Update TRENDS in Genetics Vol.23 No.

11 539

15 Davidson, E.H. and Erwin, D.H. (2006) Gene regulatory networks and 23 Hibberd, D.J. (1975) Observations on the ultrastructure of the
the evolution of animal body plans. Science 311, 796–800 choanoflagellate Codosiga botrytis (Ehr.) Saville-Kent with special
16 Marshall, C.R. (2006) Explaining the Cambrian ‘‘explosion’’ of animals. reference to the flagellar apparatus. J. Cell Sci. 17, 191–219
Annu. Rev. Earth Planet. Sci. 34, 355–384 24 Whisler, H.C. (1968) Developmental control of Amoebidium
17 Peterson, K.J. and Butterfield, N.J. (2005) Origin of the Eumetazoa: parasiticum. Dev. Biol. 17, 562–570
testing ecological predictions of molecular clocks against the 25 Leys, S.P. and Ereskovsky, A.V. (2006) Embryogenesis and larval
Proterozoic fossil record. Proc. Natl. Acad. Sci. U. S. A. 102, 9547– differentiation in sponges. Can. J. Zool. 84, 262–287
9552 26 Leys, S.P. and Eerkes-Medrano, D. (2005) Gastrulation in calcareous
18 Peterson, K.J. et al. (2004) Estimating metazoan divergence times sponges: In search of Haeckel’s gastraea. Integr. Comp. Biol. 45, 342–351
with a molecular clock. Proc. Natl. Acad. Sci. U. S. A. 101, 6536– 27 Martindale, M.Q. (2005) The evolution of metazoan axial properties.
6541 Nat. Rev. Genet. 6, 917–927
19 Douzery, E.J.P. et al. (2004) The timing of eukaryotic evolution: Does a 28 Martindale, M.Q. et al. (2004) Investigating the origins of triploblasty:
relaxed molecular clock reconcile proteins and fossils? Proc. Natl. Acad. ‘mesodermal’ gene expression in a diploblastic animal, the sea
Sci. U. S. A. 101, 15386–15391 anemone Nematostella vectensis (phylum, Cnidaria; class,
20 Steenkamp, E.T. et al. (2006) The protistan origins of animals and Anthozoa). Development 131, 2463–2474
fungi. Mol. Biol. Evol. 23, 93–106 29 Castillo-Davis, C.I. and Hartl, D.L. (2003) GeneMerge–post-genomic
21 King, N. (2004) The unicellular ancestry of animal development. Dev. analysis, data mining, and hypothesis testing. Bioinformatics 19,
Cell 7, 313–325 891–892
22 Leadbeater, B.S.C. (1983) Life-History and ultrastructure of a new
marine species of Proterospongia (Choanoflagellida). J. Mar. Biol. Ass. 0168-9525/$ – see front matter ß 2007 Elsevier Ltd. All rights reserved.
U. K. 63, 135–160 doi:10.1016/j.tig.2007.08.014

