Centriole and centrosome cycle in the early Drosophila embryo

GIULIANO CALLAINI and MARIA GIOVANNA RIPARBELLI

Department of Evolutionary Biology, University of Siena, via Mattioli 4, 53100 Siena, Italy

Summary

Centriole and centrosome cycles were examined byindirect immunofluorescence and electron mi-croscopy techniques in the early Drosophila embryo.The centrosomes, which are already divided atinterphase, appear as compact spheres during pro-phase and metaphase, expand and flatten fromanaphase to telophase and split into two units in late

telophase. Centriole separation starts in late meta-phase, becomes evident in anaphase and increasesduring telophase. Procentrioles appear during thefollowing interphase.

Key words: centrosomes, mitosis, Drosophila embryo.

Introduction

Mitotic divisions in Drosophila occur with no discernibleGi or G2 phases (see Glover, 1988; Glover et al. 1989) andcentrosome separation, which is generally considered to bea feature of late interphase in animal cells, is an earlyevent in the mitotic cycle (Warn et al. 1987; Kellogg et al.1988; Raff and Glover, 1988). Two centrosomes are alwayspresent above the interphase nuclei (Warn, 1986; Warnand Warn, 1986; Callaini and Anselmi, 1988). Thecentriole cycle also appears to be slightly modified.Huettner (1933) described the centriole cycle in Dros-ophila embryos from microscopic observation of materialfixed with formol-ethanol-acetic acid (1:1:1, by vol.) andstained with Heidenhain's iron hematoxylin. He reportedtwo centrioles near the nucleus during interphase and twowidely separated centrioles recognizable at either pole ofthe mitotic spindle at anaphase. During late anaphase andearly telophase the centrioles remain more or less in thesame position and during late telophase they begin tomove apart. This description of the centriole cycle in theDrosophila embryo does not seem to have found a place insubsequent literature. Only recently were these obser-vations taken up by Warn et al. (1987), who compared thecentriole cycle as proposed by Huettner (1933) withcentrosome division observations resulting from in vivostudies of Drosophila embryos injected with anti-tubulinantibodies. In the present study we used electron mi-croscopy and the polyclonal antibody Rbl88, whichspecifically recognizes a centrosome-associated antigen inDrosophila (Whitfield et al. 1988), to follow centrioledynamics and to elucidate the centrosome cycle inDrosophila.

Materials and methods

Collection of the embryosEmbryos of Drosophila melanogaster (Oregon-R strain) werecollected at 25 °C on agar plates, dechorionated in a 50%commercial bleach solution and washed with distilled water.

Following removal of excess of liquid by blotting on tissue paper,the embryos were processed for different preparations.

Staging of the embryosThe age of the embryos for immunofluorescence observations wasdetermined according to Campos-Ortega and Hartenstein (1985)by direct observation of the embryos with interference contrast orby counting the somatic nuclei.

The exact mitotic stage for electron microscopy preparationswas determined as follows. The dechorionated embryos weretreated with a solution containing 25% glutaraldehyde in 0 .1Msodium cacodylate buffer, pH7.2, in an equal volume of heptane(Zalokar and Erk, 1977). After 3min the embryos were trans-ferred for 30 min to a solution containing 2.5 % glutaraldehyde in0.1 M sodium cacodylate to which 5 % of sucrose was added and thevitelline membrane was removed with tungsten needles. Thenthe embryos were washed three times for 5 min in cacodylatebuffer and incubated for 3 min with 1 /igml"1 of the DNA-specificdye Hoechst 33258 in cacodylate buffer. The embryos weretemporarily mounted and observed with a Leitz Aristoplanmicroscope equipped with u.v. filter.

