Constant or variable number of chiasmata during recombination?

Constant or variable number of chiasmata during recombination?

We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

During recombination, is the number of chiasmata consistent for each gamete and are the chiasmata regions consistent within a single organism?

There is a large amount of variation in the distribution and number of chiasmata or crossing overs (COs). The total number of chiasmata per cell can vary within the same organism. Where chiasmata occur along the chromosome is also not consistent cell to cell. Hotspots are 1-2kb regions that experience elevated recombination compared to neighboring genomic regions. A way to think of the distribution of COs is that ~80% of recombination happens in hotspots which comprise ~10% of the genome.

Recombination depends on the chromatin context, so chiasmata are rare near the centromere and positive interference results in chiasmata being spaced out along the chromosome. A good prediction of total chiasmata number per cell is at least 1 per chromosome. Some researchers argue that one CO per chromosome arm is more accurate.

Below is a link to a good article that expands on some of the ways recombination varies in humans.

Chiasmata and the kinetochore component Dam1 are crucial for elimination of erroneous chromosome attachments and centromere oscillation at meiosis I

Establishment of proper chromosome attachments to the spindle requires elimination of erroneous attachments, but the mechanism of this process is not fully understood. During meiosis I, sister chromatids attach to the same spindle pole (mono-oriented attachment), whereas homologous chromosomes attach to opposite poles (bi-oriented attachment), resulting in homologous chromosome segregation. Here, we show that chiasmata that link homologous chromosomes and kinetochore component Dam1 are crucial for elimination of erroneous attachments and oscillation of centromeres between the spindle poles at meiosis I in fission yeast. In chiasma-forming cells, Mad2 and Aurora B kinase, which provides time for attachment correction and destabilizes erroneous attachments, respectively, caused elimination of bi-oriented attachments of sister chromatids, whereas in chiasma-lacking cells, they caused elimination of mono-oriented attachments. In chiasma-forming cells, in addition, homologous centromere oscillation was coordinated. Furthermore, Dam1 contributed to attachment elimination in both chiasma-forming and chiasma-lacking cells, and drove centromere oscillation. These results demonstrate that chiasmata alter attachment correction patterns by enabling error correction factors to eliminate bi-oriented attachment of sister chromatids, and suggest that Dam1 induces elimination of erroneous attachments. The coincidental contribution of chiasmata and Dam1 to centromere oscillation also suggests a potential link between centromere oscillation and attachment elimination.

1. Introduction

Faithful segregation of chromosomes during cell division is essential for genome integrity. During division of a somatic cell (mitosis), duplicated chromosomes (sister chromatids) attach to opposite spindle poles (bi-oriented attachment) via microtubules (MTs) and segregate from each other (equational segregation). Such a division generates two genetically identical daughter cells (figure 1a, mitosis) [2]. By contrast, during the first of two consecutive divisions of a germ cell (meiosis), homologous chromosomes attach to opposite poles and segregate, leading to generation of gametes containing half the original number of chromosomes (figure 1a, meiosis I).

Figure 1. The effects of error correction impairment on sister chromatid segregation at meiosis I. (a) Organization, spindle attachment and segregation of chromosomes in mitosis and meiosis I are shown. (b) Hypothesis regarding the role of the error correction mechanism in spindle attachment of sister chromatids in the presence and absence of chiasmata. Dotted circles, attachments corrected by the error correction mechanism. (c) Segregation patterns of GFP-visualized centromeres of sister chromatids in zygotes and pat1 haploid cells. (d,e) Equational segregation frequencies at meiosis I in rec12 + (d) and rec12Δ cells (e). (f) Equational segregation of chromosome 2 at meiosis I in pat1 haploid cells. The pat1 haploid cells were forced to enter meiosis by activation of the mating pheromone signalling pathway and subsequent inactivation of Pat1 kinase at the restrictive temperature, both of which are required for induction of reductional-like segregation of sister chromatids in haploid cells [1]. In (df), chromosome segregation was analysed in greater than 50 cells, and bars show averages of more than three independent experiments. cen1, chromosome 1 cen2, chromosome 2 +, no other mutations. Dotted lines in (d,e) distinguish results for cen1 and cen2. Error bars show standard errors. *p < 0.05 **p < 0.005 (Student's t-test). Statistical analysis of error correction-proficient (+) and error correction-defective (mad2Δ and ark1-so) sgo1Δ cells is shown.

Proper attachments are established through a dynamic process [3–6]. Chromosomes attach to the spindle randomly via kinetochores, which are protein complexes that assemble at centromeres [7–9]. Kinetochores initially attach to spindle MTs at their lateral side, and subsequently at their ends. Recent studies showed that Ndc80 and the Dam1/DASH complexes (Ska complex in vertebrates) form end-on attachments and couple MT disassembly to centromere movement. Because of the random nature of initial attachment, sister chromatids or homologous chromosomes often erroneously attach to the same pole (mono-oriented attachment) in mitosis or meiosis I, respectively. These erroneous mono-oriented attachments are unstable and are ultimately eliminated the chromosomes then repeat the attachment and detachment processes until they attach to the spindle properly. In addition, during establishment of attachment, centromeres continuously move back and forth between the spindle poles by forces generated by MT disassembly at the kinetochores [4–6,10]. However, it remains unclear how erroneous attachments are selectively eliminated during oscillation of centromeres.

It is widely believed that proper attachment is established through tension [11,12]. During mitosis, bi-oriented sister chromatids are pulled towards opposite poles, and the pulling forces generate tension because the sisters are held together by a protein complex called cohesin [13]. In the current model, this tension causes stabilization of bi-oriented attachment, whereas tensionless mono-oriented attachment is unstable and subject to elimination. Likewise, at meiosis I, because the homologous chromosomes are linked by recombination products called chiasmata, their bi-oriented attachment generates tension, which is thought to stabilize the attachments [2,11,12].

Several factors have been identified as components of the mechanism that corrects erroneous attachments. The spindle assembly checkpoint (SAC) factors inhibit the onset of anaphase, providing time for correction of erroneous attachments (e.g. [14]), and Aurora B kinase destabilizes erroneous attachments by phosphorylating kinetochore components including Ndc80 and Dam1 [3,15–17]. Consistently, impairment of the SAC pathway or Aurora B kinase increases the frequency of erroneous attachments in both mitosis and meiosis I [18–26]. The enrichment of Aurora B kinase at the inner centromere region, beneath the outer kinetochore, suggests that tension-dependent spatial separation of the outer kinetochores from the inner centromere impedes Aurora-dependent outer kinetochore phosphorylation, thereby stabilizing the attachments [3,19,27].

Although tension-dependent stabilization is currently the widely accepted model for attachment selection, it cannot easily account for attachment selection at meiosis. At meiosis I, homologous chromosomes are attached to opposite poles, while sister chromatids are attached to the same pole. Mono-oriented attachment of sister chromatids is also crucial for homologous chromosome segregation. In the fission yeast Schizosaccharomyces pombe, sister centromeres are often bi-oriented and undergo transient splitting [28]. Although these attachments are supposed to generate tension, they are eventually eliminated. In addition, in mouse oocytes, the establishment of correct attachments does not correlate with tension generated at the kinetochores moreover, tension does not seem to cause spatial separation of the kinetochores from the Aurora-enriched region [20]. Thus, attachment selection at meiosis I is still enigmatic.

At meiosis I, chiasmata are crucial for bi-oriented attachment of homologous chromosomes. However, we previously reported that in S. pombe, chiasmata also contribute to mono-oriented attachment of sister chromatids by preventing their bi-oriented attachment [28]. It remains unclear how chiasmata prevent bi-oriented attachment of sister chromatids. Here, we examined chiasma-dependent prevention of bi-oriented attachment in greater detail and found that chiasmata enabled error correction factors to eliminate bi-oriented attachment of sister chromatids and coordinated homologous centromere oscillation. In addition, we found that the kinetochore protein Dam1 contributed to attachment correction and was essential for centromere oscillation. These findings demonstrate the importance of chromosome organization and kinetochore activity in elimination of erroneous attachments and reveal a potential link between centromere oscillation and attachment elimination.

