Research Article

Actin protects mammalian eggs against chromosome segregation errors

See allHide authors and affiliations

Science  25 Aug 2017:
Vol. 357, Issue 6353, eaal1647
DOI: 10.1126/science.aal1647

Actin helps chromosome segregation in eggs

Spindle microtubules are well known to orchestrate the segregation of chromosomes during egg development. But the other major cytoskeletal component, actin, has not been thought to play a role in this process. Mogessie and Schuh examined how chromosomes are segregated in mammalian cells (see the Perspective by Maiato and Ferrás). Instead of using an entirely microtubule-dependent mechanism, mammalian oocytes use a second spindle that is made of F-actin to segregate their chromosomes correctly. Actin associated with the spindle bundles microtubules into functional kinetochore fibers, the key structures that drive chromosome segregation. Increasing or decreasing the number of actin filaments in the spindle causes an imbalance in kinetochore fiber bundling, which results in chromosome segregation errors and aneuploidy, a frequent cause of miscarriage and Down syndrome in humans.

Science, this issue p. eaal1647; see also p. 756

Structured Abstract

INTRODUCTION

When an egg is fertilized by a sperm, the chromosomes of the mother and the father become united, and a genetically unique embryo starts to form. A healthy embryo can only develop if both the egg and the sperm contain precisely one copy of each chromosome. However, human eggs frequently contain an incorrect number of chromosomes. Fertilization of these aneuploid eggs is a leading cause of miscarriages, infertility, and Down syndrome. Most aneuploidy results from chromosome segregation errors during the meiotic divisions of the egg. Chromosome segregation is well established to be driven by a spindle that consists of microtubules. The microtubules first capture and align the chromosomes at the spindle center. During anaphase, the spindle segregates the chromosomes and moves them to the spindle poles. The movement of chromosomes is driven by the shortening of microtubule bundles that are attached to the chromosomes’ kinetochores, called kinetochore fibers (K-fibers). Although actin has been reported in spindles of various species, it is generally not thought to be involved in chromosome segregation.

RATIONALE

We found prominent actin filaments in spindles of human, mouse, porcine, and ovine eggs. The filaments permeated the entire spindle volume and formed structures that resembled a microtubule spindle. The wide conservation of spindle actin suggested that it has an essential function, which we set out to investigate. We visualized spindle actin organization in mouse eggs using superresolution imaging of fluorescently labeled actin, microtubules, and chromosomes in living and fixed cells. We then investigated the function of spindle actin by either depleting actin from the spindle or increasing the amount of actin in the spindle. To deplete spindle actin, we used eggs that lacked formin-2, the actin nucleation factor that is required to form spindle actin, and treated eggs with an actin-depolymerizing drug. To increase spindle actin, we targeted a protein domain to the spindle that stabilizes actin filaments. We then imaged chromosomes and the spindle at high spatial and temporal resolution throughout the first and second meiotic division and quantified defects. We also investigated potential defects in spindle organization, including the dynamics of microtubules in the spindle and the formation and bundling of K-fibers, using a combination of quantitative assays in live and fixed eggs.

RESULTS

Spindle actin assembled gradually during the first meiotic division in mouse oocytes and became most prominent when the chromosomes segregated during anaphase: Dense, long actin bundles that spanned the entire spindle length permeated the microtubule spindle throughout anaphase. Unexpectedly, we found that eggs that lacked actin were more likely to suffer from chromosome segregation errors during both the first and second meiotic division. Actin-depleted eggs took longer to align their chromosomes and frequently contained chromosomes that were trapped at the spindle poles or oscillated within the spindle. Also, the acute addition of actin-disrupting drugs to spindles in which the chromosomes had already been aligned led to displacement of chromosomes from the spindle center. When chromosomes were segregated during anaphase, individual chromosomes were frequently lagging behind the main masses of chromosomes, and the overall speed of chromosome movement was reduced. Actin is likely to facilitate chromosome alignment and segregation by promoting the formation of K-fibers because K-fibers were reduced when spindle actin was absent. Consistent with this model, increasing the amount of actin in the spindle led to an increase in K-fibers as well as defects in chromosome alignment and segregation.

CONCLUSION

That actin drives the formation of K-fibers was unexpected because K-fibers are generally thought to be formed by microtubule-associated proteins and kinetochore components independently of actin. Our data therefore highlight a previously unknown mechanism of how K-fibers are generated in mammalian eggs. Mammalian eggs are highly prone to chromosome segregation errors, which are a leading cause of miscarriages and genetic disorders such as Down syndrome. Understanding the mechanisms that drive chromosome segregation in mammalian eggs is therefore particularly important. The presence of spindle actin in other mammalian eggs, including humans, hints at a conserved function in chromosome segregation. This function may even extend beyond mammals because actin filaments have been reported in spindles of a variety of species and cell types.

Spindle actin in a mouse egg.

Microtubule spindles in mammalian eggs are permeated by actin filaments (blue). Actin filaments are essential for the accurate alignment and segregation of chromosomes (magenta) during meiosis because they promote the formation of K-fibers. K-fibers are specialized microtubule fibers within the spindle that attach to the chromosomes’ kinetochores and mediate chromosome segregation.

Abstract

Chromosome segregation is driven by a spindle that is made of microtubules but is generally thought to be independent of actin. Here, we report an unexpected actin-dependent mechanism that drives the accurate alignment and segregation of chromosomes in mammalian eggs. Prominent actin filaments permeated the microtubule spindle in eggs of several mammalian species, including humans. Disrupting actin in mouse eggs led to significantly increased numbers of misaligned chromosomes as well as lagging chromosomes during meiosis I and II. We found that actin drives accurate chromosome segregation by promoting the formation of functional kinetochore fibers, the microtubule bundles that align and segregate the chromosomes. Thus, actin is essential to prevent chromosome segregation errors in eggs, which are a leading cause of miscarriages, infertility, and Down syndrome.

Chromosome segregation is well-established to be driven by a spindle that consists of microtubules. The microtubules first capture and align the chromosomes at the spindle center. During anaphase, the chromosomes are segregated and move to the spindle poles. The movement of chromosomes is driven by the shortening of microtubule bundles that are attached to the chromosomes’ kinetochores (1, 2). These kinetochore fibers (K-fibers) cooperate with a large number of microtubule-associated motor and nonmotor proteins to align and segregate the chromosomes (3, 4). Although actin has been reported in spindles of various species, it is generally not thought to be involved in chromosome segregation (5).