Fluorescence microscopyThe dechorionated embryos were fixed and their vitellinemembrane was removed essentially as described by Warn andWarn (1986), except for a final fixation with acetone for 5 min. Theembryos were then washed in phosphate-buffered saline (PBS)and incubated for l h in PBS containing 0.1% bovine serumalbumin. For double labeling the embryos were incubated for 5 hat room temperature with the Rbl88 antiserum, which specifi-cally recognizes a 185xlO3Mr antigen associated with thecentrosome of Drosophila embryos (Whitfield et al. 1988) at adilution of 1:400. The samples were then washed with PBS andincubated for 30 min at room temperature with a monoclonalantibody against a-tubulin (Amersham). After rinsing in PBS theembryos were incubated in secondary antibodies (goat anti-mouserhodamine-conjugated IgG and goat anti-rabbit fluorescein-conjugated IgG at a dilution of 1:1000 each; Cappel, West Chester,PA), washed again in PBS and mounted in 90% glycerolcontaining 2.5 % rc-propyl gallate to reduce photobleaching (Gilohand Sedat, 1982). Fluorescence observations were carried out witha Leitz Aristoplan microscope equipped with fluorescein andrhodamine filters. Photomicrographs were taken with Kodak Tri-X pan film and developed in Kodak HC 110 developer for 7 min at20°C.

Centrosome behavior in Drosophila embryo 539

Electron microscopyFor transmission electron microscopy observations the embryosat the desired mitotic stage were fixed in the trialdehyde solutionof Kalt and Tandler (1971) for 2h. After rinsing in 0 .1Mcacodylate buffer, pH7.2, the embryos were postfixed in 1%osmium tetroxide for 2h and dehydrated in a graded series ofethanol and bulk-stained in 1 % uranyl acetate for 1 h. Aftertreatment with propylene oxide the embryos were embedded in anEpon-Araldite mixture and polymerized at 60°C for 48 h.Random and serial sections cut using an LKB Nova ultramicro-tome and a diamond knife (Diatome Ltd, Switzerland) werecollected on copper grids and stained with uranyl acetate and leadcitrate. Sections were observed and photographed with a PhilipsEM 400 electron microscope.

Results

Immimofluorescence observationsDuring prophase the centrosomes, which appear ascompact spheres, move to the opposite sides of the nucleiand begin to form the mitotic spindle. The microtubulesform weak asters immediately surrounding the centro-somes (Fig. 1A).

At metaphase the centrosome material is condensed atopposite poles of the mitotic apparatus and the micro-tubules give rise to metaphase figures with a small asterand a distinct midzone (Fig. IB).

The separation of the sister chromatids and theirshifting toward the poles, which occurs in early anaphase,is marked by a further elongation of the spindle and astermicrotubules, in contrast to what was seen in metaphase,in which relatively long microtubules radiate toward theperiphery of the spindle (Fig. 1C). Spindle lengtheningends in late anaphase and the interzonal microtubulesgradually diminish in density and become progressivelyfainter near the poles (Fig. ID). During early anaphasethe centrosomes gradually lose their compactness and inlate anaphase organize in ovoid plates perpendicular tothe longitudinal axis of the mitotic apparatus (cf. Fig. 1Cand D).

Telophase is characterized by the gradual breakdown ofthe interzonal microtubules, indicated by a reduction inthe staining, and by the growth of the astral microtubules.During late telophase the spindle microtubules arereduced to a small central fluorescent area, including aninterzone region of reduced fluorescence, the midbody(Fig. IE). The centrosome material, recognized by theRbl88 antibody, splits into two units (Fig. IE).

During early interphase the centrosomes are alreadydivided and organize the microtubule network in the spacebetween the nucleus and the plasma membrane (Fig. IF).Centrosome material, as recognized by the Rbl88 anti-body, is loosely arranged near the nuclear surface. Wheninterphase progresses the centrosome material is moredensely associated and the centrosomes are seen as twowidely separated spheroidal structures (Fig. IF).

Electron microscopy observationsAs soon as prophase begins, each centrosome containsparent and daughter centrioles, closely juxtaposed and inperpendicular orientation to each other (Fig. 2A). Thecentrioles are surrounded by clouds of electron-densematerial from which microtubules originate (Fig. 2A). Asprophase progresses the centrosomes continue to migrateto their proper sites at the opposite poles of the nuclei.

