Current Biology, Vol. 15, 1847–1854, October 25, 2005, ©2005 Elsevier Ltd All rights reserved. DOI 10.1016/j.cub.2005.08.067

Quantitative Analysis of Protein Dynamicsduring Asymmetric Cell Division

Bernd Mayer,1,3 Gregory Emery,1,3 Daniela Berdnik,2

Frederik Wirtz-Peitz,1 and Juergen A. Knoblich1,*1Institute of Molecular Biotechnology of the Austrian

Academy of Sciences (IMBA)Dr.-Bohr-Gasse 3-51030 ViennaAustria2Department of Biological SciencesStanford UniversityStanford, California 94305

Summary

In dividing Drosophila sensory organ precursor (SOP)cells, the fate determinant Numb and its associatedadaptor protein Pon localize asymmetrically and seg-regate into the anterior daughter cell, where Numb in-fluences cell fate by repressing Notch signaling.Asymmetric localization of both proteins requires theprotein kinase aPKC and its substrate Lethal (2) giantlarvae (Lgl) [1–3]. Because both Numb and Pon local-ization require actin and myosin [4–6], lateral trans-port along the cell cortex has been proposed as apossible mechanism for their asymmetric distribution[5]. Here, we use quantitative live analysis of GFP-Pon and Numb-GFP fluorescence and fluorescencerecovery after photobleaching (FRAP) to characterizethe dynamics of Numb and Pon localization duringSOP division. We demonstrate that Numb and Ponrapidly exchange between a cytoplasmic pool and thecell cortex and that preferential recruitment from thecytoplasm is responsible for their asymmetric distri-bution during mitosis. Expression of a constitutivelyactive form of aPKC impairs membrane recruitmentof GFP-Pon. This defect can be rescued by coexpres-sion of nonphosphorylatable Lgl, indicating that Lglis the main target of aPKC. We propose that a high-affinity binding site is asymmetrically distributed byaPKC and Lgl and is responsible for asymmetric lo-calization of cell-fate determinants during mitosis.

Results and Discussion

Asymmetric segregation of cell-fate determinants is animportant mechanism to assign different fates to twodaughter cells [7–9]. In Drosophila neuroblasts, we knowfactors that are important for the asymmetric localiza-tion of the fate determinant Numb and the colocalizingand interacting protein Pon. A conserved protein com-plex, called the Par complex [10], localizes asymmetri-cally to one side of the neuroblast. The atypical proteinkinase C (aPKC) is a component of the Par protein com-plex [11–13]. The key phosphorylation substrate ofaPKC is the tumor-suppressor protein Lethal (2) giantlarvae (Lgl) [3], which is required for determinant local-

*Correspondence: juergen.knoblich@imba.oeaw.ac.at

3 These authors contributed equally to this work.

ization [1, 2]. Our current understanding is that phosphor-ylation of Lgl by aPKC on one side of the cell directsthe localization of Numb and Pon to the opposite sideto ensure their segregation into one of the two daughtercells. Lgl exists in two conformational states [14]: anonphosphorylated form that has an open conforma-tion and positively regulates membrane recruitment ofNumb and Pon, and a phosphorylated form that adoptsa closed conformation and is inactive. Expression ofa constitutively active form of Lgl, in which the aPKCphosphorylation sites are changed to alanine (Lgl3A),disrupts asymmetry of Numb and Pon and leads to uni-formly cortical localization [3]. Conversely, expressionof a truncated version of aPKC (aPKC-�N), which is nolonger asymmetric, leads to impaired membrane recruit-ment of Numb. Thus, Lgl is an important regulator ofasymmetric cell division although the mechanism bywhich it mediates Numb and Pon asymmetry is largelyunknown.

So far, mainly lateral diffusion or transport along thecell cortex has been suggested as a mechanism of Ponlocalization [5] (Figure 1D). This model is supported byexperiments showing that a functional actin cytoskele-ton is required for asymmetric protein localization [4, 15].Furthermore, myosin and Lgl interact in a phosphoryla-tion-dependent fashion [16], and inhibition of myosinmotor activity by treatment with 2,3-butanedione mon-oxime (BDM) leads to a uniformly cortical localizationof Pon [5]. Quantitative live-imaging techniques helpedus to gain more insight into the mechanisms governingasymmetric protein localization. Using confocal micro-scopy, we were able to image asymmetrically dividingDrosophila SOP cells within their native environment,associated with an epithelial sheet at the surface of thepupal notum. In a stereotyped lineage, SOP cells un-dergo a series of asymmetric divisions to give rise tomechanosensory organs [17, 18]. Fluorescence micro-scopy did not interfere with cell biological processesbecause fly pupae could still develop to the adult stagewith normally differentiated sensory organs after im-aging (data not shown).

Pon Localization during MitosisIn order to study the dynamics of asymmetric proteinlocalization, we recorded a time series of the divisionof an SOP cell expressing GFP-Pon [5] and Histone2B-RFP under the control of a specific promoter. His-tone2B-RFP was used to visualize DNA, thus allowingus to correlate distinct steps of GFP-Pon localizationwith other mitotic events (Figure 1A and Movie S1 in theSupplemental Data available with this article online). Ininterphase, some GFP-Pon was cortical, but a largepart localized to the cytoplasm. As the cell entered mi-tosis, it rounded up and underwent strong membraneblebbings, indicative of local rearrangements of thecortical cytoskeleton. Interestingly, similar blebbingevents were also observed in the first division of theC. elegans zygote (reviewed in [9]). Unlike in SOP cells,however, they only occur on the anterior side of the

Current Biology1848

Figure 1. Distinct Phases of Pon AsymmetricLocalization in SOP Cells

(A) Fluorescence live microscopy of GFP-Pon(green) and Histone2B-RFP (red) (stills fromthe division of a single SOP cell, Movie S1).Yellow arrowheads indicate cortical blebbing,blue arrowheads indicate random accumula-tions of GFP-Pon, and the white arrowhead in-dicates early GFP-Pon crescent; NEBD standsfor nuclear-envelope breakdown.(B) Positions relative to the site of crescentformation of transient random accumula-tions determined for four different movies.(C) Quantification of cortical (red line) andcytoplasmic (blue line) GFP-Pon fluorescence(mean and SEM, n = 3) during a mitotic se-quence (0 s corresponds to sister-chromatidseparation). Intensities are expressed as frac-tions of total fluorescence in the cell.(D) Cortical transport model for determinantlocalization.(E) Direct recruitment model.

C. elegans zygote, where Par-3/6 localize. Shortly afterblebbing had started, chromosomes condensed andGFP-Pon accumulated on random sites of the cell cor-tex. The accumulations were transient and did not nec-essarily predict the position of the final Pon crescent(Figure 1B). This suggested that the process leading toPon accumulation can take place all around the cell butis reinforced specifically in the crescent region. SomeGFP-Pon was also observed at the nucleus. This signalmight be due to GFP-Pon binding to the nuclear enve-lope or to the endoplasmic reticulum, and it disappearsslowly after nuclear-envelope breakdown. At nuclear-envelope breakdown (Figure 1A), cortical blebbingceased, the cell cortex smoothed, and first signs ofasymmetric localization of GFP-Pon into an anteriorcortical crescent were observed. As the cell progressedinto metaphase, the GFP-Pon signal in the crescentarea became stronger. Surprisingly, the intensity of thecortical area opposite of the crescent was almost notchanged during this process. Thus, GFP-Pon might ac-tually be recruited to the crescent directly from the cy-toplasm rather than being transported along the cell

c(ccGr(

rkdicgu

PEAedt

ortex. Indeed, quantification of fluorescence intensityFigure 1C) showed that GFP-Pon recruitment at theell cortex was accompanied by a comparable loss ofytoplasmic GFP-Pon. Note that local degradation ofFP-Pon in the cytoplasm is not responsible for this

eduction because total GFP-Pon remained unchangednot shown).

Subsequently, the metaphase plate was oriented withespect to the crescent [8] (Figure 1A), and during cyto-inesis, GFP-Pon segregated largely into the anterioraughter cell. We propose that GFP-Pon localization

s a two-step process involving the establishment of aortical area where the crescent will form and the pro-ressive recruitment of protein to the predefined sitentil metaphase.

on and Numb at the Cell Cortex Are in a Dynamicquilibrium with Cytoplasmic Proteinsymmetry of Numb and Pon could be created by lat-ral movement along the cell cortex (Figure 1D) or byirect recruitment from the cytoplasm to one side ofhe cell cortex (Figure 1E). To quantify the exchange of

Protein Dynamics during Asymmetric Cell Division1849

Numb and Pon between the cell cortex and the cyto-plasm, we used fluorescence recovery after photo-bleaching (FRAP) of GFP fusions to Numb [19] and Pon[5]. Numb-GFP can partially rescue the numb mutantphenotype [20], indicating that it is functional. GFP-Poncontains just the asymmetric-localization domain. Itsrescue behavior is unknown, but it colocalizes with en-dogenous Pon (Figure S1A) throughout mitosis. Whencytoplasmic GFP-Pon is photobleached, fluorescencerecovers with a half-time of 0.48 s (standard error of themean [SEM] ± 0.05 s, n = 3), indicating that diffusion isnot limiting (Figure 2E and Figure S1C). Recovery ofcortical GFP-Pon fluorescence occurred with single ex-ponential kinetics and a half-time of 35 s (SEM ± 4 s,n = 10), whereas the half-time for Numb-GFP was 27 s(SEM ± 5 s, n = 6) (Figure 2E). Therefore, Numb and Ponshowed a surprisingly dynamic association with thecell cortex.

Either cortical recruitment of cytoplasmic GFP-Ponor lateral diffusion/transport of cortical GFP-Pon couldbe responsible for fluorescence recovery. To measurethe exchange between cortical and cytoplasmic Pon,we repeatedly photobleached an area covering approx-imately 40% of the cytoplasm in an SOP cell express-ing GFP-Pon (Figure 2C#). Fluorescence intensity wassimultaneously recorded at the cortex. Cortical fluores-cence intensity dropped to less than 5% with a half-time of 52 s (SEM ± 4.8 s, n = 5). Thus, the cortical andcytoplasmic pools of GFP-Pon rapidly interchange witha mobile fraction of more than 95% (Figure 2C).

When the dynamic association with the cell cortex istaken into account, Pon asymmetry could be explainedeither by fast and continuous lateral transport or by di-rected recruitment to an asymmetric cortical bindingsite. To determine the contribution of lateral transport,we compared FRAP rates on the edge and in the centerof a photobleached region within the GFP-Pon cres-cent. The bleached region was defined such that a re-gion of nonbleached molecules was left behind at theedges of the crescent after photobleaching (Figure 2Aand Movie S2). To avoid recovery from above and be-low the image plane, we used a protocol in which theregion of interest was bleached in several planes (seeExperimental Procedures). The efficiency of this pro-cedure was confirmed by 3D reconstruction after pho-tobleaching in fixed tissue (Figure S1B). FRAP curvesfrom ten experiments were averaged. Their superposi-tion shows that the two regions recover nearly iden-tically with half-times of 32 s (SEM ± 3 s, n = 10) for aregion close to nonbleached GFP-Pon and of 35 s(SEM ± 4 s, n = 10) for a region farther away (centralregion) (Figure 2B). This is further supported by a kymo-graph plot of a cortical GFP-Pon crescent recoveringfrom photobleaching, where we did not detect move-ment of the edges of the bleached region (Figure 2D). Asimilar kymograph plot was obtained at prometaphase,although at this stage the apparent recovery is faster,presumably because additional Pon is recruited fromthe cytoplasm, and the cortical intensity actually in-creases during the experiment (Figure 2D). Taken to-gether, these observations suggest a model where Ponis preferentially recruited from the cytoplasm to the siteof crescent formation (Figure 1E). We propose that acortical high-affinity binding site for Pon is established

during mitosis and mediates specific recruitment ofPon to one side of the cell cortex. In a recent study [21],FRAP and kymograph analysis were used to analyzedynamics of S. cerevisiae polar caps of Cdc42p. In con-trast to Pon, Cdc42p exchanges more often with neigh-boring parts of the plasma membrane than with the cy-toplasm. Therefore, rapid exchange with the cytoplasmis not a general property of all asymmetrically local-ized proteins.

Lgl Phosphorylation Is Required for Pon CorticalRecruitment in SOP CellsTo test the role of Lgl in asymmetric protein localizationin SOP cells, we measured cortical recruitment of GFP-Pon in lgl1 mutant clones. In a similar experiment, Lglwas shown to be dispensable for Pon localization [20],although Pon recruitment seemed to be delayed [22].We calculated the ratio between total cortical and totalcytoplasmic fluorescence. Because GFP fluorescenceintensity is proportional to GFP-Pon concentration, thisratio should give a good estimate of the fraction ofGFP-Pon localized at the cell cortex. Although GFP-Ponwas still asymmetric (Figure 3A), quantitative analysisrevealed that the cortical GFP-Pon fraction was slightlybut significantly reduced (p < 0.002, Student’s t test) inlgl1 mutant clones (Figure 3B), consistent with previousfindings [22] describing a temporal delay of recruit-ment. This might be a hypomorphic phenotype causedby small residual amounts of Lgl protein present in themutant clones. We therefore used expression of dereg-ulated aPKC (aPKC-DN) as another means to inactivateLgl. Expression of aPKC-DN was shown to phenocopylgl mutants in embryonic tissues [3], presumably be-cause it phosphorylates and inactivates Lgl all aroundthe cell. In contrast to lgl1 mutant SOP cells, we ob-served a much stronger decrease of cortical GFP-Ponrecruitment upon aPKC-DN expression (Figures 3C and3D). Still, a slight cortical asymmetry was observed,which we think is due to the presence of endogenousaPKC. Even at anaphase, the degree of recruitmenthardly reached that of control cells in prophase. To testwhether Lgl phosphorylation was responsible for thisphenotype, we coexpressed aPKC-DN with nonphos-phorylatable lgl3A. Expression of lgl3A completely res-cued the cortical-recruitment defect (Figures 3C and3D). The observed differences are not due to increasedprotein levels because total cellular GFP-Pon fluores-cence remains constant (data not shown).

Thus, active, nonphosphorylated Lgl was needed forcortical recruitment of GFP-Pon although lgl1 mutantclones did not show a very strong phenotype. The easi-est explanation for the discrepancy between the lgl1

mutant and ectopic Lgl phosphorylation is the perdu-rance of residual Lgl protein in mutant tissue. This issupported by previous observations describing Numb-localization defects in temperature-sensitive alleles oflgl [1]. It is possible that Lgl can mediate its effects evenat protein concentrations below the detection limit ofthe antibody. Thus, Lgl may not be needed at stoichio-metric levels for asymmetric protein localization in SOPcells, but it instead plays a catalytic or signaling role.

Current Biology1850

Figure 2. Dynamic Exchange between Cortical and Cytoplasmic Pools of GFP-Pon

(A) Differential FRAP in metaphase SOP cells of two regions within a bleached area (gray rectangle) of the cortical crescent: close to non-bleached GFP-Pon (border, orange rectangle) and farther away (center, blue rectangle). Images are stills from one FRAP experiment (Movie S2).(B) Corresponding FRAP curves showed very similar characteristics: Half-times of recovery were 32 s (SEM ± 3 s, n = 10) at the border of thebleached region and 35 s (SEM ± 4 s, n = 10) in the center of the bleached region (FRAP rates are comparable, p > 0.5, Student’s t test, two-tailed equal variance). Error bars in all figures represent SEM. Numbers of experiments are shown in the diagrams.(C) Exchange between cortical and cytoplasmic GFP-Pon: Cortical fluorescence intensity decreases to less than 5% during repeated photo-bleaching of cytoplasmic GFP-Pon. (C#) Schematic representation of the area photobleached in (C).(D) Kymograph of FRAP experiment during prometaphase and metaphase (time resolution: 500 msec).(E) FRAP half-times of cortical GFP-Pon in control SOP cells and cells expressing either lgl3A (n = 4) or aPKC-DN (n = 5) are similar. FRAP ofcytoplasmic GFP-Pon is much faster, indicating that GFP-Pon diffusion is not rate limiting (n = 3). The FRAP half-time of cortical Numb-GFPis shown in red (n = 6). Brackets indicate p > 0.5.

Lgl Phosphorylation Regulates the Distributionof Numb and Pon Binding SitesHow could Lgl recruit Pon to the cell cortex? Formally,it is possible that Pon simply binds Lgl in a phosphory-lation-dependent manner. However, no direct interac-tion has been described and such a model would notexplain why Pon is cortical even when Lgl levels arestrongly reduced. Two other models are more likely:

Eeaansmb

ither cortical binding sites for Numb and Pon are pres-nt all around the cell, but their affinity depends on Lglnd its phosphorylation status and therefore varieslong the cell cortex (Model 1, Figure 4E); or a limitingumber of cortical binding sites are present only on oneide of the cell, and Lgl is responsible for their asym-etric distribution (Model 2, Figure 4F). To distinguishetween these models, we measured FRAP rates for

Protein Dynamics during Asymmetric Cell Division1851

Figure 3. Lgl Phosphorylation Regulates Cortical Recruitment of Pon

(A) Cortical recruitment of GFP-Pon in prometaphase, metaphase, and anaphase of control cells and cells within lgl1 mutant clones (grown at 25°C).(B) Quantification of the ratio between cortical and cytoplasmic GFP-Pon fluorescence intensity (mean and SEM). The relative recruitment ofGFP-Pon to the cell cortex is slightly reduced in SOP cells within lgl1 mutant clones (p < 0.002).(C) Cortical recruitment of GFP-Pon in prometaphase, metaphase, and anaphase in control cells or cells expressing either aPKC-DN alone ortogether with lgl3A (grown at 30°C).(D) Quantification as in (B) (mean and SEM). SOP cells expressing aPKC-DN show a strong defect in relative cortical recruitment of GFP-Pon(p < 0.001), which is rescued upon coexpression of lgl3A (p < 0.001). All panels in (A) and (C) are images from independent cells. Fluorescenceintensity is represented in pseudocolors according to the scale included.

cortical GFP-Pon in different genetic backgrounds. TheFRAP rate is a function of the rate constants for bothassociation and dissociation of GFP-Pon with its pos-tulated cortical binding site (see Experimental Pro-cedures for details of the proposed kinetic model). InModel 1, expression of activated lgl (lgl3A) or deregu-lated aPKC (aPKC-DN) should alter the affinity of thebinding site and therefore change the rate constants,resulting in a variation of the FRAP rate. Because theFRAP rate is independent of receptor concentration,however, it would remain constant under the same con-ditions in Model 2. We measured cortical GFP-PonFRAP rates in wild-type SOP cells [t(1/2)FRAP = 35 s,SEM ± 4s, n = 10], in cells expressing lgl3A [t(1/2)FRAP =32 s, SEM ± 3 s, n = 4], and in cells where Lgl wasinactivated by expression of aPKC-DN [t(1/2) =

FRAP

36 s, SEM ± 5 s, n = 6] (Figure 2E). Although expressionof aPKC-DN dramatically reduced the amount of GFP-Pon present at the cortex, it did not influence the kinet-ics of GFP-Pon binding to the cortical binding site.Thus, the number of Pon binding sites at the cell cortex,and not their affinity for Pon, seems to be reduced byaPKC-DN expression.

To gain independent evidence for the two models, wequantified the fraction of GFP-Pon present at the cellcortex. If Lgl regulated GFP-Pon binding site affinity,expression of lgl3A would change the entire SOP cellcortex to high affinity, and therefore it would increasethe cortical GFP-Pon fraction (Figure 4E). If Lgl onlyregulated the distribution of binding sites, however, thecortical fraction of GFP-Pon should remain the same(Figure 4F). We quantified cortical recruitment by mea-

Current Biology1852

Figure 4. Lgl Regulates the Distribution of Cortical Binding Sites

(A) Cortical recruitment of GFP-Pon in SOP cells: control and expression of lgl3A.(B) lgl3A expression does not increase the relative recruitment of GFP-Pon to the cell cortex although it allows equal GFP-Pon binding allaround the cell (mean and SEM, brackets indicate p > 0.5).(C) Cortical recruitment of Numb-GFP in SOP cells: control and expression of lgl3A. Note that Numb-GFP is not always completely symmetricafter Lgl3A expression (data not shown).(D) Expression of lgl3A does not increase the cortical pool of Numb-GFP (mean and SEM, brackets indicate p > 0.45).(E and F) Models for Pon or Numb localization and potential effect of lgl3A expression (Pon/Numb in yellow, high- and low-affinity bindingsites in orange and white, respectively). (E) Lgl regulates the affinity of cortical Pon or Numb binding sites in a phosphorylation-dependentmanner. (F) Lgl regulates the asymmetric distribution of binding sites.(G) Cortical binding sites can be saturated. Increased GFP-Pon expression (up to approximately 3.5-fold over the expression at 18°C), withhigher temperatures or two copies of the GFP-Pon transgene, leads to a lower cortical/cytoplasmic intensity ratio (mean and SEM, * p < 0.02,** p < 0.01). All panels in (A) and (C) are images from independent cells. Fluorescence intensity is represented in pseudocolors according tothe scale included.

suring the ratio of cortical to cytoplasmic fluorescencefor GFP-Pon (Figure 4A and 4B) and Numb-GFP (Figure4C and 4D) at different time points in mitosis. Com-pared to wild-type cells, expression of lgl3A did not

icac

cause a significant increase in cortical recruitment. This

s not because cytoplasmic GFP-Pon is limiting; in-reased GFP-Pon expression (by incubation of pupaet higher temperature, or by introduction of a secondopy of the GFP-Pon transgene) predominantly in-

creased the cytoplasmic signal (Figure 4G). Taken to-

Protein Dynamics during Asymmetric Cell Division1853

gether, our results favor Model 2 (Figure 4F), in whichLgl acts by asymmetrically distributing a limiting num-ber of cortical GFP-Pon binding sites. The loss of corti-cal fluorescence upon aPKC-DN expression indicatesthat—in addition to positioning the binding sites—lgl isalso required for their formation. However, this secondrole of lgl does not seem to be rate limiting under nor-mal conditions because lgl3A expression does not in-crease the cortical GFP-Pon fraction. Although our re-sults are most consistent with Model 2, more-complexmodels cannot be excluded. For example, lgl could dis-tribute a limiting adaptor protein that links Pon to a re-ceptor but is not the receptor itself.

