Growth of S. cereale plants

Secale cereale cv. `Blanco' plants were grown in the greenhouse (Berkeley, CA). Spikes containing flowers were used within 1.5 hours of harvesting. Anthers were dissected out of their florets on a dry glass microscope slide and transferred into culture medium within 1 hour. Only the pedicilate floret was used in all experiments.

Anther culture

Preliminary anther culture experiments were performed on intact anthers. We were able to bisect the anthers longitudinally prior to placing them in culture medium, thereby increasing the starting number of synchronous meiocyte populations to six. There was no difference in the success rate of anther cultures performed with whole or halved anthers, and all subsequent experiments were performed on anther halves.

Upon removing anthers from the floret, the three anthers were sequentially cut down the connective tissue joining the locules, giving rise to two anther halves, each consisting of two locules. Upon bisecting an anther, the two halves were immediately placed into culture medium. Anther culture and fixations were performed in flat-bottom 96-well plates, using 50 μl solution per well, and one anther half per well. After all anther sets had been placed into wells, 96-well plates were placed on a rotary shaker at 80 rpm and covered to keep the dishes dark. After the specified culture time at room temperature, anthers were fixed as described below. Anther culture medium consisted of 1× Murashige and Skoog medium (GIBCO), 3% sucrose, and 1× Murashige and Skoog vitamins (Sigma) at pH 7.

Duration of meiotic stages in vivo

Bennett et al. determined the duration of meiotic stages for S. cereale anthers in vivo ( Bennett et al., 1971). Anther samples were excised from the plant over a time course and the meiotic stage determined cytologically. These experiments, however, were performed on plants growing at 20°C, whereas our anther culture experiments were conducted at room temperature (25°C) and are thus not directly comparable. In earlier work, Bennett et al. determined the duration of meiosis for S. cereale, although not the substages, at several temperatures ( Bennett et al., 1972). Additionally, Bennett reported that in the angiosperms, changes in the total duration of meiosis generally result from proportional changes in all of the meiotic substages ( Bennett, 1977). To compare our timing of leptotene with the work of Bennett et al. (Bennett, 1971), we made the following calculations on the basis of the above data. At 20°C, S. cereale meiosis lasted 2. 12 days, whereas at 25°C, meiosis lasted 1.63 days ( Bennett et al., 1972); thus meiosis proceeds 1.3 times slower at the lower temperature. Bennett et al. report that leptotene occurs over 20 hours at 20°C ( Bennett et al., 1971). By dividing this value by the temperature-dependence ratio above (1.3), we would expect leptotene to last roughly 15.4 hours at 25°C.

Drug treatments

Initial drug concentrations were determined from values reported in the literature. Concentrations were tested in a spindle-function assay in cultured anthers. Either tapetal cell nuclear divisions (which occur when pairing is near completion) or meiosis I or II divisions were observed for accurate chromosome alignment (metaphase) and segregation (anaphase/telophase) by DAPI staining. Spindle function was used to determine the time required for drugs to exert their effects. Colchicine and amiprophos methyl (APM) resulted in failed mitosis after 3 hours of treatment ( Table 1); all treatments were performed for a minimum of 6 hours.

All drugs were stored as 100× stocks in DMSO at -20°C. Colchicine (Sigma) stocks ranged from 1 mM to 500 mM. For inhibition of telomere clustering, 100 μM was consistently used. Efficient MT depolymerization required at least 1 mM colchicine. β- and γ-lumicolchicine (Sigma), vinblastine sulfate (CalBiochem), podophyllotoxin (Sigma), trimethoxybenzene (ACROS) and 2-methoxy estradiol (Sigma) were prepared as 10 mM stocks. APM was a gift of L. Morejohn (University of Texas, Austin) and stored as a 2.5 mM stock. Drugs were diluted 1:100 in anther culture medium immediately prior to use. Control medium contained 1% DMSO.

