The Drosophila PAR-1 Spacer Domain Is Required for Lateral Membrane Association and for Polarization of Follicular Epithelial Cells

Current Biology, Vol. 15, 255–261, February 8, 2005, ©2005 Elsevier Ltd All rights reserved.

DOI 10.1016/j .c ub . 20 05 . 01 .0 3 3

The Drosophila PAR-1 Spacer Domain Is Required for Lateral Membrane Association and for Polarization of Follicular Epithelial Cells Thomas Vaccari,1,3 Catherine Rabouille,2 and Anne Ephrussi1,* 1 Developmental Biology Programme European Molecular Biology Laboratory Meyerhofstrasse 1 69117 Heidelberg Germany 2 The Department of Cell Biology University Medical Centre Utrecht AZU Rm G02.525, Heidelberglaan 100 3584CX Utrecht The Netherlands

Summary The Ser/Thr kinases of the PAR-1/MARK/Kin1 family are conserved regulators of polarity in epithelial and non-epithelial cells [1]. Drosophila PAR-1 localizes laterally in the follicular epithelium of the ovary [2, 3], where it has been shown to function at two distinct levels: It stabilizes the cytoskeleton [3, 4] and it regulates apical-basal polarity by directly inhibiting lateral assembly of the apical aPKC/Bazooka/PAR-6 complex [5, 6]. However, it has been unclear how lateral localization of Drosophila PAR-1 is achieved and whether this localization contributes to epithelial polarity in vivo. Here we show that, through its spacer domain, Drosophila PAR-1 accumulates on the lateral plasma membrane (PM) in cells of the follicular epithelium (FE). Rescue experiments indicate that in FE cells PAR-1 kinase activity is essential for all the described functions of PAR-1. In contrast, the spacer domain of PAR-1 is required for apical-basal polarity and growth control but is dispensable for microtubule (MT) stabilization. Our data indicate that the spacer domain of PAR-1 is required for lateral PM localization of PAR-1 kinase and for development of a polarized FE. Results and Discussion PAR-1 Localizes to the Lateral PM via Its Spacer Domain PAR-1 N1S, the predominant PAR-1 isoform in Drosophila ovaries [2, 7], is composed of an N terminus and kinase and spacer domains (Figure 1A). The kinase and spacer domains are separated by conserved T and UBA domains [8]. This isoform lacks the conserved C-terminal domain that, together with the spacer domain, is required for basolateral localization of EMK2/Par-1b in Madin-Darby canine kidney (MDCK) cells [9]. To determine which part of PAR-1 is responsible for its localization to the lateral cortex of FE cells, we generated a series of PAR-1 deletion mutants tagged with *Correspondence: [email protected] 3 Present address: Department of Molecular and Cell Biology, University of California, Berkeley, 142 Life Sciences Addition #3200, Berkeley, California 94720.

green fluorescent protein (GFP) and expressed them in the columnar FE of stage-10 egg chambers. GFP-tagged full-length Drosophila PAR-1 (GFP-PAR-1 N1S [10]; here called GFP-PAR-1 FL) genetically rescues par-1 lossof-function mutants and localizes laterally and to intracellular dots in FE cells (Figure 1A; [4]). GFP-PAR-1 lacking the spacer domain (GFP-PAR-1 ⌬SP) shows reduced lateral localization and a diffuse intracellular signal (Figure 1B). The spacer domain of PAR-1 fused to GFP (GFP-PAR-1 SP) localizes laterally and to intracellular dots and the nucleus (Figure 1C). The replacement of two conserved regulatory Ser/Thr in the catalytic domain of PAR-1 [11, 12] with Ala (GFP-PAR-1 K*) dramatically reduces kinase activity (P. Papadaki and A.E., unpublished results) but does not affect the ability of the protein to localize laterally (Figure 1D). These results indicate that the spacer domain is necessary for PAR-1 localization to the lateral cortex of follicular epithelium (FE) cells. As in FE cells, Drosophila Schneider 2 (S2) cells show GFP-PAR-1 FL and GFP-PAR-1 SP localization to the plasma membrane (PM) and to intracellular dots corresponding to the Golgi apparatus (Figure S1A in the Supplemental Data available with this article online). We therefore further dissected the role of the spacer domain in cortical localization of PAR-1 in S2 cells. The mutation of four amino acids within the T domain that are critical for lateral membrane-targeting of the low-density lipoprotein (LDL) receptor in MDCK cells [13], does not impair GFP-PAR-1 T* localization in S2 cells (Figure S1A). These data suggest that the conserved T domain does not play a role in PAR-1 membrane localization. The spacer domain of PAR-1 contains two highly conserved amino acid stretches that include predicted phosphorylation sites (sites A, B, and C; Figure S1B). In S2 cells, GFP-PAR-1 mutants in which the candidate Ser/Thr targets within these sites were independently replaced by Ala localize indistinguishably from GFP-PAR-1 FL (GFP-PAR-1 A*, B*, and C*; Figure S1A). The phosphorylation by aPKC of Thr-564 and -595 within the spacer domain of hPar-1b and hPar-1a (MARK3/C-TAK1), respectively, was recently reported to downregulate their cortical localization upon aPKC overexpression [14]. The homologous Thr in Drosophila PAR-1 is Thr-797 in site C of the spacer domain. As was the case in S2 cells (Figure S1A), mutation of Thr-797 to Ala does not alter the ability of the protein to associate to the PM in follicle cells (Figure S1C). However, GFP-PAR-1 C* shows increased accumulation on the apical membrane when compared to GFP-PAR-1 FL (Figure S1C; compare with Figure 1A), suggesting that site C within the spacer domain is critical for restriction of the spacer-dependent localization of PAR-1 to the lateral plasma membrane. To characterize the cortical localization of Drosophila PAR-1, we fractionated Drosophila ovary extracts into a high-speed cytosolic fraction (S) and a membrane pellet (P). Western-blot analysis reveals that the bulk of PAR-1 is recovered in the pellet (Figure S2A). As expected, ␣-tubulin is found almost exclusively in the cytosolic fraction. A similar proportion of PAR-1 is en-

