Female Germ Cell Aneuploidy and Embryo Death in Mice Lacking the Meiosis-Specific Protein SCP3

See allHide authors and affiliations

Science  10 May 2002:
Vol. 296, Issue 5570, pp. 1115-1118
DOI: 10.1126/science.1070594


Aneuploidy (trisomy or monosomy) is the leading genetic cause of pregnancy loss in humans and results from errors in meiotic chromosome segregation. Here, we show that the absence of synaptonemal complex protein 3 (SCP3) promotes aneuploidy in murine oocytes by inducing defective meiotic chromosome segregation. The abnormal oocyte karyotype is inherited by embryos, which die in utero at an early stage of development. In addition, embryo death in SCP3-deficient females increases with advancing maternal age. We found that SCP3 is required for chiasmata formation and for the structural integrity of meiotic chromosomes, suggesting that altered chromosomal structure triggers nondisjunction. SCP3 is thus linked to inherited aneuploidy in female germ cells and provides a model system for studying age-dependent degeneration in oocytes.

Aneuploidy is the leading known cause of pregnancy loss and has been recorded in ∼25% of all conceptions and 0.3% of all newborns (1). Aneuploidy in germ cells is predominantly due to aberrant female meiotic chromosome segregation, where increasing maternal age represents a well-documented risk factor. Homologous chromosomes (homologs) in meiotic cells undergo pairing (called synapsis) and recombination during which physical links, chiasmata, are established between them. Synapsis is aided by a meiosis-specific protein complex, the synaptonemal complex (SC), which comprises two axial lateral elements (AEs) and a central element (CE) (2). The two AEs, which colocalize with the sister chromatids of each homolog, become connected along their entire length by fine fibers called transversal filaments (TFs) at the pachytene stage of meiotic prophase I (3). The AE is composed of discrete protein filaments, organized by the cohesin complex (4) or by two meiosis-specific proteins, synaptonemal complex protein 2 (SCP2) and SCP3 (5). SCP3 is required for AE formation and for male fertility (6–8). The function of SCP3 during the meiotic cell divisions is unclear, however, becauseSCP3–/– male germ cells die around the zygotene stage of meiosis (8).

Here, we analyze the functional role(s) of SCP3 in meiotic chromosome segregation in female germ cells. Twelve-week-old wild-type and SCP3-deficient females were mated with wild-type males. In contrast to SCP3-deficient males,SCP3-deficient females were fertile and generated healthy offspring (Fig. 1A). However, theSCP3-deficient females exhibited a sharp reduction in litter size, generating on average 5.9 offspring per female, compared with 8.9 offspring for their wild-type siblings (Fig. 1A). This reduction could be due to ovarian failures, resulting in functional oocyte loss. A comparison of ovarian morphology in SCP3-deficient and wild-type females indicated, however, that follicular development is not affected in SCP3-deficient females (9). To investigate the possibility of embryonic death in utero,SCP3-deficient or wild-type females (12 weeks of age) were mated with wild-type males. An embryo count at 11.5 days post coitum (dpc) revealed that SCP3-deficient females contained an average of 2.6 resorbed embryos, whereas resorbed embryos were rarely observed in wild-type females (Fig. 1B). The reduced litter size inSCP3-deficient females is therefore due to embryo death in utero. Older (24-week-old) SCP3-deficient females showed a further reduction in offspring number (Fig. 1A). To see if this could be explained by increased embryo death in utero, the numbers of live and dead embryos in 12-, 16-, and 24-week-old females were compared (Fig. 1B). We found that older SCP3-deficient females contained more resorbed, and fewer live, embryos. We conclude that absence of SCP3 affects embryo survival in an age-dependent manner.

Figure 1

Absence of SCP3 affects embryo survival. (A) 12- and 24-week-old wild-type and SCP3-deficient females were mated with wild-type males and the number of pups born counted (13). Twenty-eight wild-type and 26SCP3-deficient 12-week-old female, as well as 15 wild-type and SCP3-deficient 24-week-old females, respectively, were used in these studies. SCP3-deficient females generated significantly fewer offspring compared to wild-type females of the same age (P < 0.001, Student's two-tailedt-test). Furthermore, 24-week-oldSCP3-deficient females produced significantly fewer offspring than 12-week-old SCP3-deficient females (P < 0.05, Student's two-tailedt-test). (B) 12-, 16-, and 24-week-oldSCP3-deficient and wild-type female siblings were mated with wild-type males. Seventeen, 11, and 16 wild-type and 16, 12, and 16SCP3-deficient females were used for each respective age group. Pregnant females were sacrificed at 11.5 dpc (13).SCP3-deficient females had significant embryo resorption compared to wild-type females of corresponding age (P< 0.001, Student's two-tailed t-test). The difference in the number of embryos being resorbed inSCP3-deficient females at different ages was significant [P < 0.05, one-way analysis of variance (ANOVA) with fixed effects], whereas no significant difference in embryo resorption number were seen for wild-type females (one-way ANOVA with fixed effects). Data represent mean ± SEM.