Fixation

For fluorescent in situ hybridization (FISH) experiments, anthers were fixed for 1 hour in 4% paraformaldehyde in 1× buffer A (80 mM KCl, 20 mM NaCl, 0.5 mM EGTA, 2 mM EDTA, 15 mM PIPES, 0.35 M sorbitol, 1 mM DTT, 0.5 mM spermidine, 0.2 mM spermine, pH 7.0) at room temperature. Anthers were subsequently washed in 1×buffer A and stored at 4°C.

For MT fixation, anthers were fixed in 8% paraformaldehyde in 1×PHEMS (60 mM PIPES, 25 mM Hepes, 10 mM EGTA, 2 mM MgCl₂, 0.32 M sorbitol, pH 6.8) for 2 hours at room temperature ( Chan and Cande, 1998). Anthers were subsequently washed in 1×PHEMS and stored at 4°C.

FISH procedure

Meiocytes and associated cells from a single anther or anther half were embedded in 5% acrylamide polymerized between two coverslips. FISH protocols were essentially as described previously ( Bass et al., 1997): prehybe I (1×SSC, 1×buffer A, 20% formamide), prehybe II (2×SSC, 35% formamide), prehybe III (2×SSC, 50% formamide). Coverslips were then incubated in hybridization solution (prehybe III plus approximately 200 ng probe) for at least 30 minutes ( Bass et al., 1997). Coverslips were heated 95°C for 5.5 minutes, transferred to a humid chamber for incubation overnight at room temperature. Coverslips were washed in 1×PBS. Chromatin was stained with 3 μg/ml DAPI in 1×PBS. Coverslips were mounted in glycerol. Telomeres were detected using a probe to the telomere repeat (CCCTAAACCCTAAACCCTAAACCCTAAA) with either 5′ Cy-5 or FITC label (Genset, Paris).

Microtubule immunofluorescence

Meiocytes and associated cells were embedded in 5% acrylamide polymerized between two coverslips. Cell walls were then digested with 1.5%β -glucuronidase (from Helix promatia, Sigma) in 1×PBS for 15 minutes at 36°C. Coverslips were washed thoroughly with 1×PBS. A monoclonal antibody against sea urchin α-tubulin (a gift of D. Asai, Purdue University) was applied in 1×PBS at 1:500 dilution and incubated at room temperature overnight. Coverslips were washed in 1×PBS. Secondary antibody, Alexa488-conjugated goat anti-mouse IgG (Molecular Probes) at 1:50 dilution, was applied in 1×PBS and incubated overnight at room temperature. Coverslips were washed in 1×PBS. Chromatin was stained with 3 μg/ml DAPI, and samples were mounted in glycerol.

Microscopy

All images were acquired with an Applied Precision DeltaVision microscope system equipped with an Olympus IX70 inverted microscope. A 40× 1.35 NA UApo oil immersion lens was used for all experiments. Cells were imaged in three dimensions (xyz); z-axis sections were collected at 0.2 μm spacing. Images were deconvolved using a standard conservative algorithm ( Chen et al., 1995).

Modeling and quantification

Models of nuclei were created using the DeltaVision/softWoRx 3DModel program. The nuclear periphery was modeled on the outer edge of DAPI-stained chromatin. A single point was picked for each telomere detected by FISH; we attempted to pick the peak intensity of the signal. 3DModel data were imported into MATLAB (version 5.1.0.420, The MathWorks, Inc.) for analyses. Two measurements were calculated to determine the degree of telomere clustering: (1) all pairwise telomere-to-telomere distances, which are referred to as telomere distances; and (2) the minimum of all possible telomere-to-telomere distances for each telomere, which are referred to as nearest neighbor distances. Nearest neighbor distances were used to judge small changes in telomere distributions. All distance measurements were normalized to the nuclear radius.

Telomere and nearest neighbor distances are presented as the distribution of means from individual nuclei in a given sample. All possible distances were calculated for a single nucleus, the mean of the distances for a single nucleus was obtained, and the mean distances from all nuclei in a sample were combined. Distributions (box-whisker plots) or means (t-test) are presented. Sample means (the mean of the mean distances) are described plus or minus the standard deviation. Differences were assessed at 99.9% confidence (P=0.001) using an unequal variance Student's t-test, unless indicated otherwise.