Current Biology 256

Figure 1. The PAR-1 Spacer Domain Is Required for Lateral Localization in FE Cells (A–D) Schematic representation of the domain structure of GFP-PAR-1 constructs. (A⬘ and A″, B⬘ and B″, C⬘ and C″, D⬘ and D″) Localization of the indicated GFP-PAR-1 construct in cells of the columnar FE. Two perpendicular confocal sections are shown. The lateral PM localization of GFP-PAR-1 K* is consistent with that observed for a mutant of the mammalian PAR-1 homolog EMK1 lacking the entire kinase domain, which localizes laterally in MDCK cells [32]. The scale bars represent 5 ␮m.

riched in the pellets of S2-cell extracts (Figure S2B). To assess the nature of PAR-1 association with membranes, we treated the S2-cell pellets under different conditions. Washing the pellets either with high salt or under denaturing conditions released a major proportion of PAR-1, suggesting that PAR-1 associates with membranes peripherally (Figure S2B). Surprisingly, fractionation experiments in S2 cells showed that GFP-PAR-1 ⌬SP, which poorly localizes to the lateral cortex (Figure 1B), is present in membrane fractions in proportions similar to GFP-PAR-1 FL (Figure S2C). To resolve this apparent contradiction, we expressed GFP-PAR-1 FL and GFP-PAR-1 ⌬SP in FE cells (Figure S2D) and used anti-GFP antibodies to compare their sub-cellular distribution by immunoelectron microscopy (IEM) (Table 1). GFP-PAR-1 FL is tightly associated to the lateral PM and decorates the intracellular-

Table 1. Quantification of the Amount of Gold Particles Detecting GFP-PAR-1 FL and GFP-PAR-1 ⌬SP per Cell Compartment in the Columnar FE of Stage-10 Egg Chambers

Cytosol Nucleus Membranes: PM Of which apical Of which lateral Of which basal Golgi apparatus ER ⫹ vesicles Endosomes

GFP-PAR-1 FL

GFP-PAR-1 ⌬SP

4 ⫾ 2% 9 ⫾ 1%

10 ⫾ 3% 9 ⫾ 1%

28 ⫾ 6% 5 ⫾ 3% 74 ⫾ 9% 21 ⫾ 7% 9 ⫾ 3% 50 ⫾ 7% 0%

5 ⫾ 3% 60 ⫾ 10% 16 ⫾ 5% 24 ⫾ 7% 8 ⫾ 1% 68 ⫾ 3% 0%

Equal numbers of gold particles were counted.

membrane organelles in FE cells (Figure S2E). IEM using the anti-PAR-1 N1 antibody [15] results in a similar decoration of the Golgi apparatus (Figure S2F) and of other intracellular membranes (data not shown), suggesting that the presence of GFP-PAR-1 FL on intracellular membranes is not due to overexpression or tagging artifacts. Compared to the distribution of GFP-PAR-1 FL, GFP-PAR-1 ⌬SP localization to the lateral PM was reduced and we also observed an increase in localization to the endoplasmic reticulum (Table 1). These data suggest that the spacer domain is dispensable for PAR-1 localization to membrane organelles. However, this domain mediates specific localization of PAR-1 to the lateral PM. The Spacer Domain of PAR-1 Is Required for Epithelial Polarization par-1 null (par-1W3) follicle cells display a much denser array of microtubules (MTs) than do wild-type cells [4], consistent with the fact that MARK kinases increase the dynamic instability of MTs by phosphorylation of MAPs [12]. However, the MTs are much less stable in par-1W3 than in wild-type FE cells, such that the brightness of MTs in immunostainings inversely correlates with the stability of the MT cytoskeleton, and brighter anti␣-tubulin staining thus indicates decreased MT stability. In addition, MTs of par-1W3 follicle cells are more sensitive to colcemid-induced depolymerization than are MTs of wild-type cells [4]. Because the spacer domain of PAR-1 is required for lateral accumulation of the protein, we determined whether this domain is required for PAR-1-dependent integrity of the FE MT cytoskeleton. To this end, we tested the ability of GFP-PAR-1 ⌬SP

Drosophila PAR-1 in Epithelial Polarity 257

Figure 2. The PAR-1 Spacer Domain Is Dispensable for Stabilization of the MT Cytoskeleton (A–D⬘) Single confocal IF sections stained to reveal the MT cytoskeleton in par-1W3 FE cells expressing the indicated GFP-PAR-1 construct. Single ␣-tubulin channels are shown in (A⬘), (B⬘), (C⬘), and (D⬘). The scale bar represents 5 ␮m.

to rescue the MT instability of par-1 null FE cells by expressing the mutant protein in par-1W3 cell clones (see Experimental Procedures) with a modification of the MARCM system [16]. We evaluated the stability of the MT cytoskeleton of the GFP-positive par-1W3 FE cells expressing GFP-PAR-1 ⌬SP (or control trangenes) by staining MTs with an anti-␣-tubulin antibody (Figure 2). As a negative control, we expressed GFP in par-1W3 FE cell clones and observed that in the mutant cells the MT cytoskeleton is less stable than that of neighboring wild-type cells (Figures 2A and 2A⬘), as previously reported [4]. In contrast, expression of the positive control, GFP-PAR-1 FL, in par-1W3 follicle cell clones results in a MT cytoskeleton as stable as that of neighboring wildtype cells (Figures 2B and 2B⬘). Expression of GFP-PAR1 K* results in unstable MTs, consistent with a role of the kinase domain in MT stabilization (Figures 2C and 2C⬘). Expression of GFP-PAR-1 ⌬SP in par-1W3 FE cell clones rescues MT stability (Figures 2D and 2D⬘). Finally, GFP-PAR-1 ⌬SP and GFP-PAR-1 FL behave as wildtype cells in colcemid-sensitivity experiments (data not shown). These data suggest that the spacer domain of PAR-1 is dispensable for PAR-1-dependent stabilization of the FE MT cytoskeleton. We next tested whether the spacer domain of PAR-1 is required for the formation and polarization of the FE. A loss of apical-basal polarity is observed in par-1W3 FE cell clones; the FE loses its integrity and becomes multilayered, and markers of polarity are lost or mislocalized. par-1 null cells eventually die or are lost from the epithelium, preventing the generation of large clones of mutant cells [3, 4]. To assess the ability of GFPPAR-1 ⌬SP to support complete development of the FE, we generated GFP-PAR-1 ⌬SP-expressing par-1W3 FE cell clones in the larval ovary and monitored the presence and morphology of GFP-positive follicle cells at stages 1–6 of oogenesis in the ovaries of the resulting adults (Figure 3). Expression of either GFP-PAR-1 FL or GFP-PAR-1 ⌬SP in par-1W3 FE cells rescues FE forma-

tion (Figures 3A–3D). GFP-PAR-1 FL supports development of a morphologically normal cuboidal monolayer of cells that differentiates into a columnar cell layer after stage 6 (Figures 3Aand 3B; [4]). Expression of GFPPAR-1 ⌬SP in par-1W3 FE cells supports development of an FE containing round cells that protrude from some portions of an otherwise normal cuboidal epithelium (Figures 3C and 3D). Multilayering of the FE is evident after stage 6 (see arrowheads in Figure 3D), correlates with a failure to differentiate a columnar FE over the oocyte (compare Figures 3D and 3B), and in severe cases results in egg-chamber degeneration around stages 8 and 9 (data not shown). GFP-PAR-1 ⌬SP is expressed at a level comparable to GFP-PAR-1 FL (Figure S2D). Therefore, the incomplete degree of rescue by GFP-PAR-1 ⌬SP is most likely not due to an insufficient overall amount of PAR-1 in GFP-PAR-1 ⌬SP-expressing cells, but rather to the absence of the spacer domain. We never observed par-1W3 follicles when expressing GFP-PAR-1 constructs in which the kinase domain was either lacking or mutated (data not shown), indicating that the kinase activity of PAR-1 is crucial to supporting the viability and development of follicle cells. Finally, GFP-PAR-1 C* expression in wild-type FE is sufficient to cause multilayering of the FE at the extremities of the egg chamber, suggesting that GFP-PAR-1 C* acts in a dominant-negative fashion (Figure S1D). We then assessed whether par-1W3 follicle cells rescued with GFP-PAR-1 ⌬SP are indeed epithelial cells by evaluating their polarization and arrest in interphase. First, we visualized by immunofluorescence (IF) the distribution of DE-cadherin, atypical kinase C (aPKC), and discs large (Dlg) (Figure 4), markers of apico-lateral adherens junctions, the apical PM domains, and the lateral PM domains, respectively [17–19]. We found that the apical-lateral localization of DE-cadherin is retained in GFP-PAR-1 ⌬SP-rescued cells that show an epitheliallike morphology and is lost in GFP-PAR-1 ⌬SP-rescued cells that lie out of the plane of the epithelium, at the

Current Biology 258

Figure 3. The PAR-1 Spacer Domain Is Required for FE Integrity Single confocal sections of egg chambers at stages 1–6 (A) and (C) or 8 (B) and (D) of oogenesis, in which par-1W3 FE cells express GFPPAR-1 FL (A) and (B) or GFP-PAR-1 ⌬SP (C) and (D). Arrowheads in (D) point to a multilayered region of the rescued FE. Sections in (B) and (D) were stained to reveal the actin cytoskeleton (phalloidin; red) and the nuclei (DAPI; blue).

extremities of the egg chambers (Figure 4A). Apical localization of aPKC and lateral localization of Dlg are frequently altered in a similar fashion (Figures 4B and 4C), suggesting that the spacer domain of PAR-1 is required for apical-basal polarization of FE cells. Next, we stained ovaries to detect the presence of cells positive for phospho-Histone 3, a marker of cells in mitosis. As a result of cell-division arrest of the FE at stage 6 of oogenesis, phospho-Histone-3-positive cells are never observed after this stage in wild-type FE or in par-1 null FE cells rescued with GFP-PAR-1 FL (data not shown). In contrast, phospho-Histone-3-positive cells are present beyond stage 6 exclusively within the multilayered portion of FE cells rescued with GFP-PAR-1 ⌬SP (Figure 4D). Taken together, these results indicate that the PAR-1 spacer domain is required for PAR-1 function in the epithelial organization of follicular cells but dispensable for stabilization of their MT cytoskeleton. We have shown that the PAR-1 spacer domain is specifically required for lateral PM localization in FE cells. PAR-1 does not contain any apparent transmembrane domain and appears to interact peripherally with membranes, probably through interaction with membrane bound proteins. Although PAR-1 lacking the spacer domain fails to localize to the lateral PM, it is detected on intracellular membranes. Therefore, two sets of interacting proteins may mediate PAR-1 localization to different types of membranes. GFP-PAR-1 FL, which does not contain the C-terminal domain, localizes to the lateral PM. This is in contrast to the mammalian homolog (EMK2/Par-1b), whose basolateral localization requires the C-terminal domain [9]. Therefore, it is probable that the mechanisms underlying PAR-1 membrane localization differ between species.

A Thr-to-Ala single-point mutation in the putative aPKC-target (C) site of the PAR-1 spacer domain causes PAR-1 mislocalization to the apical membrane. These data and the homology between site C and the aPKC phosphorylation site within the spacer domain of hPar1b and hPar-1a (MARK3/C-TAK1) suggest that apically localized aPKC might exclude PAR-1 from the apical plasma membrane by phosphorylation of site C, as recently shown in mammals [14]. We have also shown that expression of apically localized GFP-PAR-1 C* causes multilayering of the FE. PAR-1 is known to act on apical restriction of the aPKC/Bazooka/PAR-6 complex [6]. This suggests that residual PAR-1 on the apical membrane might affect FE organization by interfering with the formation of the aPKC/Bazooka/PAR-6 complex. Further analysis is required to determine whether Drosophila aPKC phosporylates site C and whether aPKC is instrumental in regulating the amount of PAR-1 on the plasma membrane. To address the role of the PAR-1 spacer domain in vivo, we assessed the capacity of GFP-PAR-1 ⌬SP, which shows a strong reduction of lateral PM localization in the FE, to suppress the phenotypes of par-1 null FE cell clones. Comparison of the phenotypes of par-1 null clones and of par-1 null clones rescued by GFPPAR-1 ⌬SP reveals that the latter display only a subset of the phenotypes of the former. Although they are viable and maintained in the FE, some of the GFP-PAR-1 ⌬SPrescued FE cells, especially those at the extremities of the egg chamber, fail to correctly localize aPKC to the apical membrane. These data suggest that, in the absence of the spacer domain, PAR-1 might fail to exclude the aPKC/Bazooka/PAR-6 complex from the lateral membrane. Multilayering at the egg chamber extremities is com-

Drosophila PAR-1 in Epithelial Polarity 259

Figure 4. The PAR-1 Spacer Domain Is Required for FE Polarization (A) Single confocal sections of stage 8 egg chambers, in which par-1W3 FE cells expressing GFP-PAR-1 ⌬SP (green) were stained to reveal DE-cadherin (A; DE-cad), aPKC (B), and Dlg (C). (A–C) Arrowheads point to the accumulation of mesenchymal-like cells in the FE. Single DE-cad, aPKC, and Dlg channels are shown. (D) Single confocal IF sections stained to reveal mitotic cells (p-Histone3; red) and the actin cytoskelton (phalloidin; purple) in egg chambers, in which some cells are par-1W3 FE cells expressing GFP-PAR-1 ⌬SP (green). Arrowheads point to dividing cells in the rescued FE. The scale bars represent 20 ␮m.

monly observed in FE mutants for other polarity genes, such as aPKC (our unpublished data), dlg, lgl, and scrib [20]. The mechanism of PAR-1-dependent lateral exclusion of the aPKC/Bazooka/PAR-6 complex is redundant with a second apical retention mechanism, provided by the CRB/STD/Patj complex [6]. This evidence suggests that physical constraints due to the geometry of the FE might demand activity of both the CRB/STD/Patj complex and of PAR-1 at the extremities of egg chambers. PAR-1 localized on the lateral PM might effect epithelial polarization by influencing other polarity molecules acting on the lateral PM, such as components of the Scribble/Discs large/Lethal(2)giant larvae complex. Consistent with this and with the fact that mutation of scrib, dlg, or lgl causes multilayering and loss of cell-cycle control of FE cells [20], a portion of GFP-PAR-1 ⌬SPrescued cells shows aberrant dlg localization and fails to arrest the cell cycle after stage 6. Alternatively, PAR-1

might act from the lateral PM to directly effect cellcycle control. In fact, one of the distantly related PAR-1 homologs, hPar-1a, regulates the cell cycle by 14-3-3dependent phosphorylation of Cdc25C [21]. GFP-PAR-1 ⌬SP-rescued cells show a stabilized MT cytoskeleton. Thus, lateral PM localization of PAR-1 might be essential for polarization but dispensable for MT stabilization. In hepatic cells, membranes of the Golgi apparatus can nucleate MTs [22]. It will be interesting to determine whether the cytosolic or the Golgiassociated Drosophila PAR-1 is responsible for MT stabilization in FE cells. Alternatively, it is possible that MT stabilization and FE polarization require different levels of PAR-1 activity on the lateral PM, such that the residual PAR-1 present on the lateral PM of GFP-PAR-1 ⌬SPrescued FE cells is sufficient to stabilize MT but not to effect polarization. Based on our data, we propose that, during the estab-

Current Biology 260

lishment of epithelial polarity, PAR-1 accumulates on the lateral PM via interactions between the spacer domain and an unknown membrane bound receptor, and possibly via exclusion from the apical PM by aPKC. From its lateral PM localization, PAR-1 might control apico-basal polarity by lateral exclusion of the aPKC/ Bazooka/PAR-6 complex and by regulating other polarity effectors. Experimental Procedures Molecular Biology pRM-GFP-PAR-1 (N1S) for expression of GFP-PAR-1 in S2 cells was generated by insertion of the PAR-1 N1S c-DNA, excised from pCaTub67C-GFP-PAR-1 (N1S) [10] via the BamHI/XhoI restriction sites, into the BamHI and SalI restriction sites of pRM-GFP (a gift of G. Jekely). The in-frame GFP fusion was reconstructed by the insertion of a KpnI/BamHI PCR GFP fragment. GFP-PAR-1 mutant constructs were created by PCR with pRM-GFP-PAR-1 (N1S) as a template; they were sequenced and cloned in the BamHI and SbfI sites of pRM-GFP. Primers were as follows: pRM-for: 5⬘-GTGCATCAGTTGTGGTCAGCAGC-3⬘, PAR-1 ⌬SP-rev: 5⬘-GATCCCTGCAGGTCAAAACCCCATGTTCATCCAC-3⬘, PAR-1 SP-for: 5⬘-GATCGGATCCATGGGGTTTGAGGAGG-3⬘, PAR-1 SP-rev: 5⬘-GATCCCTGCAGGTCAGGGACGTTTGCTAAAACG-3⬘, PAR-1 T*-for: 5⬘-CATGGGGTTTGCGGCGGCCGCACTCAAGCCCGCTATTGAG CCC-3⬘, PAR-1 K*-for: 5⬘-CAAAGCTGGACGCGTTCTGCGGTGCCCCGCCATATG-3⬘, pRM-rev: 5⬘-AATTAAGTAGAGCGTTTATCAG-3⬘. We generated the corresponding pUASp version of the PAR-1 mutants by subcloning the KpnI-SbfI fragment of the GFP-PAR-1 c-DNAs into the KpnI-SbfI sites of a modified version of pUASp [23].