Fig. 1. Indirect immunofluorescence of whole-mountDrosophila embryos during nuclear cycle 10 (A,B,C,D,E) andnuclear cycle 11 (F) double stained for microtubules (MTs, left)and centrosomes (Cent) (middle and right, for detail).(A) Prophase. The mitotic apparatus begins to form (left).Centrosomes appear as compact spheres (middle and right).(B) Metaphase. The mitotic spindle is well developed but theasters are very small (left). The centrosomes are quite compact(middle and right). (C) Early anaphase. Aster microtubulesgrow (left). The centrosomes gradually lose their compactness(middle and right). (D) Late anaphase. The spindle elongatesand microtubule staining diminishes at the poles (left).Centrosomes expand and enlarge into irregular ovoid plates(middle and right). (E) Telophase. Aster microtubules enlargeagain and spindle microtubules reduce (left). Centrosomesexpand further and split into two units (middle and right).Arrows and arrowheads indicate midbodies and dividedcentrosomes respectively. (F) Early interphase. The midbodieshave disappeared and the microtubules have reduced in length(left). The already divided centrosomes gradually condense(middle and right). Bars, 10 /an.

When the chromosomes have reached metaphase, eachspindle pole has a pair of mutually perpendicularcentrioles (Fig. 2B). During late metaphase, the daughtercentrioles, equal in length to the parent centrioles, are stilladjacent to their parents but the angle between them is nolonger a right-angle (Fig. 2C). This suggests that they arestarting to move away from each other.

In anaphase the centrosomes contain parent anddaughter centrioles, which are almost always disorien-tated and separated by several micrometers (Fig. 2D).

When the nuclear envelope begins to re-form around thedecondensing chromosomes at telophase, the centriolescontinue to move apart. Fig. 2E shows the ultrastructureof a spindle region during late telophase. This appears asan area devoid of yolk granules and mitochondria. In theregion between the two re-forming nuclei, small densebodies are visible, each composed of several denselypacked microtubules covered with an accumulation ofelectron-dense material. To confirm whether the pairs offluorescent structures observed at the pole of the telophasespindles with Rbl88 antibody contain only one full-sizedcentriole and no procentrioles, the polar region wasanalyzed ultrastructurally by serial sections. Random orincomplete sections would not be adequate because of thesmall size and the position of the procentrioles at the timeof their appearance. Serial sections through one pole of thespindle show two full-sized centrioles separated from eachother by several micrometers (Fig. 2F, G and H). Thesecentrioles measure 0.20 /on in diameter, 0.15 ,um in length,and consist of the usual nine peripheral triplets ofmicrotubules. Sections above and below those shown inFig. 2F and H contain no additional centrioles.

Once the nuclear envelope has completely re-formed,the centrosomes have migrated to a position between thenuclear envelope and the embryo surface (Fig. 21). Thecentrosomes have enlarged and consist of several smallclouds of electron-dense material surrounding the centrio-lar region (Fig. 2J). The procentriole, a small linear densebody with no internal organization, is visible near anextremity of the parent centriole wall (Fig. 2K). Theprocentrioles keep elongating during prophase and reachthe length of the parent centriole at the beginning ofmetaphase. The structure of the procentrioles changesduring this time and the microtubular triplets becomevisible.

540 G. Callaini and M. G. Riparbelli

Centrosome behavior in Drosophila embryo 541

m-

Discussion different stages of mitosis, and the second the time ofcentriole duplication. These questions, unresolved by

The centriole cycle in Drosophila as proposed by Huettner conventional light microscopy, may be answered by(1933) raises two main questions. The first concerns the ultrastructural analysis. The present results suggest aposition and number of centrioles observed during the centriole cycle that is slightly modified with respect to

542 G. Callaini and M. G. Riparbelli

Fig. 2. Ultrastructural analysis of the centriole andcentrosome cycle during the mitotic cycle in the earlyDrosophila embryo. (A) Position of the centrosomes (arrows) atprophase; n, nucleus. (B,C) During metaphase each centrosomecontains a pair of centrioles, which gradually lose theirorthogonal disposition. (D) During late anaphase the centriolesare more distant from each other. (E) Late telophase spindle;arrows and arrowhead indicate newly separated centrioles andmidbody, respectively. (F,G,H) Serial sections through the polesof a late telophase spindle; note the absence of daughtercentrioles; n, nucleus. (I) Sagittal and (J) grazing sections ofthe centrosomal region of two embryos during interphase ofnuclear cycle 11; arrows and arrowheads indicate centriolesand pericentriolar material, respectively; n, nucleus. (K) Detailof an interphase centrosome containing parent and smalldaughter centrioles; arrow indicates the procentriole. Bars: A,I, J, ljiim; B, C, D, F, G, H, K, 0.5 fan; E, 2f<m.