Telomeres are tightly clustered in the bouquet during zygotene in Secale cereale cv. `Blanco'

We first established guidelines for classifying the early meiotic stages in S. cereale cv. `Blanco' using our fixation, staining and imaging procedures. Meiotic stages were defined by chromatin appearance in accordance with classic definitions ( Wilson, 1925; Zickler and Kleckner, 1998) ( Fig. 1). We found that chromatin morphology correlated closely with telomere distribution, nucleolar number and nuclear diameter throughout meiotic prophase. During the stages we examined, cells from a single anther were synchronous with respect to meiotic stage ( Fig. 2b), as is true of most grasses, and thus the general descriptions provided are representative of all of the meiotic cells observed.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Meiotic prophase stages and timing of the bouquet in Secale cereale cv. `Blanco'. Telomeres (green) were detected by FISH using a probe to the telomere repeat, and chromatin (red) was stained with DAPI. Single optical sections of meiotic nuclei are shown. (a) Premeiotic interphase, (b) leptotene and (c) zygotene. The bouquet is evidenced by the close spacing of the telomere FISH signals. (d) Pachytene. Bar, 10 μm.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

(a) Diagram of rye plant and meiotic development and anther culture methodology. A single rye floret contains three anthers at an equivalent developmental stage. Meiosis occurs synchronously within a single anther and among the three anthers from one floret. The anthers are shown in yellow-green, and meiotic cells are depicted in blue. The meiotic cells cannot be observed through the anther wall; the diagram merely suggests their location. One anther half provides the time 0 control, whereas the halves of a second anther are put into culture. (b) Synchrony of meiosis in rye anthers. A projection of a 3D image of meiotic cells from a single anther. All cells appear to be in metaphase of the first meiotic division, as indicated by the presence of seven bivalents. Bar, 10 μm.

Premeiotic interphase nuclei ( Fig. 1a) have uncondensed chromatin, unfused nucleoli and polarized but dispersed telomeres. Telomere polarization, perhaps indicative of a Rabl chromosome configuration, was suggested by telomeres that resided in one nuclear hemisphere and were located in the nuclear periphery. Premeiotic nuclei could be distinguished from somatic nuclei in the anther because the diameter of the premeiotic nuclei was several micrometers greater than the diameter of somatic nuclei. Nuclei were defined as leptotene if they contained visible chromatin threads ( Fig. 1b) of irregular widths. Telomeres were scattered in one half of the nuclear periphery in leptotene nuclei and nucleoli were unfused.

Nuclei were classified as zygotene if they contained two types of chromatin fibers ( Fig. 1c): one regular chromatin thread with clearly visible edges, which is likely to represent the paired homologous chromosomes and a second fuzzy-edged chromatin thread of irregular width, similar to leptotene chromosomes. Fully clustered telomeres were found predominantly during zygotene. The delineated chromosome threads were located in the nuclear hemisphere containing the bouquet, whereas the leptotene-like chromatin threads were found away from the telomere cluster site. This is a classic example of zygotene ( Wilson, 1925), the chromatin thread morphologies most probably representing paired and unpaired chromosome territories. Nucleolar fusion appeared to take place during zygotene; nucleolar number varied during this stage even among meiotic cells from a single anther.

In pachytene nuclei ( Fig. 1d), the chromosome threads were distinct throughout the nucleus, being both of equal width and regularly spaced. The telomere cluster was dispersed in pachytene, and telomeric heterochromatin was barely distinguishable from the rest of the chromosome, probably due to continued chromatin condensation.

In our analyses of meiosis in S. cereale cv. `Blanco', we established that the bouquet is present during zygotene ( Fig. 1c), as defined by the appearance of the chromosome fibers. Classic examples of the bouquet in organisms such as lily and salamander as well as more recent studies using model genetic organisms such as Saccharomyces cerevisiae and maize demonstrate that the telomere cluster is generally seen during zygotene ( Scherthan, 2001). However, Mikhailova et al., studying a different inbred line of rye, place the formation of the telomere cluster in premeiotic interphase (Mikhailova, 2001). It is possible that either strain variations or the different fixation methods employed could account for this temporal discrepancy within S. cereale.