The direct cortical binding partners for Pon or Numbhave not yet been identified. Thus, we can only specu-late on the molecular mechanisms of their postulatedasymmetric distribution. Although our results are incon-sistent with lateral transport of GFP-Pon, they do notexclude lateral transport of its cortical anchor. Similarto what has been proposed for asymmetric cell divisionin C. elegans [9], a possible mechanism could be localtearing and contraction of the cortical actin cytoskele-ton. Lgl was shown to inhibit the cortical localization ofmyosin II, and it has been proposed that cortical myo-sin II might exclude asymmetrically segregating pro-teins [6]. These data could be integrated with our modelif myosin II excludes the cortical binding sites ratherthan influencing determinant localization directly. Alter-natively, transmembrane receptors for Pon or Numbcould be delivered to the position of crescent formationby vesicle transport (as proposed before for determi-nants [2]). Such a mechanism in which transmembranereceptors are present on vesicles that dock at themembrane in an Lgl-dependent fashion would be con-sistent with our quantitative observations. It would alsoexplain why Lgl is essential for crescent formation butnot needed in metaphase for maintenance of asymmet-ric protein localization [1]. It is remarkable that lateraldiffusion of transmembrane proteins is slow enough toallow a stable asymmetric distribution, if the delivery ofthe protein is asymmetric, both in yeast [23] and in SOPcells (F.W.-P. and J.A.K., unpublished data). The yeastLgl orthologs Sro7p and Sro77p have been implicatedin plasma-membrane fusion of secretory vesicles [24],and Lgl has been proposed to regulate vesicular tar-geting to specific membrane domains [2]. Furthermore,asymmetric protein localization in Drosophila requiresmyosin VI, a motor whose main function is vesiclemovement [25], suggesting that vesicle trafficking playssome role.

Our experiments show how quantitative analysis ofbiological processes can lead to new mechanistic in-sights. In cultured cells, quantitative measurements ofGFP fluorescence have been used to determine associ-ation constants of protein complexes [26]. Althoughsuch experiments cannot yet be done in whole livingmulticellular organisms, improved bio-optical tools willcertainly expand the territory for quantitative biologicalanalysis in the future.

ConclusionsOur data provide insight into the dynamic protein move-ments of cell-fate determinants and their associated

adaptor proteins during asymmetric cell division. We

propose that these determinants are preferentially re-cruited from the cytoplasm to a high-affinity bindingsite during late prophase. Establishment of this bindingsite is regulated by the phosphorylation status of Lgl.The role of Lgl is more to concentrate binding sites onone side of the cell than to act as a receptor itself orchange the affinity of another Numb or Pon binding site.

Experimental Procedures

Flies and Antibodylgl1 clones were generated with the Ubx-flp MARCM [27] system.Ubx-Flp was generated by inserting two copies of the Ubx en-hancer fragment PBX-41 (gift from M. Bienz) into pCaSpeR-hsFlp,which carries the Flp recombinase under control of a completehsp70 promoter. Ubx-flp induces recombination in all imaginaldiscs. To generate Histone2B-RFP, we PCR-cloned the Histone2Bcoding region into pDONR (Invitrogen) by BP recombination. LRrecombination with the destination vector pUAST-DEST15 con-taining monomeric red fluorescent protein (RFP) [28] generatedpUAST-His2B::mRFP1, which was used for transgenic flies. neur-Gal4 [29] was used to express UAS-histone2B-rfp, UAS-numb-gfp[19, 20], UAS-gfp-pon [5], and UAS-lgl3A [3] in SOP cells. Expres-sion of UAS-aPKC-DN [3] under the control of neur-Gal4 and tub-Gal80ts [30] was achieved by shifting pupae between 0 and 13–14hr APF to 30°C before imaging.

Rabbit anti-Pon is a gift from Y.N. Jan [31]; it does not recognizeGFP-Pon and was visualized with a Cy5 coupled secondary anti-body (Jackson).

Confocal Microscopy and FRAPLive imaging of Drosophila pupae was as described [29]. Imageswere recorded on a Zeiss LSM510 confocal microscope. All photo-bleaching experiments were done by scanning the bleach region atmaximum laser intensity with a high numerical aperture objective.Cortical photobleaching was performed in a Z stack of 4–5 planeswith a distance of 2 �m each with bleach regions of identical geom-etry. During the recovery period, images from a single focal plane(see Figure S1B) were recorded at lower laser intensity with timeintervals of 5 s. Average intensities of the region of interest werecorrected for background and quantified with Metamorph (Univer-sal Imaging). The resulting curves were fitted to a single exponen-tial function y = A(1 − e−kt) with Origin 6.1 (OriginLab) from which theFRAP half-time t(1/2) = ln(2)/k was calculated. Kymographs weregenerated with Metamorph along a line that followed the cell cortexand was defined for a time series of images (time interval approxi-mately 500 msec).

The kinetic model is based on the following binding equilibrium:

GFP− Pon + Binding site ⇄koff

kon

GFP− Ponbound

In this model, the apparent FRAP rate depends not only on kon butalso on koff because bleached GFP-Pon molecules have to leavetheir binding sites before unbleached GFP-Pon can bind. The FRAPrate can be calculated as: kFRAP = [GFP-Pon] × kon + koff [32].