2. Results

2.1. Chiasmata change the effects of the error correction mechanism on sister chromatid segregation

We previously reported that sister chromatids frequently attach to opposite spindle poles and that chiasmata prevent these attachments [28]. We hypothesized that chiasmata change the attachment types that are eliminated by the error correction mechanism such that, in the presence of chiasmata, the mechanism selectively eliminates bi-oriented attachment of sister chromatids, thereby retaining mono-oriented attachment in contrast with the situation in mitosis (figure 1b, with chiasmata). If so, impairment of the error correction mechanism should increase the frequency of bi-oriented attachment in chiasma-forming cells due to defective elimination. Conversely, in cells lacking chiasmata, the error correction mechanism would selectively eliminate mono-oriented attachment of sister chromatids, as in mitosis (figure 1b, without chiasmata), and error correction impairment should increase their mono-oriented attachment.

To test this idea, we compromised the error correction mechanism and evaluated spindle attachments of sister chromatids in chiasma-forming and chiasma-lacking cells. To compromise the error correction mechanism, we introduced the mad2Δ mutation or the ark1-so mutation, which represses meiotic expression of the ark1 gene that encodes fission yeast Aurora B kinase [25]. The mad2Δ mutation is thought to reduce the time available for correcting improper attachments by compromising the SAC pathway, whereas the ark1-so mutation impairs attachment elimination and prevents SAC activation. To evaluate attachments, we examined sister chromatid segregation by visualizing the centromere-proximal locus of one of the homologues of chromosome 1 (cen1) or 2 (cen2) using the lacI/lacO recognition system (figure 1c) [1,29]. In this analysis, we deleted the sgo1 gene, which encodes one of two Shugoshin proteins of fission yeast that specifically functions as a centromere cohesion protector during meiosis and prevents separation of sister chromatids attached to opposite poles [28,30–33].

Segregation of sister chromatids to the opposite poles (equational segregation) was very rare in all types of cells in the sgo1 + background (figure 1d), confirming that Sgo1 prevents separation of sister chromatids [28,31,32]. By contrast, in the sgo1Δ background, although equational segregation was rare in chiasma-forming cells (figure 1d), it was very frequent in diploid rec12Δ cells [28], which do not form chiasmata because of the lack of Rec12 (an Spo11 homologue in fission yeast), a factor required for the formation of DNA double-strand breaks (figure 1e, +) [34]. Frequent equational segregation was not caused by a loss of Rec12 functions, as demonstrated by the observation that equational segregation was also frequent in haploid meiotic cells, which do not form chiasmata (figure 1c and f, +) [28]. Importantly, in the sgo1Δ background, introduction of the mad2Δ or ark1-so mutation significantly increased the percentages of equational segregation in chiasma-forming diploid cells (figure 1d), but decreased the segregation percentages in chiasma-lacking diploid rec12Δ (figure 1e) or haploid meiotic cells (figure 1f). These results indicate that the error correction mechanism decreases bi-oriented attachment of sister chromatids in the presence of chiasmata, but conversely increases bi-oriented attachment (thereby decreasing mono-oriented attachment) in the absence of chiasmata. In our previous study, the equational segregation frequencies of cen1 were somewhat higher in the rec12Δ background [28], although the reason for this is unknown. However, the mad2Δ mutation similarly decreased equational segregation, being consistent with our current results.

2.2. Chiasmata change the effects of the error correction mechanism on metaphase sister centromere splitting

We next sought to assess bi-oriented attachment of sister chromatids in a more direct manner by examining transient splitting of sister centromeres, a probable outcome of their bi-oriented attachment, which often occurs irrespective of chiasma formation [28]. To precisely evaluate sister centromere splitting during metaphase, we specified metaphase stage by visualizing S. pombe securin Cut2, a biochemical marker of the pre-anaphase stage that localizes at the prometa/metaphase spindle (figure 2a) [35]. In addition, to follow fine centromere dynamics, we determined three-dimensional positions of cen2 on the spindle at high temporal resolution by acquiring images of cen2 and the spindle pole body (SPB the fungal centrosome) in several different focal planes every approximately 3 s (figure 2b). We visualized the SPB using a GFP-tagged SPB component, Sid4 (Sid4-GFP) [36], but left one copy of Sid4 untagged in diploid cells because in our live cell analyses, cells containing two copies of GFP-tagged Sid4 exhibited high frequencies of equational segregation of sister chromatids even in the sgo1 + background (data not shown), which probably resulted from impairment of the function of Sid4 in the regulation of anaphase exit [37,38]. Leaving one copy of untagged Sid4 largely eliminated the segregation abnormalities, as shown by sister chromatid segregation patterns in cells containing one copy of untagged Sid4, which were similar to those observed in cells lacking Sid4 tagging (electronic supplementary material, figure S1).

Figure 2. Centromere and SPB dynamics at metaphase I. (a) Localization of GFP-visualized centromeres and SPBs and RFP-visualized Cut2 at metaphase I (left), and an enlarged image of centromeres and the SPBs (right), shown by the white box in the left image. Bars, 2 µm. (b) Centromere and SPB dynamics at metaphase I in wild-type cells. Maximum projection images of centromeres and SPBs taken every approximately 3 s were arranged from left to right. Arrowheads indicate splitting sister centromeres. Bar, 2 µm. (c) Changes in centromere and SPB positions at metaphase I in various cells. Left images show kymograph of centromeres (cen2) and SPBs and right graphs show changes in centromere–SPB and SPB–SPB distances. Vertical and horizontal bars of kymographs show 2 µm and 30 s, respectively. Arrows indicate co-localization of centromeres with the SPB. (d) Mean observation frequencies of centromere splitting per centromere obtained by the bootstrap method. Graph shows the mean frequencies per centromere. The dotted line distinguishes the results of rec12 + and rec12Δ cells. Numbers in parentheses at the top of the graph indicate the number of centromeres examined. Error bars indicate 95% confidence intervals. *p < 0.05 **p < 0.01 (Studentized bootstrap analysis). n.s., not significant. p-values > 0.05 and smaller than 0.1 are shown in parentheses.

During metaphase I, centromeres oscillated between the spindle poles in both chiasma-forming wild-type and chiasma-lacking rec12Δ diploid cells, as previously reported (figure 2c electronic supplementary material, movies S1 and S2) [28]. In both types of cells, sister centromeres often underwent transient splitting, confirming our previous finding that bi-oriented attachment of sister chromatids occurs irrespective of chiasma formation (figure 2b,c) [28]. The frequencies of sister centromere splitting varied (electronic supplementary material, figure S2A), but the mean splitting frequencies per centromere obtained by the bootstrap method were significantly higher in rec12Δ cells than in rec12 + cells (the difference was also significant in the usual t-test, p < 0.01 figure 2d, +). In our previous report, the mean splitting frequencies were not significantly different [28] this discrepancy may be due to the presence of an untagged copy of Sid4 in this study. The elevated frequency of sister centromere splitting in chiasma-lacking cells confirms that chiasmata prevent bi-oriented attachment of sister chromatids.

In ark1-so rec12 + cells, the mean splitting frequency was significantly higher than in rec12 + cells, supporting the notion that bi-oriented attachment of sister chromatids was increased in these cells (figure 2d). In addition, although the differences were not significant in other mad2Δ or ark1-so cells, the mean splitting frequencies tended to be higher in the rec12 + background and lower in the rec12Δ background (figure 2d electronic supplementary material, movies S3–S6). These tendencies were consistent with the idea that when the error correction mechanism is impaired, bi-oriented attachment of sister chromatids increases in chiasma-forming cells, but conversely decreases in chiasma-lacking cells. The mechanism driving centromere movement is intact in these mutant cells, as demonstrated by the lack of significant differences in the mean centromere velocities, the mean standard deviations of centromere positions and the plots of the mean square displacement (MSD) of centromeres (electronic supplementary material, figure S2B–D).

