News & Views | Published:


Crossover control in two steps

Nature volume 462, pages 4647 (05 November 2009) | Download Citation

During meiotic cell division, chromosome pairs exchange genetic material in a tightly controlled crossover process. Higher-order chromosome structure may regulate this genetic reshuffling at two distinct stages of meiosis.

Meiosis is a cell-division program that reduces the chromosome complement by half, generating haploid gametes (eggs and sperm) from diploid germ cells. Meiosis also has another important function. It drives genetic diversity in a population by allowing genetic exchanges between chromosomes, in a sense 'reshuffling the chromosomal deck'. This genetic exchange takes place through the formation of double-strand breaks (DSBs) in the DNA which, when mended, can result in a crossover (CO) of genetic material between chromosome pairs (bivalents). Such COs are essential for generating physical attachment between chromosome pairs, which ensures that the sperm and eggs get the correct complement of chromosomes when the germ cell divides. CO formation is highly regulated, but how this process is controlled at a genome-wide or a region-specific level is not fully understood. A recent study published in Cell by Mets and Meyer1 implicates higher-order chromosome structure in the regulation of the frequency and distribution of DSBs, and therefore of COs. The authors' analysis in the roundworm Caenorhabditis elegans also reveals that CO control operates in at least two temporally distinct stages of the meiotic recombination pathway.

After DSB formation in meiotic chromosomes, the broken ends of DNA are resected, creating 3′ single-stranded DNA overhangs that are coated by the Rad51 recombination protein. This nucleoprotein filament invades an intact homologous sequence on the partner chromosome, creating a D-loop structure (Fig. 1). Capture of the second free DNA end, DNA synthesis and ligation result in the formation of a recombination intermediate known as a double Holliday junction. This, when resolved, results in either the reciprocal exchange of genetic information between chromosomes (CO) or transfer of information to the repaired chromosome only (non-crossover, NCO)2 (Fig. 1). Most NCOs, however, probably stem from an earlier repair mechanism termed synthesis-dependent strand annealing2.

Figure 1: Crossover formation and control.
Figure 1

a, During meiosis, chromosomes undergo double-strand break (DSB) formation. b, The broken ends of DNA are resected, producing 3′ single-stranded DNA tails with which the Rad51 protein (not shown) associates. c, The single-stranded DNA invades a homologous DNA sequence on the chromosome pair, creating a D-loop structure. d, Capture of the second free DNA end, DNA synthesis and ligation result in a double Holliday junction (dHJ) intermediate structure. e, dHJs can be resolved by cleavage either as a crossover (CO) (if cleaved at the arrowheads shown in d) with reciprocal exchange of genetic information between chromosomes, or as a non-crossover (NCO) with transfer of genetic information to the repaired chromosome only. However, NCOs probably occur mostly through synthesis-dependent strand annealing (SDSA). Mets and Meyer1 propose that, in the roundworm Caenorhabditis elegans, CO control can operate at two steps. The first step (1) is controlled by condensins I and II, protein complexes that compact chromosome structure during meiosis (green clamp). The inset shows how condensin protein complexes seem to regulate the number and distribution of DSBs (scissors), and hence of COs. The increase in chromosome axis length in condensin mutants (broken clamp) might reflect changes in the number and size of chromatin loops. This may, in turn, lead to changes in the frequency and distribution of DSBs (which are perhaps more common at loops), and consequently in changes in the number and distribution of COs. The second CO control step (2) may follow DSB formation, when the CO-versus-NCO decision is made.

Studies in yeast, plants and other multicellular organisms3,4 suggest that CO formation is controlled at several levels. First, COs are not evenly distributed along chromosomes: some genomic regions (recombination hotspots) are more prone to undergoing CO than others4,5. Second, each bivalent undergoes at least one CO (obligate CO)6. Third, a CO in one region of the chromosome decreases the likelihood that additional COs will occur in that vicinity (CO interference)7. Last, when DSB formation is reduced, the number of NCO events are sacrificed in favour of maintaining CO levels (CO homeostasis)8.

Mets and Meyer1 take advantage of the extreme case of CO interference observed in C. elegans to investigate the mechanisms of CO regulation. Usually in this worm, only one CO occurs in each bivalent9 and so a loss of CO interference is easy to detect. Previous work from the same group10 had shown that the C. elegans protein DPY-28, a component of the dosage-compensation protein complex involved in the regulation of X-chromosome gene expression, also controls CO formation during meiosis. Mutations in the dpy-28 gene increase the number of DSBs and the frequency of COs, and reduce CO interference, so that more than one CO can take place per bivalent. Mets and Meyer's work1, as well as a study by Csankovszki and colleagues11, reveal three biochemically distinct condensin complexes in C. elegans: condensin IDC, involved in dosage compensation; condensin II, involved in mitotic and meiotic chromosome compaction; and a newly described condensin I complex, which includes DPY-28 and comprises five proteins in total, all of which are also found in either condensin IDC or condensin II (refs 12, 13). Similarly to mutations in dpy-28, mutations in the genes encoding any of the other four components of condensin I change the distribution of COs, and increase the number of double and triple COs in the C. elegans X chromosome. Moreover, this is not observed in a mutant that affects only the condensin IDC complex, and so reveals a specific function for the novel condensin I complex in CO regulation.

Mets and Meyer report that, in condensin-I-mutant worms, chromosome regions that have more COs have more DSBs (they measured these breaks by detecting the number of RAD-51 foci). Changes in the position of COs also correlate with changes in the sites of DSB formation. These results imply that condensin I might regulate CO formation on a chromosome-wide basis by affecting DSB production.

How might the condensin I complex affect DSB and CO formation? Chromosomes become condensed during meiosis, adopting an organization in which chromatin loops extend from the chromosome's axis or structural core14. Condensin protein complexes are a central contributor to this compacted organization15. In line with this, Mets and Meyer1 observe that mutations in condensin I subunits increase the axis lengths of meiotic chromosomes (importantly, induction of DSBs with γ-irradiation does not). Mutations in subunits of the condensin II complex also increase the length of the chromosome axes, the number of DSBs and the amount of double COs. However, the two condensin complexes probably work independently to control COs by controlling chromosome structure and hence the position and frequency of DSB formation.

In C. elegans, the authors found an average of 2.1 DSBs per meiotic chromosome pair, with a range of between 0 and 9. But the number of bivalents with no DSBs was markedly lower than expected if the distribution of breaks occurs at random (as predicted by a Poisson distribution). This implies that an active mechanism ensures that there is at least one DSB per bivalent — a prerequisite to achieve an obligate CO. In addition, most bivalents have about 2–6 DSBs, but only one CO, which suggests that CO control must also be enforced at a step after DSB formation, during the CO-versus-NCO decision.

Many questions surround the process of meiotic CO control. Why and how do CO hotspots appear and disappear during population evolution? What is the molecular basis for the process controlling obligate CO and CO interference? Are these processes connected? If so, how is CO control executed at both the whole chromosome level and locally within short genomic intervals? Mets and Meyer's work1 supports a model that starts to address some of these questions.

Changes in chromosome structure imposed by condensins I and II may alter both the frequency and the distribution of recombination hotspots, perhaps by exposing different loci to the DSB machinery. Given that DSBs have been proposed to occur preferentially at chromatin loops16, altering the density and position of loops could change the frequency and distribution of DSBs, and consequently COs (Fig. 1). This suggests that CO control — for example, ensuring the formation of the obligate CO — is exerted early in meiosis, either prior to or at DSB formation, and implicates condensins I and II in mediating this control on a chromosome-wide basis. Changes in chromosome structure might also create and relieve tension along the chromosome, which could be translated into the CO-versus-NCO decision, accounting, at least in part, for CO interference, as has been suggested previously17. However, the precise molecular basis for the execution of this decision remains to be defined. The effect of condensins and other elements that control chromosome structure may differ among individuals, perhaps because of varying levels of expression of these elements, and interplay with different cofactors. This could explain both the variance and dynamics of CO control.

But there are still outstanding questions. Are condensin complexes involved in CO control in organisms such as yeast and mammals? In these organisms, central features of meiosis differ from those of C. elegans — for example, more DSBs occur per chromosome, and CO interference is weaker. Are condensin complexes also involved in the control of CO homeostasis? How do the two distinct condensin complexes coordinate the CO control they exert? Continued studies of CO control in various organisms will be crucial to shedding more light on these questions.


  1. 1.

    & Cell 139, 73–86 (2009).

  2. 2.

    & Microbiol. Mol. Biol. Rev. 63, 349–404 (1999).

  3. 3.

    & Curr. Opin. Genet. Dev. 19, 105–112 (2009).

  4. 4.

    , & Nature Rev. Genet. 5, 413–424 (2004).

  5. 5.

    Nature Rev. Genet. 2, 360–369 (2001).

  6. 6.

    Symp. Soc. Exp. Biol. 38, 293–320 (1984).

  7. 7.

    Am. Nat. 50, 193–221 (1916).

  8. 8.

    , , & Cell 126, 285–295 (2006).

  9. 9.

    , & Genetics 162, 1169–1177 (2002).

  10. 10.

    et al. Genes Dev. 22, 194–211 (2008).

  11. 11.

    et al. Curr. Biol. 19, 9–19 (2009).

  12. 12.

    , & J. Cell Biol. 167, 613–625 (2004).

  13. 13.

    , , & Genes Dev. 16, 729–742 (2002).

  14. 14.

    & Annu. Rev. Genet. 33, 603–754 (1999).

  15. 15.

    & Nature Rev. Genet. 4, 520–534 (2003).

  16. 16.

    , , & Cell 111, 791–802 (2002).

  17. 17.

    et al. Proc. Natl Acad. Sci. USA 101, 12592–12597 (2004).

Download references

Author information


  1. Yonatan B. Tzur and Monica P. Colaiácovo are in the Department of Genetics, Harvard Medical School, Boston, Massachusetts 02115, USA.

    • Yonatan B. Tzur
    •  & Monica P. Colaiácovo


  1. Search for Yonatan B. Tzur in:

  2. Search for Monica P. Colaiácovo in:

About this article

Publication history




By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing