Kinetochore and chromosome separation during abnormal meiosis
The present study and previous reports show that all the abnormal meiosis have interesting features although each have some its own characteristics in distant F1 hybrids (Zhou et al. 2008; Luo et al. 2013), allotriploid (Cui et al. 2022) and odd-allotetraploid lilies (Xiao et al. 2022). 1) Univalents are quite common at metaphase I and their sister-chromatids usually separated and moved to the two opposite poles at anaphase I. At metaphase I, in distant F1 hybrids, such as LA and OA, most homoeologous chromosomes do not pair and thus form univalent (Barba-Gonzalez et al. 2004; Zhou et al. 2008); In allotriploid (LLO), all the homologous L-chromosomes prefer to pair each other and O-chromosomes tend to form univalent (Cui et al. 2022); Similarly, odd-allotetraploid (LAAA), all the homologous A-chromosomes prefer to associate multivalents and L-chromosomes tend to be univalent (Xiao et al. 2022). Intriguingly, the sister-chromatids of such univalents are clearly separated and moved to the opposite poles at anaphase I. The same phenomenon is also observed in allotriploid Alstroemeria (Kamstra et al. 2004). This is totally different from normal meiosis I, in which the chromosomes of bivalents separate rather than sister-chromatids. Univalents are commonly reported in plant F1 hybrids, and they are the reason for 2n gametes resulting from an abnormal meiosis in many plants (Ramanna and Jacobsen 2003), such as Trifolium (Ansari et al. 2022), wheat/rye hybrids (Silkova et al. 2013), etc. Interestingly, the sister-chromatid separations of univalents, which are formed by premature separation of bivalents at meiosis I, are observed in aging-related oocytes of mice and human, and the cohesion loss and splitting of sister kinetochores are regarded as the reasons for the abnormal phenomenon (Zielinska et al. 2015; Nakagawa and FitzHarris 2017). Akera and Lampson (2016) suggested that sister kinetochores should have been fused during normal meiosis I and sister kinetochores of univalents were split at abnormal meiosis I (Fig. 6(1)). However, according to GISH analysis on lily abnormal meiosis, it is more plausible to explain the abnormal meiosis I in a modified way shown as in Fig. 6(2). This is because a centromere on each chromosome has two kinetochores which are attached to spindle microtubules in normal mitosis; however, during normal meiosis, a centromere on each chromosome of a bivalent only has one kinetochore attached to spindle microtubules (www.wikipedia.org). Possibly, one chromosome has one kinetochore in its centromere, and kinetochore is possibly duplicated accompanying DNA replicating in mitosis (Fig. 6(2d)); however, during normal meiosis I, kinetochore duplication is hindered by synapses (Fig. 6(2e)) or the two duplicated kinetochores on sister chromatids are so close and function as one due to cohesion and chiasmata; and during abnormal meiosis I (Fig. 6(2f)), some bivalents premature and become univalent, or some chromosomes do not pair and remain univalents, and their kinetochores could replicate because no chiasmata or cohesion suppress kinetochore duplication. Since then, it is reasonable that both univalents and bivalents are disjoined and move to opposite poles during abnormal meiosis I in F1 distant lily hybrids (Zhou et al. 2008). 2) Bivalents are also common in allotriploid lilies, and are also formed in F1 distant LA or OA hybrids at metaphase I. They are usually disjoined at anaphase I as normal meiosis (Zhou et al. 2008; Xiao et al. 2022; Cui et al. 2022). The similar result is also reported in allotriploid Alstroemeria (Kamstra et al. 2004). 3) Trivalents or other multivalents are often found in odd-allotetraploid (Xiao et al. 2022), and occasionally occur in allotriploid (Cui et al., 2022) and distant hybrids (Zhou et al., 2008). They are disjoined evenly or unevenly at anaphase I (Zhou et al. 2015). Besides, lagging chromosomes and micronuclei are common in abnormal meiosis not only in lily (Zhou et al. 2008; Zhang et al., 2017; Cui et al. 2022; Xiao et al. 2022), but also in rice hybrids (Liu et al. 2021), Saccharum hybrids (Li et al. 2021), Populus hybrids (Wang et al. 2015), and Musa (Ahmad et al. 2021).
The partial female fertile of aneuploid lilies
It is confirmed that F1 distant hybrids can be female parents to be backcrossed and produce allotriploid lilies regardless their male sterility (Zhou 2007; Liu et al. 2021); Similarly, triploid, allotriploid and odd-allotetraploid can be female parents to hybridize with appropriate males though they are highly male sterile (Lim et al. 2000; 2003; Khan et al. 2009; Natenapit et al. 2010; Xie et al. 2010; Zhou et al. 2011, 2012, 2014; Chung et al. 2013; Wang et al. 2015; Suzuki and Yamagishi 2015; Xi et al. 2015; Xiao et al. 2019; Cui et al. 2022). The phenomena seemingly look strange, not only because triploids are usually seedless, but also one plant’s male fertility should be similar to its female fertility due to same genetic materials, i.e., same meiosis of gametogenesis. So, why are there such big difference between male and female fertility in these lilies? The question was well explained by comparative analysis between Fritillaria embryo sac and Polygonum embryo sac (Zhou 2007). According to megasporogenesis, in a polygonum embryo sac, the nucleic DNA of its central cell is twice that of its egg cell, while in a fritillaria embryo sac, the nucleic DNA of its central cell is twice that its somatic cell (Fig. 7) (Zhou 2007; Zhou et al. 2011, 2012). So, for triploid watermelon as an example of polygonum-type plant, in its embryo sac, both egg and central cell are aneuploid, once double fertilized with a diploid, both embryo and endosperm are aneuploid; thus, triploid watermelon is seedless. By contrast, for a triploid lily, in its fritillaria-type embryo sac, its egg is aneuploid, but its central cell is hexaploidy; Once double fertilized with a diploid or tetraploid, endosperm is euploid and it develops well, and then make some aneuploid embryos survival (Zhou 2007; Zhou et al. 2011, 2012). This is the reason why all the male sterile lilies have some partial female fertility. In the present study, the fertility of the aneuploid lily is similar to that of triploid or odd-allotetraploid lilies. Once embryo sacs of an aneuploid lily are double fertilized by diploid or tetraploid male, both embryo and endosperm are aneuploid; surprisingly, some seeds develop well rather than seedless in triploid watermelon. How to explain its partial female fertility? Aneuploid lilies are usually less vigor and die early, however some grow well, like J1614 in the present research, indicating that it has balanced genes though it is an aneuploid. The nucleic DNA of each central cell in its embryo sac is invariably twice that of a somatic cell, meaning that all its central cells have balanced genes. Once fertilized with tetraploid male, its endosperm has the balanced genes and could develop and make aneuploid embryos survival (Fig. 7).
Chromosome numbers of progenies or functional gametes of aneuploid lilies
In the present study, J1614, containing 48 chromosomes, is aneuploid because of its unbalanced chromosomal composition. Its progenies have variable chromosomes, ranging from 46–53, when tetraploid as male. This is like aneuploid progenies obtained from 2x/4x × aneuploid (Zhong et al. 2022) and 3x × 2x/4x hybridizations (Lim et al. 2000, 2003; Khan et al. 2009; Xie et al. 2010; Natenapit et al. 2010; Zhou et al. 2011, 2012, 2014; Chung et al. 2013; Wang et al. 2015; Suzuki and Yamagishi 2015; Xi et al. 2015; Cui et al. 2022). All of them indicate that functional gametes produced by aneuploid or triploid lilies usually have much higher chromosome numbers when tetraploid used as male than when diploid used as male (Lim et al. 2003; Khan et al. 2009; Zhou et al. 2011, 2012; Wang et al. 2015; Xi et al. 2015). Besides, 3x × 4x is much more successful than 3x × 2x in Lilium (Zhou et al. 2011, 2012). The present study also shows that aneuploid × 4x is much more successful than aneuploid × 2x in Lilium.