We postulated that the embryo death observed inSCP3-deficient females could be due to generation of aneuploid oocytes resulting from chromosomal segregation errors. Karyotypic status was monitored in one-cell zygotes, which contain discrete maternally and paternally derived pronuclei, enabling independent assessment of the respective karyotypes. Seven of 17 one-cell zygotes derived from SCP3-deficient females displayed chromosomal aberrations, compared with 0 of 90 one-cell zygotes from wild-type females (Fig. 2A). We propose that SCP3 deficiency in oocytes during meiosis results in inherited aneuploidy leading to embryo death. Chromosomal abnormalities in one-cell zygotes from SCP3-deficient oocytes are likely caused by recombinational errors during meiosis. To find out when these errors occur, we first investigated metaphase I (MI) oocytes from wild-type and SCP3-deficient females (Fig. 2B). Forty-seven of 93 MI oocytes from 13- to 14-week-old mutant females contained univalents (i.e., homologs not held together by chiasmata), whereas no univalents were found in 75 oocytes from wild-type females. Most of the affected MI oocytes contained one to three pairs of univalents (Fig. 2B). Furthermore, chiasmata distribution in 18 MI wild-type and 54SCP3-deficient oocytes, indicated that the number of bivalents having a double exchange decreased two- to threefold inSCP3-deficient oocytes (0.6 ± 1.0, SEM) compared to wild-type (1.6 ± 1.3, SEM). These data show that chiasmata number is slightly reduced in SCP3-null oocytes. Because short chromosomes are held together by only one chiasma (10), they are at greater risk of being converted into univalents. Indeed, analysis of SCP3-null MI oocytes showed that a majority of the identified univalents are short chromosomes (Fig. 2B).

Figure 2

One-cell zygotes and oocytes from SCP3-deficient females are chromosomally abnormal. 12- to 16-week-old wild-type or SCP3-deficient females were paired with wild-type males. Cells derived from wild-type females are indicated by +/+, whereas cells taken from SCP3-deficient females are indicated by –/–. (A) Zygotes at the one-cell stage were isolated and the mitotically arrested chromosomes karyotyped (13). The individual pronuclei are indicated by a dotted line; the arrow indicates a pronucleus with one extra chromosome (a total of 21 chromosomes). MI (B) or diakinesis (C) oocytes were analyzed for achiasmatic chromosomes (13). Arrows in (B) and (C) indicate univalent chromosomes.

To define whether formation of univalent chromosomes is due to a failure to establish or to maintain chiasmata during meiosis, we looked for achiasmatic homologs in oocytes at the diakinesis stage of meiosis. At this stage, the spindle has not yet become attached to the chromosomes, reducing the risk of losing preformed chiasmata. Approximately 50% of the SCP3-deficient diakinesis oocytes contained univalents, whereas virtually no univalents were observed in wild-type oocytes (Fig. 2C). This strongly suggests that the achiasmatic homologs formed in SCP3-deficient MI oocytes are due to a failure to establish chiasmata between homologous chromosomes.

Absence of SCP3 thus appears to have a subtle but distinct effect on meiotic recombination. The recombination protein DMC1 (11) forms similar arrays of foci in wild-type andSCP3–/– zygotene oocytes (9), however, indicating that early meiotic recombination activities were not substantially affected in mutant oocytes. Recombination in wild-type and SCP3-deficient pachytene oocytes was monitored, using the recombination protein MLH1 (Fig. 3A), a marker for the chiasma structure (12). MLH1 foci are formed in the absence of SCP3 and the foci associate with synapsed meiotic chromosomes labeled by SCP1, a core component of the TF of the SC (2) (Fig. 3A). We observed a slight increase in the average number of MLH1 foci associated with SCP1 fibers in mutant oocytes (26.0 ± 1.5, SEM) compared to wild-type cells (24.2 ± 0.8, SEM), as well as a larger variability in foci number between individual mutant cells (9). Finally, we measured crossing-over frequency in offspring from SCP3–/– and wild-type females, using polymorphic microsatellite markers. The F2 offspring and their parents were typed using four pairs of polymorphic microsatellite markers on four separate chromosomes (Table 1). Despite a slight variability in the recombination rates for the different chromosomes analyzed, the crossing-over process is clearly active in mutant oocytes at about wild-type levels. We conclude that the recombination process inSCP3-null oocytes is functional and only subtly affected by the absence of SCP3.

Figure 3

Altered MLH1 and SCP1 expression inSCP3-deficient oocytes. Cells derived from wild-type females are indicated by +/+, whereas cells taken fromSCP3-deficient females are indicated by –/–. (A) SCP3-deficient or wild-type oocytes were isolated, fixed and analyzed by immunofluorescence microscopy (13). The distribution of MLH1 (red/white dots) and SCP1 (blue fibers) are shown for pachytene oocytes. Arrowheads indicate axial interruptions of the SCP1 fiber. Because of the axial interruptions seen in the SCP1 fiber, it was not possible to determine with certainty whether some homologs completely lacked MLH1 foci. (B) SCP1 fiber length increases in SCP3-deficient oocytes. Total SCP1 fiber length in 10 individual wild-type orSCP3-deficient pachytene oocytes was measured and average fiber length calculated. Data is presented as mean ± SEM.

Table 1

Loss of SCP3 in oocytes has a subtle, but distinct effect on crossing-over frequency. BackcrossedSCP3 +/– females (C57BL/6NCrlBR) were mated toSCP3 +/– males (C57BL/6NCrlBR×129/Ola) to generate F1 hybrid mice (13). F1hybrid female mice nullizygous or wild type for SCP3 were mated with wild-type males (C57BL/6NCrlBR) to generate F2offspring. Tail DNA from F2 offspring and their parents was PCR typed using four pairs of polymorphic microsatellite markers for chromosomes 12, 15, 17, and 18. Crossing-over frequency was evaluated by χ2 testing derived from the 2 × 2 contingency test.

View this table:

Lack of SCP3 could alter SC structure, thereby disturbing chiasma formation and chromosome segregation. SCP1 formed linear filaments inSCP3-null pachytene oocytes. However, these filaments frequently contained axial gaps (Fig. 3A), in some cases separating the SCP1 fiber from the terminally-located centromeric region of the chromosomes (9). These axial gaps suggest that short stretches of the homologs synapse incorrectly without SCP3. SCP1 fiber length in SCP3-deficient pachytene oocytes increased twofold over the wild-type (Fig. 3B), indicating that chromatin compaction of the axial core is severely affected in the absence of SCP3. To monitor the effects of SCP3 deficiency on SC formation, silver-stained wild-type and SCP3-deficient pachytene oocytes were compared. SC structures seen in wild-type oocytes were found to be absent in SCP3-deficient oocytes (13). Furthermore, immunofluorescence microscopy analysis of SCP2 did not produce the expected filamentous signal inSCP3-deficient pachytene oocytes, instead forming nuclear foci, approximately half of which colocalized with a centromere marker (13). We found that SCP3 is required for AE formation, SCP2 integration into the AE, and uninterrupted SCP1-mediated synapsis.

Despite the prevalence of aneuploidy (1), very little is known about the causes of chromosomal nondisjunction in female germ cells. We show here, in a mouse model system, a direct link between SCP3-deficiency in females and the fate of their offspring. We demonstrate that a slight decrease in chiasma number leads to increased recombination failure for the smallest chromosomes and to a corresponding increase in aneuploid embryos in SCP3-null females.

What then is the role of SCP3 and how does its absence cause chromosomal non-disjunction? SCP3 expression is restricted to meiotic prophase I cells, disappearing during diakinesis in female oocytes (14). The structural changes seen in meiotic chromatin due to SCP3 absence strongly suggest that nondisjunction is caused by errors that initially occur at pachytene. It was also found that the SCP1 fiber in SCP3-deficient pachytene oocytes contained axial gaps, indicating incomplete homolog synapsis. Incorrect homolog alignment could affect the crossing-over process that generates chiasmata. SCP1 homologs in yeast (Zip1) and Drosophila[C(3)G] have been shown to be important for both synapsis and crossing-over, and their absence results in a high level of nondisjunction (15, 16). Mutations in these two proteins that introduce noncontiguous axial gaps in their respective TF structures reduce the levels of meiotic exchange and crossover interference. In addition, mutations in genes that affect crossover interference randomize crossover distribution between chromosomes, such that chromosomes may fail to cross over and therefore nondisjoin (2, 17, 18). The observed changes in chromatin structure and SCP1 organization along the homologs in SCP3-null oocytes could therefore affect crossover interference, affecting both chiasma positioning and frequency.

The increased embryo death in older SCP3–/– females is notable. A reduced oocyte pool in humans has been proposed to result in a premature increase in aneuploidy and embryo death (19). By this model, a shortened reproductive life span, resulting from a substantial loss of oocytes in the developingSCP3–/– ovary could explain the increased embryo death. Alternatively, SCP3 absence could have two effects on chromosomal nondisjunction, an early effect seen as a reduction in chiasmata number and a later effect that only manifests itself in older females. The second model could be explained as a two-hit mechanism, as previously suggested for aging human oocytes (20). The first hit would be due to a reduction in crossover interference resulting in suboptimal positioning of chiasmata along the chromosomes, whereas the second hit would occur due to an inability of older oocytes to resolve chiasmata located at certain chromosomal regions. The model system described here should thus be valuable toward understanding the basis for meiotic nondisjunction and age-dependent aneuploidy in human oocytes.

  • * Present address: Molecular Sciences, Astra Zeneca R&D, SE-151 85 Södertälje, Sweden.

  • To whom correspondence should be addressed. E-mail: christer.hoog{at}


View Abstract

Stay Connected to Science

Navigate This Article