Fluorescence Ratio MeasurementsFor calculating the cortical recruitment of GFP-Pon and Numb-GFP,integrated fluorescence intensities of the whole SOP cell (Itot) andthe cytoplasm (Icyt) were measured and corrected for backgroundfluorescence. The ratio R of cortical and cytoplasmic fluorescencewas calculated as: R = (Itot − Icyt)/Icyt. After the measurements,images were processed with Adobe Photoshop as follows. After aGaussian blur was performed, the levels tool was used to increasedynamic range. Background fluorescence was subtracted, and apseudocolor lookup table was applied in Metamorph (the scale barshown in the figures corresponds to the original intensities).

Supplemental DataSupplemental Data include one supplemental figure and two mov-

ies and are available with this article online at http://www.current-biology.com/cgi/content/full/15/20/1847/DC1/.

Current Biology1854

Acknowledgments

We want to thank Karin Paiha for excellent help with microscopyand image processing; Vic Small for helpful suggestions on FRAP;and Y.N. Jan, F. Roegiers, and the Bloomington Drosophila StockCenter for antibodies and flystocks. Work in J.A.K.’s lab is sup-ported by Boehringer Ingelheim, the Austrian Academy of Sci-ences, and the Austrian Research Fund (FWF). G.E. is supportedby the Swiss National Science Foundation, and F.W.-P. is supportedby the Boehringer Ingelheim Fonds.

Received: June 10, 2005Revised: August 30, 2005Accepted: August 31, 2005Published: October 25, 2005

References

1. Ohshiro, T., Yagami, T., Zhang, C., and Matsuzaki, F. (2000).Role of cortical tumour-suppressor proteins in asymmetric divi-sion of Drosophila neuroblast. Nature 408, 593–596.

2. Peng, C.Y., Manning, L., Albertson, R., and Doe, C.Q. (2000).The tumour-suppressor genes lgl and dlg regulate basal pro-tein targeting in Drosophila neuroblasts. Nature 408, 596–600.

3. Betschinger, J., Mechtler, K., and Knoblich, J.A. (2003). The Parcomplex directs asymmetric cell division by phosphorylatingthe cytoskeletal protein Lgl. Nature 422, 326–330.

4. Knoblich, J.A., Jan, L.Y., and Jan, Y.N. (1997). The N terminusof the Drosophila Numb protein directs membrane associationand actin-dependent asymmetric localization. Proc. Natl. Acad.Sci. USA 94, 13005–13010.

5. Lu, B., Ackerman, L., Jan, L.Y., and Jan, Y.N. (1999). Modes ofprotein movement that lead to the asymmetric localization ofpartner of Numb during Drosophila neuroblast division. Mol.Cell 4, 883–891.

6. Barros, C.S., Phelps, C.B., and Brand, A.H. (2003). Drosophilanonmuscle Myosin II promotes the asymmetric segregation ofcell fate determinants by cortical exclusion rather than activetransport. Dev. Cell 5, 829–840.

7. Bardin, A.J., Le Borgne, R., and Schweisguth, F. (2004). Asym-metric localization and function of cell-fate determinants: Afly’s view. Curr. Opin. Neurobiol. 14, 6–14.

8. Betschinger, J., and Knoblich, J.A. (2004). Dare to be different:Asymmetric cell division in Drosophila, C. elegans, and verte-brates. Curr. Biol. 14, R674–R685.

9. Cowan, C.R., and Hyman, A.A. (2004). Asymmetric cell divisionin C. elegans: Cortical polarity and spindle positioning. Annu.Rev. Cell Dev. Biol. 20, 427–453.

10. Ohno, S. (2001). Intercellular junctions and cellular polarity: ThePAR-aPKC complex, a conserved core cassette playing funda-mental roles in cell polarity. Curr. Opin. Cell Biol. 13, 641–648.

11. Izumi, Y., Hirose, T., Tamai, Y., Hirai, S., Nagashima, Y., Fuji-moto, T., Tabuse, Y., Kemphues, K.J., and Ohno, S. (1998). Anatypical PKC directly associates and colocalizes at the epithe-lial tight junction with ASIP, a mammalian homologue of Caeno-rhabditis elegans polarity protein PAR-3. J. Cell Biol. 143, 95–106.

12. Wodarz, A., Ramrath, A., Grimm, A., and Knust, E. (2000). Dro-sophila atypical protein kinase C associates with Bazooka andcontrols polarity of epithelia and neuroblasts. J. Cell Biol. 150,1361–1374.

13. Rolls, M.M., Albertson, R., Shih, H.P., Lee, C.Y., and Doe, C.Q.(2003). Drosophila aPKC regulates cell polarity and cell prolif-eration in neuroblasts and epithelia. J. Cell Biol. 163, 1089–1098.

14. Betschinger, J., Eisenhaber, F., and Knoblich, J.A. (2005). Phos-phorylation-induced autoinhibition regulates the cytoskeletalprotein Lethal (2) giant larvae. Curr. Biol. 15, 276–282.

15. Shen, C.P., Knoblich, J.A., Chan, Y.M., Jiang, M.M., Jan, L.Y.,and Jan, Y.N. (1998). Miranda as a multidomain adapter linkingapically localized Inscuteable and basally localized Staufenand Prospero during asymmetric cell division in Drosophila.Genes Dev. 12, 1837–1846.

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

6. Strand, D., Jakobs, R., Merdes, G., Neumann, B., Kalmes, A.,Heid, H.W., Husmann, I., and Mechler, B.M. (1994). The Dro-sophila lethal(2)giant larvae tumor suppressor protein formshomo-oligomers and is associated with nonmuscle myosin IIheavy chain. J. Cell Biol. 127, 1361–1373.

7. Bellaiche, Y., and Schweisguth, F. (2001). Lineage diversity inthe Drosophila nervous system. Curr. Opin. Genet. Dev. 11,418–423.

8. Roegiers, F., and Jan, Y.N. (2004). Asymmetric cell division.Curr. Opin. Cell Biol. 16, 195–205.

9. Roegiers, F., Younger-Shepherd, S., Jan, L.Y., and Jan, Y.N.(2001). Two types of asymmetric divisions in the Drosophilasensory organ precursor cell lineage. Nat. Cell Biol. 3, 58–67.

0. Justice, N., Roegiers, F., Jan, L.Y., and Jan, Y.N. (2003). Lethalgiant larvae acts together with numb in notch inhibition andcell fate specification in the Drosophila adult sensory organprecursor lineage. Curr. Biol. 13, 778–783.

1. Wedlich-Soldner, R., Wai, S.C., Schmidt, T., and Li, R. (2004).Robust cell polarity is a dynamic state established by couplingtransport and GTPase signaling. J. Cell Biol. 166, 889–900.

2. Langevin, J., Le Borgne, R., Rosenfeld, F., Gho, M., Schweis-guth, F., and Bellaiche, Y. (2005). Lethal giant larvae controlsthe localization of notch-signaling regulators numb, neuralized,and Sanpodo in Drosophila sensory-organ precursor cells.Curr. Biol. 15, 955–962.

3. Valdez-Taubas, J., and Pelham, H.R. (2003). Slow diffusion ofproteins in the yeast plasma membrane allows polarity to bemaintained by endocytic cycling. Curr. Biol. 13, 1636–1640.

4. Lehman, K., Rossi, G., Adamo, J.E., and Brennwald, P. (1999).Yeast homologues of tomosyn and lethal giant larvae functionin exocytosis and are associated with the plasma membraneSNARE, Sec9. J. Cell Biol. 146, 125–140.

5. Petritsch, C., Tavosanis, G., Turck, C.W., Jan, L.Y., and Jan,Y.N. (2003). The Drosophila myosin VI Jaguar is required forbasal protein targeting and correct spindle orientation in mi-totic neuroblasts. Dev. Cell 4, 273–281.

6. Rabut, G., Doye, V., and Ellenberg, J. (2004). Mapping the dy-namic organization of the nuclear pore complex inside singleliving cells. Nat. Cell Biol. 6, 1114–1121.

7. Lee, T., and Luo, L. (1999). Mosaic analysis with a repressiblecell marker for studies of gene function in neuronal morpho-genesis. Neuron 22, 451–461.

8. Campbell, R.E., Tour, O., Palmer, A.E., Steinbach, P.A., Baird,G.S., Zacharias, D.A., and Tsien, R.Y. (2002). A monomeric redfluorescent protein. Proc. Natl. Acad. Sci. USA 99, 7877–7882.

9. Bellaiche, Y., Gho, M., Kaltschmidt, J.A., Brand, A.H., andSchweisguth, F. (2001). Frizzled regulates localization of cell-fate determinants and mitotic spindle rotation during asymmet-ric cell division. Nat. Cell Biol. 3, 50–57.

0. McGuire, S.E., Mao, Z., and Davis, R.L. (2004). Spatiotemporalgene expression targeting with the TARGET and gene-switchsystems in Drosophila. Sci. STKE 2004, pl6.

1. Lu, B., Rothenberg, M., Jan, L.Y., and Jan, Y.N. (1998). Partnerof Numb colocalizes with Numb during mitosis and directsNumb asymmetric localization in Drosophila neural and muscleprogenitors. Cell 95, 225–235.

2. Robia, S.L., Ghanta, J., Robu, V.G., and Walker, J.W. (2001).Localization and kinetics of protein kinase C-epsilon anchoringin cardiac myocytes. Biophys. J. 80, 2140–2151.