Microtubule-determined plastid distribution during microsporogenesis in

Lilium longiflorum

ICHIRO TANAKA

Department of Biology, Yokohama City University, Seto 22-2, Kanazawa-ku, Yokohama 236, Japan

Summary

The relationship between organelle distribution andthe cytoskeleton was examined during microsporo-genesis in Ulium longiflorum. The distribution pat-tern of plastid and mitochondrial nucleoids wasfollowed by fluorescence microscopy after stainingwith 4',6-diamidino-2-phenyUndole (DAPI). Althoughthe plastid nucleoids gradually enlarged duringprophase I, by anaphase I of meiosis they wererandomly distributed in the cytoplasm of eachmicrosporocyte. At telophase I the plastid nucleoidswere aggregated in the equatorial region of the cell.After entering prophase II the plastid nucleoids wererandomly distributed in the cytoplasm, and attelophase II they had reaggregated to the equators ofthe two cells. After the completion of meiosis theywere located at the two poles of each youngmicrospore. This distinct cell polarity of plastidnucleoids was preserved in isolated protoplasts. Inall cells where the distribution of plastid nucleoidswas non-random, the nucleoids were invariablysituated furthest away from the interphase andtelophase nuclei. However, the distribution of mito-chondrial nucleoids throughout meiotic divisionshowed little cell polarity.

Analysis of the microtubule and actin cytoskel-etons during microsporogenesis revealed that themicrotubules radiated out from the cell nuclei only at

the stages when the distribution of plastids showedpolarity, whereas the actin filaments were usuallyrandomly oriented throughout the cytoplasm, inde-pendent of the plastid arrangement and of theorganization of microtubule cytoskeleton. Theradiating microtubules seemed to exclude the plas-tids from around the cell nuclei. Treatment ofcultured pollen tetrads with colchicine disrupted theplastid polarity, probably by depolymerizing theradiating microtubules, resulting in a random distri-bution of the plastid nucleoids. Treatment withcytochalasin B, however, had no effect on thearrangement of plastids.

These results demonstrate that microtubules func-tion in the movement and distribution of plastids inmale reproductive cells of higher plants. Further, it isassumed that the system of radiating microtubulesthat controls the distribution of plastids during malemeiosis is also involved in the subsequent formationof male gametes, which are deficient in plastids inmany angiosperm plants, including this lily.

Key words: Ldium, meiosis, microtubule cytoskeleton, plastid,polarity.

Introduction

The structure of the microtubule and actin cytoskeletonsduring male meiosis in higher plants has been studied in anumber of species by immunocytochemistry and rhoda-mine-phalloidin staining. Van Lammeren, et al. (1985),Sheldon and Dickinson (1986), Hogan (1987) and Traaset al. (1989) used immunofluorescence to show the patternof microtubules, while Sheldon and Hawes (1988), Traasetal. (1989) and Van Lammeren et al. (1989) usedfluorescent-dye-labelled phallotoxins to visualize thestructure of actin filaments (F-actin). These studiesconcentrated mainly on the roles of the two cytoskeletonsin division processes, in particular their function innuclear division and cytokinesis through formation ofspindles and phragmoplasts. However, during microsporo-genesis the microtubules and actin filaments are thoughtto be involved in other functions, including cytoplasmic

Journal of Cell Science 99, 21-31 (1991)Printed in Great Britain © The Company of Biologists Limited 1991

transport, positioning of the nucleus and organelledistribution.

The cytoplasmic events occurring during male meiosishave largely been demonstrated by electron microscopy(Maruyama, 1968; Dickinson and Heslop-Harrison, 1970;Dickinson, 1981; Bird etal. 1983). Although the behaviorof organelles is not the same in all higher plants, theappearance of cell polarity in organelle distribution hasbeen reported in several cases. In Lilium, Dickinson (1981)showed that plastids were located at the poles of eachmicrospore, whereas in Cosmos mitochondria were associ-ated with the nucleus at the same tetrad stage (Dickinsonand Potter, 1979). Such cell polarity in the distribution ofplastids and mitochondria during microsporogenesis isalso seen in other species, including gymnosperms andferns (Willemse, 1972; Wolniak, 1976; Bednara et al. 1986;Rodkiewicz et al. 1986). These studies also suggest that thedistribution of organelles might be controlled by the

21

cytoskeleton. However, the evidence is insufficient toconclude that there is a definite relationship between thetwo.

It has generally been difficult to observe these organ-elles by light microscopy. However, fluorescence mi-croscopy after staining with 4'6-diamidino-2-phenylindole(DAPT), a DNA-specific fluorochrome, enables smallamounts of DNA such as chloroplast DNA to be detected(Coleman, 1978, 1979; James and Jope, 1978;Kuroiwa etal. 19816). Using DAPI staining, Miyamura et al. (1987)showed the presence of plastid and mitochondrial nu-cleoids in cells during microsporogenesis. This techniqueis superior to electron microscopy in visualizing thedistribution of organelles such as plastids and mitochon-dria throughout the cell.

I have therefore used fluorescence microscopy, withthree kinds of fluorescent probes, to examine the role of themicrotubule and actin cytoskeletons in organelle move-ment and distribution in lily microsporocytes and micro-spores. These cells have many advantages for such studies.First, the developmental process during microsporo-genesis is highly synchronized and the successive stagesare closely correlated with bud length (Ito and Stern, 1967;Tanaka et al. 1979). The cells are also large, andprotoplasts can be obtained at each stage (Ito, 1973;Tanaka et al. 1987). Further, the organization of themicrotubule and actin cytoskeletons during microsporo-genesis has been described previously in Lilium by otherinvestigators (Dickinson and Sheldon, 1984; Sheldon andDickinson, 1986; Sheldon and Hawes, 1988).