Additional support for a higher frequency of bi-oriented attachment of sister chromatids in error correction-defective, chiasma-forming cells came from an analysis of homologous centromere positions (figure 3a). In wild-type cells, homologous centromeres were positioned mostly nearly parallel to the spindle axis (at angles less than 10° relative to the spindle axis), whereas in mad2Δ or ark1-so cells, their positions were often tilted and non-parallel (greater than or equal to 10° figure 3b). In addition, homologous centromeres tended to be positioned closer, as demonstrated by a reduction in the mean homologous centromere distances (figure 3b,c). Notably, homologous centromere colocalization was observed during 9.2% of total observation time in ark1-so cells but was never observed in wild-type cells (figure 3d electronic supplementary material, table S1). These phenotypes indicate a reduction in opposing poleward forces exerted on the homologous centromeres, which likely reflects an increase in the frequency of bi-oriented attachment of sister chromatids. In support of this view, the homologous centromeres occasionally switched their positions relative to the SPB by oscillating independently of each other in ark1-so cells (6 switches during 1643 s observation time) (figure 3d electronic supplementary material, table S1), indicating direct exertion of opposite forces on each of the homologous centromeres.

Figure 3. Homologous centromere angles and distances and splitting sister centromere angles. (a) Angle of homologous centromeres relative to the spindle and the distance between them. Photo shows the representative image of tilted homologous centromeres. White and red dotted lines in photo show the axes of the spindle and homologous centromeres, respectively. (b) Distribution of homologous centromere angles relative to the spindle (upper graphs) and 2D plots of the mean angles and distances of homologous centromeres in each cell (lower graphs). (c) Means of the mean distances between homologous centromeres per cell. Bars show means of the mean distances in each cell. Drawings show the predicted relationship between attachments and homologous centromere distance. Black arrows indicate forces exerted on chromosomes. *p < 0.05 (Student's t-test). n.s., not significant (number in parenthesis shows p-value). Numbers in parentheses at the top of the graph show the numbers of cells examined. Error bars indicate standard errors. (d) Colocalization and position switching of homologous centromeres in ark1-so cells. Photos show kymographs of homologous centromeres (cen2) and the SPBs. The red dotted square in the left photo indicates colocalization of centromeres. Red arrows in the right photo show switching of homologous centromere positions relative to the SPBs. Horizontal bar: 30 s vertical bar: 1 µm. Note that SPB#2 was out of focus around the end of observation in the right photo. (e) Angle of splitting sister centromeres and the distance between them. Image shows splitting sister centromeres (arrowheads) and the SPBs. Dotted lines indicate axes of the spindle (white) and the splitting centromeres (yellow and red). Bar, 1 µm. Upper drawing demonstrates the angle of the splitting centromeres to the spindle axis, ‘θ’. Lower drawing demonstrates predicted MT attachments of tilted splitting sister centromeres. (f) Distribution of angles of splitting sister centromeres. (g) Increased tilting of splitting sister centromeres in ark1-so cells. Green bars: 0° ≤ θ < 20° brown bars: 20° ≤ θ < 60° blue bars: 60° ≤ θ ≤ 90°. +, no other mutations. Sixty, 58, 132, 169, 22 and 109 splitting sister centromeres were examined in wild-type, mad2Δ, ark1-so, rec12Δ, mad2Δ rec12Δ and ark1-so rec12Δ cells, respectively.

In addition to these findings, analysis of the positions of splitting sister centromeres raised the possibility that in addition to sister chromatids, a single chromatid is often attached to both poles. We noted that splitting sister centromeres were frequently tilted in a variety of directions, sometimes almost perpendicular to the spindle axis in both rec12 + and rec12Δ cells (figure 3e–g electronic supplementary material, figure S3), suggesting that spindle attachments of these centromeres are frequently different from simple bi-oriented attachment. Introduction of the ark1-so mutation increased the proportion of splitting sister centromeres that tilted towards a perpendicular angle relative to the spindle axis in both rec12 + and rec12Δ cells (figure 3f,g electronic supplementary material, figure S3), suggesting that the error correction mechanism contributes to elimination of these attachments irrespective of chiasma formation. Given the correlation between tilted positioning of homologous centromeres and bi-oriented attachment of sister centromeres, the tilted positions of splitting sister centromeres probably reflected bi-oriented attachments of a single centromere (merotelic attachment) together with bi-oriented attachment of sister centromeres (figure 3e, bottom drawing).

Considering all these results together, we concluded that chiasmata alter attachment correction patterns by enabling error correction factors to eliminate bi-oriented attachment of sister chromatids.

2.3. Chiasmata coordinate homologous centromere oscillation

Because tensionless attachments are thought to be eliminated by the error correction factors, we sought to determine whether chiasmata decrease tension across sister centromeres by measuring inter-sister centromere distances projected on the spindle axis, which probably reflect tension generated by poleward pulling forces (electronic supplementary material, figures S3 and S4). The projected distances in wild-type cells revealed two distinct distribution groups that covered distinct distance ranges, whereas the distances in rec12Δ cells revealed a single distribution group (electronic supplementary material, figure S4C). The distribution group that covered the longer distance range in wild-type cells mostly originated from a dataset of a single cell shown in figure 2 (electronic supplementary material, figure S4C, WT w/o). Irrespective of the presence or the absence of this particular dataset, however, the mean projected distance in wild-type cells did not significantly differ from that of rec12Δ cells (electronic supplementary material, figure S4D). Therefore, we could not draw a conclusion about whether the tension across sister centromeres is changed by chiasmata.

Since tension changes could not be assessed in our analyses, we next examined centromere dynamics in more detail to obtain a clue about chiasma-dependent elimination of bi-oriented attachment of sister chromatids. We noted that in chiasma-forming wild-type cells, homologous centromeres tended to move in the same direction and reverse movements in a coordinated manner, whereas this tendency was not apparent in chiasma-lacking rec12Δ cells. Indeed, homologous centromeres moved in the same direction during 66.2% and 46.7% of the observation time in wild-type and rec12Δ cells, respectively (figure 4a) the latter percentage is close to that of random movement. In addition, in ark1-so cells, the centromeres often reached the SPB without reversing their movements and co-localized with it (co-localization was observed during 6.4% of observation time in ark1-so cells, but never in wild-type cells figure 2c, ark1-so, arrows), indicating that centromere reversal is affected by decreased Aurora B activities. These observations suggest a relationship between centromere oscillation and attachment correction.

Figure 4. Relationship between homologous centromere movements and correlation analysis of centromere dynamics at metaphase I. (a) Movement direction of homologous centromeres. Blue and yellow bars indicate proportions of a pair of homologous centromeres moving in the same and opposite directions, respectively. (b) Auto-correlation (AC) and cross-correlation (CC) of centromere velocities. Relationship between correlation values (AC or CC) and centromere movements is shown. Arrows above chromosomes show the direction of centromere movements. (c) ACF and CCF of centromere velocities in wild-type and rec12Δ cells. Blue lines and red dotted lines, respectively, indicate the mean correlation values and 95% confidence intervals obtained by the bootstrap method. (d) p-values of ACF and CCF values at the indicated time points. (e) p-values between CCF values of wild-type and rec12Δ cells at the indicated time points. Asterisks in (d,e) indicate significant differences. The numbers of examined cells and total observation times (shown in parentheses) are the following. Wild-type: 14 cells (1326.8 s) rec12Δ: 11 cells (1193.4 s).

To confirm chiasma-dependent coordination of homologous centromere oscillation, we characterized centromere movements in a quantitative manner using the correlation function, which is useful for characterizing centromere oscillation (e.g. [39–41]). Specifically, we collected centromere velocities relative to the SPB obtained from all observed cells and calculated the correlation coefficients between velocities separated by various lag times (figure 4b). The auto-correlation plot of wild-type cells decreased and reached a negative value albeit small, the negative value was statistically significant around 35 s (figure 4c,d, WT). This indicates that the centromere movements tend to reverse around 35 s in chiasma-forming wild-type cells (figure 4b, upper right). The auto-correlation plots of rec12Δ cells also exhibited a similar pattern, indicating that centromeres oscillate in chiasma-lacking rec12Δ cells in a similar manner (figure 4c,d, rec12Δ). Thus, centromeres oscillate between the two SPBs with a weak but significant periodicity in a chiasma-independent manner.

