Figure 1.
Centriole Migration in Primary Spermatocytes
(A) Time-lapse series of confocal images from a wild-type primary spermatocyte expressing GFP-PACT (centrioles) and His2AvD–GFP (chromosomes). The centrioles (arrows) can be seen moving away from the plasma membrane (0) towards the nucleus (N) and then migrating diametrically apart as the chromatin condenses. The chromosomes are fully condensed at timepoint 121 min.
(B–D) The two centriole pairs (green) projected over the phase-contrast view (grey) can be seen close to the fenestrated NE and away from the plasma membrane (pm) in control cells (B), while they remain plasma membrane-bound in asp (C) and in colcemid-treated wild-type cells (D). In asp spermatocytes (C), the position of the membrane-bound centrioles correlates tightly with the pointed end of phase-dark protrusions (arrows) that are not present in colcemid-treated cells. These reflect the distribution of phase-contrast membranes known to overlap microtubules in these cells.
(E–J) XY projections (E–G) and their corresponding optical sections (H–J) of control (E and H), asp (F and I), colcemid-treated spermatocytes (G and J) expressing an endogenous GFP–α-tubulin confirm that the two major MTOCs in control cells are close to the nucleus, but remain near the plasma membrane in the two experimental conditions. MTOC activity in colcemid-treated spermatocytes was assayed following a 1-s pulse of 350 nm light to inactivate the drug, thus allowing microtubule regrowth. The yellow bar in the XY projections (E–G) marks the position of the corresponding XZ optical sections (H–J).
Figure 2.
Time-Lapse Series of Meiosis Progression in Control, asp, and Colcemid-Treated Spermatocytes
Timepoint 0 coincides with the time of NEB revealed by the sudden entry of GFP signal into the nucleus. In control cells (Video S2), microtubules are mainly organised around the centrosomes (arrows). However, when centrosomes are kept away from the nuclear region by mutation in asp (Video S3) or colcemid treatment (Video S4), microtubule nucleation and growth are clearly revealed over the nuclear region (N), well away from the centrosomes. Such noncentrosomal microtubules may form bundles that eventually are sorted into spindlelike bipolar microtubule arrays. Microtubules were labelled with an endogenous GFP–α-tubulin fusion.
Figure 3.
The Timing of Noncentrosomal Microtubule Nucleation
Referred to NEB (timepoint 0), the timing of chromosome condensation and of onset of chromosome segregation is essentially identical in control, asp, and colcemid-treated spermatocytes. In control spermatocytes, aster microtubules can be seen entering the nuclear region 3–6 min after NEB. They do not in asp or following colcemid treatment. In these two cases, however, centrosome-independent microtubule polymerisation can be seen over the nuclear region. It starts between 9 and 13 min after NEB, coinciding with or very shortly after the end of chromosome condensation.
Figure 4.
The Place of Noncentrosomal Microtubule Nucleation
The initial stages of noncentrosomal microtubule nucleation revealed by an endogenous GFP–α-tubulin fusion (left) and phase contrast (right). Following the corresponding videos, it is possible to unmistakably tell the chromosomes (arrows) apart form the other phase-dark objects that are present over the nuclear region (asterisks). The cell in (A) is shown as a single timeframe and the cell in (B) as a time-lapse series. In both cells, noncentrosomal microtubule nucleation (arrowheads) takes place close to the remains on the NE and does not overlap with the major chromosomes. Nucleation sites can be clustered (A) or dispersed (B). In the time-lapse series (B), only the chromosomes that are in focus are labelled. Timepoint 0 min in these series corresponds to the first sign of noncentrosomal microtubule nucleation, around 11 min after NEB. A white bar marks the growing end of a microtubule bundle that at timepoint 93 min reaches one of the bivalents.
Figure 5.
Chromosome Segregation in Anastral Spindles in Drosophila Spermatocytes
(Control [Video S5]) At metaphase I (0), the bivalents (revealed by a His2Avd–YFP fusion, shown by double arrowheads) are aligned in the middle of the spindle (revealed by a GFP–α-tubulin fusion), at the metaphase plate. At the onset of anaphase (3 min), the homologue chromosomes start to migrate towards opposite poles (single arrowheads) and to decondense. During anaphase B (4 min and 6 min), the spindle poles move apart from each other and the two sets of decondensed chromosomes become further separated.
(asp [Video S6]) At timepoint 0, the bivalents align at the metaphase plate. Homologue chromosomes split apart at the onset of anaphase I (4 min). However, anaphase A migration is highly impaired. By the time the chromosomes start to decondense, they have barely moved towards the spindle poles (8 min and 14 min), and often homologue chromosomes end up included in the same daughter nucleus.
(Colcemid [Video S8]) As in asp spermatocytes, the asters (arrows) remain at the plasma membrane at metaphase I in colcemid-treated cells, and the bivalents align in a metaphase plate-like within the acentrosomal spindles (0 min). Homologue chromosomes split apart at the onset of anaphase (upper cell, 6 min) and significantly segregate from one another (upper cell, 8 min; lower cell, 3 min). Further separation of the daughter nuclei during anaphase B is very limited in these cells (8 min), and cytokinesis does not occur.
Figure 6.
Correlation between the Orientation of the Cytokinesis Furrow, the Asters, and the Spindle in asp Spermatocytes
Schematic representation (A–C) of the relative position of the asters (red), the spindle (yellow), and the cytokinesis furrow (blue), corresponding to a control cell (D) and two examples of asp mutant spermatocytes (E and F), respectively. Asters (arrows) and spindles are labelled with a GFP–α-tubulin fusion. The position of the cleavage furrow (double-headed arrow) was determined by time-lapse imaging of these cells (data not shown). In wild-type cells (n = 10), plotting spindle and furrow orientation relative to the interastral axes shows that asters and spindle are tightly aligned, and cleavage occurs at an angle of 90° ± 10° with respect to them (G). In asp spermatocytes (n = 10), the plane of cleavage occurs at a 90° ± 10° angle with respect to the asters and does not correlate with the orientation of the anastral spindle.
Figure 7.
Noncentrosomal Microtubules and Spindle Assembly
(Central column) Spindle assembly in Drosophila spermatocytes with membrane-bound centrosomes. At the time of NEB, the chromatin (pale blue) starts to condense, and the membrane-bound centrosomes (red) organise asters (yellow) at a significant distance from the nuclear region. Around 12 min after NEB, the first noncentrosomal microtubules (green) start to nucleate near the remnants of the NE (grey), as the chromosomes achieve full condensation (dark blue). These microtubules then bundle, associate with the chromosomes, and eventually end up organised into a bipolar anastral array whose shape is reminiscent of the female meiotic spindle.
(Left column) Spindle assembly in wild-type Drosophila oocytes (Theurkauf and Hawley 1992; Matthies et al. 1996). NEB starts at the beginning of stage 13 of oocyte development. At this stage, the oocyte does not contain centrosomes and the chromosomes (karyosome) are tightly condensed (dark blue). Microtubules (green) appear 11–15 min after NEB within the nuclear region in association with the karyosome. These microtubules form bundles and are sorted around the chromatin into a bipolar spindle. Evidence suggests that ER components may be required for spindle assembly in these cells (Kramer and Hawley 2003). At metaphase I, recombined bivalents are aligned at the spindle equator, while those that have not recombined are found closer to the spindle poles. Meiosis remains arrested at this point (stage 14) until oocyte activation. Despite the obvious morphological similitude, the equivalence between these and the anastral spindles organised in spermatocytes with membrane-bound centrosomes is unclear.
(Right column) Hypothesis regarding the contribution of centrosomal and noncentrosomal microtubules to spindle assembly during meiosis I in wild-type Drosophila spermatocytes. Before NEB, the centrosomes are located at opposite positions near the nucleus. Shortly after NEB, astral microtubules enter the nuclear region and make the first contact with the condensing chromatin. No evidence of noncentrosomal microtubule polymerisation near the nuclear region at this stage has been found yet. Once chromosomes are fully condensed, microtubule bundles of centrosomal origin (yellow) connecting centrosomes to chromosomes already exist. At this stage, noncentrosomal microtubules (green) start to polymerise in association with the remnants of the NE. These microtubules form bundles that interact with the chromosomes and intermingle with the microtubules of centrosomal origin. The fully mature spindle in these cells would therefore contain a spindle-shaped structure made of microtubules of noncentrosomal origin (green) embedded in another spindle-shape array made of two overlapping asters (yellow). We propose that each of these subsets may perform to a certain extent some of the functions carried out by normal spindles, but neither of them can on its own mediate robust cell division.