Crossover interference underlies sex differences in

In many organisms, recombination rates differ between on a chromosome reduces the possibility of a second event
the two sexes. Here we show that in mice, this is because nearby [1–3], a phenomenon known as crossover interfer-
of a shorter genomic interference distance in females ence. In many species, recombination rates differ in the two
than in males, measured in Mb. However, the interfer- sexes. The female recombination map is 1.7 times longer
ence distance is the same in terms of bivalent length. We than that of males in humans [4,5] and 1.3 times longer in
propose a model in which the interference distance in mice [6]. Several mechanisms have been proposed to play
the two sexes reflects the compaction of chromosomes important roles: haploid selection [7]; different epistatic
at the pachytene stage of meiosis. interactions among genes expressed during male and
female meiosis [8]; presence of X-linked modifiers [9];
and regional differences in the chromatin structure of male
Introduction and female gametocytes [10]. However, experimental evi-
Meiosis consists of two consecutive cell divisions after a dence in support of these suggestions has remained elu-
single round of DNA replication, thereby ensuring reduction sive. Here we show that crossover interference in meiosis is
of the chromosome number to produce haploid gametes. the main factor underlying sex differences of recombina-
This reduction occurs in the first meiotic division, when tion rates, and that the average intercrossover distance is
homologous chromosomes are joined together in prophase to the same in both sexes when measured in micrometers of
form bivalents and eventually separate in anaphase. In synaptonemal complex length.
mammals, higher plants and yeast, chromosome recognition
and formation of the synaptonemal complex is initiated by Distribution of recombination events along mouse
double-strand breaks on one chromatid. These breaks are chromosome 1
repaired by homologous recombination, leading to genetic Recombination rates in each sex were measured in
crossing over and/or gene conversion when a non-sister backcrosses of C57BL/6JxCAST/EiJ F1 male and female
chromatid is used as a template. Given the segregation of mice to C57BL/6J. The entirety of mouse chromosome 1
chromatids into haploid gametes, only half of the genetically (Chr 1) was examined at 7 Mb resolution, which ensured
recombinant chromosomes that result from molecular the detection of virtually all crossovers taking into account
recombination events will be detected. the strong positive interference in mouse recombination
Crossover events are not randomly spaced along [11]. In total, we detected 2715 recombination events
chromosomes. Instead, the presence of one crossover event in 2762 progeny of female F1 parents and 1509 recombina-
Corresponding author: Petkov, P.M. ([email protected]). tion events in 1881 progeny of male F1 parents. The average
Available online 26 October 2007. recombination rates were 0.51 cM per Mb (cM/Mb) in
www.sciencedirect.com
540 Update TRENDS in Genetics Vol.23 No.11

Table 1. Distribution of crossover classes along mouse Chr 1 in progeny of female and male F1
Zero crossovers Single Double Triple Average number of
crossovers crossovers crossovers chiasmata per bivalent
Female Number 742 1367 611 42 1.97
Relative Frequency 0.27 0.49 0.22 0.015
Male Number 657 941 281 2 1.60
Relative Frequency 0.35 0.50 0.15 0.001

females and 0.41 cM/Mb in males, which corresponds to a females. Note that chromosomes lacking a crossover are
female-to-male ratio of 1.23. In these parameters, Chr 1 did not the product of a meiosis in which no chiasmata (the
not differ significantly from the genome wide sex-averaged cytologically visible manifestations of crossing over)
recombination rate of 0.55 cM/Mb and female-to-male ratio formed on Chr 1; there must be at least one chiasma on
of 1.3. every chromosome for a successful meiosis. Because chias-
There was a maximum of three recombination events on mata formation involves only two of the four available
an individual chromosome in both crosses; however, the chromatids, one-half of the chromosomes are non-crossover
relative frequencies of single, double and triple crossovers when there is a single chiasma and one-quarter when there
were markedly different in male and female meiosis are two. Using the data in Table 1 and assuming equal
(p = 1016, by a x2 test). No crossovers were found in probability for chiasma formation between any two non-
27% of the progeny of F1 females and 35% of F1 males. sister chromatids (no chromatid interference), we esti-
Single crossovers were found in 50% of the progeny of mated (by maximum likelihood) that for Chr 1 in female
both sexes, which is consistent with the expectation of an meiosis, 11.9% of bivalents will have three chiasmata,
obligate crossover on each chromosome (see Online Supple- 72.4% two chiasmata and 15.6% a single chiasma. For
mentary Material). The difference lay in the frequencies of the male meiosis, these figures are 0.8%, 58.6% and
multiple crossovers (Table 1). The progeny of female F1 40.5%, respectively. It is evident that female bivalents
had 1.5 times higher frequency of double crossovers and form two or three chiasmata more frequently than male
14 times higher frequency of triple crossovers than the bivalents.
progeny of male F1. Triple crossovers were extremely rare The distribution of recombination events along the
in male meiosis; only two such events were found in all entirety of Chr 1 showed similar trends in both female
male F1 progeny compared with 42 in the progeny of and male meioses, with elevated recombination rates