Fly Genetics Fly lines carrying UAS GFP-PAR-1 mutant transgenes were established with standard techniques. For in vivo localization experiments (Figure 1), crossing GFP-PAR-1 flies with flies carrying the CY2GAL4 driver [24] or the GR1-GAL4 driver [25] allowed GAL4-mediated expression of the transgenes in the columnar FE. Between two and four independent GFP-PAR-1 lines per transgene were used. In vivo rescue experiments were performed with the following modification of the MARCM technique [26]: Flies carrying the par-1 null allele, par-1W3, recombined on the FRT G13 [2], and the appropriate UAS GFP-PAR-1 transgenes were crossed with flies carrying the heat-shock-inducible recombinase FLP, FRT G13 tubulin GAL80, and tubulin GAL4. After heat shock, expression of the rescue transgenes is obtained only in par-1W3 homozygous-mutant cells. These cells are the only ones devoid of the GAL80 repressor, which controls the expression of the transgenes by modulating GAL4 activity, and are positively marked by GFP as a result of the expression of the GFP-PAR-1 transgenes. To generate small clones (Figure 2), after eclosion we heat shocked flies twice for 1 hr at 37.5⬚C during 2 days, and dissected their ovaries for analysis 3 days after heat shock. To generate large follicle cell clones (Figures 3 and 4), we heat shocked first-instar larvae twice for 1 hr at 37.5⬚C during 2 days and dissected the ovaries of the resulting adult females 15 days after heat shock. No GFP expression was ever observed in non-heat-shocked controls.

S2 Cells For the localization studies in S2 cells (Figure S1), 1 ␮g pRM-GFPPAR-1 (N1S) or derivatives/106 S2 cells was transfected, and induction of expression of the transgenes was obtained by addition of 0.7 mM CuSO4 in the culture medium for 24 hr. S2 cells were propagated in 1⫻ Drosophila SFM medium (Gibco) containing 18 mM L-Glutamine, 50 U/ml penicillin, and 50 g/ml streptomycin.

Membrane Fractionations, Western Blotting, and IF Membrane-fractionation experiments were performed according to Kondylis and Rabouille 2003 [27] and began with 107 S2 cells or 20 ovaries per sample. Western blottings were performed as in Markussen et al., 1995 [28]. Primary antibodies used for Western blotting were as follows: affinity-purified rabbit anti-PAR-1 spacer (1:300) [7], affinity-purified rabbit anti-GFP (1:1000, Santa Cruz), and monoclonal mouse anti-tubulin DM1 (1:2500, Sigma). The blots were incubated with secondary antibodies coupled to HRP (1:250) and visualized by ECL (Amersham). Antibody stainings on ovaries were performed according to Tomancak et al., 2000 [7], with the exception of MT stainings, which were performed according to Doerflinger et al., 2003 [4]. IF on S2 cells was performed as follows: Cells were fixed in 4% paraformaldehyde (Polysciences) in PBS for 10 min, washed, and extracted in 0.5% Triton X-100 in PBS for 10 min, and incubated for 45 min with a dilution of the primary antibody in 0.5% BSA and 0.1% Triton X-100 in PBS. Cells were washed and incubated for 45 min with FITC- and rhodamine-conjugated secondary antibodies and phalloidin (1:500, Molecular Probes) in 0.1% Triton X-100 in PBS. Antibodies used for IF are: Mouse anti GP120 (1:100) [29], mouse anti-␣-tubulin DM1 (1:500 Sigma), rat anti-DE-cadherin DCAD-2 (1:100) [30], rabbit anti-aPKC (1:500, Santa Cruz), mouse anti-Dlg (1:100 Developmental Studies Hybridoma Bank), and rabbit anti-phospho-Histone 3 (1:500, Sigma). Confocal images (Leica NT) were edited with Photoshop (Adobe). IEM Ovaries dissected from 3-day-old Oregon R females were fixed in 2% PFA and 0.2% paraformahedyde in 0.2 M Phosphate buffer (pH 7.4) for 3 hr and processed for cryosectioning [31]. Ultrathin cryosections were labeled with rabbit polyclonal anti-GFP antibody diluted 1/100 (Molecular Probes) or the affinity-purified rabbit anti PAR-1 N1 [15], followed by 10 nm protein A gold secondary antibodies and embedding in methyl-cellulose. Sections were viewed under a Jeol transmission electron microscope. The percentage of gold particles on the different subcellular compartments shown in Table 1 was estimated in FE cells according to the following criteria. The Golgi area is defined by nonstacked cisternae and surrounding vesicles and tubules. Endosomes are defined as vacuoles, being electronlucent, or being filled with either vesicles or multiple sheets of membranes. The ER was recognized by the cisternae of about 90 nm width. The nucleus was recognized by the nuclear envelope. The free cytoplasmic pool represented the gold particles not associated with membranes. About 500 gold particles were assigned to the different compartments in two different experiments. Supplemental Data Supplemental figures are available with this article online at http:// www.current-biology.com/cgi/content/full/15/3/255/DC1/. Acknowledgments We thank Daniel St Johnston for the GFP-PAR-1 (N1S) c-DNA, Tadashi Uemura for the DCAD-2 antibody, and Ga´spar Je´kely for the pRM-GFP plasmid. We thank Vangelis Kondylis, Adrian Oprins, Elly Van Donselaar, and Marisa Oppizzi for technical assistance. We especially thank Piyi Papadaki for measuring the kinase activity of GFP-PAR-1 FL and GFP-PAR-1 K*, helping to generate the GFPPAR-1 C* flies, and for critically reading of the manuscript. We are grateful to David Bilder for the possibility of performing the final experiments of this paper in his lab. T.V. was supported by a fellowship from the European Molecular Biology Laboratory International Ph.D. Programme. Received: April 28, 2004 Revised: October 28, 2004 Accepted: November 11, 2004 Published: February 8, 2005 References 1. Pellettieri, J., and Seydoux, G. (2002). Anterior-posterior polarity in C. elegans and Drosophila–PARallels and differences. Science 298, 1946–1950.

Drosophila PAR-1 in Epithelial Polarity 261

2. Shulman, J.M., Benton, R., and St Johnston, D. (2000). The Drosophila homolog of C. elegans PAR-1 organizes the oocyte cytoskeleton and directs oskar mRNA localization to the posterior pole. Cell 101, 377–388. 3. Cox, D.N., Lu, B., Sun, T.Q., Williams, L.T., and Jan, Y.N. (2001). Drosophila par-1 is required for oocyte differentiation and microtubule organization. Curr. Biol. 11, 75–87. 4. Doerflinger, H., Benton, R., Shulman, J.M., and Johnston, D.S. (2003). The role of PAR-1 in regulating the polarised microtubule cytoskeleton in the Drosophila follicular epithelium. Development 130, 3965–3975. 5. Benton, R., and Johnston, D.S. (2003). A conserved oligomerization domain in drosophila Bazooka/PAR-3 is important for apical localization and epithelial polarity. Curr. Biol. 13, 1330–1334. 6. Benton, R., and Johnston, D.S. (2003). Drosophila PAR-1 and 14–3-3 inhibit Bazooka/PAR-3 to establish complementary cortical domains in polarized cells. Cell 115, 691–704. 7. Tomancak, P., Piano, F., Riechmann, V., Gunsalus, K.C., Kemphues, K.J., and Ephrussi, A. (2000). A Drosophila melanogaster homologue of Caenorhabditis elegans par-1 acts at an early step in embryonic-axis formation. Nat. Cell Biol. 2, 458–460. 8. Drewes, G., Ebneth, A., and Mandelkow, E.M. (1998). MAPs, MARKs and microtubule dynamics. Trends Biochem. Sci. 23, 307–311. 9. Bohm, H., Brinkmann, V., Drab, M., Henske, A., and Kurzchalia, T.V. (1997). Mammalian homologues of C. elegans PAR-1 are asymmetrically localized in epithelial cells and may influence their polarity. Curr. Biol. 7, 603–606. 10. Huynh, J.R., Shulman, J.M., Benton, R., and St Johnston, D. (2001). PAR-1 is required for the maintenance of oocyte fate in Drosophila. Development 128, 1201–1209. 11. Timm, T., Li, X.Y., Biernat, J., Jiao, J., Mandelkow, E., Vandekerckhove, J., and Mandelkow, E.M. (2003). MARKK, a Ste20-like kinase, activates the polarity-inducing kinase MARK/PAR-1. EMBO J. 22, 5090–5101. 12. Drewes, G., Ebneth, A., Preuss, U., Mandelkow, E.M., and Mandelkow, E. (1997). MARK, a novel family of protein kinases that phosphorylate microtubule-associated proteins and trigger microtubule disruption. Cell 89, 297–308. 13. Matter, K., Hunziker, W., and Mellman, I. (1992). Basolateral sorting of LDL receptor in MDCK cells: the cytoplasmic domain contains two tyrosine-dependent targeting determinants. Cell 71, 741–753. 14. Hurov, J.B., Watkins, J.L., and Piwnica-Worms, H. (2004). Atypical PKC Phosphorylates PAR-1 Kinases to Regulate Localization and Activity. Curr. Biol. 14, 736–741. 15. Vaccari, T., and Ephrussi, A. (2002). The fusome and microtubules enrich Par-1 in the oocyte, where it effects polarization in conjunction with Par-3, BicD, Egl, and dynein. Curr. Biol. 12, 1524–1528. 16. Lee, T., and Luo, L. (1999). Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis. Neuron 22, 451–461. 17. Oda, H., Uemura, T., and Takeichi, M. (1997). Phenotypic analysis of null mutants for DE-cadherin and Armadillo in Drosophila ovaries reveals distinct aspects of their functions in cell adhesion and cytoskeletal organization. Genes Cells 2, 29–40. 18. Cox, D.N., Seyfried, S.A., Jan, L.Y., and Jan, Y.N. (2001). Bazooka and atypical protein kinase C are required to regulate oocyte differentiation in the Drosophila ovary. Proc. Natl. Acad. Sci. USA 98, 14475–14480. 19. Goode, S., and Perrimon, N. (1997). Inhibition of patterned cell shape change and cell invasion by discs large during Drosophila oogenesis. Genes Dev. 11, 2532–2544. 20. Bilder, D., Li, M., and Perrimon, N. (2000). Cooperative regulation of cell polarity and growth by Drosophila tumor suppressors. Science 289, 113–116. 21. Peng, C.Y., Graves, P.R., Ogg, S., Thoma, R.S., Byrnes, M.J., 3rd, Wu, Z., Stephenson, M.T., and Piwnica-Worms, H. (1998). C-TAK1 protein kinase phosphorylates human Cdc25C on serine 216 and promotes 14–3-3 protein binding. Cell Growth Differ. 9, 197–208. 22. Chabin-Brion, K., Marceiller, J., Perez, F., Settegrana, C.,

23. 24.

25.

26.

27.

28.

29.

30.

31.

32.

Drechou, A., Durand, G., and Pous, C. (2001). The Golgi complex is a microtubule-organizing organelle. Mol. Biol. Cell 12, 2047– 2060. Rorth, P. (1998). Gal4 in the Drosophila female germline. Mech. Dev. 78, 113–118. Queenan, A.M., Ghabrial, A., and Schupbach, T. (1997). Ectopic activation of torpedo/Egfr, a Drosophila receptor tyrosine kinase, dorsalizes both the eggshell and the embryo. Development 124, 3871–3880. Gupta, T., and Schupbach, T. (2003). Cct1, a phosphatidylcholine biosynthesis enzyme, is required for Drosophila oogenesis and ovarian morphogenesis. Development 130, 6075–6087. Lee, T., and Luo, L. (2001). Mosaic analysis with a repressible cell marker (MARCM) for Drosophila neural development. Trends Neurosci. 24, 251–254. Kondylis, V., and Rabouille, C. (2003). A novel role for dp115 in the organization of ⫹ER sites in Drosophila. J. Cell Biol. 162, 185–198. Markussen, F.H., Michon, A.M., Breitwieser, W., and Ephrussi, A. (1995). Translational control of oskar generates short OSK, the isoform that induces pole plasma assembly. Development 121, 3723–3732. Stanley, H., Botas, J., and Malhotra, V. (1997). The mechanism of Golgi segregation during mitosis is cell type-specific. Proc. Natl. Acad. Sci. USA 94, 14467–14470. Oda, H., Uemura, T., Harada, Y., Iwai, Y., and Takeichi, M. (1994). A Drosophila homolog of cadherin associated with armadillo and essential for embryonic cell-cell adhesion. Dev. Biol. 165, 716–726. Liou, W., Geuze, H.J., and Slot, J.W. (1996). Improving structural integrity of cryosections for immunogold labeling. Histochem. Cell Biol. 106, 41–58. Cohen, D., and Musch, A. (2003). Apical surface formation in MDCK cells: regulation by the serine/threonine kinase EMK1. Methods 30, 269–276.