those proposed by Huettner (1933) and more recently byStafstrom and Staehelin (1984). Two centrosomes areobserved at the beginning of the mitotic cycle, eachconsisting of parent and daughter centrioles, whichmigrate to the opposite poles of the nucleus to form themitotic spindle. At metaphase the centrosomes containtwo equal and perpendicularly orientated centrioles thatgradually lose their orthogonal arrangement in thetransition to anaphase. At telophase the centrioles moveapart, and during interphase they undergo duplication.The procentriole maturation follows the conventionalpathway observed in other animal cells (see Vorobjev andNadezhdina, 1987).

The distribution of the centrosome-associated antigenrecognized by the Rbl88 antibody changed during themitotic cycle. Combined immunofluorescence and electronmicroscopy observations suggested a close relationshipbetween pericentriolar material and centrioles in theassemblage of functional spindles. When parent anddaughter centrioles were orthogonally arranged in pro-phase and metaphase, the centrosomes appeared ascompact spheres. From late metaphase the centrioles losttheir orthogonal arrangement and moved away from eachother until late telophase. At this time the centrosomesspread in a flat ovoid plate and divided into two separateunits each containing only one centriole. The splitting ofthe centrosomal material at telophase in Drosophila wasrecently also confirmed with antibodies directed againstmicrotubule-associated proteins (Kellogg et al. 1989).

The Drosophila centrosome cycle reported here differsfrom the special case of mitosis in sea urchin eggs, inwhich the centrosome divides before the nuclear envelopehas formed (Paweletz et al. 1984, 1987). However, in seaurchin eggs parent and daughter centrioles shift andduplicate in telophase (Sluder and Rieder, 1985), whereasin the Drosophila embryo the shifting of the centriole pairsis an event that is distinct from centriole replication andoccurs at different times during the mitotic cycle. Serialsections demonstrated that in late telophase the centro-somes have already separated into two solitary centrioles,each surrounded by an aster of microtubules, and theprocentrioles appear later in interphase. These ultrastruc-tural observations are in agreement with the findings inmercaptoethanol-treated sea urchin eggs in which cen-triole replication is not a prerequisite for the splitting offof daughter from parent centrioles (Sluder and Rieder,1985). Centriole replication in the early Drosophilaembryo is not a prerequisite for the splitting of thecentrosome-associated antigen recognized by the Rbl88antibody. This centrosomal material splits into two

separate units, each containing only one centriole and ableto nucleate microtubules. However, centriole pairs areapparently needed to start regular mitotic progression.This agrees with the observations of Sluder and Rieder(1985) that the reproductive capacity of sea urchincentrosomes is correlated with the number of centriolesthey contain and that only centrosomes with centrioleduplexes reproduce between mitosis and form regularmitotic spindles (Sluder et al. 1989).

The authors are indebted to Dr W. Whitfield of the Departmentof Biochemistry, Imperial College of Science, Technology andMedicine, London, for his generous gift of the Rbl88 antibody.This research project was supported in part by grants from theMinistero della Pubblica Istruzione.

References

CALLAINI, G. AND ANSELMI, F. (1988). Centrosome splitting duringnuclear elongation in the Drosophila embryo. Expl Cell Res 178,415-425.

CAMPOS-ORTEGA, J. A. AND HARTENSTEIN, V. (1985) The EmbryonicDevelopment of Drosophila melanogaster, Springer-Verlag, Berlin.

GILOH, I. I. AND SEDAT, J. W. (1982). Fluorescence microscopy: reducedphotobleaching of rhodamine and fluorescein protein conjugates by n-propyl gallate. Science 217, 1252-1255.