Telomere clustering can proceed in an in vitro anther culture system

There are three anthers in each S. cereale floret ( Fig. 2a), and each single anther contains several hundred meiotic cells. In rye, as in many grasses, the meiotic cells of a single anther are developmentally synchronous throughout meiosis ( Fig. 2b); the three anthers of one floret are also meiotically synchronous ( Fig. 2a). To establish the starting stage, one anther half was fixed immediately (time 0). The remaining anther halves were allowed to progress through meiosis in culture.

Experiments were initially evaluated for success by examining the chromatin morphology at time 0 and after various time periods of in vitro culture. Experiments were assessed for progression in meiotic stage between time 0 and the end of the culture period. Nuclei were scored for chromatin appearance, nucleolar number and relative nuclear diameter. Additionally, mitosis in the tapetal cells was monitored.

Indicators of unsuccessful anther culture were micronuclei, collapsed or shrunken nuclei, burst nuclei or aberrant chromatin appearance, most often decondensed chromatin. Distinguishing successful anther culture experiments was straightforward, as gross abnormalities were easily detected at the light microscope level. Anthers that were cultured for longer than 36 hours very rarely yielded normal-appearing nuclei in preliminary experiments; hence all experimental results described here were from experiments of 24 hours or less in duration. Since meiosis in S. cereale is more than one and a half days in duration at 25°C ( Bennett et al., 1971), we were unable to obtain a complete meiotic sequence in any one experiment. However, the interval from premeiotic interphase to zygotene could be completed within 10-16 hours in our in vitro culture system (for example, Figs 3 and 6) (C.R.C. and W.Z.C., unpublished), and thus bouquet formation could be studied readily in cultured anthers.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

The effect of colchicine on telomere clustering. To convey the 3D architecture of the cell, the data corresponding to the entire nuclear volume (approximately 100 z-sections) was divided into quarters, and each quarter was projected into a single image ( Bass et al., 1997). The resulting images are displayed sequentially. Control and 100 μM colchicine-treated anthers were cultured for 16 hours. Telomeres (green) were detected by FISH using a probe to the telomere repeat, and chromatin (red) was stained with DAPI. Bar, 10 μm.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

The effect of APM, a plant-specific microtubule-depolymerizing agent, on telomere clustering. Representative nuclei are displayed as sequential projections of the three-dimensional nucleus, as in Fig. 3a. Control and 25 μM APM-treated anthers were cultured for 14 hours. Telomeres (green) were detected by FISH using a probe to the telomere repeat, and chromatin (red) was stained with DAPI. Bar, 10 μm.

Over the course of 95 experiments using optimized medium conditions and culture periods less than 24 hours, 29 successful experiments were performed, giving a success rate of approximately 30%. Successful anther culture experiments accurately approximated in vivo meiotic progression. Meiotic chromosome morphology of cells from cultured anthers appeared to be identical to that of meiotic cells from anthers taken directly from the plant (compare Fig. 3 control with Fig. 5 time 0). Meiotic cells of a single cultured anther (or anther half) remained synchronous throughout the culture period, evidenced both cytologically (data not shown) and by the small standard deviations for telomere distance measurements (discussed below). Failed anther culture experiments may result from mishandling of the tissue, leading to damage and general necrosis, or may be correlated with some unknown environmental factor, such as growth conditions in the greenhouse that affect general anther health. The failure of meiotic progression in such cases appears to be an all or nothing response, affecting the entire meiotic cell population of an anther as opposed to a subset of cells. Our anther culture experiments suggested that the interval between premeiotic interphase and the bouquet stage, corresponding to leptotene, could be completed in 10-16 hours (see above). Importantly, the duration of leptotene we determined is in agreement with that obtained previously ( Bennett et al., 1971) for the duration of S. cereale meiotic stages in vivo (15.4 hours; see Materials and Methods).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

The effect of colchicine on bouquet dispersal. Representative nuclei are displayed as sequential projections of the 3D nucleus as in Fig. 3. Control and 100 μM colchicine-treated anthers were cultured for 18 hours. Telomeres (green) were detected by FISH using a probe for the telomere repeat, and chromatin (red) was stained with DAPI. Bar, 10 μm.

