Introduction

Cytomixis, a migration of cytoplasmatic or nuclear materials between adjacent cells, is an intricate phenomenon well reported in plants (Lattoo et al. 2006; Singhal and Kumar 2008; Pierre and Sousa 2011; Guan et al. 2012; Mursalimov et al. 2013; Mursalimov and Deineko 2015). This natural process was reported for the first time around 100 years ago by Kornicke (1901) although the phenomenon was named by Gates (1911). Cytomixis was observed in different plant groups including Pteridophyta, Gymnosperms, Monocots, and Dicots. More specifically, the process was mainly reported in hybrids, aneuploids, mutants, polyploids, and apomictic individuals (de Nittancourt and Grant 1964; Gottschalk 1970; Mantu and Sharma 1983; Li et al. 2009; Pierre and Sousa 2011). Cytomixis was most frequently observed during meiosis I of microsporogenesis (Koul 1990; Lattoo et al. 2006; Kumar et al. 2010). Although cytomixis is not common in somatic cells, the phenomenon was also reported in root tips (Jacob 1941; Sarvella 1958; Tarkowska 1960; Kostritsyna and Soldatov 1991; Guzicka and Wozny 2005) and tapetal cells, among others (Cooper 1952).

The transference of DNA occurs by cytoplasmatic connections called cytomictic channels. These channels derive from an atypical plasmodesmata structure, which persists throughout meiosis (Risueno et al. 1969; Pierre and Sousa 2011). The origin of cytomictic channels is still unknown and different explanations have been described: failure in cellular wall formation during cytokinesis (Sarvella 1958; Bisalpufra and Stein 1966); adjacent plasmodesmata fusion (Wang et al. 1998) and defects in callose synthesis (Sheidai and Fadaei 2005; Sidorchuk et al. 2007). In general, cytomictic channels connect two or more cells in the same division stage (Souza and Pagliarini 1997) but some reports also indicate the exchange of genetic materials between cells in different stages (Song and Li 2009; Lone and Lone 2013).

Cytomixis may be influenced by stress conditions such as physical and chemical factors (Bobak and Herich 1978; Narain 1979; De and Sharma 1983; Dwivedi et al. 1988; Souza and Pagliarini 1997; Bhat et al. 2006; Gulfishan et al. 2010). Independently of its origin, cytomixis results in serious genetic consequences. When it occurs in pollen mother cells (PMCs) it frequently affects gametogenesis and fertility (Mursalimov et al. 2013; Mursalimov and Deineko 2015). Whereas the normal product of meiosis division must possess half of the DNA of the mother cells, the transference of genetic material between cells may result in gametes with increased or decreased DNA content leading to unbalanced gametes (Mursalimov and Deineko 2015). Many reports indicate the presence of unbalanced or unreduced pollen originated from meiotic irregularities, mainly in individuals with cytomixis (Bell 1964; Lattoo et al. 2006; Mursalimov and Deineko 2015). In an evolutionary context, cytomixis has been studied over time (Kamra 1960) assuming an important role in new species formation, especially by polyploidization via unreduced gametes (Bell 1964; Falistocco et al. 1995; Lattoo et al. 2006).

Lippia alba (Verbenaceae), an important aromatic shrub used in Brazilian folk medicine, is characterized by various chemotypes as well as large phenotypic plasticity and genome variation (Pierre et al. 2011; Reis et al. 2014; Viccini et al. 2014). A recent study published by Reis et al. 2014 indicates the occurrence of a putative polyploid complex naturally formed in this species with diploids (2n = 30; 64 individuals), aneuploid (2n = 38; one individual), triploids (2n = 45; 16 individuals), tetraploids (2n = 60; 24 individuals), and one hexaploid individual (2n = 90), collected in 14 states of the five major geographical regions of Brazil (North, Northeast, Center-West, Southeast, and South). The authors hypothesized an autopolyploid origin of each chromosome number by unilateral and bilateral crossing between cytotypes via unreduced gametes. Nevertheless, the mechanism involved in the formation of unreduced pollen is still unclear and presents an interesting point to be elucidated, which could help understand the origin of the cytotypes.

Here, we describe, for the first time, the cytomixis phenomenon in microsporogenesis of L. alba, one possible mechanism associated with the formation of unreduced gametes.

Materials and methods

Biological material

The tetraploid accession of L. alba (BGEN-92, 2n = 4x = 60) was cultivated in standard conditions at the Plant Experimental Station of the Universidade Federal de Juiz de Fora (UFJF), Brazil. Flower buds were collected in the morning and immediately fixed in cold ethanol: acetic acid solution 3:1 (v/v) for a minimum of 24 h.

Meiotic preparation

Slides were prepared by squash technique. The anthers were excised from the flower buds in a drop of 45 % acetic acid, squashed using a coverslip and stained with 5 % acetocarmine. Chromosomal abnormalities of different meiotic stage cells (zygotene, metaphase I, anaphase I, metaphase II, anaphase II, and telophase II) were quantified considering at least 100 cells per stage. The images were registered using the CellSens software (Olympus).

Pollen viability

Pollen viability was investigated according to Heslop-Harrison and Heslop-Harrison (1970) with minor modifications. Mature and fresh anthers were excised and stained with fluorescein diacetate dissolved in acetone (2 mg/mL) combined with 10 % sucrose for 30 min. The fluorescent pollen were considered viable and the non-fluorescent ones as unviable. At least 1000 pollen grains were analyzed using a BX 51 microscope (Olympus) with appropriate filter. The diameter of ca. 100 grains was measured. The images were digitalized using the CellSens software (Olympus).

Results and discussion

Here, we describe for the first time the cytomixis phenomenon in a tetraploid individual of L. alba. In this accession, the process occurs specifically in prophase I and among meiocytes only during the zygotene stage (Fig. 1a–f). Approximately 80 % of zygotene cells showed an exchange of genetic material (Table 1). Although previous reports revealed that this species showed a high index of irregularities during meiosis and unviable pollen (mainly in polyploid cytotypes), no event of cytomixis has been observed so far (Brandão et al. 2005; Pierre et al. 2011; Reis et al. 2014). In our study, in addition to cytomictic cells observed at zygotene stage, the analysis of meiotic behavior revealed a high percentage of irregularities at subsequent stages of prophase I: precocious migration and lost chromosomes (metaphase I); laggard chromosomes, chromosome bridge and multipolar anaphase (anaphase I); chromosomal losses (metaphase II); late migration (anaphase II); dyads, triads, polyads and micronuclei (telophase II) (Table 1).

Fig. 1
figure 1

Meiotic irregularities in a tetraploid accession of Lippia alba. a–f cytomixis among PMCs at zygotene stage cell; g PMC with micronucleus; h multipolar anaphase I with chromosome loss; i multipolar anaphase I with chromosome bridge and chromosome losses; j anaphase I with lagging chromosomes; k triad; l dyad; m viable pollen; n unviable pollen with reduced size; o unviable pollen with irregular shape. Black arrows indicate micronuclei; asterisk indicates empty PMC after transference of chromatin; white arrow indicates chromosome bridge; arrowheads indicate chromosome losses. Bar 5 µm (a–l); bar 10 µm (m–o)

Full size image
Table 1 Meiotic irregularities observed at different stages in a tetraploid accession of L. alba
Full size table

Many studies reported the formation of cytomictic channels in PMCs, mainly in prophase I (Falistocco et al. 1995; Wang et al. 2002; Bellucci et al. 2003; Sidorchuk et al. 2004; Imeri-Buiza 2007; Negron-Ortiz 2007; Song and Li 2009). However, species such as Nicotiana tabacum, Chlorophytum comosum, Medicago sativa, and Meconopsis aculeate showed communications between nuclei in interphase, in different stages of meiosis I and rarely in meiosis II (Bellucci et al. 2003; Lattoo et al. 2006; Singhal and Kumar 2008; Guan et al. 2012; Mursalimov et al. 2013). In the individual investigated here, cytomixis process involved from two to 16 adjacent cells and the cytomictic channels varied in length and number (Fig. 1a–f). In some cases the nuclei migration resulted in almost empty cells and others apparently showed twice the amount of DNA content (Fig. 1f). After chromatin transfer, the receptor cell frequently shows condensed micronucleus and unorganized chromatin (Fig. 1a). The migration of chromatin from one cell to another directly interferes with its function and development. The increase or reduction of DNA content in PMCs may cause problems in microsporogenesis and consequently failure in the formation of balanced gametes (Bellucci et al. 2003; Singhal and Kumar 2008; Ferreira et al. 2009; Kumar et al. 2010). The most common abnormalities observed were condensed micronucleus, desynapsis, chromosome stickiness, incorrect pairing, unequal segregation, laggard chromosomes, precocious migration, and chromosome bridge (Lattoo et al. 2006) as was observed in a tetraploid accession of L. alba investigated here (Table 1; Fig. 1g). All of these abnormalities can result in unreduced or unbalanced gametes (Bretagnolle and Thompson 1995).

The frequency of cytomixis can vary within the same species with different ploidy levels (Lavia et al. 2011; Guan et al. 2012). Jeelani et al. (2011), for example, observed different cytomixis indices in diploid and tetraploid cytotypes of Silene conoidea, S. edgeworthii, and S. vulgaris (Caryophyllaceae). In Ranunculus hirtellus (Ranunculaceae) the tetraploid individuals showed transfers of chromatin and abnormal microsporogenesis, while diploids exhibited normal meiosis (Kumar and Singhal 2011). In L. alba, up to now, only the tetraploid cytotype showed chromatin migration in PMCs.