I show that microtubules radiating from the cell nuclei,which were first reported in Lilium by Dickinson andSheldon (1984), and which commonly appear during malemeiosis in other higher plants (Van Lammeren et al. 1985;Hogan, 1987; Traas et al. 1989), play an important role inplastid arrangement during male meiosis. The signifi-cance of the specific distribution of plastids is discussedfurther with respect to male gamete formation aftermeiosis.

Materials and methods

Plant materialThe experimental material was Lilium longiflorum cv. 'Georgia',grown in a greenhouse. A correlation between the successivestages of microsporogenesis and bud length in this lily haspreviously been reported (Ito and Stern, 1967; Tanaka et al. 1979),but it varies slightly with culture conditions. Under greenhouseconditions, meiosis begins when the bud is 13 mm in length and iscompleted by the time the bud reaches 27 mm (pollen tetrads). Inthis study, premeiotic cells (11mm buds), microsporocytes takenduring meiotic division (13-27 mm buds), pollen tetrads(27-28 mm buds) and uninucleate microspores just after theliberation from tetrads (28 mm buds) were isolated from anthersof each bud. Each anther was cut open at one end with sharpforceps, and cells were extruded by gentle squeezing from the enddistal to the cut. In the anthers of buds 27-28 mm in length, atleast two kinds of pollen tetrads can be recognized under anordinary microscope; one is the early-mid tetrad in which theformation of pollen exine consisting of sporopollenin does notoccur, the other is the late tetrad in which exine formation hasbegun.

Protoplast isolationThe microsporocytes extruded during meiotic division and pollentetrads at the early-mid stage were suspended directly in enzymesolution, which contained 1 % (w/v) Macerozyme R-10 (Yakult

Honsha Co., Ltd, Tokyo, Japan), 1% (w/v) Cellulase Onozuka R-10 (Yakult), 0.1 % (w/v) pectolyase Y23 (Seishin PharmaceuticalCo., Ltd, Tokyo, Japan), 0.1 % (w/v) Zymolyase (Seikagakukogyo,Co., Ltd, Tokyo, Japan), 0.5% (w/v) potassium dextran sulphateand 0.3 M sucrose in White's modified solution (White, 1963) withthe pH adjusted to 5.8. After treatment with enzyme for 1 h at30°C under stationary conditions, isolated protoplasts werefiltered through 50 inn nylon mesh and washed three times withWhite's medium containing 0.3 M sucrose (washing solution) bycentrifugation at 100 # for 5min. In some experiments centrifu-gation was not employed.

Cell culture and drug treatmentsPollen tetrads extruded from anthers were cultured in White'smedium supplemented with 0.05 % (w/v) yeast extract and 0.3 Msucrose, at pH5.8 (Ito and Stern, 1967).

Anticytoskeletal drugs were added from the start of culture.Cytochalasin B (Sigma Chemical Co., St Louis, MO, USA) wasdissolved in dimethyl sulfoxide (DMSO) at a concentration of2 mgrnl"1. The stock solution was added to the culture medium togive final concentrations of 1, 10 and lOOjigml"1. Addition ofDMSO alone to the medium (below 5 % (v/v)) had no appreciableeffect on the cultured cells. Colchicine (Wako Pure ChemicalIndustries, Ltd, Osaka, Japan) dissolved in distilled water wasused at concentrations of 0.1,1 and lOmgml"1. All cultures weremaintained at 25 °C for 2 days.

DNA stainingThe cells extruded from anthers, the isolated protoplasts and thecultured tetrads were fixed in 1 % (v/v) glutaraldehyde inwashing solution for more than 30min, and were then stainedwith l^gml"1 4',6-diamidino-2-phenylindole (DAPI, Sigma)dissolved in NS buffer containing lmM CaCl2, lmM MgCl2,0.1 mM ZnS04, 0.25 M sucrose, 0.8 mM phenylmethylsulfonylfluoride, 6mM 2-mercaptoethanol, lmM EDTA and 20mMTris-HCl, pH7.6 (Kuroiwa etal. 1981a).

Microtubule stainingThe intact and cultured cells were fixed in 3 % (w/v) paraformal-dehyde in microtubule-stabilizing buffer containing 50 mM Pipes(pH 7.3), 10 mM EGTA, 5 mM MgS04 and 0.3 M sucrose for 30 min.They were allowed to settle and washed twice in buffer withoutfixative. Samples were then affixed to coverslips coated with 0.1 %(w/v) poly-L-lysine (MT 70 000-150000, Sigma) and were oftensquashed beneath another non-coated coverslip. Freshly isolatedprotoplasts were affixed directly to the poly-L-lysine coatedcoverslips and fixed in methanol (-20°C). After washing inborate-buffered saline (BBS) containing 162 mM H3BO3, 35 mMNaOH and 144 mM NaCl (pH 8.1) supplemented with 0.05 % (w/v)Triton X-100, the coverslips were exposed to mouse monoclonalanti-chick brain o<-tubulin antibody (Amersham, Buckingham-shire, England) at a final concentration of 1:500 (v/v) for 1 h at37 °C in a moist chamber. Following further washing in BBS with0.05% (w/v) Triton X-100, the samples were labeled with a 1:25(v/v) dilution of fluorescein isothiocyanate (FITC)-conjugatedanti-mouse Ig (Amersham, Buckinghamshire, England) for 1 h at37 °C in a moist chamber. After washing in BBS, samples werestained with l/igml"1 DAPI for 15 min, and finally washed inphosphate-buffered saline (PBS), and mounted in glycerol (10%glycerin in PBS).