The reliable segregation of chromosomes is of particular importance during meiosis, the specialized cell division that leads to the formation of eggs and sperm. The egg and the sperm need to have precisely one copy of each chromosome to give rise to a healthy embryo upon fertilization. But surprisingly, human eggs frequently contain an incorrect number of chromosomes. Depending on the age of the woman, around 10 to 50% of eggs are chromosomally abnormal (68). This makes aneuploidy in human eggs the most common cause of pregnancy losses and several genetic disorders such as Down syndrome (911).

Aneuploidy in eggs arises from chromosome segregation errors during the meiotic divisions of the egg’s progenitor cell, the oocyte (11, 12). Oocytes are stored in the ovary for prolonged periods, where they are arrested in prophase of the first meiotic division (13). Once every menstrual cycle, an oocyte resumes meiosis and matures into a fertilizable egg. First, the nuclear envelope breaks down, and a microtubule spindle assembles in the center of the oocyte. Next, the spindle aligns the chromosomes at its equator and moves to the oocyte surface. It then segregates the homologous chromosomes during anaphase I and eliminates half of them into a very small cell, the polar body. Subsequently, the metaphase II spindle assembles and aligns the other half of the chromosomes at its equator. The egg then stays arrested in metaphase II until it is fertilized. Upon fertilization, the sister chromatids are segregated during anaphase II, and half of them are eliminated into a second polar body. This completes the second meiotic division. In the newly formed and genetically unique embryo, two pronuclei assemble that contain the male and female DNA, and the mitotic divisions of the embryo begin (13). The mechanisms that drive and control chromosome segregation during the two meiotic divisions as well as the causes of high levels of aneuploidy in meiosis are still incompletely understood.

Results

Actin filaments permeate the microtubule spindle in oocytes of many mammals, including humans

We found prominent actin filaments in spindles of human, mouse (14, 15), porcine, and ovine eggs by staining actin with fluorescent phalloidin (fig. S1A). The filaments permeated the entire spindle volume and formed structures that resembled a microtubule spindle. We will refer to these actin structures as spindle actin throughout the manuscript. The wide conservation of spindle actin suggested that it has an essential function, which we set out to investigate.

Spindle actin assembled gradually during the first meiotic division in mouse oocytes and became most prominent when the homologous chromosomes segregated during anaphase I: Dense, long actin bundles that spanned the entire spindle length permeated the microtubule spindle throughout anaphase (Fig. 1A and movies S1 and S2). Superresolution imaging of actin, microtubules, and chromosomes upon drug-induced microtubule depolymerization revealed that spindle actin is microtubule-dependent (fig. S2, A and B, and movie S3). The prominence of spindle actin during anaphase suggested a potential role in chromosome segregation. To investigate this possibility, we examined the behavior of chromosomes and microtubules in the presence and absence of spindle actin in live oocytes.

Fig. 1 Actin filaments permeate the microtubule spindle during anaphase in mouse oocytes.

(A) Stills from a single confocal section time-lapse movie showing spindle actin [enhanced green fluorescent protein (EGFP)–UtrCH, gray] and chromosomes [H2B–monomeric red fluorescent protein (mRFP), magenta] during anaphase I in a mouse oocyte. (B and C) Single confocal section images of phalloidin-stained spindle actin in (B) dimethyl sulfoxide (DMSO)– or cytochalasin D–treated (CytoD) and (C) Fmn2+/+ or Fmn2−/− eggs. Actin (phalloidin) is shown in gray, and chromosomes (Hoechst) are shown in magenta. Scale bars, 10 μm.

Actin prevents lagging chromosomes during anaphase I

Spindle actin was disrupted by two different approaches: isolation of oocytes from Fmn2−/− mice, in which lack of the actin nucleator formin-2 disrupts spindle actin (Fig. 1C) (14, 15), as well as treatment of oocytes with the actin-depolymerizing drug cytochalasin D (Fig. 1B, fig. S3, and movies S4 and S5). Disruption of actin had no notable effect on the timing of meiotic maturation events (fig. S4A), alignment of homologous chromosomes on the metaphase I spindle (fig. S4B), efficiency of anaphase I onset (fig. S4C), or spindle elongation during anaphase I (fig. S4, D and E). Thus, oocytes are generally healthy and capable of progressing through meiosis when actin is absent. But strikingly, oocytes were significantly more likely to show chromosome segregation defects during anaphase I; instead of segregating all chromosomes simultaneously, individual chromosomes were frequently lagging behind the two main masses of chromosomes (movies S6 to S9). The overall speed of chromosome movement was also reduced (Fig. 2, C and F).

Fig. 2 Disruption of F-actin causes lagging chromosomes during anaphase I.

(A) Scheme illustrating the criteria that were used to quantify mildly and severely lagging chromosomes and chromosome misalignment in live mouse oocytes. (B) Frequency of mildly lagging chromosomes in DMSO- or cytochalasin D–treated oocytes quantified as in (A). Data are from nine independent experiments. (C) Chromosome segregation speeds in DMSO- or cytochalasin D–treated oocytes. Data are from three independent experiments. (D) Stills from representative time-lapse movies of anaphase I in DMSO- or cytochalasin D–treated oocytes. Microtubules (EGFP-MAP4-MTBD) are shown in gray, and chromosomes (H2B-mRFP) are shown in magenta. Dashed line boxes mark regions that are magnified in insets. (E) Frequency of mildly lagging chromosomes in Fmn2+/+ or Fmn2−/− oocytes quantified as in (A). Data are from seven independent experiments. (F) Chromosome segregation speeds in Fmn2+/+ or Fmn2−/− oocytes. Data are from three independent experiments. (G) Maximum intensity projection stills from representative time-lapse movies of anaphase I in Fmn2+/+ or Fmn2−/− oocytes. Microtubules (EGFP-MAP4-MTBD) are shown in gray, and chromosomes (H2B-mRFP) are shown in magenta. Dashed line boxes mark regions that are magnified in insets. (H) Frequency of mildly lagging chromosomes in BSA- or exoenzyme C3–microinjected oocytes quantified as in (A). Data are from three independent experiments. (I) Chromosome segregation speeds in BSA- or exoenzyme C3–microinjected oocytes. Data are from three independent experiments. (J) Maximum intensity projection stills from representative time-lapse movies of anaphase I in BSA- or exoenzyme C3–microinjected oocytes. Microtubules (EGFP-MAP4-MTBD) are shown in gray, and chromosomes (H2B-mRFP) are shown in magenta. Dashed line boxes mark regions that are magnified in insets. Scale bars, 10 μm. Error bars represent SEM. Box plots show median (horizontal white lines), mean (small white squares), 25th and 75th percentiles (boxes), 5th and 95th percentiles (whiskers), and where present, outliers (small lines). The number of analyzed oocytes is specified in italics. Fisher’s exact test was used to test for significance in (B), (E), and (H). Two-tailed Student’s t test was used to test for significance in (C), (F), and (I). Z-projections, 16 sections every 1.5 μm.