The drugs we initially tested in our investigation are known MT-depolymerizing agents, and monitoring spindle function provided a straightforward means of testing drug effectiveness. Anthers were split and the halves were placed in control or drug-containing medium. After 0.5, 1, 2, 3, 4 or 5 hours in culture, anthers were fixed and examined for indications of MT function. Colchicine and APM inhibited cell division after 3 hours of culture, as indicated by the failure of metaphase chromosomes to align in the spindle midzone or undergo anaphase chromosome separation in the drug-treated cells ( Table 1). Control cells either exhibited normal metaphase alignment or had already completed division ( Table 1).

Colchicine inhibits meiotic telomere clustering

We investigated the effect of colchicine on bouquet formation. Colchicine is an effective MT depolymerizing agent in plants at millimolar or greater concentrations ( Morejohn, 1991), while at 100 μM, its effects on MTs are less pronounced or undetectable ( Liu et al., 1995). Previous experiments examining the effect of colchicine on meiotic prophase have used high colchicine concentrations (1-25 mM), which have the potential of toxic side effects. We assessed the concentration dependence of the effect of colchicine on bouquet formation in the range from 25 μM-5 mM (data not shown). Colchicine was effective in inhibiting telomere clustering at concentrations as low as 25 μM in anther culture, although only a subset of cells were affected. All meiotic cells in an anther were affected at concentrations greater than 75 μM, and thus 100 μM colchicine was used in all subsequent experiments examining bouquet formation. Anthers containing premeiotic interphase or early leptotene cells were cultured in colchicine for 10-16 hours, allowing sufficient time for bouquet formation ( Fig. 3). At the beginning of these experiments, all meiotic cells exhibited dispersed telomeres. The telomeres were associated with the nuclear periphery and were polarized within the nucleus. At the end of the culture period, untreated nuclei showed a prominent telomere cluster, indicated by closely associated telomere FISH signals. Nuclei treated with either 5 mM or 100 μM colchicine did not form a bouquet; telomeres remained dispersed. The telomere distribution in colchicine-treated cells was similar to that observed in time 0 cells; the peripheral localization and polarization of telomeres within the nucleus were retained after colchicine treatment. The results were similar for both colchicine concentrations.

Telomere distributions in colchicine-treated nuclei were analyzed by calculating mean telomere and nearest neighbor distances (see Materials and Methods) ( Fig. 4a,b). The telomere distance ( Fig. 4a) in nuclei treated with 100 μM colchicine for 10-16 hours (1.13±0.07, n=20) was not significantly different from telomere distances in time 0 nuclei (1.15±0.09, n=26). The mean nearest neighbor distance ( Fig. 4b), however, differed slightly but significantly between time 0 (0.30±0.05, n=26) and 100 μM colchicine-treated (0.26±0.05, n=20) nuclei after 10-16 hours, indicating that limited telomere juxtaposition occurred in colchicine-treated cells.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Telomere distributions in colchicine-treated nuclei. Box-whisker plots of the (a) mean telomere distances and (b) nearest neighbor distances are shown. Time 0, n=26; control, n=37; 100 μM colchicine, n=20. Boxes represent the second and third quartiles (50th and 75th percentiles); the bold line through the box is the median, and whiskers extend to the range.

Colchicine does not inhibit telomere dispersal from the bouquet

To determine whether colchicine affected telomere dispersal as cells enter pachytene, anthers containing bouquet-stage cells were cultured with 100 μM colchicine for 14-18 hours, allowing sufficient time for telomere migration out of the bouquet ( Fig. 5). Time 0 anthers showed a prominent telomere cluster. At the end of the culture period, both untreated and colchicine-treated nuclei were marked by dispersed telomeres. Thus, telomere movement resulting in bouquet dissolution did not appear to be affected by colchicine.

Depolymerization of cytoplasmic microtubules does not inhibit telomere clustering

We wished to determine whether colchicine's inhibitory effect on bouquet formation was a result of cytoplasmic MT depolymerization. The concentration of colchicine that exerts a maximal effect on bouquet formation (100 μM) was much lower than the concentrations commonly employed to affect MTs in plants ( Morejohn, 1991), and cytoplasmic MT arrays were present in the meiocytes. However it was possible that these arrays were subtlety disturbed. Thus, we used APM, a highly effective MT depolymerizing agent in higher plants ( Bajer and Mole-Bajer, 1986) and green algae ( Collis and Weeks, 1978), to assess whether cytoplasmic microtubules were required for bouquet formation.

To investigate the effect of MT depolymerization on telomere clustering, anthers in early meiotic prophase were cultured with 25 μM APM for 12-16 hours, allowing for meiotic progression ( Fig. 6). Time 0 nuclei displayed unclustered telomeres; telomeres were distributed over approximately half of the nuclear surface. Telomeres were fully clustered in both untreated and APM-treated nuclei after culture ( Fig. 6). To determine whether colchicine and APM depolymerized cytoplasmic MTs during meiotic prophase at the concentrations used, we observed the organization of MTs after 15 hours of treatment with colchicine or APM ( Fig. 7). Untreated meiotic cells exhibited an array of MTs asymmetrically distributed around the nucleus ( Fig. 7a). Cells treated with 5 mM colchicine ( Fig. 7c) or 25 uM APM did not have visible MTs ( Fig. 7b). Cells treated with 100 μM colchicine ( Fig. 7d) appeared to contain a normal MT array, although a few cells within the sample (less than 10%) appeared to have aberrant MT distributions (data not shown). We conclude that cytoplasmic MT arrays are not required for telomere clustering. This suggests that the inhibitory effects of colchicine on bouquet formation may not be mediated through cytoplasmic MT depolymerization.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

The effect of colchicine and APM on cytoplasmic microtubule organization. MTs were detected using an antibody against sea urchin α-tubulin. The cells shown are of roughly equivalent meiotic stages. (a) Control, (b) 25μ M APM, (c) 5 mM colchicine and (d) 100 μM colchicine. Bar, 10μ m.

Bouquet inhibition is specific to colchicine and podophyllotoxin

We tested compounds structurally related to colchicine and additional MT-depolymerizing drugs for effects on the bouquet. The selected drugs ( Table 2) were used to address three questions. (1) Are the bouquet target and tubulin pharmacologically identical? The β-tubulin molecule contains several distinct binding sites for anti-mitotic drugs. Colchicine and podophyllotoxin bind to one site, whereas vinblastine binds to another site ( Wilson and Jordan, 1994; Hamel, 1996). (2) Is the colchicine A-ring sufficient for bouquet inhibition? Colchicine and podophyllotoxin contain identical methoxybenzene rings, and we wished to determine whether this ring was the only requirement for bouquet inhibition. 3-methoxy estradiol and trimethoxybenzene share the conserved ring, and 3-methoxy estradiol binds weakly to the colchicine-binding site onβ -tubulin ( Hamel, 1996). (3) Do toxic colchicine side effects, for example, interactions with membranes, result in a failure of bouquet formation? Lumicolchicine is a non-MT binding derivative of colchicine, with similar lipophilic properties, which should act as a control for the potential membrane-disrupting actions of colchicine.

Anthers containing meiotic cells in early prophase were treated separately with the various drugs and cultured for 12-18 hours. Telomere distribution was determined by heterochromatin; MTs were judged by immunofluorescence, as described in Fig. 7. Time 0 anthers contained meiotic nuclei with dispersed telomeres. Untreated nuclei that exhibited clustered telomeres were compared with the drug-treated nuclei to determine the presence or absence of the bouquet and cytoplasmic microtubules. Of the drugs tested, we found that only colchicine and the related compound podophyllotoxin inhibited telomere clustering. The remaining drugs had no obvious effects on bouquet formation even though several of the drugs disrupted cytoplasmic microtubules ( Table 2). Importantly, the non-MT depolymerizing derivatives of colchicine, β- andγ -lumicolchicine, did not affect telomere clustering, suggesting that lipophylic-based side effects are an unlikely cause of bouquet inhibition.