Figure 1. Interference as a function of intercrossover distances. (a) Cumulative rates of double crossovers in female and male meioses. Intercrossover distances are
measured in megabases of DNA length. (b) Coefficient of coincidence as a function of distance between two crossovers in female and male meioses in a sliding window of
30 intervals. Intercrossover distances are measured in megabases of DNA length. (c) Cumulative rates of double crossovers in female and male meioses. Intercrossover
distances are expressed in physical length of pachytene bivalents (mm), calculated from [13]. (d) Coefficient of coincidence as a function of distance between two crossovers
in female and male meioses in a sliding window of 30 intervals. Intercrossover distances are expressed in physical length of pachytene bivalents (mm). A comparison
between panels (a) and (c), as well as (b) and (d) shows that interference in both sexes is the same in terms of bivalent length (mm) rather than genomic length (Mb). In
panels (a–d): female, red line; male, blue line.

www.sciencedirect.com
 Update TRENDS in Genetics Vol.23 No.11 541

around 40 Mb, 75–80 Mb, 120–130 Mb, 165–170 Mb and in 1 (MLH1) foci (1.85 and 1.54) [13] lends support to this
the 10 Mb region near the centromere-distant telomere interpretation, assuming that MLH1 foci mark >95% or
(Figure S1a). Relative recombination rates (cM/Mb) were virtually all crossovers [14].
higher in males, compared with females, in the centro- Our data provide genetic evidence that the crucial
mere-distant telomeric region between 40 Mb and 80 Mb. parameter in the relative positioning of crossovers along
The reverse trend was observed in most of the remaining bivalents is the interference distance, which is, in turn,
regions of the chromosome. Although the ratio of recombi- related to physical distance along the synaptonemal com-
nation events involved in single and double crossovers did plex. A correlation of synaptonemal complex length with
not vary along the chromosome in female meiosis (Figure interchiasma distance has already been suggested by
S1b), this was not true in males. Except for the 10 Mb Tease and Hultén [12], who did not detect significant
region that borders the telomere, crossovers involved in differences in mean distances between adjacent MLH1 foci
double recombinants were significantly reduced along the that mark chiasmata in female and male germ cells. How-
rest of the chromosome (Figure S1c). ever, although it is likely that all MLH1 foci mark sites of
Figure 1a shows the cumulative rates of double crossing over, there remains some uncertainty whether
crossovers as a function of the intercrossover distance, some crossovers are processed through an alternative, non-
expressed in Mb. On average, the distance between the MLH1 pathway [14] and how the latter will influence
two events was 102 Mb in females and 122 Mb in males. interference. Our genetic data provide the opportunity to
Only 2% of double crossovers were spaced closer than measure interference between inherited double crossovers
40 Mb apart in females and closer than 57 Mb apart in and to describe how the strength of interference varies as a
males. function of distance.
The coefficient of coincidence (Z) is a traditional As evident in Figure 1b, the coefficient of coincidence
measure of interference [1,2]. It is expressed as a ratio exceeds unity for distances between 75 and 120 Mb in
of the frequency with which crossovers occur in a pair of female meiosis and then drops to around or below unity
intervals relative to the marginal frequencies for crossover for distances above 120 Mb; in male meiosis, it exceeds
events in the two intervals. Plotting the coefficient of unity for distances above 112 Mb. Such dependence has
coincidence as a function of intercrossover distance, we been reported [15] using Drosophila data for nine loci.
found substantial differences between female and male Confidence bounds on the coincidence curves (Figure S2)
meioses (Figure 1b). In female meiosis, the interference indicate that this trend is real and not because of sampling
was complete up to 40 Mb and then faded away between 40 variation. Values of Z above unity simply reflect that,
and 77 Mb, with Z = 0.5 at 62 Mb; in males, complete under strong positive crossover interference, crossovers
interference was found up to 57 Mb fading away between tend to be more evenly spaced than random.
57 and 112 Mb with Z = 0.5 at 95 Mb. Investigations in yeast have shown that two groups of
The difference in interference distances explains why genes are important for crossover interference. Mutations
triple crossovers are common in female meiosis but very in the genes of the first group, which consists of ZIP1 [16],
rare in males. To position three crossover events that are MER3 [17] and MSH4 [18], reduce the frequency of cross-
not subjected to strong interference (Z>0.5) in female ing over in addition to abolishing crossover interference,
meiosis requires at least 124 Mb, which is quite possible whereas mutations among the second group, NDJ1/
within the 197 Mb span of Chr 1. However, to accommodate TAM1, TID1 and DMC1, abolish interference without
three such events in males requires spacing that reducing crossing over [19]. It has been shown in mice
approaches the entire length of the chromosome. that mutations in axial elements of the synaptonemal
complex do not affect interference [20], whereas
Interference correlates with the length of bivalents at mutations in proteins of the central element result in
the pachytene stage of meiosis absence of completed crossing over [21,22], which points
The recombination events begin by the initiation of to the possibility that as yet unknown proteins might
double-strand breaks in the leptotene stage of meiosis I, affect both interference distance and synaptonemal com-
and the repair process that resolves intermediates into plex length. If so, these proteins must be arranged line-
crossovers or convertants is completed in the pachytene arly along the chromosome or bivalent to explain
stage. Differences in bivalent lengths between female and interference action over distances on the order of microns,
male meioses have been reported in both humans [12] and either by creating a physical barrier or by controlling
mice [13]. De Boer et al. [13] measured the average length mechanical stress along the bivalents [23]. An additional,
of bivalent 1 in the pachytene stage of meiosis I in mice of interesting candidate has been suggested by a recent
mixed B6/129 genetic background to be 13.7 mm in females study that found an entirely new class of short piRNA
and 10.2 mm in males. Using these estimates, the distri- molecules of 30 nt in length that are abundantly
bution of double crossovers over interference distances in expressed in zygotene and/or pachytene of male meiosis
females and males is not different in terms of physical [24]. Some of them are implicated in maintaining trans-
length of pachytene bivalents, with a minimum interfer- poson silencing in the germline genome [25], but other
ence distance of 2.8 microns in both sexes (Figure 1c and classes could conceivably have a function in regulating
1d). The fact that the average numbers of chiasmata per recombination and/or interference.
bivalent calculated from our data in female and male The relationship between bivalent length at the
meioses (1.97 and 1.60; see Table 1) are in striking agree- pachytene stage and interference distances raises a sig-
ment with the cytological data inferred from mutL homolog nificant issue. The decision about which double-strand
www.sciencedirect.com
542 Update TRENDS in Genetics Vol.23 No.11

breaks will become crossovers and thus the interference 6 Roderick, T.H. et al. (1996) Recombination Percentages and
Chromosomal Assignments. In Genetic Variants and Strains of the
pattern is made early in meiosis, prior to the leptotene-
Laboratory Mouse (Lyon, M.F. and Searle, A.G., eds), pp. 929–1255,
zygotene transition [20] when proteins involved in double- Oxford University Press
strand break repair and Holiday junction resolution 7 Lenormand, T. and Dutheil, J. (2005) Recombination difference
localize between the aligned cores of the two homologous between sexes: a role for haploid selection. PLoS Biol. 3, e36
chromosomes [26]. The correlation between interference 8 Lercher, M.J. and Hurst, L.D. (2003) Imprinted chromosomal regions of
the human genome have unusually high recombination rates. Genetics
and bivalent length implies a relationship between chro-
165, 1629–1632
mosomal organization at the time of double-strand break 9 de La Casa-Esperon, E. et al. (2002) X chromosome effect on maternal
formation and what is seen at the pachytene stage. This recombination and meiotic drive in the mouse. Genetics 161, 1651–1659
could be the case if double-strand breaks are initiated at 10 Gerton, J.L. and Hawley, R.S. (2005) Homologous chromosome
sites that eventually will be aligned between the two interactions in meiosis: diversity amidst conservation. Nat. Rev.
Genet. 6, 477–487
homologous chromosomes. It has been shown that males 11 Broman, K.W. et al. (2002) Crossover Interference in the Mouse.
have longer chromatin loops than females at the pachytene Genetics 160, 1123–1131
stage of meiosis [12], which in turn suggests that they have 12 Tease, C. and Hultén, M.A. (2004) Inter-sex variation in synaptonemal
fewer interloop regions involved in synapsis. If recombina- complex lengths largely determine the different recombination rates in
tion occurs only in interloop regions but not in loops, factors male and female germ cells. Cytogenet. Genome Res. 107, 208–215
13 de Boer, E. et al. (2006) Two levels of interference in mouse meiotic
that determine loop size at earlier stages (before the recombination. Proc. Natl. Acad. Sci. U. S. A. 103, 9607–9612
leptotene-zygotene transition) would affect both inter- 14 Guillon, H. et al. (2005) Crossover and Noncrossover Pathways in
crossover distance and synaptonemal complex length in Mouse Meiosis. Mol. Cell 20, 563–573
a similar manner. 15 McPeek, M.S. and Speed, T.P. (1995) Modeling interference in genetic
recombination. Genetics 139, 1031–1044
16 Sym, M. and Roeder, G.S. (1994) Crossover interference is abolished
Crossover interference and other factors affecting sex in the absence of a synaptonemal complex protein. Cell 79, 283–
differences in recombination: concluding remarks 292
Our data do not preclude the possibility that additional 17 Nakagawa, T. and Kolodner, R.D. (2002) Saccharomyces cerevisiae
factors, such as substantial differences in sex specific rates Mer3 is a DNA helicase involved in meiotic crossing over. Mol. Cell.
Biol. 22, 3281–3291
of recombination on regional scale and at the level of
18 Novak, J.E. et al. (2001) The budding yeast Msh4 protein functions in
individual recombination hotspots [27], X-linked modifiers chromosome synapsis and the regulation of crossover distribution.
[9], or postmeiotic events, such as gametic selection [7], Genetics 158, 1013–1025
might influence sex differences in recombination rates. 19 Shinohara, M. et al. (2003) Crossover interference in saccharomyces
However, the major factor underlying genome-wide sex cerevisiae requires a TID1/RDH54- and DMC1-dependent pathway.
Genetics 163, 1273–1286
differences in recombination rates is crossover interfer- 20 de Boer, E. et al. (2007) Meiotic interference among MLH1 foci requires
ence, which acts on physical rather than genomic dis- neither an intact axial element structure nor full synapsis. J. Cell Sci.
tances. In this sense, the fundamental processes that 120, 731–736
regulate positioning of multiple crossovers along the mam- 21 Bolcun-Filas, E. et al. (2007) SYCE2 is required for synaptonemal
malian chromosomes appear to be the same in female and complex assembly, double strand break repair, and homologous
recombination. J. Cell Biol. 176, 741–747
male meiosis. 22 de Vries, F.A. et al. (2005) Mouse Sycp1 functions in synaptonemal
complex assembly, meiotic recombination, and XY body formation.
Supplementary data Genes Dev. 19, 1376–1389
Supplementary data associated with this article can be 23 Kleckner, N. et al. (2004) A mechanical basis for chromosome function.
Proc. Natl. Acad. Sci. U. S. A. 101, 12592–12597
found, in the online version, at doi:10.1016/j.tig.2007.08.015.
24 Girard, A. et al. (2006) A germline-specific class of small RNAs binds
mammalian Piwi proteins. Nature 442, 199–202
References 25 Carmell, M.A. et al. (2007) MIWI2 is essential for spermatogenesis and
1 Muller, H.J. (1916) The mechanisms of crossing-over. Am. Nat. 50, 193– repression of transposons in the mouse male germline. Dev. Cell 12,
221 503–514
2 Sturtevant, A.H. (1915) The behavior of the chromosomes as studied 26 Moens, P.B. et al. (2007) Initiation and resolution of interhomolog
through linkage. Z. Indukt. Abstammungs. Vererbungsl. 13, 234–287 connections: crossover and non-crossover sites along mouse
3 Zhao, H. et al. (1995) Statistical analysis of crossover interference using synaptonemal complexes. J Cell Sci 120, 1017–1027
the chi-square model. Genetics 139, 1045–1056 27 Shiroishi, T. et al. (1990) Recombinational hotspot specific to
4 Broman, K.W. et al. (1998) Comprehensive human genetic maps: female meiosis in the mouse major histocompatibility complex.
individual and sex-specific variation in recombination. Am. J. Hum. Immunogenetics 31, 79–88
Genet. 63, 861–869
5 Donis-Keller, H. et al. (1987) A genetic linkage map of the human 0168-9525/$ – see front matter ß 2007 Elsevier Ltd. All rights reserved.
genome. Cell 51, 319–337 doi:10.1016/j.tig.2007.08.015

www.sciencedirect.com