We next examined the correlation between the velocities of homologous centromeres by cross-correlation analysis. The cross-correlation plot of wild-type cells increased from a negative value and reached a peak at a positive value around 35 s (figure 4c), and both the negative and the positive values were statistically significant (figure 4d, WT). This shows that homologous centromeres tend to move in the same direction (figure 4b, lower left) and reverse their movements every approximately 35 s (figure 4b, lower right). Thus, homologous centromere oscillation is coordinated. By contrast, the plot of rec12Δ cells differed significantly from the wild-type plot and exhibited no significant correlation at any lag times (figure 4ce, rec12Δ). This result indicates a lack of coordination in homologous centromere movements. These results demonstrate that chiasmata coordinate homologous centromere oscillation.

We also characterized centromere movements in ark1-so cells. The cross-correlation value at 0 s was negative, as in wild-type cells, indicating that homologous centromeres tended to move in the same direction (electronic supplementary material, figure S5A,B). However, the statistical significance of the correlation values of the auto- and cross-correlation plots was lost at around 35 s (electronic supplementary material, figure S5A–C). This observation suggests weakened periodicity of centromere oscillation and is consistent with altered centromere reversal in ark1-so cells. By contrast, no such alteration was observed in mad2Δ cells, whose major defect is in the SAC-dependent inhibition of precocious anaphase onset (electronic supplementary material, figure S5D–F). The facts that chiasmata coordinate homologous centromere oscillation and that oscillation dynamics are altered in ark1-so cells raise the possibility that centromere oscillation is linked to attachment elimination.

2.4. Dam1 is required for chiasma-dependent attachment correction

To further explore the relationship between attachment elimination and centromere oscillation, we next searched for kinetochore mutants defective in chiasma-dependent attachment elimination. Among kinetochore proteins, we suspected that Dam1 might contribute to attachment correction and centromere oscillation because Dam1 is an Aurora B substrate and a component of the Dam1/DASH complex that is thought to drive chromosome segregation by coupling MT shortening with kinetochore movements [42–50] we found that chiasma-dependent attachment elimination was impaired in cells bearing the dam1 deletion (dam1Δ).

dam1Δ cells exhibited largely normal spore formation and only a slight decrease in spore viability (electronic supplementary material, figure S6A,B). This observation indicates that meiotic chromosome segregation is not severely impaired in dam1Δ cells, being consistent with the fact that Dam1 is not essential for chromosome segregation in mitosis [51]. In addition, both crossover and non-crossover recombination occurred at a wild-type level in dam1Δ cells (electronic supplementary material, figure S6c), indicating that recombination-dependent chiasma formation is normal in dam1Δ cells.

Importantly, the dam1Δ mutation impaired disjunction of homologous chromosomes (figure 5a), as seen in mad2Δ and ark1-so mutants [25,52]. Furthermore, the dam1Δ mutation increased equational segregation of sister chromatids in rec12 + cells (figure 5b, left) but it decreased equational segregation in sgo1Δ rec12Δ or haploid meiotic sgo1Δ cells (figure 5b,c). Consistent with the equational segregation frequencies, the mean splitting frequencies of sister centromeres decreased in dam1Δ rec12Δ cells, and although the difference was not significant, the mean splitting frequencies tended to be higher in dam1Δ cells (figure 5d,e electronic supplementary material, figure S2A). In addition, opposing forces exerted on homologous centromeres decreased, as demonstrated by the increase in tilted, non-parallel positioning of homologous centromeres and the reduction in their distances (figure 5f,g). These phenotypes mirror those of error correction-defective mad2Δ and ark1-so mutants (figure 3b,c). Furthermore, as seen in ark1-so cells, the proportion of splitting sister centromeres that were tilted towards a perpendicular direction increased irrespective of chiasma formation (figure 5h,i electronic supplementary material, figure S3). These similarities to error correction-defective mutants strongly suggest that attachment correction is defective in dam1Δ cells. The dam1Δ mutation significantly shortened the mean projected distances (electronic supplementary material, figure S4C,D), suggesting a reduction in opposing forces exerted on sister centromeres.

Figure 5. Effects of the dam1Δ mutation on chromosome attachment to the spindle at meiosis I. (a) Frequencies of non-disjunction of homologous chromosomes (cen2). (b,c) Equational segregation frequencies of sister chromatids (cen2) in diploid (b) and pat1 haploid cells (c) at meiosis I. In (a–c), chromosome segregation was analysed in more than 50 cells, and bars show averages of more than three independent experiments. Error bars show standard errors. *p < 0.01 **p < 0.005 ***p < 0.0001 (Student's t-test). (d) Separated sister centromeres in dam1Δ cells. Maximum projection images of centromeres and SPBs taken every approximately 3 s were arranged from left to right. Arrowheads show separated sister centromeres. Bar, 2 µm. (e) Observation frequencies of centromere splitting obtained by the bootstrap method. Graph shows the mean frequencies per centromere. *p < 0.05 **p < 0.01 n.s., not significant (p-value is shown in parentheses). Error bars show 95% confidence intervals. Numbers in parentheses at the top of the graph indicate the numbers of centromeres examined. +, no other mutations. Means of wild-type and rec12Δ cells were adopted from figure 2. Dotted line distinguishes results of rec12 + and rec12Δ cells. (f) Distribution of angles of homologous centromeres relative to the spindle (upper graph), and 2D plots of the mean angles and distances of homologous centromeres in each cell (lower graph). (g) Distances between homologous centromeres. Graph shows means of the mean distances in each cell. Drawings show the predicted relationship between attachments and homologous centromere distances. Black arrows indicate forces exerted on chromosomes. Numbers in parentheses indicate the number of cells examined. *p < 0.001 (Student's t-test). Error bars indicate standard errors. (h) Distribution of angles of splitting sister centromeres. Dotted red lines show the distributions of wild-type and rec12Δ cells. (i) Increased tilting of splitting sister centromeres in dam1Δ cells. Green bars: 0° ≤ θ < 20° brown bars: 20̊ ≤ θ < 60° blue bars: 60° ≤ θ ≤ 90°. +, no other mutations. Eighty-three and 56 splitting sister centromeres were examined for dam1Δ and dam1Δ rec12Δ cells, respectively.

The attachment changes did not result from premature anaphase onset or impaired Aurora B localization, as dam1Δ cells exhibited delayed anaphase onset (metaphase duration: 11.4 ± 0.7 min in wild-type cells (n = 10) and 31.5 ± 2.3 min in dam1Δ cells (n = 6)) and Aurora B localization patterns similar to those seen in wild-type cells (pre-anaphase punctate localization adjacent to kinetochores and anaphase spindle localization electronic supplementary material, figure S7A,B).

2.5. Dam1 is required for metaphase centromere oscillation and anaphase A

We also found that centromere oscillation was abolished in dam1Δ cells. At metaphase I in both dam1Δ and dam1Δ rec12Δ cells, spindle length tended to increase (electronic supplementary material, figure S8A), perhaps reflecting the loss of inward pulling forces generate by kMT depolymerization [53], and although centromeres were on the spindle, as in wild-type cells, they remained largely still without undergoing oscillation (figure 6a,b electronic supplementary material, Movies S7–S11). Accordingly, centromere velocities and the standard deviation of centromere positions decreased markedly, and the MSD plots shifted downward (figure 6c electronic supplementary material, figure S8B,C). The auto- and cross-correlation plots exhibited different patterns from those in wild-type and rec12Δ cells (electronic supplementary material, figure S8D–F).

Figure 6. Centromere and spindle dynamics at meiosis I in dam1Δ cells. (a) Behaviour of centromeres (cen2) and the spindle at meiosis I. Green indicates cen2 and SPB, and magenta indicates the spindle. Numbers indicate time (in minutes). Arrows and arrowheads indicate the SPB and cen2, respectively. White lines indicate cell outlines. Bar, 5 µm. (b) Changes in centromere and SPB positions at metaphase I in dam1Δ and dam1Δ rec12Δ cells. Left images show kymograph of centromeres (cen2) and SPBs and right graphs show changes in centromere–SPB and SPB–SPB distances. Vertical and horizontal bars of kymographs show 2 µm and 30 s, respectively. (c) Mean centromere velocities. Dotted lines distinguish results of rec12 + and rec12Δ cells. (d) Changes in centromere–SPB and SPB–SPB distances during anaphase I. (e) Loss of anaphase A centromere movements in dam1Δ cells. Photos show changes in centromere positions during anaphase I. Numbers show time in minutes from the start of anaphase. Dotted lines and arrowheads in photos show positions of SPB and cen2, respectively. Bar, 2 µm. Graph shows the mean centromere–SPB distances during metaphase I (Meta) and anaphase I (Ana). Numbers of centromeres examined are shown in parentheses. A dotted line in graph distinguishes results of wild-type and dam1Δ cells. Error bars show standard error. *p < 0.05 **p < 0.01 ***p < 0.0005 n.s., not significant (Student's t-test).

In addition to centromere oscillation, anaphase poleward centromere movements were impaired in dam1Δ cells. Upon anaphase onset, homologous centromeres normally separate from each other by poleward movements (anaphase A), and then by spindle elongation (anaphase B) (figure 6a,d,e, WT) [23]. In dam1Δ cells, however, anaphase poleward centromere movements rarely occurred, and their distances from the nearest SPBs were largely unchanged (figure 6a,d,e electronic supplementary material, figure S8G,H). Nonetheless, homologous centromeres successfully separated from each other as the spindle elongated (figure 6a,d), indicating that homologous centromeres undergo anaphase B without undergoing anaphase A. Notably in this regard, anaphase A movements were impaired in rec12Δ cells due to bi-oriented attachment of sister chromatids (electronic supplementary material, figure S8H) [28]. The lack of centromere oscillation and anaphase A are consistent with previously reported dam1Δ phenotypes: kinetochores remain at the MT ends for a long time (more than 20 min) and inhibit MT disassembly after kinetochore attachment to the MT ends [43,54], and chromosomes frequently fail to reach the SPB during mitotic anaphase [51,55]. The coincidental impairment of centromere oscillation, together with the attachment correction defect, provides further support for a link between attachment correction and centromere oscillation.

3. Discussion

3.1. Roles of chiasmata and Dam1 in attachment correction

In this study, we investigated whether chiasmata cause elimination of bi-oriented attachment via the error correction mechanism by analysing the attachments in mad2Δ or ark1-so cells forming or lacking chiasmata. Analysis of sister chromatid segregation, transient sister centromere splitting and homologous centromere positioning collectively revealed that in the mad2Δ and ark1-so backgrounds, bi-oriented attachment of sister chromatids increased in chiasma-forming cells, but mono-oriented attachment increased in chiasma-lacking cells. Presumably, sister chromatids are initially frequently bi-oriented irrespectively of chiasma formation, but the error correction mechanism eliminates bi-oriented attachment specifically in chiasma-forming cells. Less prominent impacts of the mad2Δ mutation on attachments than those of the ark1-so mutation probably resulted from distinct mutation-dependent defects. The mad2Δ mutation decreases the time available for correction but does not compromise attachment elimination unlike the ark1-so mutation in mad2Δ cells, despite attachment elimination, the shortage of the correction time probably resulted in incomplete attachment correction. Together, these observations indicate that chiasmata enable error correction factors to eliminate bi-oriented attachment of sister chromatids that are otherwise not eliminated.

A previous study reported that mad2 and ark1 mutations increased bi-oriented attachment of sister chromatids in the absence of chiasmata [31], contradicting the results of this study. The ark1-as mutation and a different centromere marker used in that previous study may have affected attachments. In addition, it is important to note that pairing of homologous centromeres, which occurs during meiotic prophase, may affect the initial kinetochore–MT interaction. However, the contribution of centromere pairing to the observed differences between chiasma-forming and chiasma-lacking cells is probably negligible because centromere pairing occurs at a largely normal level in rec12Δ cells [56] and equational segregation frequencies in haploid meiotic cells, which lack pairing, were almost the same as those in rec12Δ cells (figure 1e,f).

In addition to bi-oriented attachment of sister chromatids, our analysis of sister centromere tilting suggested frequent occurrence of merotelic attachment and its elimination by the error correction mechanism, irrespective of chiasma formation. Indeed, merotelic attachment is frequently observed in chiasma-forming meiotic cells as well as in chiasma-lacking mitotic cells [57,58]. Tilted centromere positioning may have resulted from only one of the sister centromeres interacting with MT (monotelic attachment) or from detachment of sister centromeres from MTs after separation. However, given the active mobilities of splitting sister centromeres (electronic supplementary material, figure S2B–D) and the stable end-on attachment seen in dam1Δ cells [43,54], these events are likely to be rare. It is also possible that both sister centromeres interact with MTs extending from the same pole (syntelic attachment), one laterally and the other at the ends, as in vertebrate cells [59], and that bi-directional lateral sliding of one of the centromeres causes tilted positioning. However, the consistency of the frequencies of sister centromere splitting with those of equational segregation suggests that syntelic attachment is probably transient, if it occurs at all.

We also found that the dam1Δ mutation increased bi-oriented attachment of sister chromatids in chiasma-forming cells and increased mono-oriented attachment in chiasma-lacking cells. This suggests that Dam1 plays a crucial role in the attachment correction. The increased tilting of splitting sister centromeres seen in rec12 + dam1Δ cells further supports this notion. Because dam1Δ cells exhibited delayed anaphase onset, Dam1 must contribute to attachment elimination via an SAC-independent pathway. Given that Dam1 plays a crucial role in kinetochore–MT interaction and is a substrate of Aurora B [42,51,60], we speculate that the Dam1 complex is directly involved in the attachment elimination process as a component of the Aurora B regulatory pathway.

3.2. Functions of chiasmata and Dam1 in centromere oscillation

Using the correlation functions, we showed that centromeres oscillated irrespectively of chiasma formation and that chiasmata coordinated homologous centromere oscillation. Various experimental- and/or simulation-based studies demonstrated that coordinated oscillation of sister chromatids depends on the tension generated by sister centromere cohesion [10,61–71]. Given this notion, it is likely that coordinated homologous centromere oscillation similarly depends on tension generated by chiasmata. During oscillation of mitotic sister centromeres, kinetochore-interacting MT bundles (kMTs) extending forward of the moving centromeres (leading kMTs) drive centromere movements by their disassembly, whereas those extending rearward (trailing kMTs) assemble (electronic supplementary material, figure S9A, centromere movement and kMT dynamics) [4–6,10], and switching of either the leading or trailing kMTs induces reversal of centromere movements (electronic supplementary material, figure S9A, centromere movement and kMT dynamics) [72]. In fission yeast, Kinesin-8 motors, Klp5 and Klp6, that promote disassembly of longer kMTs induce kMT switching in the proximity of the equator [41,71,73,74]. The kMT switching causes transient centromere relaxation or stretching (electronic supplementary material, figure S9A, mitosis, transition state), which likely induces coordinated assembly/disassembly switching of the other kMTs through force-dependent changes in MT dynamics [75–78]. During homologous centromere oscillation, the kMTs probably undergo assembly and disassembly in a similar manner, and the chiasma-dependent link generates a temporal tensile change that coordinates assembly/disassembly switching of the kMTs (electronic supplementary material, figure S9A, meiosis I).

We also found that Dam1 was essential for centromere oscillation and anaphase A poleward centromere movements. This finding indicates that centromere oscillation and anaphase A depend on the same mechanism. The Dam1 complex probably drives centromere movements by coupling MT disassembly to centromere movement [49,50]. It may also promote MT disassembly, as kMT shortening is inhibited after kinetochore capture in dam1Δ cells [43,54,79]. Despite the lack of centromere oscillation and anaphase A, kinetochore–MT interactions were not eliminated in dam1Δ cells, as demonstrated by sister centromere splitting and anaphase B centromere movements. Therefore, a factor(s) in addition to Dam1 may cooperatively drive centromere movement. The factors that are likely to mediate kinetochore–MT interaction in dam1Δ cells include the Ndc80 complex, a distinct kinetochore–MT interface factor and the Klp5/Klp6 complex, which can couple MT disassembly to cargo movement like the Dam1 complex and play crucial roles in centromere oscillation and/or chromosome segregation [71,73,74,80–87]. Indeed, kinesin-8 motors are essential in dam1Δ cells [51] and contribute to stable kinetochore–MT interaction during centromere oscillation [73].

3.3. Functions of chiasma and Dam1 in attachment elimination

The precise mechanism of chiasma-dependent attachment elimination and the mechanism of Dam1-dependent attachment elimination itself remain unclear. It was previously proposed that chiasmata eliminate bi-oriented attachment of sister chromatids by generating a chromosome configuration that arranges sister kinetochores outward and the Aurora B-enriched region inward [31,88]. In this model, poleward pulling brings improper attachment sites close to the Aurora B-enriched region, leading to the elimination of the attachments. However, this configuration requires firm association of sister centromeres. When sister centromeres separate as seen in our study, improper attachment sites would barely approach the Aurora B-enriched region (electronic supplementary material, figure S9B). Therefore, the validity of this model remains debatable.

It is possible that chiasmata eliminate bi-oriented attachments by reducing tension across sister centromeres. Indeed, although we could not observe significant difference between projected inter-sister centromere distances of rec12 + and rec12Δ cells, a subset of the distances in the rec12 + cells covered a slightly shorter distance range than those covered in rec12Δ cells, suggesting a reduction in tension. However, in a rec12 + cell shown in figure 2, the projected centromere distances were comparable to or greater than those in rec12Δ cells (electronic supplementary material, figure S4C), suggesting that at least in this particular cell, tension was probably not reduced. Nonetheless, the bi-oriented attachment was eliminated as demonstrated by re-association of the splitting sister centromeres (figure 2b,c, WT). Therefore, we speculate that chiasma-dependent attachment elimination is not solely dependent on the reduction in overall tension.

Our finding that both chiasmata and Dam1 contribute to centromere oscillation raises the possibility that attachment elimination is coupled to centromere oscillation. This possibility is supported by our finding that homologous centromere oscillation dynamics are altered in the error correction-defective ark1-so mutant. One possible explanation is that centromere oscillation causes elimination of erroneous attachments. In chiasma-forming cells, when improperly attached homologous centromeres undergo oscillation via coordinated disassembly and assembly of the leading and the trailing kMTs (electronic supplementary material, figure S9C, Meiosis I), stochastic and/or length-dependent initiation of assembly/disassembly MT switching may give assembling kMT ends a chance to experience a chiasma-dependent minus end-directed load at improper attachment sites. The minus end-directed load applied to the assembling kMT ends may bring attachment sites close to the inner centromere region, inducing Aurora kinase-dependent kMT detachment. Alternatively, the minus end-directed load may cause kMT detachment due to the intrinsically weak resistance of the interaction between assembling MTs and the kinetochore to a minus end-directed load. Indeed, attachment of Stu2, a XMAP215/Dis1 family kinetochore component in budding yeast, to assembling MT ends can withstand a tensile force of approximately 4 pN under plus end-directed load [89], whereas attachment of Xenopus XMAP215 can withstand a force of only approximately 1 pN force under minus end-directed load [90]. In this model, chromosome oscillation actively eliminates improper attachments by applying a load to the improper attachment sites, and the lack of chiasmata or centromere oscillation impairs attachment elimination.

An alternative, but not mutually exclusive, possibility is that attachment elimination requires an attachment property that enables centromere oscillation. We found that centromere oscillation is impaired in ark1-so cells. In addition, expression of mutant forms of the kinetochore component Ndc80 that cannot be phosphorylated by Aurora B inhibits centromere oscillation [81,83]. These observations indicate that centromere oscillation requires Aurora B-dependent kinetochore phosphorylation. In addition, variants of Dam1 or Ndc80 bearing Aurora B-phosphomimetic mutations form diffusible attachment to the MT lattice and exhibit diffusion-like movements on the MT [44,46,91]. Therefore, centromere oscillation probably depends on Aurora B-dependent diffusible attachment. Because diffusion activities of kinetochore components increase as their MT affinities decrease, only diffusible attachments may allow attachment elimination. In the absence of Dam1, attachment may become non-diffusible and non-eliminable, resulting in impairment of both centromere oscillation and elimination of erroneous attachments. Further close investigation of the effects of the phospho-mutant forms of Dam1/Ndc80 on attachment correction and centromere oscillation would be important to verify this possibility.

The chromosome oscillation-dependent mechanism can also account for the elimination of merotelic attachment of a single chromatid in a chiasma-independent manner and in both mitosis and meiosis. Perhaps, cohesin-dependent coordinated oscillation of sister chromatids eliminates merotelic attachment in the same way that coordinated homologous chromosome oscillation eliminates bi-oriented attachment of sister chromatids (electronic supplementary material, figure S9C, mitosis). Consistently, a loss of sister chromatid cohesion causes merotelic attachment in mitosis [92,93], and in dam1Δ cells, impaired centromere oscillation was accompanied by increased sister centromere tilting, a probable outcome of merotelic attachment, irrespectively of chiasma formation (figure 5i). Furthermore, the oscillation-dependent mechanisms likely contribute to attachment elimination in vertebrates, although the mechanisms are not completely the same, as demonstrated by additional contribution of Aurora A kinase to centromere oscillation and attachment correction [94,95].

The oscillation-dependent mechanism does not deny the importance of overall tension across sister centromeres in attachment elimination. It is clear that as the number of properly attached kMTs increases, tension across sister centromeres increases. Gradual elevation of tension may incrementally decrease kinetochore phosphorylation levels, resulting in an incremental increase in MT binding affinity of kinetochores, as in the case of Ndc80 phosphorylation [91]. An alternative, but not mutually exclusive, possibility is that tension greater than some threshold alters the kinetochore phosphorylation state. Kinetochore phosphorylation is regulated by phosphatases in addition to Aurora B [96], and the antagonistic actions of these enzymes may robustly maintain the kinetochore phosphorylation state during metaphase, as suggested for antagonistic regulatory systems [97]. However, once tension exceeds the threshold, the kinetochores may be completely dephosphorylated by the kinetochore regulatory system [96,98,99]. In either case, tension converts the metaphase-type diffusible, correctable attachments to the anaphase-type non-diffusible, non-correctable attachments, linking establishment of proper attachments with the metaphase-to-anaphase transition. This scenario is consistent with the well-established relationship between tension and loss of Aurora-dependent phosphorylation and attachment stabilization (e.g. [12,27,100, 101]). In addition, in this scenario, a reduction in tension leads to an increase in kinetochore phosphorylation, making the kinetochore state preferable for attachment correction this can account for correction of merotelic attachment induced by reduced tension [102].

In this study, we showed that chiasmata and Dam1 differentially contribute to the elimination of erroneous attachments and centromere oscillation. These findings raise the possibility that attachment elimination is coupled with centromere oscillation. The precise mechanisms of chiasma-dependent or Dam1-dependent attachment elimination, as well as the relationship between centromere oscillation with attachment elimination, remain to be elucidated. However, our findings and the suggested relationship will undoubtedly shed new light on understanding of the mechanisms of attachment correction and centromere oscillation. Because chromosome missegregation is associated with various diseases or disorders, including tumorigenesis and birth-related Down's syndrome, our findings may be of clinical importance.

4. Material and methods

4.1. Yeast strains, media and basic genetical methods

Strains used in this study are shown in electronic supplementary material, tables S2 and S3. Media and basic genetical methods used in this study were described previously [103]. The dam1 deletion strain was generated by replacing the entire dam1 gene with the G418-resistance gene by a PCR-based method [104,105].


Although the number of reciprocal recombination events at meiosis is similar for organisms with widely different genome sizes such as the mouse and lily (which have between 20 and 50 events), the number of DNA-DNA interactions that are recognized by RAD51/DMC1p immunocytology at prophase of meiosis is much higher (250 for the mouse and more than 2000 for lily)(Anderson et al., 2001). It follows that most or all of these early interactions do not necessarily function in the formation of reciprocal events. Conceivably, the numbers relate to genome size, 3.3 pg for the mouse and 141 pg for lily. However, the genome sizes differ by a factor of 43, whereas the number of RAD51/DMC1 foci in the mouse differs only by a factor of 10 from the lily. A better correlate appears to be the length of paired prophase chromosomes, as measured by the total length per nucleus of the synaptonemal complexes, SCs, which are the axes of the bivalents (about 200 μm for the mouse and about 3000 μm for lily). The tentative conclusion is that these foci, which are associated with the chromosome cores and SCs, function in SC formation(Anderson et al., 2001). Synaptic failure in the absence of RAD51/DMC1 foci in SPO11 -/- mice also suggests a role for early nodules in synapsis(Baudat et al., 2000Romanienko and Camerini-Otero,2000).

Molecular models for the resolution of DNA-DNA interactions without reciprocal recombination have been discussed by Gilbertson and Stahl(Gilbertson and Stahl, 1996)and involve helicase-topoisomerase activity to resolve joint molecules. The acquisition of replication protein A, RPA and Bloom mutated protein, BLM, at the RAD51/DMC1 sites might be the physical manifestation of the models —RPA may stimulate BLM-RecQ helicase activity(Brosh et al., 2000) in concert with topoisomerase IIIa (Johnson et al., 2000) to resolve the early DNA-DNA interactions at the RAD51/DMC1 sites.

The development of reciprocal genetic exchange events at meiosis in many fungi, plants and animals can be monitored at several levels: (1) at the chromosomal level by chiasmata, which are the sites of reciprocal recombination (Jones, 1987)(2) by the recombination nodules, RNs, which correlate with genetic and cytological patterns of recombination(Carpenter, 1975Carpenter, 1979Albini and Jones, 1988) and(3) by the MLH1p sites, which are associated with crossover sites(Anderson et al., 1999)

Nodules are SC-associated, electron-microscope-defined structures that have been reported in the meiocytes of protists, fungi, plants and animals. In early meiotic prophase, `nodules' refer to the several hundreds of small dense bodies about 100 nm in diameter that are associated with chromosome cores and SCs and contain the RAD51/DMC1 proteins in lily(Anderson et al., 1997), mouse and human (Haaf et al., 1995Moens et al., 1997Barlow et al., 1997). Since the correlation of these structures with reciprocal recombination is tenuous, they are usually referred to as `early nodules', EN. The late RNs, as originally defined by Carpenter (Carpenter,1975) in Drosophila melanogaster oocyte SCs, correspond in number and location to reciprocal recombinant events in normal and mutant D. melanogaster. In the rat, `late RNs' are well defined electron-dense bodies of variable shape and size, 100 to 200 nm, located on the mature SC at pachytene stage VII of the spermatogenic pathway but not in earlier pachytene stages I to V (Clermont,1972 Moens,1978). Cross-sectioned SCs and whole-mount, shadow-cast EM preparations show that RNs are located on the surface of the SC either along the central element or obliquely across the SC(Fig. 7). The reported number of late RNs (19 to 22 per nucleus) in complete EM-reconstructed rat-pachytene nuclei, their non-random distribution and their association with MLH1 (Mut L homolog) protein in rat and mouse (this report), agree with their proposed function in reciprocal recombination by Carpenter(Carpenter, 1975Carpenter, 1979). A number of publications have assigned various proteins to RNs but the assignments have not previously been verified by EM demonstration of these proteins on the RNs.

Definition of a rat recombination nodule, RN, in a shadow-cast preparation of an SC. (A) The SC consists of two lateral elements, le, and a median central element. (B) Occasionally RPA can be detected at these RNs. Mouse RNs are somewhat smaller and less distinct and are therefore visualized with PTA in the other illustrations. The width of the SC is about 200 nm.

Definition of a rat recombination nodule, RN, in a shadow-cast preparation of an SC. (A) The SC consists of two lateral elements, le, and a median central element. (B) Occasionally RPA can be detected at these RNs. Mouse RNs are somewhat smaller and less distinct and are therefore visualized with PTA in the other illustrations. The width of the SC is about 200 nm.

We report the events in individual mouse and rat spermatocyte nuclei from early to late meiotic prophase in terms of chromosome core behavior and associated protein complexes. The immunofluorescence observations are refined and detailed by immunoelectron microscopy of the recombination-related proteins and by EM visualization of RNs. These observations are interpreted in the context of the chromosome synapsis and reciprocal recombination and discussed in relation to reports by others on these events.

Perspectives and Future Aims

The mechanisms behind inverted meiosis in holocentric organisms are currently unknown. The occurrence of inverted meiosis demands modification in the conserved mechanisms of meiotic cohesion and chromosome segregation. New adaptations and differential regulation of meiotic cohesions such as REC8 and centromere cohesion guardians such as shugoshins are expected to have happened. Additionally, modification of the spindle attachment machinery also should be expected due to an alternative centromeric organization. Furthermore, the observed chiasmata formation between holocentric chromosomes demands adaptations of the mechanisms that prevent recombination at or around centromeres. The limited knowledge of holocentromeres and close relatives of Cyperaceae limits us to speculate about what to expect in terms of adaptations of the meiotic recombination machinery to holocentricity. Future studies aiming the molecular characterization of such mechanisms will be of interest for evolutionary and comparative biology studies.


Division of Evolutionary Biology, LMU Munich, Planegg-Martinsried, Germany

Joshua V. Peñalba & Jochen B. W. Wolf

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar


J.V.P. wrote the manuscript. J.B.W.W. edited the manuscript before submission. Both authors researched data for the article and substantially contributed to the discussion of content.

Corresponding authors

Variability of Chiasma Frequencies in Different Tomato Species

This article presents the results of comparative studies of the frequency and distribution of chiasmata in pollen mother cells (PMCs) in five diploid tomato species, Solanum lycopersicum, S. pimpinellifolium, S. peruvianum, S. habrochaites, and S. neorickii, and one autotetraploid species, S. pimpinellifolium. It was established that under the same growing conditions, the total chiasma frequency in the cell depended on the species. At the same time, the green-fruited species S. peruvianum, S. neorickii, and S. habrochaites differed in distal chiasma frequency, while the red-fruited species S. lycopersicum and S. pimpinellifolium differed in interstitial chiasma frequency. It was shown that the total chiasma frequency in PMCs of plants of one species is a stable index of recombination potential that does not depend on the growing conditions. The redistribution between distal and interstitial chiasmata was found to be more variable, depending on the species, year, geographic growth conditions. In autotetraploid, the chiasma frequency per bivalent was lower than that in diploid S. pimpinellifolium plants, primarily due to interstitial chiasmata, the frequency of which remained at the level characteristic for diploid plants. It was concluded that the recombination plasticity of the tomato genomes was due to the redistribution of chiasmata along bivalents, and not to the change in their number in the cell.