Oocytes were scored to have mildly or severely lagging chromosomes when individual chromosomes were still separated from the main chromosome mass 12 or 18 min after anaphase onset, respectively (Fig. 2A). These manual quantifications (Fig. 2, B, D, E, and G, and fig. S5A) as well as the automated detection of lagging chromosomes in Imaris (fig. S5, G and H) confirmed that disruption of actin led to a significant increase in chromosomes that moved more slowly than the two main chromosome masses.

Formin-2 also nucleates a network of actin filaments in the cytoplasm that transports Rab11a-positive vesicles and mediates asymmetric spindle positioning in mouse oocytes (1417). Thus, we tested whether the lack of actin-dependent transport in Fmn2−/− oocytes contributes to chromosome segregation errors. The formation of Rab11a-positive vesicles and their transport along actin as well as asymmetric spindle positioning can be blocked by brefeldin A (BFA) (17). BFA treatment did not lead to lagging chromosomes, suggesting that the actin-dependent transport of the spindle and Rab11a-positive vesicles are dispensable for chromosome segregation (figs. S1C and S5, C, E, and F).

We also tested whether the failure of cytokinesis in cytochalasin D–treated and Fmn2−/− oocytes contributes to lagging chromosomes (18, 19). To this end, we microinjected oocytes with exoenzyme C3, a specific inhibitor of the small guanosine triphosphatase Rho A that is essential for cytokinesis (20) but dispensable for the formation of spindle actin (fig. S1B). The frequency of lagging chromosomes was not significantly different from bovine serum albumin (BSA)–injected control oocytes (Fig. 2, H to J, and figs. S4, C and F, and S5, D and G). Thus, lagging chromosomes in cytochalasin D–treated and Fmn2−/− oocytes are not caused by cytokinetic failure. Collectively, these data establish a role for actin in preventing lagging chromosomes.

Actin drives chromosome alignment during metaphase II

Spindle actin was also present during the second meiotic division (Fig. 1, B and C, and movie S10). To investigate the function of spindle actin during meiosis II, we added cytochalasin D to eggs that had just progressed through anaphase I. Cytochalasin D addition significantly delayed the congression of chromosomes on the metaphase II spindle (Fig. 3, A to C). Consistent with this observation, cytochalasin D treatment throughout oocyte maturation also caused severe chromosome alignment defects in meiosis II. Instead of being tightly clustered on the metaphase plate, chromosomes were scattered across the entire spindle or trapped at the spindle poles (Fig. 3, D and E; fig. S6, A and B; and movies S7 and S9). These defects were independent of actin-mediated vesicle transport because treatment of oocytes with BFA did not have an effect on the efficiency of chromosome alignment (Fig. 3F and fig. S6D).

Fig. 3 Actin promotes chromosome alignment during metaphase II.

(A) Stills from representative time-lapse movies of chromosome congression in DMSO- or cytochalasin D–treated mouse eggs. Eggs were treated with DMSO or cytochalasin D after completion of anaphase I. Microtubules (EGFP-MAP4-MTBD) are shown in gray, and chromosomes (H2B-mRFP) are shown in magenta. (B) Quantification of chromosome congression on the metaphase II spindle in eggs treated with DMSO or cytochalasin D after completion of anaphase I. (C) Average time of chromosome congression in DMSO- or cytochalasin D–treated eggs. (D to G) Representative images and frequency of severe chromosome misalignment in (D) DMSO- or cytochalasin D–treated, (E) Fmn2+/+ or Fmn2−/−, (F) MeOH- or BFA-treated, and (G) BSA- or exoenzyme C3–microinjected eggs. Microtubules (EGFP-MAP4-MTBD) are shown in gray, and chromosomes (H2B-mRFP) are shown in magenta. Arrowheads highlight severely misaligned chromosomes. (H) Frequency of severe chromosome misalignment in eggs that were acutely treated with DMSO or cytochalasin D. (I) Stills from representative time-lapse movies of chromosome misalignment in acutely DMSO- or cytochalasin D–treated eggs. Dashed line boxes mark regions that are magnified in insets. Microtubules (EGFP-MAP4-MTBD) are shown in gray, and chromosomes (H2B-mRFP) are shown in magenta. Scale bar, 10 μm. Error bars represent SEM. Data are from two [(B) and (C)], four [(D) to (F)], and three [(G) and (H)] independent experiments. The number of analyzed eggs is specified in italics. Chromosome misalignment was quantified as in Fig. 2A. Fisher’s exact test was used to test for significance. Z-projections, 16 sections every 1.5 μm.

Cytokinetic failure in cytochalasin D–treated and Fmn2−/− oocytes frequently gave rise to two spindles that later merged into a larger spindle containing twice the number of chromosomes (fig. S6, G to I, and movies S7 and S9). However, misalignment of chromosomes in these eggs was not due to the increased number of chromosomes on their metaphase II spindles; chromosomes aligned efficiently when cytokinesis was blocked by exoenyzme C3 or BFA treatment, which led to the same increase in chromosome number without affecting spindle actin (Fig. 3, F and G, and fig. S6, C and D).

When we acutely removed actin from metaphase II spindles, accurately aligned chromosomes became displaced from the spindle equator and moved toward the poles, often oscillating between the two spindle poles (Fig. 3, H and I, and movies S11 and S12). Thus, actin is required to both align and maintain the correct position of chromosomes on the metaphase II spindle.

Actin prevents lagging chromosomes during anaphase II

Because spindle actin persists during anaphase II (fig. S7), we investigated whether it is also essential for accurate chromosome segregation in meiosis II. Eggs progressed into anaphase II even when chromosomes were trapped at the spindle poles (Fig. 4G), indicating that disruption of actin can indeed give rise to aneuploidy.

Fig. 4 Disruption of F-actin causes lagging chromosomes during anaphase II.

(A) Frequency of lagging chromosomes in DMSO- or cytochalasin D–treated eggs quantified as in Fig. 2A. Data are from five independent experiments. (B) Stills from representative time-lapse movies of anaphase II in DMSO- or cytochalasin D–treated eggs. Dashed line boxes mark regions that are magnified in insets. Microtubules (EGFP-MAP4-MTBD) are shown in gray, and chromosomes (H2B-mRFP) are shown in magenta. (C) Frequency of lagging chromosomes in Fmn2+/+ or Fmn2−/− eggs quantified as in Fig. 2A. Data are from four independent experiments. (D) Stills from representative time-lapse movies of anaphase II in Fmn2+/+ or Fmn2−/− eggs. Dashed line boxes mark regions that are magnified in insets. Microtubules (EGFP-MAP4-MTBD) are shown in gray, and chromosomes (H2B-mRFP) are shown in magenta. (E) Frequency of lagging chromosomes in eggs acutely treated with DMSO or cytochalasin D, quantified as in Fig. 2A. Data are from three independent experiments. (F) Stills from representative time-lapse movies of anaphase II in eggs acutely treated with DMSO or cytochalasin D. Dashed line boxes mark regions that are magnified in insets. Microtubules (EGFP-MAP4-MTBD) are shown in gray, and chromosomes (H2B-mRFP) are shown in magenta. (G) Frequency of DMSO- or cytochalasin D–treated and Fmn2+/+ or Fmn2−/− eggs with chromosomes trapped at spindle poles at anaphase II onset. Data are from four independent experiments. Scale bars, 10 μm. Error bars represent SEM. The number of analyzed eggs is specified in italics. Fisher’s exact test was used to test for significance. Z-projections, 16 sections every 1.5 μm.

The frequency of lagging chromosomes during anaphase II was increased in Fmn2−/− eggs (Fig. 4, C and D) and in eggs that had been treated with cytochalasin D throughout oocyte maturation (Fig. 4, A and B). The increase in lagging chromosomes was not due to cytokinetic failure in these eggs because blocking cytokinesis by inhibiting Rho A did not lead to lagging chromosomes (fig. S6, E and F). Chromosome segregation errors in anaphase II were also independent of other previous meiotic defects because lagging chromosomes were also induced by the acute addition of cytochalasin D in meiosis II (Fig. 4, E and F). Together, these data suggest that actin is also essential for accurate chromosome segregation during anaphase II.

Mechanism of actin-dependent chromosome alignment and segregation

Why do chromosomes misalign and missegregate when actin is disrupted? To address this, we examined whether key mechanisms that drive chromosome alignment and segregation are perturbed. The turnover of microtubules in the spindle as well as the flux of microtubules toward the spindle pole, which is thought to promote the movement of chromosomes during anaphase (2123), were not significantly altered (fig. S8). More subtle changes in K-fiber dynamics may be masked, though, by the high abundance of nonkinetochore microtubules in acentrosomal spindles (24). Misaligned chromosomes were also positive for MAD1 (fig. S9), suggesting that they are still recognized by the spindle assembly checkpoint, which monitors the attachment of kinetochores to microtubules (25). The presence of misaligned chromosomes did not block progression of oocytes into anaphase (Fig. 4G) . This is consistent with previous work demonstrating that mouse oocytes progress into anaphase efficiently in the presence of small numbers of misaligned chromosomes (2630).

Actin promotes the formation of K-fibers

K-fibers are instrumental in creating the forces that are required for chromosome segregation (3, 4). Thus, we investigated whether the formation of K-fibers or their correct interaction with kinetochores relies on spindle actin. To this end, we used a cold-mediated microtubule depolymerization assay, which selectively preserves kinetochore-bound, stable microtubules (K-fibers) (Fig. 5B).

Fig. 5 Actin promotes the formation of kinetochore fibers.

(A) Quantification of monotelic (only one kinetochore is attached to microtubules), merotelic (a single kinetochore is attached to K-fibers originating from both spindle poles), and syntelic (both kinetochores are attached to K-fibers originating from the same spindle pole) kinetochore-microtubule attachments in DMSO- or cytochalasin D–treated eggs. (B and C) To investigate whether disruption of actin perturbs the formation of K-fibers, we selectively depolymerized the more dynamic, nonkinetochore-bound microtubules by briefly placing the oocytes on ice (B). We then fixed the cold-treated oocytes and immunostained the remaining population of kinetochore-bound, stable microtubules (K-fibers) with an antibody against α-tubulin (B). As a measure of K-fiber density, we quantified the mean fluorescence intensities of three-dimensional isosurface reconstructions of K-fibers that were generated from high-spatial-resolution images using Imaris (Bitplane) [(C), isosurface reconstruction]. (D) Representative immunofluorescence images of K-fibers in DMSO- or cytochalasin D–treated eggs. Microtubules (tubulin) are shown in gray, and chromosomes (Hoechst) are shown in magenta. Insets are magnifications of regions marked by dashed line boxes. (E) Quantification of normalized fluorescence intensity of K-fibers in DMSO- or cytochalasin D–treated eggs. (F to M) Representative immunofluorescence images of K-fibers and normalized fluorescence intensity quantifications in [(F) and (G)] BSA- or exoenzyme C3–microinjected, [(H) and (I)] Fmn2+/+ or Fmn2−/−, [(J) and (K)] acutely DMSO- or cytochalasin D–treated, and [(L) and (M)] CLASP1-expressing Fmn2+/+ or Fmn2−/− eggs. Microtubules (tubulin) are shown in gray, and chromosomes (Hoechst) are shown in magenta. Scale bars, 10 μm. Box plots are as described in Fig. 2. Data are from three [(A), (E), (I), (K), and (M)] and two (G) independent experiments. Error bars represent SEM. The number of analyzed oocytes is specified in italics. Two-tailed Student’s t test was used to test for significance. Z-projections, 20 sections every 1.5 μm.

Lagging chromosomes are frequently caused by merotelic microtubule attachments, an incorrect type of kinetochore-microtubule attachment in which a single kinetochore is attached to both spindle poles (Fig. 5A) (31). However, the frequency of merotelic attachments was not increased upon disruption of actin (Fig. 5A). Instead, the vast majority of chromosomes was correctly attached to microtubules. Only chromosomes that were trapped at the spindle poles were incorrectly attached, with both kinetochores being either attached to the same spindle pole (syntelic) or one kinetochore being unattached (monotelic) (Fig. 5A).

Next, we assessed whether the formation of K-fibers was affected. To measure K-fiber density, we quantified the mean fluorescence intensity of K-fiber isosurface reconstructions that were generated from high-resolution three-dimensional image stacks in Imaris (Bitplane) (Fig. 5C). This revealed that the intensity of K-fibers in meiosis I was significantly reduced when actin was absent in Fmn2−/− or cytochalasin D–treated oocytes (fig. S10, A to D) but not in BFA-treated oocytes (fig. S10, E and F). Also, the acute addition of cytochalasin D to metaphase II–arrested eggs led to a significant reduction in K fiber intensity (Fig. 5, J and K). Consistent with these results, treatment of oocytes with cytochalasin D throughout oocyte maturation reduced the K-fiber intensity of the metaphase II spindle by ~40% (Fig. 5, D and E). This was particularly unexpected because these eggs had twice the number of chromosomes and hence twice the number of kinetochore fibers because of cytokinetic failure. In stark contrast, inhibiting cytokinesis by exoenzyme C3, which does not affect spindle actin (fig. S1B), led to a ~1.6-fold increase in K-fiber intensity (Fig. 5, F and G), which is consistent with an increase in chromosomes on the spindle. Fmn2−/− eggs also have twice the number of chromosomes on the spindle but did not show the expected increase in K-fiber intensity (Fig. 5, H and I). Overexpression of the K-fiber dynamics regulating protein CLASP1 (cytoplasmic linker–associated protein 1) (32) in Fmn2−/− eggs rescued the K-fiber intensity to levels that were similar to those of exoenzyme C3–microinjected metaphase II eggs (Fig. 5, L and M), providing further evidence that K-fibers are decreased in Fmn2−/− eggs. Together, these data suggest that actin promotes the formation of K-fibers in mammalian eggs.

Stabilization of actin leads to chromosome segregation errors

Next, we investigated whether actin needs to be dynamic to promote K-fiber formation, chromosome alignment, and segregation. To this end, we treated oocytes with high concentrations of SiR-Actin, a derivative of the actin-stabilizing drug Jasplakinolide (Fig. 6A) (33). SiR-Actin–treated oocytes were significantly more likely to have chromosome alignment and segregation defects during meiosis I and II (fig. S6, B and C). High levels of SiR-Actin also blocked cytokinesis and led to the formation of spindles with twice the number of chromosomes. However, the increase in chromosome numbers was not linked to an increase in K-fiber intensity (fig. S6, D and E), suggesting that actin needs to be dynamic to promote the formation of K-fibers. Consistent with this observation, treatment of oocytes throughout maturation with BFA, which also blocks cytokinesis and vesicle-dependent actin dynamics (17), did not lead to an increase in metaphase II K-fiber intensity (fig. S10, G and H). Together, these data indicate that actin needs to be dynamic to promote K-fiber formation.

Fig. 6 Stabilization of actin by jasplakinolide leads to chromosome alignment and segregation errors.

(A) Spindle actin in a live mouse egg labeled with SiR-Actin, a derivative of the actin-stabilizing drug Jasplakinolide. (B) Frequency of mildly lagging chromosomes during anaphase I in DMSO- or SiR-Actin–treated oocytes. Data are from three independent experiments. (C) Representative images and frequency of severe chromosome misalignment in DMSO- or SiR-Actin–treated eggs. Arrowheads indicate severely misaligned chromosomes. Scale bar, 10 μm. Data are from three independent experiments. Z-projections, 16 sections every 1.5 μm. (D) Representative immunofluorescence images of K-fibers (tubulin, gray) and chromosomes (Hoechst, magenta) in DMSO- or SiR-Actin–treated eggs. Insets are magnifications of regions marked by dashed line boxes. Z-projections, 20 sections every 1.5 μm. (E) Quantification of normalized fluorescence intensity of K-fibers in DMSO- or SiR-Actin–treated eggs. Data are from three independent experiments. Error bars represent SEM. Box plots are as described in Fig. 2. The number of analyzed oocytes is specified in italics. Fisher’s exact test [(B) and (C)] and two-tailed Student’s t test (E) were used to test for significance.

Increasing spindle actin enhances K-fiber bundling

How does actin promote the formation of K-fibers? Several in vivo and in vitro studies suggest that actin can organize, bundle, and stabilize microtubules (3438). It was thus conceivable that actin drives the formation of K-fibers via similar mechanisms. The actin filaments in the spindle were on average 1.2 ± 0.4 μm apart (fig. S11A), suggesting that they are indeed dense enough to bundle and stabilize K-fibers in meiotic spindles.

To test this hypothesis further, we artificially enriched meiotic spindles in actin by targeting the actin-stabilizing calponin-homology (CH) domain of utrophin to the spindle (Fig. 7B and fig. S11, B and C). The actin-enriched spindles were typically elongated with sharply focused poles, and their microtubules appeared more bundled (Fig. 7, A and B). Cold-mediated microtubule depolymerization revealed a striking increase in the degree of K-fiber bundling in actin-enriched spindles: In the majority of wild-type spindles, K-fibers appeared as individual microtubule bundles (Fig. 7, C to E, and fig. S11H). In stark contrast, over 90% of actin-enriched spindles had K-fibers that were not discernible as individual fibers but were instead clustered into thick bundles (Fig. 7, C, D, and F; and fig. S11I). Also, the intensity of K-fibers was significantly increased (Fig. 8, E and F), whereas tubulin turnover and microtubule flux were decreased (Fig. 8, A to D).

Fig. 7 Increasing spindle actin enhances K-fiber bundling.

(A) Microtubule bundling and spindle pole organization in control (MAP4) and actin-enriched (MAP4-UtrCH) spindles. Microtubules are shown in gray, and chromosomes (H2B-mRFP) are shown in magenta. Insets are 5× magnifications of dashed boxes. (B) Representative phalloidin staining of actin in control and actin-enriched spindles presented as overexposed (with visible cytoplasmic network) and nonoverexposed (bottom) images. Actin (phalloidin) is shown in gray, and chromosomes (Hoechst) are shown in magenta. (C) Categories for quantification of individualized or bundled microtubules. (D) Quantification of K-fiber bundling in control or actin-enriched meiosis II spindles, using categories shown in (C). Data are from three independent experiments. Error bars represent SEM. The number of analyzed eggs is specified in italics. Fisher’s exact test was used to test for significance. [(E) and (F)] EGFP fluorescence profile analyses of K-fiber bundling in (E) control and (F) actin-enriched spindles. ImageJ was used to obtain the fluorescence profile of lines (shown in white) that were drawn perpendicularly to the long axes of spindles at the indicated positions. Scale bars, 10 μm. Z-projections, 16 sections every 1.5 μm [(A) and (B)].

Fig. 8 Increasing spindle actin affects microtubule dynamics and leads to lagging and misaligned chromosomes.

(A) Rate of microtubule flux (quantified as shown in fig. S8A) in control (MAP4) or actin-enriched (MAP4-UtrCH) spindles. Data are from three independent experiments. (B) Half-life of fluorescence signal dissipation (quantified as shown in fig. S8A) in control (MAP4) or actin-enriched (MAP4-UtrCH) spindles. Data are from three independent experiments. (C) Distribution of half-life of fluorescence signal dissipation values in control (MAP4) or actin-enriched (MAP4-UtrCH) spindles. Data are from three independent experiments. (D) Representative single-section confocal stills from time-lapse movies of tubulin photoactivation in control (MAP4) or actin-enriched (MAP4-UtrCH) spindles. (E) Representative immunofluorescence images of K-fibers in control or actin-enriched meiosis II spindles. Microtubules (tubulin) are shown in gray, and chromosomes (Hoechst) are shown in magenta. Insets are magnifications of regions marked by dashed-line boxes. Z-projections, 20 sections every 1.5 μm. (F) Quantification of normalized fluorescence intensity of K-fibers in control or actin-enriched spindles. Data are from three independent experiments. (G) Stills from representative time-lapse movies of anaphase I in oocytes expressing MAP4 or MAP4-UtrCH. Microtubules are shown in gray, and chromosomes (H2B-mRFP) are shown in magenta. Z-projections, 16 sections every 1.5 μm. (H) Frequency of mildly lagging chromosomes in oocytes expressing MAP4 or MAP4-UtrCH quantified as in Fig. 2A. Data are from three independent experiments. (I) Frequency of misaligned chromosomes in eggs expressing MAP4 or MAP4-UtrCH quantified as in Fig. 2A. Data are from three independent experiments. (J) Model for the function of actin in protecting eggs against aneuploidy. Bundling of microtubules into functional K-fibers by actin promotes correct alignment of chromosomes in metaphase and their accurate segregation during anaphase. Disruption of actin compromises K-fibers and leads to chromosome alignment and segregation defects and subsequently to aneuploidy. Scale bars, 10 μm. Error bars represent SEM. Box plots are as described in Fig. 2. The number of analyzed oocytes is specified in italics. Two-tailed Student’s t test [(A), (C), and (F)] and Fisher’s exact test [(H) and (I)] were used to test for significance.

Although timing and efficiency of anaphase onset were unaffected (fig. S11, D and E), actin-enriched spindles had gross chromosome alignment and segregation defects (Fig. 8, G to I, and movie S13). These defects could be due to excessive bundling of K-fibers (Fig. 7) and reduced turnover and flux of microtubules in these spindles (Fig. 8, A to D). Although K-fibers need to be stable enough to pull chromosomes apart during anaphase, they also need to be dynamic enough to allow destabilization of incorrect kinetochore-microtubule attachments (3941). Together, our data establish that reducing or increasing the amount of actin in spindles leads to a decrease or increase in K-fibers, respectively, both of which cause chromosome segregation errors in mammalian eggs.

Discussion

Using a combination of high-resolution live-cell imaging and loss-of-function assays, we have revealed an unexpected function of actin in segregating chromosomes in mammalian eggs, a process that is generally thought to be entirely microtubule-dependent. In particular, we found that prominent actin filaments permeate the microtubule spindle in eggs of many mammalian species, including humans. Actin is essential for the accurate alignment and segregation of chromosomes during meiosis (Fig. 8J). Interfering with actin led to a marked increase in lagging chromosomes and a decrease in chromosome speed during anaphase. The lagging chromosomes that were caused by disrupting actin most frequently moved to the correct spindle pole but sometimes remained in the spindle center (Fig. 4B), where they might be partitioned incorrectly by the cytokinetic furrow. Actin was also essential for the accurate alignment of chromosomes during metaphase II. In many eggs, chromosomes remained located at the spindle poles instead of aligning on the metaphase plate. Despite having severe chromosome alignment defects, these eggs still progressed into anaphase, resulting in aneuploidy.

We report that actin promotes the formation of kinetochore fibers. Our data suggest a model in which actin helps to bundle and stabilize K-fiber microtubules in the spindle (Fig. 8J). Decreasing or increasing actin in the spindle leads to a reduction or increase in K-fibers, respectively, precluding the accurate alignment and segregation of chromosomes.

Chromosome alignment in meiosis I was independent of actin (fig. S4B). This is consistent with the observation that spindle actin assembles only shortly before anaphase during meiosis I (14, 15), whereas the alignment of chromosomes is largely completed several hours before anaphase onset (42). The alignment of chromosomes in meiosis I is thus likely to be primarily microtubule- and kinesin-dependent (42).

The often slightly weaker phenotypes in Fmn2−/− versus cytochalasin D–treated oocytes may be due to residual actin filaments in the spindles of Fmn2−/− eggs (Fig. 1C). These may help to form better functional K-fibers than in cytochalasin D–treated eggs, where spindle actin is completely disrupted (Fig. 1B).

That actin filaments are required to form K-fibers was unexpected because K-fibers are generally thought to rely entirely on the cooperative action of microtubule-associated proteins and kinetochore components (3, 4). There is extensive evidence, though, that actin mediates bundling and organization of microtubules in vivo and in vitro, which is consistent with our model (3438). Functionally relevant interactions between actin and microtubules have been reported in various systems to date (43). Although our data strongly suggest that the defects in chromosome alignment and segregation are directly due to disrupting spindle actin, it is also still possible that actin promotes alignment and segregation of chromosomes via additional, less direct mechanisms.

Our data add to the emerging picture that actin has many essential functions during meiosis (44). For example, spindle positioning, a process that is orchestrated by microtubules in many cell types (4547), is primarily actin-dependent in mouse oocytes (14, 15, 17, 48). Actin also mediates the long-range transport of vesicles in mouse oocytes (16) and has been implicated in chromosome segregation in grasshopper spermatocytes (49). In starfish oocytes, where microtubules are too short to capture chromosomes after nuclear envelope breakdown, a meshwork of contractile actin filaments collects the chromosomes from the large nucleus and delivers them to the assembling spindle (50).

The presence of spindle actin in other mammalian eggs, including humans (fig. S1A), hints at a conserved function in chromosome segregation. This surprising function may even extend beyond mammals; F-actin has been reported in spindles of a variety of species and cell types (5, 5053).

Materials and Methods

Preparation and microinjection of oocytes

All mice were maintained in a specific pathogen-free environment according to UK Home Office regulations and the guidelines of the MPI-BPC animal facility. Oocytes were isolated from ovaries of 8-12 week old 129 S6/SvEvTac or Fmn2−/− mice (54), cultured, and microinjected as previously described (55). Final concentrations of 48 nM BSA (Sigma) and Exoenzyme C3 (Merck) were calculated by dividing the total amount of injected protein by the total volume of the oocyte. In some experiments, meiotic progression into anaphase II was induced by activating oocytes with 10 mM SrCl2 in calcium free M2 medium (56).

Preparation and culturing of the human oocyte shown in fig. S1A is described in (57) and was approved by the UK’s National Research Ethics Service under REC reference 11/EE/0346.

Porcine and ovine ovaries were obtained from a local slaughterhouse and rapidly transported to the laboratory at 37°C in M2 medium. Oocytes covered with several layers of cumulus cells were collected from ovaries by aspiration with an 18-gauge needle and cultured in M2 medium at 37°C for 48 hours prior to fixation and processing.

In vitro fertilization (IVF)

B6CBAF1 female mice were superovulated by injection of pregnant mare serum gonadotropin (PMSG) followed 48 hours later by injection of human chorionic gonadotropin (hCG). 14 hours later, meiosis II-arrested eggs were collected from oviducts and IVF was performed in HTF medium (Millipore MR-070-D) as previously described (58). After 2 to 3 hours, zygotes were fixed in anaphase II.

Expression constructs and mRNA synthesis

To mark microtubules, the microtubule binding domain of MAP4 (MAP4-MTBD, 659-1125 aa) was cloned by PCR using primers CAATGTACACCCCGCCAAAC and GTCGACTTAGATGCTTGTCTCC and the full length mouse MAP4 ORF (59) as a template. The EGFP variant of MAP4-MTBD was generated by inserting the corresponding PCR product into pEGFP-C1 (Clontech). MAD1 was cloned by PCR from mouse cDNA using primers CAAGCTTATGGAAGACCTCGGGG and GAATTCCTAGATAGAGGTCTGGCGGC and inserted into pEGFP-C3 vector (Clontech) to generate a fluorescently-tagged variant. SNAP-tagged MAP4-MTBD was cloned by inserting MAP4-MTBD ORF into pSNAPf (NEB). The coding sequence of human CLASP1 (NM_015282.2) was cloned by PCR using primers GGACTCAGATCTCGAGCTCAATGGAGCCTCGCATGGAG and CCGTCGACTGCAGAATTCGATTAGCTGTGCGTGGAGAC and a pEGFP-CLASP1 (laboratory of Helder Maiato) plasmid as a template. Tagged coding sequences were subsequently inserted into pGEMHE (55) for in vitro transcription and linearized with AscI. Expression constructs for labeling chromosomes (H2B-mRFP) (55) and actin (EGFP- UtrCH) in live cells (60) were previously described. EGFP-MAP4-UtrCH was cloned by assembling AgeI-KpnI flanked MAP4-MTBD-2xlinker and KpnI-SalI flanked 2xlinker- UtrCH into the AgeI-SalI site of pGEMHE thereby reconstituting a 4xlinker repetitive nucleotide sequence between MAP4 and UtrCH, each linker encoding amino acids Gly-Gly-Ser. With the exception of Fig. 8, E and F where it was expressed at similar levels to MAP4, EGFP-MAP4-UtrCH in spindle actin enrichment assays was expressed at much lower levels than EGFP-MAP4 (fig. S11, F and G).

Capped mRNA was synthesized using T7 polymerase (mMessage mMachine kit, following the manufacturer’s instructions, Ambion). mRNA concentrations were determined on agarose gels by comparison with an RNA standard (Ambion).

Confocal and superresolution microscopy

Images were acquired with Zeiss LSM710, LSM780, LSM800 and LSM880 microscopes at 37°C. Oocytes were imaged in M2 medium with or without cytoskeletal inhibitors under mineral oil using a 40x C-Apochromat 1.2 NA water-immersion objective. Superresolution time-lapse images were acquired using the Airyscan module on Zeiss LSM800 and LSM880 microscopes and processed post-acquisition using ZEN2. For chromosome alignment and segregation analyses, images were typically acquired at a temporal resolution of 5-6 min and optical slice thickness of 1.5 μm confocal sections covering ~20 μm.

Manual and automated quantification of misaligned and lagging chromosomes

For analyses of chromosome misalignment in acute drug addition experiments, only chromosomes that became misaligned within two hours after drug treatment were quantified. For analyses of lagging chromosomes, only meiotic spindles that were parallel to the imaging plane were considered. To manually quantify segregation errors, we classified chromosomes that failed to clear the central spindle region within 12 or 18 min of anaphase I onset (time 0 refers to last metaphase frame before anaphase onset) as lagging or severely lagging, respectively (Fig. 2A). Chromosomes that lagged severely were also included in the quantification of mildly lagging chromosomes. Quantification of severely lagging chromosomes exclusively contained those chromosomes that lagged for at least 18 min after anaphase onset. Similarly, quantification of mildly misaligned chromosomes as shown in Fig. 2A included chromosomes that were also severely misaligned. To automate the quantification of lagging chromosomes, we reconstructed three-dimensional isosurfaces of fluorescently-labeled chromosomes in Imaris (Bitplane). We then assessed if surfaces other than the two main chromosome masses were detected by the software upon anaphase onset. Surfaces that were detected separately from the two main chromosome masses 12 min after the onset of anaphase were scored as lagging. For quantification of chromosome segregation speeds, both lagging and non-lagging chromosomes were analyzed. All quantitative analyses of chromosome alignment defects in meiosis II were performed at least four hours after completion of anaphase I.

Photoactivation experiments

For analyses of microtubule flux and turnover, meiotic spindles that were parallel to the focal plane were selected using H2B-mRFP-labeled chromosomes in oocytes co-expressing photoactivatable-GFP (PAGFP)-Tubulin. Line segments were then used to mark regions of interest at least two-thirds or a half-spindle length away from chromosomes. Regions of interest were photoactivated using the 488 nm laser line and images were acquired at an interval of 3 s for 2 min. Photobleaching not due to microtubule turnover was corrected by monitoring bleaching of a non-activated region outside the spindle. Microtubule flux rates were measured in ImageJ by following the position of intensity maxima in the photoactivated region over time. Microtubule turnover rates were obtained by plotting time versus maximum intensity plots in OriginPro (OriginLab) and fitting single exponential curves to plots, thereby determining the half-life of fluorescence decay. Single exponential fitting was used because it produced curves with better goodness-of-fit than double exponential fitting.

Immunofluorescence microscopy

Mouse, human, porcine and ovine oocytes were fixed for 25 to 30 min at 37°C with 100 mM HEPES, 50 mM EGTA, 10 mM MgSO4, 2% formaldehyde, and 0.2% Triton X-100 and extracted in PBS supplemented with 0.1% Triton X-100 at 4°C overnight. Antibody, F-actin and chromosome staining were performed for 1-2 hours in PBS, 3% BSA, and 0.1% Triton X-100. Microtubules were stained using primary rat anti-α-tubulin (MCA78S, Serotec; 1:200) and Alexa-Fluor-546- or Alexa-Fluor-647-labeled secondary anti-rat (Molecular Probes 1:200) antibodies. F-actin was stained with Rhodamine-phalloidin (Molecular Probes; 1:20). DNA was stained with 5 μg/ml Hoechst 33342 (Molecular Probes).

Images were acquired with Zeiss LSM710, LSM780, LSM800 and LSM880 confocal microscopes equipped with a 63x C-Apochromat 1.2 NA water-immersion objective as previously described (55). Images in control and perturbed situations were acquired with identical imaging conditions. Care was taken that images were not saturated during acquisition.

Spindle volume measurements of fixed oocytes were performed in three dimensions using the Isosurface function in Imaris (Bitplane).

Drug addition experiments

To disrupt actin, oocytes were treated with cytochalasin D (Calbiochem) at a final concentration of 5 μg/ml. To stabilize actin filaments, oocytes were treated with 5 μM SiR-Actin (Spirochrome) immediately before imaging. To acutely depolymerize microtubules, eggs were treated with 5 μM Nocodazole (Sigma) immediately before imaging. In spindle actin reassembly assays, mouse oocytes expressing fluorescently labeled actin, microtubules and chromosomes were allowed to mature to meiosis II in the presence of cytochalasin D and washed out into cytochalasin D free M2 medium immediately before imaging. In acute drug addition experiments, oocytes matured in M2 medium were washed into cytochalasin D containing M2 medium immediately before imaging. For MII activation of acutely drug-treated eggs, oocytes were incubated with cytochalasin D for two hours before activation. For chromosome alignment analyses in meiosis II, acute drug additions were performed more than four hours after completion of anaphase I. To inhibit vesicle biogenesis, oocytes were treated with 10 μM BFA (Brefeldin A, Sigma).

Cold-mediated microtubule depolymerization assays

To determine K-fiber stability, a previously described assay for specific depolymerization of non-kinetochore associated microtubules was adapted (61). Briefly, non-kinetochore microtubules were depolymerized by placing oocyte containing culture dishes at 4°C for 15 min. Cells were immediately fixed following cold treatment and processed for immunofluorescence microscopy as described above. K-fiber fluorescence intensities were measured from three-dimensional isosurface reconstructions of spindles using Imaris (Bitplane).

Statistics

Average (mean), standard error of the mean, standard deviation and statistical significance based on two-tailed Student’s t test or Fisher’s exact test were calculated in OriginPro (OriginLab). P values are designated as *P < 0.05, **P < 0.005, and ***P < 0.0005. Nonsignificant values are indicated as N.S.

Supplementary Materials

www.sciencemag.org/content/357/6353/eaal1647/suppl/DC1

Materials and Methods

Figs. S1 to S11

Movies S1 to S13

References and Notes

  1. Acknowledgments: We thank the staff of the Medical Research Council Laboratory of Molecular Biology’s Animal, Genotyping and Microscopy Facilities for technical assistance, and members of the Schuh laboratory for discussions. We thank K. Scheffler for help with in vitro fertilization and mouse embryo imaging experiments. We are thankful to K. Elder and M. Blayney from Bourn Hall Clinic, Cambridge, for providing the human oocyte shown in fig. S1A. We thank H. Maiato for generously providing a plasmid-encoding human CLASP1. Research leading to these results has received financial support from the European Research Council under grant agreement no. 337415. Plasmids are available from M.S. under a materials transfer agreement with the Max Planck Society. B.M. and M.S. conceived the study and designed the experiments and methods for data analysis; B.M. performed all experiments, analyzed the data, and prepared figures; B.M. and M.S. wrote the manuscript; and M.S. supervised the study. The authors declare no competing financial interests.
View Abstract

Navigate This Article