GLOVER, D. M. (1988). Mitosis in Drosophila. J. Cell Sci. 92, 137-146.GLOVER, D. M., ALPHEY, L , AXTON, J. M., CHESHIRE, A.. DALBY, B.,

FREEMAN, M., GIRDHAM, C, GONZALEZ, C, KARESS. R. E , LEIBOWITZ,M. H., LLAMAZARES, S., MALDONADO-CODINA, M. G., RAFF, J. W.,SAUNDERS, R., SUNKEL, C. E. AND WHITFIELD, W. G. F. (1989). Mitosisin Drosophila development. J. Cell Sci. Suppl. 12, 277-291.

HUETTNER, A F. (1933). Continuity of the centrioles in Drosophilamelanogaster Z. Zellforsch mikrosk. Anat. 19, 119-134.

KALT, M. R. AND TANDLER, B. (1971). A study of fixation of earlyamphibian embryos for electron microscopy. J. Ultrastruct. Res. 36,633-645.

KELLOGG, D. R., FIELD, C. M. AND ALBERTS, B. M. (1989). Identificationof microtubule-associated proteins in the centrosome, spindle, andkinetochore of the early Drosophila embryo. J. Cell Bwl. 109,2977-2991.

KELLOGG, D R., MITCHISON, T. AND ALBERTS, B M. (1988) Behaviour ofmicrotubules and actin filaments in living Drosophila embryosDevelopment 103, 675-686

PAWELETZ, N., MAZIA, D AND FINZE, E.-M. (1984). The centrosome cyclein the mitotic cycle of sea urchin eggs. Expl Cell Res. 152, 47-65.

PAWELETZ, N., MAZIA, D. AND FINZE, E.-M. (1987). Fine structuralstudies of the bipolarization of the mitotic apparatus in the fertilizedsea urchin egg. I. The structure and behavior of centrosomes beforefusion of the pronuclei. Eur. J. Cell Biol. 44, 195-204.

RAFF, J. W. AND GLOVER, D. M. (1988). Nuclear and cytoplasmic mitoticcycles continue in Drosophila embryos in which DNA synthesis isinhibited with aphidicolin. J. Cell Biol. 107, 2009-2019.

SLUDER, G., MILLER, F. J. AND RIEDER, C. L. (1989). Reproductivecapacity of sea urchin centrosomes without centrioles. Cell Motil.Cytoskel. 13, 264-273.

SLUDER, G AND RIEDER, C. L. (1985). Centriole number and thereproductive capacity of spindle poles. J. Cell Biol. 100, 887-896.

STAFSTROM, J. P. AND STAEHELIN, L. A. (1984). Dynamics of the nuclearenvelope and of nuclear pore complexes during mitosis in theDrosophila embryo. Eur. J Cell Biol. 34, 179-189.

VOROBJEV, I. A. AND NADEZHDINA, E. S. (1987). The centrosome and itsrole in the organization of microtubules Int. Rev. Cytol. 106, 227-293.

WARN, R. M. (1986). The cytoskeleton of the early Drosophila embryo. J.Cell Sci Suppl. 5, 311-328.

WARN, R. M., FLEGG, L. AND WARN, A. (1987). An investigation ofmicrotubule organization and functions in living Drosophila embryosby injection of a fluorescently labeled antibody against tyrosinated a-tubulin. J. Cell Biol. 105, 1721-1730.

WARN, R. M. AND WARN, A. (1986). Microtubule arrays present duringthe syncytial and cellular blastoderm stages of the early Drosophilaembryo. Expl Cell Res. 163, 201-210.

WHITFIELD, W G. F., MILLAR, S. E., SAUMWEBER, H., FRASH, M. ANDGLOVER, D. M. (1988). Cloning of a gene encoding an antigenassociated with the centrosome in Drosophila. J. Cell Sci. 89,467-480.

ZALOKAR, M. AND ERK, I. (1977). Phase-partition fixation and staining ofDrosophila eggs Stain Technol 52, 89-95.

{Received 29 June 1990 - Accepted 13 August 1990)

Centrosome behavior in Drosophila embryo 543