A cyclase-associated protein regulates actin and cell polarity during Drosophila oogenesis and in yeast

A cyclase-associated protein regulates actin and cell polarity during Drosophila oogenesis and in yeast
of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  964 Research Paper  A cyclase-associated protein regulates actin and cell polarityduring Drosophila oogenesis and in yeast Buzz Baum, Willis Li and Norbert Perrimon Background: A polarised cytoskeleton is required to pattern cellular space, andfor many aspects of cell behaviour. While the mechanisms ordering the actincytoskeleton have been extensively studied in yeast, little is known about theanalogous processes in other organisms. We have used Drosophila oogenesisas a model genetic system in which to investigate control of cytoskeletalorganisation and cell polarity in multicellular eukaryotes. Results: In a screen to identify genes required for Drosophila oocyte polarity, weisolated a Drosophila homologue of the yeast cyclase-associated protein, CAP.Here we show that CAP preferentially accumulates in the oocyte, where it inhibitsactin polymerisation. CAP also has a role in oocyte polarity, as cap mutants fail toestablish the proper, asymmetric distribution of mRNA determinants within theoocyte. Similarly in yeast, loss of CAP causes analogous polarity defects, alteringthe distribution of actin filaments and mRNA determinants. Conclusions: This study identifies CAP as a new effector of actin dynamics in Drosophila . As CAP controls the spatial distribution of actin filaments andmRNA determinants in both yeast and Drosophila , we conclude that CAP hasan evolutionarily conserved function in the genesis of eukaryotic cell polarity. Background Multicellular organisms contain a plethora of cellular formsand functions. In order to pattern an undifferentiated cel-lular space, symmetry must first be broken, for exampleby a localised extracellular signal. Then polar filaments arerequired to communicate this spatial information to allparts of the cell [1]. In multicellular eukaryotes, theprecise mechanisms by which cells organise the actincytoskeleton and generate polarity are poorly understood,in part because of the difficulty of generating mutants thatlack important cell-biological functions. To identify suchmutations, we conducted a genetic screen in  Drosophila which relies on the ability to generate mutant germlinetissue in a mosaic animal [2].During  Drosophila oogenesis, anterior–posterior (A–P) anddorsal–ventral (D–V) axes are established in a process thatrequires both actin and microtubule cytoskeletons. Thedynamic polarity of the developing oocyte is easily visu-alised by determining the asymmetric localisation of mRNA determinants, for example bicoid and  oskar  [3]. Atearly stages of oogenesis, a single microtubule-organisingcentre situated at the posterior of the nascent oocytenucleates a microtubule network which supports thepolarised traffic of cellular materials, including mRNAs,from nurse cells into the oocyte [4]. Then at stage6–8, theoocyte microtubule cytoskeleton is reorganised inresponse to a signal from the overlying posterior folliclecells [5–7]. The resulting polarised microtubule array isthought to mediate the transport of bicoid  and oskar  mRNAs to opposite poles of the oocyte, where they areused to pattern the future A–P axis of the embryo [3]. Atlate stages, the polar A–P microtubule array is disassem-bled while microtubules form at the oocyte cortex, drivingmixing of the oocyte cytoplasm in a process termed cyto-plasmic streaming [3]. While it is well established thatmicrotubules mediate the polar transport of mRNAsduring  Drosophila oogenesis, the role of actin in mRNAlocalisation is less clear. Mutations or drugs that perturbthe actin cytoskeleton disrupt mRNA localisation, however,by inducing premature cytoplasmic streaming [8–10]. Fur-thermore, at late stages of oogenesis and in the earlyembryo, tropomyosin, a protein that stabilises actin fila-ments, is required to correctly localise oskar  mRNA at theposterior pole [11–13]. Therefore, although actin filamentsappear evenly distributed within the wild-type oocyteduring establishment of polarity, an intact actin cytoskele-ton is required to maintain polarisation of the microtubulearray. Then, following