Actin stainingAll samples were suspended in the washing solution and weredirectly stained with 0.33/IM rhodamine-phalloidin (MolecularProbes, Evgene, OR, USA) dissolved in PBS and 1 ^gmP 1 DAPIon a glass slide for more than 1 h.

Fluorescence microscopyStained cells were observed under an Olympus BHS-RFKepifluorescence microscope equipped with phase contrast. Prep-arations were viewed under ultraviolet irradiation (a UG 1excitation filter and an L 435 barrier filter) for DAPI, blue-light

22 I. Tanaka

irradiation (BP 490 and EY 455 excitation filters and an O 515barrier filter) for FITC, and green-light irradiation (BP 545 andEO 530 excitation filters and an 0 590 barrier filter) forrhodamine. Fluorescence micrographs were usually taken on FujiNeopan F film (ASA 32) for DAPI and Kodak Tri-X Pan film (ASA400) for FITC and rhodamine.

Results

Behavior of plastid nucleoids during meiotic divisionFig. 1 shows the presence and distribution pattern ofnuclear and organellar DNA in a microsporocyte at eachstage of meiotic division, as visualized by fluorescencemicroscopy after staining with DAPI. Since the cell wall ofthe microsporocytes had practically no fluorescence, theappearance of cell nuclei and the numbers of cells asvisualized by light microscopy were used to identify eachstage. The fluorescence from plastid nucleiods could not becompletely distinguished from that of mitochondrialnucleoids. However, it is known that plastid nucleoids arelarger and fluoresce more intensively than mitochondrialones (Miyamura et al. 1987) and, in fact, the fluorescencefrom mitochondrial nucleoids was only detectable here insquash preparations at high magnification (Fig. 2). Be-cause of the very faint fluorescence from the mitochondrialnucleoids, it was concluded that the distinct fluorescentdots observed in the cytoplasm in Fig. 1 indicated thepresence of plastid nucleoids.

The size and distribution of plastid nucleiods changed,in parallel with the behavior of the cell nuclei duringmeiosis. Plastid nucleoids were present in premeiotic cells(Fig. 1A), but their number seemed to increase after theonset of meiosis. In leptonema and zygonema, manysmaller-sized plastid nucleoids were observed throughoutthe cytoplasm (Fig. 1B,C). During pachytene and diplo-tene stages the size of the plastid nucleoids graduallyincreased (Fig. 1D,E), remaining at a maximum untildiakinesis (Fig. IF). At metaphase I, anaphase I andprophase I the larger nucleoids were distributed atrandom in the cytoplasm of the nearly spherical cells(Fig. 1G,H)- From anaphase I to telophase I the plastidnucleoids migrated between two daughter nuclei. Attelophase I, when the two daughter nuclei were polarizedat the ends of the cell, the majority of plastid nucleoidswere aggregated in the equatorial region, where the cellplate was being formed (Fig. II). This was the firstappearance of cell polarity in the distribution of plastidnucleoids.

During interphase II, after completion of meioticdivision I, the polarized distribution of the plastidnucleoids was maintained for a while (Fig. 1J). Frominterphase II to prophase II, when the nucleus had movedto the center of the cell, the plastid nucleoids becamescattered around the periphery of each ellipsoidal cell, butthereafter their distribution became random again(Fig. IK). The behavior of plastid nucleoids from meta-phase to telophase during the second meiotic division(Fig. 1L-N) was very similar to that during the firstdivision (Fig. 1G-I), showing polarity at telophase II(Fig. IN). At early tetrad stage, just after completion ofthe second division, the distribution of the plastidnucleoids was polarized to one side of each microspore,distant from the cell nucleus and close to the site of cellplate formation. Following migration of the nucleus to thecenter of the cell, the plastid nucleoids were once againscattered around the periphery of the cell and migrated tothe two ends of the ellipsoidal cell (Fig. 10).

Cell polarity of plastid nucleoids in young microsporesA distinct cell polarity in the distribution of plastidnucleoids was established in young microspores from mid-tetrad stage (Fig. 2A). The plastid nucleoids were polar-ized at the two poles of each microspore, with the nucleusin the center of the cell. High magnification revealed thepresence of larger plastid nucleiods and smaller mitochon-drial ones (Fig. 2B). Although the plastid and mitochon-drial nucleoids were not fully distinguished from eachother under phase-contrast optics, fluorescent opticsshowed that the population of plastid nucleoids wasclearly located away from the nucleus (Fig. 2C,D). Themitochondrial nucleoids showed little cell polarity,although there was some tendency to concentrate aroundthe nucleus. This was also true of microsporocytes duringmeiosis.

Protoplasts was easily prepared from pollen tetradswithout exine by enzyme treatment. The distributionpattern of plastid nucleoids in the isolated protoplasts wasnot uniform, but cell polarity was apparent in nearly allcells (Fig. 2E-G). The plastid nucleoids were located awayfrom the nuclei. Similar polarity was seen in protoplastsisolated from microsporocytes at telophase I, interphase IIand telophase II. Cell polarity was also unaffected bycentrifugation. However, the organelle nucleoids could notbe recognized in free microspores soon after liberationfrom pollen tetrads, because of the presence of the thickpollen exine.

Microtubule organization during microsporogenesisThe organization of the microtubule cytoskeleton duringmeiotic division was usually examined using isolatedprotoplasts. The results were similar to those previouslydescribed in Lilium (Sheldon and Dickinson, 1986). Inaddition to cortical, cytoplasmic and spindle microtubules(Fig. 3A-C), peculiar microtubules radiating out from theentire nucleus were observed at telophase I, interphase II,telophase II and the pollen tetrad stage (Fig. 3D,F,H andJ). At telophase I and II when the two daughter nucleiwere polarized at the ends of each protoplast, the radiatingmicrotubules were predominantly arrayed towards thecell plate where phragmoplasts were formed (Fig. 3D,H).Double staining for microtubules and DNA suggested thatthe radiating microtubules played a role in the specificplastid arrangements (Fig. 3E,I). At interphase II andpollen tetrad stage the presence of radiating microtubulesalso seemed to result in the exclusion of the plastids fromaround the cell nuclei (Fig. 3F,G,J and K).

In squashed microspores without enzyme treatment theradiating microtubules were very striking (Fig. 3L). Theyseemed to radiate out uniformly from the nuclear surfaceto the plasma membrane and to be less dense at the twopoles that were furthest away from the nucleus. Doublestaining showed that the plastid nucleoids were concen-trated in the pole regions where the radiating micro-tubules were less dense (Fig. 3L,M). Soon after liberationfrom tetrads, visualization of the microtubule cytoskeletonwas difficult because of the presence of the thick pollenexine.

Actin organization during microsporogenesis (Fig. 4)Throughout meiotic division actin filaments were ob-served in the cytoplasm of intact cells not treated withfixative. This was consistent with previously reportedresults in Lilium (Sheldon and Hawes, 1988). They usuallyformed a network, regardless of the behavior of thechromosomes (Fig. 4A-C). At telophase I, interphase II,

Plastid distribution during nude meiosis 23

Fig. 1. Behavior of plastid nucleoids during male meiosis in Lilium longiflorum. Fluorescent dots in the cytoplasm show thepresence of plastid nucleoids, visualized with DAPI staining. (A) Premeiotic interphase. (B) Leptotene. (C) Zygotene. (D) Pachytene.(E) Diplotene. (F) Diakinesis. (G) Metaphase I. (H) Anaphase I. (I) Telophase I. (J) Interphase II. (K) Prophase II. (L) Metaphase II.(M) Anaphase II. (N) Telophase II. (0) Early-tetrad stage. Arrows in I and N show the cell plate. Non-random distribution of theplastid nucleoids is evident in I, J, N and O. Bar, 20 fira.

24 /. Tanaka

Fig. 2. Cell polarity of plastid nucleoids in young microspores of Lilium longiflorum. All cells are stained with DAPI, and viewedby fluorescence microscopy except in C. (A) Mid-tetrad stage. (B) Young microspores with thin exine at late tetrad stage. (C, D) Ayoung microspore under phase-contrast (C) and fluorescence (D) optics. In B and D plastid nucleoids (large arrows) are polarized atthe two poles, whereas mitochondrial nucleoids (small arrows) are scattered throughout the cytoplasm. (E-G) Protoplasts isolatedfrom pollen tetrads. The plastid nucleoids are absent from the periphery of the cell nuclei. Bars, 20 |Um.

telophase II and pollen tetrad stage when distinct plastidpolarity was recognized, the actin filaments were ran-domly distributed in the cytoplasm (Fig. 4D,E). Therandom distribution of actin filaments was also preservedin isolated protoplasts (Fig. 4F), although the plastidshowed polarized distribution. This indicates that theactin cytoskeleton was not involved in the organizationand distribution of the plastid nucleoids.

Effects of anticytoskeletal drugs on the plastidarrangements (Fig. 5)When pollen tetrads were cultured in a nutrient mediumfor 2 days, plastid polarity was retained in nearly all themicrospores, although the cells were somewhat enlarged(Fig. 5A). Addition of 1 or lOmgml"1 colchicine to thenutrient medium disrupted plastid polarity, resulting in arandom distribution of plastids (Fig. 5B). Double staining

Plastid distribution during male meiosis 25

Fig. 3. Organization of microtubule cytoskeleton during microsporogenesis in Lilium longiflorum. All cells were doubly stainedwith anti-tubulin antibody and DAPI. (A-D, F, H, J, L) Immunofluorescence; (E, G, I, K, M) DAPI staining. Images for DAPI areeliminated in A-C. (A-K) Protoplasts isolated at prophase I (A), metaphase I (B), metaphase II (C), telophase I (D, E), interphaseII (F, G), telophase II (H, I) and pollen tetrad stage (J, K). (L, M) A young microspore. Microtubules radiating from the cell nucleiare seen in D, F, H, J and L. In D there are distinct radiating microtubules independent of phragmoplast microtubules. In L someof the radiating microtubules reach the plasma membrane. Thick arrows show the presence of plastid nucleoids, where theradiating microtubules from the nuclei are less dense. Bars, 20 /.an.

26 /. Tanaka

Fig. 4. Organization of the actin cytoskeleton during microsporogenesis in Lilium longiflorum. All cells were doubly stained withrhodamine-phalloidin and DAPI. (A-D, F) Rhodamine-phalloidin staining; (E) DAPI staining. (A) Metaphase I. (B) Metaphase II.(C) Pollen tetrad. (D, E) Interphase II. (F) An isolated protoplast from pollen tetrads. Actin filaments form an extensive cytoplasmicnetwork irrespective of the behavior of the chromosomes and the plastid polarity. Bar, 20 ̂ m.

for microtubules and DNA showed that the disruption ofplastid polarity was due to the disappearance of theradiating microtubules, caused by the colchicine(Fig. 5C-F). Treatment with O.lmgml"1 colchicine hadlittle effect on plastid arrangement, and the radiatingmicrotubules appeared the same as in the control cells.When colchicine was added during meiotic division, itusually caused arrest of the division process or multi-polardivision by destroying the spindle microtubules. However,when added after metaphase II, pollen tetrads withrandomly distributed plastids were frequently formed(data not shown).

On the other hand, addition of cytochalasin B had noeffect on plastid polarity in pollen tetrads. At concen-trations of 10 and lOO^gml"1 actin filaments becamefragmented, or disappeared, but the arrangement ofplastids and microtubules was little affected (Fig. 5G-J).

Discussion

Visualization of plastid nucleoids by DAPI stainingIn this study, sequential changes in the distribution ofplastid and mitochondrial nucleoids during male meiosiswere followed using fluorescence microscopy after stainingwith DAPI. The behavior of these organelles during male

meiosis has mainly been documented ultrastructurally(Maruyama, 1968; Dickinson and Heslop-Harrison, 1970;Bird et al. 1983). Observations of ultrathin sections are notable to show the polarity and distribution of the totalorganellar population. Observation of these organelles inthe light microscope is also limited. The use of DAPIstaining enables plastids and mitochondria to be clearlyvisualized by the fluorescence of the DNA in theirnucleoids.

It is difficult to distinguish between plastid andmitochondrial nucleoids by differences in the intensity offluorescence, although it has been reported that plastidDNA fluoresces more strongly (Coleman, 1978, 1979;James and Jope, 1978; Kuroiwa et al. 19816). Miyamuraetal. (1987) also reported that in Triticum aestivum theplastid nucleoids could be distinguished from mitochon-drial ones on the basis of differences in size using phase-contrast optics, and that during male meiosis the plastidnucleoids fluoresced more intensely than mitochondrialnucleoids. Also, enlargement of the larger dots duringprophase I seemed to coincide with the fact that activeDNA synthesis took place in plastids at the pachytenestage (Smyth, 1982). Further, localization of plastids at thetwo poles in young microspores of Lilium has beendemonstrated by electron microscopy (Dickinson, 1981).This was also shown in the lily studied here (data not

Plastid distribution during male meiosis 27

Fig. 5. Effects of anticytoskeletal drugs on the distribution of plastids in cultured pollen tetrads of Lilium longiflorum. Pollentetrads were cultured for 2 days in drug-free medium (A, C, D), lOmgml"1 colchicine-containing medium (B, E, F) and 100(igml"cytochalasin B-containing medium (G-J). All cells were doubly stained with DAPI and anti-tubulin antibody orrhodamine-phalloidin. (A, B, D, F, H, J) DAPI staining; (C, E, I) immunofluorescence; (G) rhodamine-phalloidin staining. In thecontrol culture, plastid polarity due to the presence of radiating microtubules was evident (A, C, D). Colchicine disrupted theplastid polarity by depolymerizing the radiating microtubules (B, E, F). In the colchicine-treated cell the microtubules arerecognized only near the nuclear envelope. Disappearance of actin filaments caused by cytochalasin B had no effect on the plastidpolarity (G, H), and the microtubules remained in the presence of cytochalasin B (I, J). Bars, 20,um.

shown). Therefore it was concluded that the larger, moreintensively fluorescing dots were plastid nucleoids.

Appearance of plastid polarity during male meiosisAlthough slightly squashed preparations were used toexamine plastid nucleoids, a non-random pattern of

distribution was distinct at specific stages, and this wasnot affected by cell wall digestion and centrifugationprocedures. At telophase most plastid nucleoids aggre-gated at the equatorial region of the cell. Such a migrationof plastids to the equator at telophase has been reported inother plant species (Wolniak, 1976; Bednara etal. 1986;

28 /. Tanaka

Rodkiewicz et al. 1986), but in these mitochondria migratetogether with the plastids. In this study polarity in thedistribution of the plastids was recognized only attelophase I, interphase II, telophase II and tetrad stage(Figs 1, 2).

The finding that from telophase I the appearance ofplastid polarity was restricted to telophase and interphasesuggested that the phenomenon was not related to theformation of spindles and chromosomes, but to thepresence of the nuclear envelope. Further, it is assumedthat the position of the nucleus in the cell is an importantfactor in the distribution of plastids.

Function of microtubules in the distribution of plastidsThe dynamics of microtubule organization during malemeiosis have already been demonstrated, using immuno-fluorescence, in several angiosperms (Van Lammerenetal. 1985; Sheldon and Dickinson, 1986; Hogan, 1987;Traas et al. 1989) and mosses (Brown and Lemmon,1987a,6; Busby and Gunning, 1988). These studies showedthat microtubules functioned in nuclear division andcytokinesis during male meiosis through the formation ofspindles and phragmoplasts. This was also found to be thecase in the present study.

Microtubules that radiate out from the cell nuclei havebeen observed during microsporogenesis in several angio-sperms, but no clear function has been demonstrated. Thisradial system of microtubules was first observed inprotoplasts isolated from young microspores of Lilium byDickinson and Sheldon (1984). They suggested that theseradiating microtubules may function in cytoplasmictransport. Subsequently it was suggested that the radiat-ing microtubules were involved in nuclear positioning(Van Lammeren et al. 1985; Hogan, 1987) or in theestablishment of the future division plane (Traas et al.1989).