Cytomixis is usually observed in plants with some level of genetic, physiological, or biochemical instability such as haploids, triploids, hybrids, mutants, aneuploids, and apomicts individuals (De Nittancourt and Grant 1964; Gottschalk 1970; Mantu and Sharma 1983). In L. alba, meiotic instability, genome restructuring, and chromosome rearrangements were already reported in polyploids (Reis et al. 2014). Genomic rearrangements are common in neopolyploids and tend to stabilize after successive generations of karyotype “adjustment” (Soltis and Soltis 2009; Chester et al. 2010, 2013; Lipman et al. 2013). This nuclear restructuring is an adaptation to the new conditions imposed by the increase of DNA content (Chester et al. 2013). Therefore, for these individuals, considering that cytomixis allows a large scale of DNA losses and addition, this process can help genome adjustments, accelerating the adaptation/stability of new polyploids.

As outlined above, cytomixis is a potential mechanism to generate meiotic instability and to form abnormal gametes. Thus, the chromatin transfer between adjacent PMCs can result in unreduced or unbalanced gametes that could contribute to the formation of aneuploids and polyploid individuals (Bell 1964; Lattoo et al. 2006). It is well known that polyploidy is the major driver of speciation in plants (Masterson 1994; Soltis et al. 2003; Mittelbach et al. 2007; Soltis et al. 2009; Soltis and Soltis 2009; Jiao et al. 2011), and cytomixis might be one of mechanisms that contributes to the formation of new polyploids via unreduced gametes (Bretagnolle and Thompson 1995; Falistocco et al. 1995; Ramsey and Schemske 1998; Ghaffari 2006; Lattoo et al. 2006). Several papers reported the formation of viable 2n pollen in cytomictic plants and also that cytomixis was directly associated with the formation of new cytotypes (Falistocco et al. 1995; Zhou 2003; Ghaffari 2006; Lattoo et al. 2006; Negron-Ortiz 2007; Singhal and Kumar 2008; Song and Li 2009; Singhal et al. 2011; Mursalimov and Deineko 2015).

In general, unreduced pollen grains show larger volume and diameter than haploid ones. On the other hand, cytomixis can also produce pollen grains with reduced size (pollen with less DNA) (Ramsey and Schemske 2002; Ramanna and Jacobsen 2003; Soltis et al. 2004; Larrosa et al. 2012). In wild potatoes (Solanum tuberosum), for example, the authors suggested that pollen grains larger and smaller than the average size formed diploids and aneuploids gametes contributing to the formation of new hybrids (Larrosa et al. 2012). In our results with L. alba, the fluorescein diacetate test revealed a high percentage of unviable pollen (ca. 87 % of 1069 grains analyzed). Moreover, most of the pollen grains showed irregular shape and heterogeneous size (Fig. 1m–o). A large variation was also observed in pollen diameters both for viable (from 7.65 to 65.44 µm) and unviable (from 4.87 to 54.84 µm) ones (Fig. 2), suggesting the existence of variation in the DNA content of pollen grains (Ramsey and Schemske 2002; Ramanna and Jacobsen 2003; Larrosa et al. 2012).

Fig. 2
figure 2

Variation of diameter length for viable and unviable pollen grains in a tetraploid accession of Lippia alba. The box plots show the median (central horizontal line inside the boxes); the interquartile range (white box) and the outlier . The boxes comprise 50 % of the values (upper and lower; 25 % quartile around the median). Measured values larger than the upper quartile are plotted as outlier

Full size image

In L. alba complex, a hypothesis to explain the formation of different cytotypes involves unilateral and bilateral crossings via unreduced pollen (Reis et al. 2014). We suggest that the irregularities at telophase II (dyad, triad, and polyads) observed in our data constitute an evidence of unreduced and unbalanced pollen formation. According to the hypothesis of polyploid complex formation, the tetraploid individuals can originate hexaploids (6x = 90) via union of unreduced and reduced gametes (n = 60 + n = 30) (Fig. 3). Triploid individuals (3x = 45) could be a result of crossing involving reduced gametes between diploids and tetraploids cytotypes (n = 15 + n = 30) (Fig. 3). In addition, aneuploid individual might be formed by unbalanced gametes with aneuploid chromosome numbers, as observed in the individual with 38 chromosomes (Fig. 3; Reis et al. 2014). In L. alba, giant pollen (possibly with somatic DNA amount) were previously quantified and the authors suggested that they were a consequence of dyads and triads observed in meiosis (Reis et al. 2014). Here, we also observed pollen grains that were bigger and smaller than those previously reported for L. alba (Pierre et al. 2011), giving an additional evidence to aneuploid gametes formation.

Fig. 3
figure 3

Representative scheme for the origin of Lippia alba cytotypes from tetraploid individuals. Continuous arrows represent reduced gametes; discontinuous arrows represent unreduced gametes; gray arrows represent unbalanced gametes. Chromosome numbers in each box represent the somatic number observed for each individual with different ploidy levels. (1) tetraploid (2n = 60) individuals can originate the hexaploid via crossing involving reduced and unreduced gametes; (2) tetraploid can form triploid cytotype by crossing with diploid; and (3) tetraploid can originate the aneuploid individual (2n = 38) from unbalanced gametes which are resulted from cytomixis process

Full size image

Although additional studies should be conducted in a higher number of cytotypes to evaluate the occurrence of cytomixis in the species, the present data open a new perspective to better understand the reproductive biology and the origin of the polyploid complex in L. alba.