Historically, chromosomal rearrangements have long been recognized as a major driving force promoting divergence (White, 1978), but it was only recently that the role of rearrangements as recombination modifiers has been recognized and formalized in the recombination suppression models of chromosomal speciation (Noor et al., 2001 Rieseberg, 2001 Navarro and Barton, 2003 Ayala and Coluzzi, 2005). Thus, the emphasis is no longer on the decrease in fitness of chromosomal hybrids, but on the suppression of recombination between the rearranged chromosomes, resulting in a partial barrier to gene flow between populations. The association with speciation events lies in the tight linkage between the rearrangements and reproductive isolation or locally adaptive genes owing to the suppression of recombination in chromosomal heterozygotes. Thus, rearrangements can contribute to the persistence of reproductive isolation genes in the face of gene flow for a longer time than if they were absent (Noor et al., 2001), and extend the action of linked isolation genes over larger genomic regions than if no rearrangements were present (Rieseberg, 2001 Butlin, 2005). The emergence of the recombination suppression models stimulated a considerable amount of both empirical (for example, Lowry and Willis, 2010 Ayala et al., 2011 MacGaugh and Noor, 2012 Farré et al., 2013) and theoretical (for example, Kirkpatrick and Barton, 2006 Feder and Nosil, 2009 Faria and Navarro, 2010) studies exploring the operational conditions, mechanisms and processes involved. Most, if not all, of these analyses have focused on inversions and reciprocal translocations, whereas the most widespread rearrangement, Robertsonian (Rb) fusion/fission, has received considerably less attention (King, 1993).

The aim of the present study is to assess the effect of heterozygosity for Rb fusions on recombination patterns. These rearrangements consist of the joining of two nonhomologous acrocentric chromosomes by the centromere to form one biarmed chromosome. The house mouse, Mus musculus domesticus, was chosen as the biological model for several reasons. First, extensive chromosomal diversity through fixation of Rb fusions occurs throughout its distribution and all chromosomes, except the sex pair, are involved in Rb fusions in wild populations (Piálek et al., 2005). Second, underdominance levels of simple Rb heterozygotes, that is, carrying trivalents formed by the pairing of Rb fusions with the two homologous acrocentrics, are well documented in this subspecies. In particular, the data indicate that heterozygosity for a single-Rb fusion has a limited effect on hybrid fitness (Hauffe and Searle, 1998 Sans-Fuentes et al., 2010) in this case, there will be almost no constraint on its fixation probability, but conversely, its contribution to postmating isolation will be extremely reduced unless it is associated with changes in recombination pattern. Finally, several studies have indicated a reduction in crossover rates in wild homozygous Rb vs standard mice in the proximal centromeric regions (Bidau et al., 2001, Castiglia and Capanna, 2002 Dumas and Britton-Davidian, 2002 Merico et al., 2003, 2013). Analyses of recombination rates in wild Rb heterozygotes have also been assessed but involve, in most cases, polymorphic individuals, that is, carrying variable numbers of Rb bivalents and trivalents, making it difficult to disentangle the effect of each meiotic type of configuration (Wallace et al., 1992 Bidau et al., 2001 Castiglia and Capanna, 2002 Capilla et al., 2014). In contrast, extensive analyses of recombination patterns in single-Rb heterozygotes have been performed by genetic assays in crosses between different laboratory strains (Davisson and Akeson, 1993). Recombination suppression was recorded in almost all of the Rb fusions tested, but the extent of the suppression varied between chromosomes and crosses, suggesting an influence of the genetic background on this trait. The present analysis is performed on male and female laboratory-bred F1 hybrids (2n=31) between two chromosomal races of the house mouse (2n=40 and 2n=22) from Tunisia, as well as wild-caught mice spanning the hybrid zone between two North Italian races (2n=40 and 2n=22). In both cases, structural genomic differences exist between races, as all chromosomes except two pairs (a small acrocentric pair and the sex chromosomes) are involved in Rb fusions. The Tunisian F1 hybrids exhibit a homogeneous meiotic architecture in which all Rb chromosomes are present as trivalents, whereas the number of Rb fusions (heterozygous and homozygous) was variable in the Italian mice. Although recent cytogenetic estimates of recombination rely on the analysis of MLH1 foci (mismatch repair protein) on synaptonemal complexes (Anderson et al., 1999), recombination was assessed in the present study by meiotic chiasma analyses permitting direct comparisons with previously published data (Dumas and Britton-Davidian, 2002). The effects of Rb heterozygosity on recombination patterns were identified by comparing the chiasma-based assessment of the Tunisian F1 mice with those previously recorded for the Tunisian parental standard and Rb races from which they originated (Dumas and Britton-Davidian, 2002). An estimate of variation in recombination rates was approached by comparison with the data from the Italian samples. From this, we infer the extent and genomic distribution of the barrier to gene flow between chromosomal races and discuss the relevance of Rb rearrangements to reproductive isolation and divergence processes.

Genome and Gene Structure∗

4.2.4 Meiotic Recombination

Meiotic recombination [1] refers to the reciprocal physical exchange of chromosomal DNA between the parental chromosomes and occurs at meiosis during spermatogenesis and oogenesis, serving to ensure proper chromosome segregation. During the four-strand stage of meiosis, two duplex DNA molecules (one from each parent) form a hybrid, and a single strand of one duplex is paired with its complement from the other duplex. Single-stranded DNA is exchanged between the homologous chromosomes, and the process involves DNA strand breakage and resealing, resulting in the precise recombination and exchange of DNA sequences between the two homologous chromosomes. This process is highly efficient and does not usually result in mutations at the sites of recombination. Recombination thus shuffles genetic material between homologous chromosomes, generating much of the genetic diversity that characterizes differences between individuals, even within the same family.

The frequency of recombination between two loci along a chromosome is proportional to the physical distance between them, and historically, this provided the basis for defining the genetic distance between loci, allowing genetic maps to be constructed. The genetic proximity of two loci is measured by the percentage of recombination between them a map distance of 1 centimorgan (cM) indicates 1% recombination frequency between the two loci. The human genome sequence has made it possible to compare genetic and physical distances and to analyze variations in recombination frequency in different chromosomal regions. On average, 1 million bp (1 Mb) correspond to 1 cM (1% recombination frequency). However, there is a tremendous local variation between individual chromosomes and among particular chromosomal regions. For example, the average recombination rate is higher in the short arms of chromosomes and at the distal segments of the arms but overall is suppressed near the centromeres. There is also a significant variation in the recombination rates between the sexes, with 1.6-fold more recombination on average in females relative to males. On average, female recombination is higher at the centromeres and male recombination is higher at the telomeres [2] .

Antibody variation

In B cells, the variable region of the light chain gene has 40 variable (V) and five joining (J) segments. An enzyme called DNA recombinase randomly excises most of these segments out of the gene, splicing one V segment to one J segment. During RNA processing, all but one V and J segment are spliced out. Recombination and splicing may result in over 10 6 possible VJ combinations. As a result, each differentiated B cell in the human body typically has a unique variable chain. The constant domain, which does not bind to an antibody, is the same for all antibodies. The large diversity of antibody structure translates into the large diversity of antigens that antibodies can bind and recognize.

Similar to TCRs (T cell receptors) and BCRs (B cell receptors), antibody diversity is produced by the mutation and recombination of approximately 300 different gene segments encoding the light and heavy chain variable domains in precursor cells that are destined to become B cells. The variable domains from the heavy and light chains interact to form the binding site through which an antibody can bind a specific epitope on an antigen. The numbers of repeated constant domains in Ig classes (discussed below) are the same for all antibodies corresponding to a specific class. Antibodies are structurally similar to the extracellular component of the BCRs. The maturation of B cells into plasma cells occurs when the cells gain the ability to secrete the antibody portion of its BCR in large quantities.

$ left(egin (1-2 ext

) (alpha (gamma -sigma (1- ext) (gamma +(-gamma alpha -alpha +1) kappa +1))-2 ext sigma +sigma) & sigma ext

(1- ext

) left(-kappa (gamma +1) alpha^<2>+(gamma +kappa +1) alpha -2 ight) sigma ext (1- ext) (-(1-alpha) kappa alpha -alpha +2) & sigma (1-2 ext) ( ext

(-kappa (1-alpha) alpha -alpha +2)+alpha kappa -1) end ight), . $

At the interior fixed point ((h^<*>_<1>,p^<*>_<1>)) , the matrix is given by (J(h^<*>_<1>,p^<*>_<1>)=)

Watch the video: What is VVT, how does it work? A cool guide to the VVT system. Subtitles! (August 2022).