In this study, it was clear that the radiating micro-tubules observed at telophase I, interphase II, telophase IIand pollen tetrad stage functioned in maintaining thespecific distribution of plastids. Although double stainingwith anti-tubulin antibody and DAPI reduced the resol-ution of the plastid nucleoids, there was obviously a closerelationship between the radiating microtubule arraysand the distribution of the plastid nucleoids. The corre-lation between the radiating microtubules and plastidarrangement was clearest in young microspores from themid-tetrad stage (Fig. 3L,M). The microtubules seemed toradiate out uniformly from the nuclear surface to theplasma membrane and to be less dense at the two poles,which would result in the peculiar arrangement of plastidsin the ellipsoidal cell, in which they were situated at thetwo poles, furthest away from the nucleus. The role of theradiating microtubules in the arrangement of the plastidswas ascertained by colchicine treatment of pollen tetrads,which lacked spindle and phragmoplast microtubules.Colchicine treatment resulted in the disappearance of theradiating microtubules, and the disruption of the plastidarrangement (Fig. 5B,E and F). It was therefore concludedthat the presence of radiating microtubules resulted inexclusion of plastids from around the cell nuclei. However,it was unclear how the radiating microtubules excludedthe plastids. The plastids might be transported to thedistal ends of the microtubules.

In telophase cells, the function of radiating micro-tubules was complicated because of the presence ofphragmoplast microtubules. The plastids were aggregatedin the equatorial regions of the cells, coincident with the

phragmoplast microtubules. The radiating microtubulesfrom the cell nuclei were predominantly arrayed towardsthe equators of the cells where phragmoplasts wereformed, and were longer between the two nuclei(Fig. 3D,H). However, there were distinct radiating micro-tubules, independent of the phragmoplast microtubules.Such microtubule arrangement at telophase is also seen inother microsporocytes (Sheldon and Dickinson, 1986;Hogan, 1987; Traas et al. 1989). Euteneuer and Mclntosh(1980) showed that the phragmoplast in Haemanthusendosperm was formed at the junction where the distalends of the microtubules from two daughter nuclei wereoverlapping. Although the phragmoplast microtubuleshad originated from the daughter nuclei, they wereshortened so as to become concentrated at the equator atlate telophase in mitotic cells (De Mey et al. 1982). Unlikethe somatic cell phragmoplast, the radiating microtubulesin meiotic cells remained stable after completion of cellplate formation. Further, the radiating microtubulesappeared between two daughter nuclei at telophase intomato when phragmoplast formation did not occur(Hogan, 1987). Therefore, it was assumed that the plastidswere scattered in the equatorial region, the furthestdistance from the daughter nuclei, due to the presence ofthe radiating microtubules. The relationship between theradiating microtubules and the phragmoplast micro-tubules was uncertain. Because the migration of plastidsinto the equator seemed to occur before phragmoplastformation, the radiating microtubules arrayed towardsthe equator might form the phragmoplast after transport-ing the plastids to the equator.

The radiating microtubules were detected in telophaseand interphase cells. This suggests that the radiatingmicrotubules originated from the nuclear envelope andthat the nuclear envelope functioned as the microtubuleorganizing center (MTOC). De May et al. (1982) andDickinson and Sheldon (1984) have suggested thatmicrotubule organizing activity is associated with thenuclear envelope. It has also been suggested thatmicrotubules in higher plant cells are organized fromzones or areas of the nucleus (Clayton et al. 1985). In thepresent study it was clear that the radiating microtubulesfor the most part originated from the entire nucleus(Fig. 3L). The extensive organization of the radiatingmicrotubules would also account for the change in plastidpolarity caused by migration of the cell nuclei from theperiphery to the center during interphase II and pollentetrad stage.

Such an arrangement of organelles based on micro-tubules is well established in animal cells (Stebbings,1990), and in lower plant cells an interaction betweenplastids and microtubules has also been shown (Menzel,1985; Busby and Gunning, 1988). To the author's knowl-edge this study is the first demonstration of a microtubule-based plastid arrangement in higher plant cells.

Independence of actin filaments in the distribution ofplastidsIt has been suggested that actin filaments might functionin the movement and distribution of organelles (Menzeland Elsner-Menzel, 1989; Van Lammelen et al. 1990). Thedynamics of actin organization during male meiosis wererecently revealed by rhodamine-phalloidin staining(Sheldon and Hawes, 1988; Traas et al. 1989; VanLammelen et al. 1990). I therefore examined the organiz-ation of the actin cytoskeleton throughout microsporo-genesis. My results were not always consistent with

Plastid distribution during male meiosis 29

previous reports in terms of the association of actinfilaments with the spindle and the phragmoplast. How-ever, the organization of actin in dyads and pollen tetradswas found to be the same as previously reported, in thatthe actin filaments were oriented at random throughoutthe cytoplasm. In the dyads and pollen tetrads though,polarity of plastid distribution was clear. Further, disap-pearance of actin filaments caused by cytochalasin B hadno effect on the plastid polarity in the cultured pollentetrads. Therefore, it was concluded that actin filamentsdo not function in the distribution of plastids.

Significance of the radiating microtubules in malegamete formationThe microspores derived from male meiosis soon divide togive generative and vegetative cells. The composition ofthe cytoplasm of the generative cell, which is theprogenitor of two male gametes in angiosperm plants, isvery different from that of the vegetative cell. Ultrastruc-tural studies have shown that the generative cells ofseveral species, including the lily studied here, lackplastids (Heslop-Harrison, 1968; Sanger and Jackson,1971; Clauhs and Grun, 1977; Rodriguez-Garcia andGarcia, 1978; Nakamura and Miki-Hirosige, 1985;Schroder, 1986; Tanaka et al. 1989). These studies sugges-ted that plastids were maternally inherited in thesespecies. Although there may be a number of ways by whichplastids are excluded from the male gametes (Sears, 1980;Hagemann and Schroder, 1989), it is clear that it is thefirst mitosis that results in all or almost all of the plastidsbeing absent from the generative cell. The first mitosis is apolar one, in which the microspore nucleus migrates to oneend of the cell. If the radial system of microtubules wereinherited through the first mitosis, and accounted for theexclusion of plastids from around the nucleus, it wouldseem likely that the generative cell formed at one endwould not contain plastids. The arrangement of plastidsand organization of the microtubule cytoskeleton duringthe first mitosis in microspores are therefore now beinginvestigated, using protoplasts isolated from pollen grains(Tanaka et al. 1987).

The author is grateful to Dr P. Lumsden for critically readingthe manuscript. This work was supported in part by a Grant-in-Aid for Special Research on Priority Areas (Project no. 02242101,Cellular and Molecular Basis for Reproductive Processes inPlants) from the Ministry of Education, Science and Culture,Japan.

References

BEDNARA, J., GIELWANOWSKA, I. AND RODKIEWICZ, B. (1986). Regulararrangement of mitochondria and plastids during sporogeneais inEquisetum. Protoplasma 130, 145-152.

BED, J., POTER, E. K. AND DICKINSON, H. G. (1983). Events in thecytoplasm during male meiosis in Lilium. J. Cell Sci. 59, 27-42.

BROWN, R. C. AND LKMMON, B. E. (1987a). Division polarity,development and configuration of microtubule arrays in bryophytemeiosis I. Meiotic prophaae to metaphase I. Protoplasma 137, 84-99.

BROWN, R. C. AND LKMMON, B. E. (19876). Division polarity,development and configuration of microtubule arrays in bryophytemeiosis n. Anaphase I to the tetrad. Protoplasma 138, 1-10.

BUSBY, C. H. AND GUNNING, B. E. S. (1988). Development of thequadripolar meiotic cytoskeleton in spore mother cells of the mossFunana hygrometrica. J. Cell Sci. 91, 127-137.

CLAUHS, R. P. AND GRUN, P. (1977). Changes in plastid andmitochondrial content during maturation of generative cells ofSolanum (Solanaceae). Am. J. Bot. 64, 377-383.

CLAYTON, L , BLACK, C M. AND LLOYD, C. W. (1985). Microtubulenucleating sites in higher plant cells identified by an autoantibodyagainst pericentriolar material. J. Cell Bwl. 101, 319-324.

COLEMAN, A. W. (1978). Visualization of chloroplast DNA with twofluorochromes. Expl Cell Res. 114, 95-100.

COLEMAN, A. W. (1979). Use of the fluorochromes 4'6-diamidino-2-phenylindole in genetic and developmental studies of chloroplastDNA. J. Cell Biol. 82, 299-305.

DK MEY, J., LAMBERT, A. M., BAJER, A. S., MOEREMANS, M AND DEBRABANDER, M. (1982). Visualization of microtubules in interphaseand mitotic plant cells of Haemanthus endoserm with the immuno-gold staining method. Proc. natn. Acad. Sci. U.SA. 79, 1898-1902.

DICKINSON, H. G. (1981). The structure and chemistry of plastiddifferentiation during male meiosis in Lilium henryi. J. Cell Sci. 52,223-241.

DICKINSON, H. G. AND HESLOP-HARRISON, J. (1970). The behaviour ofplastids during meiosis in the microsporocyte of Lilium longiflorumThunb. Cytobios 6, 103-118.

DICKINSON, H. G. AND POTTER, U. (1979). Post-meiotic nucleo-cytoplasmic interaction in Cosmos bipinnatus. Planta 145, 449-457.

DICKINSON, H. G. AND SHELDON, J. M. (1984). A radial system ofmicrotubules extending between the nuclear envelope and the plasmamembrane during early male haplophase in flowering plants. Planta161, 86-90.

EUTENEUER, U. AND MCINTOSH, J. R (1980). Polarity of midbody andphragmoplast microtubules. J Cell Biol 87, 509-515.

HAOEMANN, R. AND SCHRODER, M.-B. (1989). The cytological basis of theplastid inheritance in angiosperms. Protoplasma 152, 57-64.

HESLOP-HABHISON, J. (1968) Synchronous pollen mitosis and theformation of the generative cell in maBsulate orchids. J. Cell Set. 3,457-466

HOOAN, C. J. (1987). Microtubule patterns during meiosis in two higherplant species. Protoplasma 138, 126-136.

ITO, M. (1973) Studies on the behavior of meiotic protoplasts I. Isolationfrom microsporocytes of liliaceous plants. Bot. Mag. Tokyo 86,133-141.

ITO, M. AND STERN, H. (1967). Studies of meiosis in vitro I. In vitroculture of meiotic cells. Devi Biol. 16, 36-53.

JAMES, T. W. AND JOPE, C. (1978). Visualization by fluorescence ofchloroplast DNA in higher plants by means of the DNA-specific probe4'6-diamidino-2-phenylindole. J. Cell Bwl. 79, 623-630.

KUROIWA, T., NISHIBAYASHI, S., KAWANO, S. AND SUZUKI, T. (1981a).Visualization of DNA in various phage (T4, x, T7, 029) by ethidiumbromide epifluore3cent microscopy. Expenentia 37, 969-970.

KUROIWA, T., SUZUKI, T., OGAWA, K. AND KAWANO, S. (19816). Thechloroplast nucleus: Distribution, number, size, and Bhape, and amodel for the multiplication of the chloroplast genome duringchloroplast development. PI. Cell Physwl. 22, 381-396

MAHUYAMA, K. (1968). Electron microscopic observation of plastids andmitochondria during pollen development in Tradescantia paludosa.Cytologia 33, 482-497.

MENZEL, D. (1986). Fine structure study on the association of thecaulerpalean plastid with microtubule bundles in the siphonaleangreen alga Chlorodesmis fastigiata Ducker (Udoteaceae). Protoplasma125, 103-110.

MENZKL, D. AND ELSNER-MENZEL, C. (1989). Actin-based chloroplastrearrangements in the cortex of the giant coenocytic alga Caulerpa.Protoplasma 150, 1-8.

MIYAMURA, S., KUROIWA, T. AND NAOATA, T. (1987). Disappearance ofplastid and mitochondrial nucleoids during the formation ofgenerative cells of higher plants revealed by fluorescence microscopy.Protoplasma 141, 149-159.

NAKAMURA, S. AND MIKI-HIROSIGE, H. (1985). Fine-structural study onthe formation of the generative cell wall and intine-3 layer in agrowing pollen grain of Lilium longiflorum. Am. J. Bot 72, 365—375.

RODKIEWICZ, B., BEDNARA, J., MOSTOWSKA, A., DUDA, E. AND STOBIECKA,H (1986). The change in disposition of plastids and mitochondriaduring microsporogenesis and sporogenesis in some higher plantsAda bot. neerl. 35, 209-215.

RODRIGUEZ-GAHCIA, M. I AND GARCIA, A. (1978). Differentiation of theplastid population during microsporogenesis and the development ofthe pollen grain in the Liliaceae. Biol. Cell 33, 63-70.

SANGER, J. M. AND JACKSON, W. T. (1971). Fine structure study of pollendevelopment in Haemanthus katherinae Baker LT. Microtubules andelongation of the generative cells. J. Cell Sci. 8, 303-315.

SCHRODER, M.-B. (1986). Ultrastructural studies on plastids ofgenerative and vegetative cells in Liliaceae 5. The behavior of plastidsduring pollen development in Chlorophytum comosum (Thunb.)Jacques. Theor. appl. Genet. 72, 840-844.

SEARS, B. B. (1980). Elimination of plastids during spermatogenesis andfertilization in the plant kingdom. Plasmid 4, 233-255.

SHELDON, J. M. AND DICKINSON, H. G. (1986) Pollen wall formation inLdium: The effect of chaotropic agents, and the organization of themicrotubular cytoskeleton during pattern development Planta 168,11-23.

30 /. Tanaka

SHELDON, J. M. AND HAWES, C. (1988). The actin cytoskeleton duringmale meiosis in Lilium. Cell Biol. Int. Rep. 12, 471-476.

SMYTH, D. R. (1982) Proplastid DNA synthesis at pachytene in pollenmother cells of Lilium henryi. J. Cell Sci. 56, 293-302.

STKBBINOS, H. (1990). How is microtubule-based organelle tranalocationregulated? J. Cell Sci. 96, 5-7.

TANAKA, I., KITAZUME, C. AND ITO, M. (1987). The isolation and cultureof lily pollen protoplasts. Plant Sci. 50, 205-211.

TANAKA, I., NAKAMURA, S AND MIKI-HIBOSIGE, H. (1989). Structuralfeatures of isolated generative cells and their protoplasts from pollenof some liliaceous plants. Gamete Res. 24, 361-374.

TANAKA, I., TAGUCHI, T. AND ITO, M. (1979). Studies on microsporedevelopment in liliaceous plants I. The duration of the cell cycle anddevelopmental aspects in lily microspores Bot. Mag. Tokyo 92,291-298.

TRAAS, J A , BURGAIN, S. AND DUMAS Da VAULX, R. (1989). Theorganization of the cytoskeleton during meiosis in eggplant (Solanummelongena (L.)): microtubules and F-actin are both necessary forcoordinated meiotic division. J. Cell Sci. 92, 641-550.

VAN LAMMBRKN, A. A. M., BEDNARA, J AND WILLEMSE, M. T. M (1989).

Organization of the actin cytoskeleton during pollen development inGastena uerrucosa (Mill.) H. Duval visualized with rhodamine-phalloidin. Planta 178, 531-539.

VAN LAMMEREN, A. A. M., KEIJZER, C. J., WILLEMSE, M. T. M. ANDKIEFT, H. (1985). Structure and function of the microtubularcytoskeleton during pollen development in Gastena verrucosa (Mill.)H. Duval. Planta 165, 1-11.

WHITE, P. R. (1963). The Cultivation of Animal and Plant Cells, 2ndedn, p. 60. New York: Ronald Press

WILLEMSK, M. T. M. (1972). Morphological and quantitative changes inthe population of cell organelles during microsporogenesis of Gastenavernicosa. Acta bot. neerl. 21, 17-31.

WOLNIAK, S M. (1976). Organelle distribution and apportionmentduring meiosis in the microsporocyte of Ginkgo bdoba L. Am. J. Bot.63, 251-258.

(Received 21 November 1990 - Accepted 24 January 1991)

Plastid distribution during male meiosis 31