181) bound to Sepharose beads overnight at 4 °C. Sepharose beads were washed and immunoblotted as for immunoprecipitations. RECEIVED 10 SEPTEMBER 2002; REVISED 6 NOVEMBER 2002; ACCEPTED 25 NOVEMBER 2002; PUBLISHED 27 JANUARY 2003.

1. Bhat, M. A. et al. Discs Lost, a novel multi-PDZ domain protein, establishes and maintains epithelial polarity. Cell 96, 833–845 (1999). 2. Hong, Y., Stronach, B., Perrimon, N., Jan, L. Y. & Jan, Y. N. Drosophila Stardust interacts with Crumbs to control polarity of epithelia but not neuroblasts. Nature 414, 634–638 (2001). 3. Bachmann, A., Schneider, M., Thellenberg, E., Grawe, F. & Knust, E. Drosophila Stardust is a partner of Crumbs in the control of epithelial cell polarity. Nature 414, 638–643 (2001). 4. Roh, M. H. et al. The Maguk protein, Pals1, functions as an adapter, linking mammalian homologues of Crumbs and Discs Lost. J. Cell Biol. 157, 161–172 (2002). 5. Wodarz, A., Ramrath, A., Grimm, A. & Knust, E. Drosophila atypical protein kinase C associates with Bazooka and controls polarity of epithelia and neuroblasts. J. Cell Biol. 150, 1361–1374 (2000). 6. Petronczki, M. & Knoblich, J. A. DmPAR-6 directs epithelial polarity and asymmetric cell division of neuroblasts in Drosophila. Nature Cell Biol. 3, 43–49 (2001). 7. Joberty, G., Petersen, C., Gao, L. & Macara, I. G. The cell-polarity protein Par6 links Par3 and atypical protein kinase C to Cdc42. Nature Cell Biol. 2, 531–539 (2000). 8. Lin, D. et al. A mammalian PAR-3–PAR-6 complex implicated in Cdc42/Rac1 and aPKC signalling and cell polarity. Nature Cell Biol. 2, 540–547 (2000). 9. Yeaman, C., Grindstaff, K. K. & Nelson, W. J. New perspectives on mechanisms involved in generating epithelial cell polarity. Physiol. Rev. 79, 73–98 (1999). 10. Hung, A. Y. & Sheng, M. PDZ domains: structural modules for protein complex assembly. J. Biol. Chem. 277, 5699–5702 (2002). 11. Wodarz, A. Establishing cell polarity in development. Nature Cell Biol. 4, E39–E44 (2002). 12. Tepass, U. & Knust, E. Crumbs and stardust act in a genetic pathway that controls the organization of epithelia in Drosophila melanogaster. Dev. Biol. 159, 311–326 (1993). 13. Doerks, T. et al. L27, a novel heterodimerization domain in receptor targeting proteins Lin-2 and Lin-7. Trends Biochem. Sci. 25, 317–318 (2000). 14. Tepass, U. Adherens junctions: new insight into assembly, modulation and function. Bioessays 24, 690–695 (2002). 15. Gao, L., Joberty, G. & Macara, I. G. Assembly of epithelial tight junctions is negatively regulated by Par6. Curr. Biol. 12, 221–225 (2002). 16. Suzuki, A. et al. Atypical protein kinase C is involved in the evolutionarily conserved par protein complex and plays a critical role in establishing epithelia-specific junctional structures. J. Cell Biol. 152, 1183–1196 (2001). 17. Qiu, R. G., Abo, A. & Martin, S. G. A human homolog of the C. elegans polarity determinant Par-6 links Rac and Cdc42 to PKCζ signaling and cell transformation. Curr. Biol. 10, 697–707 (2000). 18. Tabuse, Y. et al. Atypical protein kinase C cooperates with PAR-3 to establish embryonic polarity in Caenorhabditis elegans. Development 125, 3607–3614 (1998). 19. Izumi, Y. et al. An atypical PKC directly associates and colocalizes at the epithelial tight junction with ASIP, a mammalian homologue of Caenorhabditis elegans polarity protein PAR-3. J. Cell Biol. 143, 95–106 (1998). 20. Makarova, O., Roh, M. H., Liu, C. J., Laurinec, S. & Margolis, B. Mammalian Crumbs3 is small transmembrane protein linked to Pals1. Gene (in the press). 21. Lemmers, C. et al. hINADl/PATJ, a homolog of Discs lost, interacts with crumbs and localizes to tight junctions in human epithelial cells. J. Biol. Chem. 277, 25408–25415 (2002). 22. Itoh, M. et al. Junctional adhesion molecule (JAM) binds to PAR-3: a possible mechanism for the recruitment of PAR-3 to tight junctions. J. Cell Biol. 154, 491–497 (2001). 23. Ebnet, K. et al. The cell polarity protein ASIP/PAR-3 directly associates with junctional adhesion molecule (JAM). EMBO J. 20, 3738–3748 (2001).

NATURE CELL BIOLOGY VOL 5 FEBRUARY 2003 www.nature.com/naturecellbiology

©2003 Nature Publishing Group

141

letters 24. Ohno, S. Intercellular junctions and cellular polarity: the PAR–aPKC complex, a conserved core cassette playing fundamental roles in cell polarity. Curr. Opin Cell Biol. 13, 641–648 (2001). 25. Horne-Badovinac, S. et al. Positional cloning of heart and soul reveals multiple roles for PKC λ in zebrafish organogenesis. Curr. Biol. 11, 1492–1502 (2001). 26. Etienne-Manneville, S. & Hall, A. Integrin-mediated activation of Cdc42 controls cell polarity in migrating astrocytes through PKCζ. Cell 106, 489–498 (2001). 27. Gotta, M., Abraham, M. C. & Ahringer, J. CDC-42 controls early cell polarity and spindle orientation in C. elegans. Curr. Biol. 11, 482–488 (2001). 28. Kay, A. J. & Hunter, C. P. CDC-42 regulates PAR protein localization and function to control cellular and embryonic polarity in C. elegans. Curr. Biol. 11, 474–481 (2001).

142

ACKNOWLEDGMENTS We thank S. Laurinec for antibody generation and A. Liu for assistance with tissue culture and confocal microscopy. We thank A. Saltiel for helpful discussions. We acknowledge the University of Michigan Microscopy and Image Analysis Core for allowing us to use the confocal microscope. This work was partially supported by grant CA40042 from the National Institutes of Health, DHHS (to I.G.M). B.M. is an investigator of the Howard Hughes Medical Institute.

COMPETING FINANCIAL INTERESTS The authors declare that they have no competing financial interests.

NATURE CELL BIOLOGY VOL 5 FEBRUARY 2003 www.nature.com/naturecellbiology

©2003 Nature Publishing Group