disassembly of the A–P-polarisedmicrotubule array, an actin-based structure may be neces-sary to anchor mRNA determinants at the posterior poleduring streaming. Similarly, F-actin is required for theasymmetric distribution of mRNAs in other  Drosophila tissues, for example neuroblasts [14,15]. In a screen to isolate mutations perturbing the proper organ-isation of the actin cytoskeleton and oocyte polarity, weidentified a  Drosophila homologue of the cyclase-associated Address: Department of Genetics, Howard HughesMedical Institute, Harvard Medical School, Boston,Massachusetts 02215, USA.Correspondence: Norbert PerrimonE-mail: 14 June 2000 Revised: 12 July 2000 Accepted: 13 July 2000 Published: 31 July 2000Current Biology 2000, 10:964–9730960-9822/00/$ – see front matter ©2000Elsevier Science Ltd. Allrights reserved.  proteins (CAPs), which we have named capulet ( cap ). CAPwas first cloned from yeast [16,17], and CAP homologuesfrom a variety of organisms have been shown to associatewith adenylate cyclase [18–20], Abl tyrosine kinase [21], andmonomeric actin (G-actin) [22–25]. As the domain structureof CAP is similar in yeasts, plants and animals, CAP mayhave a conserved role, linking signal transduction to reor-ganisation of the actin cytoskeleton. In this paper wepresent an analysis of CAP function in the control of actinorganisation and cell polarity in both the  Drosophila germline and in yeast. We show that CAP is a major regula-tor of actin dynamics in  Drosophila , and that CAP functionsin both animal cells and fungi to organise the polarised dis-tribution of F-actin and mRNAs. Results Identification of a Drosophila homologue of the yeastcyclase-associated protein In a mosaic screen to isolate mutations that perturb actinorganisation in germline clones we identified a mutationin a novel gene capulet ( cap ) (Figure1a). cap was indepen-dently identified by two other labs ([26], and Z. Wills et al. , unpublished data) and encodes a protein of 424amino acids (Figure1a; GenBank accession numberAF132566) that is ~50% identical and ~65% similar toboth human CAPs [27].  Drosophila CAP is represented bya single ~2.4kb transcript during development, which isabsent from homozygous cap 10  mutant larvae (Figure1c).CAPs have been shown to inhibit actin polymerisation in vitro , by sequestering monomeric actin [22–25]. Thisactin-binding activity has been mapped to the carboxy-ter-minal region of CAP [18,23]; however, our analysis identi-fied a ‘verprolin homology’-related domain [28] in allCAPs, just carboxy-terminal of the polyproline-rich domain(Figure1b). In members of the verprolin/WASP family,this motif binds actin monomers in vitro , but catalysesactin polymerisation in vivo [29,30]. Therefore, in CAPhomologues, this region of the protein may be used tofacilitate actin binding. As CAP proteins have also beenfound associated with Abl tyrosine kinase [21] and withadenylate cyclase [18], it is possible that CAP represents anintermediary in these signal transduction cascades, perhapsaltering actin dynamics in response to extracellular cues. CAP is required in the germline for proper actinorganisation To characterise the cap mutant phenotype in detail, wecompared actin filaments in wild-type ovaries and inmosaic egg chambers carrying cap germline clones(Figures2a,b,3a,b). F-actin organisation within the wild-type egg chamber does not appear to change dramaticallyuntil stage10B when actin ‘dumping’ fibres appear innurse cells. In cap germline clones, however, the distribu-tion of F-actin appears relatively normal during the veryearly stages of oogenesis, but becomes highly polarisedand dynamic as the egg chamber develops. Ectopic F-actinis first seen at stage5–6 of oogenesis in a dense structureat the posterior pole of the oocyte (Figure2a). Bystages6–8, ectopic actin filaments appear to shift to theanterior of the oocyte, where they are found close to ringcanals (Figure2a). At stage10B, dumping fibres can beseen forming on schedule in the nurse cells of cap clones(data not shown). Finally, in eggs, extensive filamentousactin structures form close to the cortex (Figure2c), andectopic F-actin is visible in the few embryos that are pro-duced (data not shown). Interestingly, the change in thedistribution of F-actin, from posterior cortical to anteriorvesicle-like structures, mirrors the reorganisation of themicrotubule cytoskeleton seen in the wild type duringstages 6–8 [31,32]. Repolarisation of the microtubule arrayis thought to be induced by a signal from posterior folliclecells, dependent upon the prior action of Gurken in thegermline [6,7]. Therefore, to determine whether the samesignal also affects actin organisation, we reduced Gurkenfunction in the cap mutant background. We find that thelocalisation of ectopic F-actin is unaltered in  gurken cap double germline clones, so further work will be requiredto identify the cues responsible for the dynamic distribu-tion of actin aggregates in the cap mutant.As ectopic actin structures are formed in cap mutant eggchambers, other F-actin-rich structures are lost. In partic-ular, cortical F-actin underlying the nurse cell mem-branes disappears prematurely at stages8–9 of oogenesis(Figures2a,3). Therefore CAP may simultaneouslyinhibit actin polymerisation at some sites and facilitate theformation of F-actin at others. If the pool of actin withinthe egg chamber is limited, an alternative hypothesis can Research Paper CAP organises F-actin and cell polarity Baum etal. 965 Figure 1 cap is homologous to the cyclase-associated proteins. (a) Weidentified cap in a screen for mutations that perturb F-actin in germlineclones. CAP is ~50% identical and ~65% similar to both human CAPs[27]. CAPs have a conserved structure with an amino-terminal domain(AC) that binds adenylate cyclase in yeast [18], a polyproline-richdomain (PP) which interacts with c-Abl [21], followed by a carboxy-terminal monomeric actin-binding domain [18,23] and a dimerisationdomain (D) [43]. (b) CAP contains a region found in verprolin (V) familymembers [28]. This sequence lies outside the defined CAP carboxy-terminal region [18,23,25] (368–524 in yeast). (c) A single mRNAspecies at ~2.4kb was observed in a developmental northern blot.Thiswas absent from homozygous mutant cap 10  third instar larvae. (a)(c)(b)  Current Biology mRNArRNANACPPVActinDC MajorityHuman CAPDrosophila CAPHuman WASPN WASP 1stN WASP 2ndVerprolinWave WT cap  be imagined, in which actin filaments are lost from nursecell cortices to compensate for the formation of actinaggregates within the oocyte. In order to test whether theactin cytoskeleton is similarly polarised in other mutantsthat have excess accumulation of F-actin, we analysed twinstar  germline clones. twinstar  inhibits actin filamentformation in vivo and encodes the  Drosophila homologueof an actin-severing protein, cofilin [33]. We find thatalthough ectopic actin filaments form in twinstar  germlineclones, as in the cap mutant, ectopic actin aggregates format sites throughout the early twinstar mutant egg chamber(Figure2d). Therefore, CAP has the specific function of inhibiting actin polymerisation within the oocyte. As CAP inhibits actin polymerisation within the oocyte,but not in nurse cells, we generated an antibody to  Drosophila CAP to see if this localised function isreflected in the wild-type distribution of the protein.This antibody is specific, as it recognises CAP in tissueextracts (~45kilodaltons (kD)) (data not shown). CAP ispresent throughout the follicle cells and in the germline,but at early stages of oogenesis the protein preferentially 966 Current Biology Vol 10 No 16 Figure 2 In cap germline clones, F-actin accumulates in the oocyte, where CAP isconcentrated in the wild type. (a) The ovaries of females carrying cap mutant germline clones were stained with TRITC–phalloidin to visualiseF-actin. In cap germline clones, actin accumulates in the oocyte atstage5–6 of oogenesis. Actin filaments then shift at stages6–8 to theoocyte–nurse cell boundary. Subsequently, F-actin is lost from nurse cellcortices as germ cells begin to fuse (see also Figure 3). (b) A wild-typeovariole is shown for comparison. Following germ-cell differentiation, thelevel of F-actin appears slightly higher within the oocyte at the posteriorof the egg chamber (to the right in all panels). (c) Ectopic actin filamentsfirst form at the posterior of the oocyte at stage 5–6 of oogenesis, inboth cap 10  and cap 1 germline clones. In cap 7 mutant tissue, F-actincables form in the oocyte at a similar time, only to disappear atstage8–9. Actin aggregates later accumulate at the cortex of cap mutant eggs. (d) We compared the distribution of F-actin accumulationin the cap mutant with that seen in twinstar  germline clones (using either tsr  2 and tsr  1 alleles). At early stages of oogenesis, actin filaments areseen in clumps throughout the twinstar  mutant egg chamber, while themorphology of mutant tissue often appears very disrupted. Later inoogenesis, F-actin is present at highest levels at the cortices of thenurse cells and oocyte, and in aggregates and filamentous structureswithin tsr  mutant eggs. (e) A polyclonal anti-CAP peptide antibody wasgenerated and used to stain wild-type ovaries. The staining was notseen with pre-immune sera and was confirmed using an independentantibody generated against the whole protein (Wills etal. , inpreparation). CAP is present at low levels throughout the egg chamberbut, following germ cell differentiation, the protein accumulatespreferentially at the posterior of the oocyte (compare F-actin and CAPstaining), and at the oocyte cortex in later egg chambers. (a)(c)(d) Wild type cap 10  cap 10  cap 7 tsr cap (b)  Current Biology (e) F-actinCAP Figure 3 Protein kinase A (PKA) and CAP have related functions in thegermline. (a) TRITC-labelled phalloidin was used to visualise F-actinwithin wild-type egg chambers. (b) Aggregates of actin filamentsaccumulate in cap mutant ovaries, whereas F-actin is lost from nursecell cortices. This later phenotype may cause the cell fusion eventsobserved, allowing occasional nurse cell nuclei to enter the oocyte(indicated by broken line). (c)  pka and  pka cap double germline cloneswere stained to visualise F-actin.  pka mutants both lose nurse cellcortical actin [49,50] (leading to cell fusion, nurse-cell nuclei indicatedby a broken line) and accumulate ectopic actin structures within theoocyte, close to the nurse cell–oocyte interface. The double mutanthas an exaggerated phenotype, with complete fusion of the germlineand ectopic actin aggregates at the posterior pole of the syncytium.This implies a related function for CAP and PKA in the germline. cap and  pka clones exhibit distinct phenotypes in other tissues, however,which are not accentuated in the double mutant (data not shown). (d) Loss of one copy of CAP enhances a  pka mutant phenotype. (a)(c)  pka pkacap pkacap+ pkacap+/–  Current Biology (d)(b) Wild type  cap  accumulates in the oocyte (Figure2e). Later in oogenesis,CAP appears to be localised at the oocyte cortex. Thus,CAPis concentrated in the oocyte, where it functions toinhibit actin accumulation.Our screen also identified a mutation in the catalyticsubunit of protein kinase A (PKA). Therefore, we com-pared  pka and cap mutant phenotypes in the  Drosophila germline. Like the cap mutant,  pka germline clones losenurse cell cortical actin, while simulataneously accumulat-ing ectopic actin structures (Figure 3c). In addition, the  pka mutant phenotype is sensitive to the dosage of CAP,and actin defects are dramatically enhanced in  pka cap double germline clones (Figure 3c,d). These data suggestthat PKA and CAP functionally cooperate in the germlineto control actin organisation. CAP is required for oocyte polarity In cap germline clones, F-actin accumulates in a highlypolarised fashion within the egg chamber and oocyte.Thus, we investigated whether loss of CAP perturbs otheraspects of normal polarity, including the asymmetric local-isation of mRNAs within the oocyte. We examined thedistribution of bicoid  and oskar  mRNAs, which localise toanterior and posterior poles of the oocyte, respectively(Figure4a). We find that although oskar  mRNA is concen-trated in one region of the oocyte in over 90% of eggchambers, oskar  mRNA is mislocalised in 76% of stage8–10 cap germline clone egg chambers ( n =184).Moreover, in 28% of cases, oskar  transcripts are localised tothe anterior or lateral part of the oocyte (Figure4b). Inaddition, in 64% of stage-10 egg chambers that maintaincorrect overall polarity, oskar  mRNA has a diffuse distribu-tion and is not tightly focused at the posterior pole(Figure4b bottom right panel). We also examined thelocalisation of bicoid  transcripts ( n =184). bicoid  mRNAaccumulates at an aberrant site in 65% of cap mutant eggchambers (Figure4c), and is localised to the posterior polein 36% of stage8–10 egg chambers. Thus, cap germlineclones display two related mRNA polarity defects. First, Research Paper CAP organises F-actin and cell polarity Baum etal. 967 Figure 4 CAP is required for proper oocyte polarity.Wild-type and cap egg chambers werestained for oskar  mRNA to assess cellpolarity. (a) In the wild type, oskar  RNA istightly localised to the posterior of the oocyte,while bicoid  mRNA localises to the anteriormargin. (b) In over 90% of cap germlineclones , oskar  mRNA accumulates in adiscrete region of the oocyte. oskar  RNAappears aberrantly localised in 76% of mutantegg chambers, however, and in 28% of casesis concentrated in anterior or lateral regions ofthe oocyte. In addition, in 64% of stage-10egg chambers that properly concentrate oskar  mRNA in the posterior region of the oocyte, oskar  transcripts appear to diffuse away fromthe pole ( n =184). (c) In 65% of cap mutantegg chambers, bicoid  mRNA is mislocalised,and in 50% of egg chambers appearsconcentrated at an alternative site, forexample the lateral, central or posterior regionof the oocyte ( n =184). (d,e) Yolk particleswere visualised in fixed tissue under (d) lightmicroscopy or (e) in live ovarioles byautofluorescence. Yolk accumulates in theoocytes of control egg chambers which lack profilin function (in which prematurecytoplasmic streaming occurs [8]), and inwild-type egg chambers (data not shown), butis not seen in the oocytes of cap mutant eggchambers. (e) In live cap mutant eggchambers, autofluorescent yolk particlesaccumulate at the nurse cell–oocyte boundaryand are not properly transported into theoocyte. These particles appear abnormallylarge in the mutant. In stage7–9 cap mutantegg chambers, yolk particles do not appear toexhibit the movements characteristic ofpremature cytoplasmic streaming [10] (datanot shown). (f) F-actin (red) and β -galactosidase (green) were visualised inwild-type and cap mutant egg chambersexpressing kinesin–lacZ. At this stage,kinesin–lacZ is transported to the posteriorpole of the wild-type oocyte (arrows). In many cap mutant egg chambers, kinesin–lacZ isconcentrated at the anterior cortex, often at asite where the oocyte membrane seems toforce its way into the nurse cell cluster. (g)  pka germline clones exhibit mRNA polaritydefects similar to those observed in cap mutant egg chambers. oskar  transcripts aremisplaced in almost 50% of egg chamberslacking PKA activity (mRNA accumulates at acentral position within the oocyte in fewerthan 10% of cases). In late egg chambers, adiffuse gradient of oskar  mRNA isoccasionally observed at the posterior of theoocyte. Also, yolk fails to form within theoocyte of many  pka mutant egg chambers(data not shown). These defects resemblethose seen in cap germline clones, implyingthat  pka and cap mutant phenotypes have acommon aetiology.  Current Biology (a)(c)(g)(d)(f) oskar mRNA bicoid mRNA oskar mRNA bicoid mRNAKinesin–lacZ capcap WildtypeWTControl capcap (b)(e) Wildtype oskar   mRNA  pka  although oocytes are able to concentrate oskar and  bicoid  mRNAs locally within the oocyte, they appear unable tocoordinate mRNA polarity with the morphological polarityof the egg chamber. Second, in the majority of egg cham-bers in which oskar  mRNA is correctly transported to theposterior pole of the oocyte, oskar  message is not tightlylocalised at the cortex. Mutations in several actin-related genes disrupt mRNAlocalisation, by inducing microtubule-based cytoplasmicstreaming [8–10]. To test whether loss of CAP also disruptsthe distribution of mRNAs by inducing premature cyto-plasmic streaming, we looked at the movement of yolk par-ticles within cap mutant egg chambers [10]. Interestingly,yolk often fails to form in the oocyte in cap germline clones(Figure4d). Instead, the analysis of yolk autofluorescencein live cap mutant egg chambers reveals abnormal yolk par-ticles accumulating at the nurse cell–oocyte boundary(Figure4e). Therefore CAP may be required for a rela-tively late step in the formation of yolk granules and/or forthe directional transport of yolk into the oocyte. Moreover,in a time-lapse analysis of yolk particles in stage7–9 cap mutant egg chambers, we did not observe the movementscharacteristic of cytoplasmic streaming (data not shown),This is not unexpected because streaming disrupts mRNAlocalisation completely [8], whereas cap mutant oocytesaccumulate mRNA determinants at discrete sites. Alterna-tively, the mRNA and yolk localisation defects observed inthe cap mutant could result from a misoriented micro-tubule array. To test this hypothesis we used kinesin–lacZ(kin–lacZ, a fusion between β -galactosidase and a plus-end-directed microtubule motor [34]) to assay microtubulepolarity in cap germline clones. In the wild type, kin–lacZtranslocates to the posterior pole of stage8–9 oocytes [34].Within cap germline clones, kin–lacZ often becomes con-centrated at specific but aberrant sites in the oocyte, indi-cating that microtubules are polarised but misaligned inthe absence of CAP (Figure4f). Interestingly, in caseswhere kin–lacZ is found at the anterior cortex, this alteredmicrotubule polarity is accompanied by a change in mor-phology of the cap mutant oocyte, which appears to invadethe nurse cell cluster (Figure4f). This may in turn con-tribute to the fusion of nurse cells and oocyte observed inthe mutant. Finally, in later egg chambers, kin–lacZappears delocalised, as it does in the wild type followingthe onset of cytoplasmic streaming. In conclusion, earlydefects in oskar and  bicoid  localisation in the cap mutant arelikely to reflect underlying defects in the microtubulecytoskeleton. Interestingly,  pka germline clones exhibitmRNA polarity and yolk defects like those of the cap mutant (Figure 4g, and data not shown). These datasupport the notion that CAP and PKA have relatedgermline functions. CAP is required for yeast cell polarity and control of theactin cytoskeleton To investigate whether CAP has an evolutionarily con-served function to control the spatial organisation of F-actin and mRNAs, we turned to  Saccharomyces cerevisiae  (budding yeast), where F-actin structures and a mRNAdeterminant,  ASH1, are asymmetrically localised withinthe bud [35]. In yeast, in contrast to  Drosophila oogenesis,microfilament and microtubule cyoskeletons functionindependently [36], and polarity is organised primarily byactin filaments, simplifying the analysis. Yeast cells deleted for cap ( cap ∆ ) exhibit several morpholog-ical defects [37]. Mutant cells vary in size and shape whencompared to the wild type (Figure5a,b). This reflectsunpolarised growth, probably arising from actin-relateddefects in vesicle targeting [38]. cap ∆ cells also exhibit a dis-organised actin cytoskeleton (Figure5a). The majority of  cap mutant cells, however, are still able to generate a polaractin organisation, with filaments concentrated in the bud.Moreover, the wild-type F-actin distribution appears accen-tuated in many cap ∆ cells, implying that, in yeast, CAPprevents hyperpolarisation of the actin cytoskeleton, as itdoes in  Drosophila cap germline clones. To visualise cell polarity in yeast, we used  ASH1 mRNA asa reporter. In yeast,  ASH1 mRNA is a determinant of celldifferentiation and is asymmetrically localised by myosinmotors tracking along polar actin cables. Upon cell division,the daughter cell derived from the bud inherits  ASH1 message, preventing it from switching mating type [39–41].Therefore, using green fluorescent protein (GFP) to label 968 Current Biology Vol 10 No 16 Figure 5 CAP is required in yeast for control of the actin cytoskeleton and forproper cell polarity. (a) Actin was visualised in cells lacking the cap gene using TRITC–phalloidin. cap ∆ cells exhibit morphological andF-actin defects. (b) cap mutant and wild-type yeast expressingMS2–GFP together with ASH1 mRNA with MS2-binding sites amassa single GFP-labelled mRNA particle (g ASH1 [42]). g ASH1 is foundat the bud tip in 90% of wild-type cells. g ASH1 enters the bud in 85%of cap ∆ cells, but in the majority of cases it fails to localise at the budtip. g ASH1 movement within the bud was monitored in live wild-typeand cap ∆ cells. In 10sec intervals, visible changes in g ASH1 positionwere noted and averaged. In films of wild-type cells, g ASH1 remainedat the bud tip 75% of the time, but in the cap mutant, g ASH1 waspresent at the bud tip in only 19% of frames (data not shown).  Current Biology (a) Wild type cap F-actinF-actingASH1 (b)
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks