INTRODUCTION
The genus Zea (Poaceae - Maydeae) includes the perennial species Z. perennis (Hitch.) (Reeves & Mangelsdorf) and Z. diploperennis Iltis (Doebley & Guzman), and the annual species Z. luxurians (Durieu & Ascherson) (Bird), Z. nicaragüensis (Iltis & Benz) and Z. mays L. The latter has four subspecies: Z. mays ssp. mays (the maize), Z. mays ssp. parviglumis, Z. mays ssp. huehuetenanguensis and Z. mays ssp. mexicana (Doebley, 1990; Iltis and Benz, 2000). All taxa have 2n=20, except for Z. perennis with 2n=40.
Cytogenetic evidence indicates that maize and its wild relatives (the teosintes) are cryptic polyploids. In the pioneering studies of Anderson (1945) maize was considered as allotetraploid 2n=4x=20, derived from extinct ancestors with 2n=10. Meiotic studies of intra- and interspecific hybrids confirmed their allopolyploid nature, indicating that maize and its wild relatives are tetraploids, except for the octoploid Z. perennis, with x=5 being the basic chromosome number (revisited in Poggio et al., 2005; Poggio and González, 2018). Further molecular studies provided compelling evidence for the allopolyploid nature of maize (Moore et al., 1995; Gaut and Doebley, 1997; White and Doebley, 1998; Swigonova et al., 2004).
Our research group has carried out numerous cytogenetic studies on Zea species, mainly focused on native Argentinian and Bolivian maize landraces (Tito et al., 1991; Quintela Fernández et al., 1995; Poggio et al., 1998, 1999a, 1999b, 2000a, 2000b, 2005; Rosato et al., 1998; González et al., 2004, 2006, 2013; González and Poggio, 2011, 2015, 2021; Realini et al., 2016, 2018, 2021; Fourastié et al., 2017; Poggio and González, 2018).
Maize is one of the most important cereal crops worldwide, growing in a broad range of agro-ecological regions. In South America, many landraces are adapted to a great variety of climatic conditions at different growing altitudes (Roberts et al., 1957; Wellhausen et al., 1957; Timothy et al., 1961; Rodríguez et al., 1968; Salhuana and Machado, 1999; Sánchez et al., 2007; Cámara Hernández et al., 2011; Orozco-Ramírez et al., 2016). Cámara Hernández et al. (2011) described more than 50 morphological native landraces from Northern Argentina, which are cultivated using ancestral farming practices. Of these, 36 landraces grow from lowlands to highlands in Northwestern Argentina (NWA), while 15 are cultivated at lower altitudes in Northeast Argentina (NEA). They are potential sources of genetic variation, constituting useful reservoirs of tolerance and resistance alleles for biotic and abiotic stresses.
The objectives of this work are to analyze genome size variation in the genus Zea based on studies carried out by our research group, to discuss the causes of such variation and to explore the relationship between genome size and cytological, phenological and environmental characteristics.
INTER- AND INTRASPECIFIC VARIATION IN THE GENOME SIZE OF ZEA
Genome size plays an important role in the cytogenetic variation of a taxon or group of taxa. This character is measured by microdensitometry with Feulgen’s stain or, most often, using flow cytometry (Tito et al., 1991; Dolezel et al., 2007). DNA content is usually expressed in picograms (pg) as the 2C-value, which refers to the amount of genomic DNA in unreplicated somatic cells (Poggio et al., 2010).
In genus Zea, the interspecific variation in the mean DNA content has been reported to range between 4.2 and 11.36 pg (Tables 1 and 2) (Laurie and Bennett, 1985; Tito et al., 1991; Fourastié et al., 2017). In maize, intraspecific variability in the DNA content has been recorded in native landraces from the following sites: a) NEA, cultivated in lowlands from 98 to 591 m.a.s.l.; b) NWA, growing along a broad altitudinal cline from 80 to 3900 m.a.s.l.; and c) Bolivia, cultivated from 200 to 3250 m.a.s.l. (Table 2) (Quintela Fernández et al., 1995; Rosato et al., 1998; Realini et al., 2016; Fourastié et al., 2017; González and Poggio, 2021). Table 2 shows the range of mean 2C-values (4.20-6.75 pg) for these landraces. A similar genome size variation has been obtained for other American maize populations growing along altitudinal clines (Laurie and Bennett, 1985; Rayburn and Auger, 1990; Díez et al., 2013; Bilinsky et al., 2018).
CAUSES OF GENOME SIZE VARIATION
In Zea, genome size variation has been mainly attributed to differences in the heterochromatin located in knobs and to interspersed DNA amount (e.g. retrotransposon families), making up over 70% of the nuclear genome (SanMiguel et al., 1998; Meyers et al., 2001; Realini et al., 2016; Fourastié et al., 2017). The presence of accessory chromosomes (B chromosomes; Bs) also contribute to DNA content variation, as discussed later (Rosato et al., 1998; Fourastié et al., 2017; González and Poggio, 2021).
Chromosomes of Zea species have blocks of heterochromatin called knobs (Kato, 1976; McClintock, 1978). This heterochromatin is highly condensed and composed of highly repeated DNA sequences (Peacock et al., 1981). Knobs can be visualized in dividing cells as well as in interphase nuclei as chromocenters (Wan and Widholm, 1992). During pachytene of Zea species, knobs can be found at 34 different chromosomal positions in the karyotype (Kato, 1976). They are detected using C and DAPI chromosome banding (González et al., 2013). At the molecular level, knobs are composed of two sequence families of 180 base pairs (180-bp) and 350 base pairs (TR-1), and may contain several retrotransposons (Dennis and Peacock, 1984; Ananiev et al., 1998). The sequence families, which are represented by thousands to millions of tandemly arranged copies, conform the different knob types (knobs exclusively containing 180- bp repeats, knobs exclusively containing TR-1 repeats and knobs with different proportion of both sequences) (Ananiev et al., 1998).
As mentioned above, DAPI banding identifies knobs as DAPI-positive bands that are A-T rich (Figure 1 A, B, C, F). However, this technique does not provide information on knob sequences. The fluorescent in situ hybridization (FISH) allows the detection and localization of specific sequences on interphase nuclei and metaphase chromosomes (Poggio et al., 1999b; 2005; González et al., 2006; González and Poggio, 2011; 2015). The hybridization of the 180-bp and TR-1 knob sequences on mitotic metaphases of Zea is used to determine the sequence composition of each DAPI-positive band (i.e. each knob) (Figure 1 D, E) (Albert et al., 2010; González et al., 2013).
Different maize landraces and teosintes show a wide variation in the size, number, chromosome position and sequence composition of heterochromatic knobs (Kato, 1976; McClintock et al., 1981; González et al., 2013), thus serving as valuable cytological markers. For example, they have been used for the cytogenetic characterization of landraces from Northern Argentina (Realini et al., 2016; Fourastié, 2017).
In Zea, DNA content is positively correlated with the number and size of knobs, as well as with the percentage of heterochromatin in the karyotype (Laurie and Bennett, 1985; Tito et al., 1991; Poggio et al., 1998; Realini et al., 2016; Fourastié et al., 2017; González and Poggio, 2021). The percentage of heterochromatin in a Zea karyotype is calculated by summing all chromosomal portions occupied by the DAPI-positive bands detected in the chromosomal complement. Z. luxurians has the highest DNA content of the 2n=20 species within the genus (Table 1), possibly due to the larger number and size of knobs, which are at terminal position on almost all chromosomes (Figure 1 C, D) (González and Poggio, 2011; González et al., 2013). On the contrary, DAPI-banding and FISH experiments did not detect conspicuous knobs in the octoploid Z. perennis, showing the lowest DNA content per basic genome (Cx) among Zea species (Kato and López, 1989; Tito et al., 1991; González et al., 2013) (Table 1; Figure 1 F). The small quantity of knob sequences in Z. perennis was postulated to be a consequence of the genome downsizing occurring during the process of secondary polyploidization in this species (Poggio et al., 2005; González and Poggio, 2015).
In maize landraces, the percentage of heterochromatin is positively correlated with the genome size (González and Poggio, 2011; Realini et al., 2016; Fourastié et al., 2017), suggesting that the variation in DNA content is mainly due to differences in the size, and to a lesser extent, in the number of knobs. However, some authors provided evidence for the presence of other sources than heterochromatin knobs contributing to genome size (revisited in Realini et al., 2016; 2018). Transposable elements (TEs) play a role in the dynamics of the nuclear genome, either through polymorphic insertions and deletions or by mediating ectopic recombination events leading to structural variation in the genome (SanMiguel and Bennetzen, 1998; Meyers et al., 2001). Recently, Coutinho Silva et al. (2020) demonstrated that in maize a higher 2C-value is associated with a more abundant distribution of LTR-retrotransposons in the karyotype, mainly from the Grande family. Further studies will enhance the knowledge on the differential composition of retrotransposon families in native maize landraces from Argentina, and its influence on genome size variation.
NEGATIVE CORRELATION BETWEEN DNA CONTENT AND ALTITUDE OF CULTIVATION IN MAIZE LANDRACES
In NWA, Rosato et al. (1998) reported a significant negative correlation between DNA content and altitude of cultivation in landraces growing along an altitudinal cline. This was further supported by Fourastié et al. (2017) for other NWA populations. Recently, González and Poggio (2021) observed that the Bolivian landraces cultivated at higher altitudes have lower DNA content than those growing at lower altitudes (Figure 2). Negative correlations between genome size and altitude of cultivation were also detected in different altitudinal clines from the American continent (Rayburn and Auger, 1990; Díez et al., 2013; Bilinsky et al., 2018).
HETEROCHROMATIN HAS ADAPTIVE SIGNIFICANCE ALONG AN ALTITUDINAL CLINE
Realini et al. (2016) observed a positive correlation between the length of the vegetative cycle and the percentage of heterochromatin in maize landraces from lowlands in NEA. Knob heterochromatin is the latest component to finish DNA replication because increased DNA packaging extends DNA synthesis, leading to a longer cell cycle, which may affect the rate of cell division and plant development (Pryor et al., 1980; Buckler et al., 1999; Greilhuber and Leitch, 2013). On this basis, length of the vegetative cycle was proposed to be optimized through artificial selection for an appropriate percentage of heterochromatin (Realini et al., 2016; 2021; Bilinsky et al., 2018).
As mentioned above, the percentage of heterochromatin is positively correlated with genome size. In addition, DNA content is higher in inbred lines with long vegetative cycles than in precocious ones, and their F1 hybrids have an intermediate genome size (González and Poggio, in prep.). Jian et al. (2017) also observed a high correlation between genome size and flowering time in inbred lines growing under tropical conditions. These results may explain the negative correlation detected between genome size and percentage of heterochromatin with cultivation altitude (Figure 2) (Tito et al., 1991; Poggio et al., 1998; Fourastié et al., 2017; González and Poggio, 2021). Thus, the fact that maize landraces at high altitudes are precocious and show a reduction in heterochromatin percentage most likely represents an adaptation to a shorter growing season typical of highlands, with natural selection acting on the flowering time across altitudinal clines (Bilinsky et al., 2018). This reinforces the hypothesis that the percentage of heterochromatin has adaptive value along altitudinal clines.
B-CHROMOSOME POLYMORPHISMS AS SOURCES OF GENOME SIZE VARIATION
In Zea, B chromosomes (Bs) are also regarded as a source of genome size variation (McClinctock et al., 1981; Kato and López, 1990; Poggio et al., 1998; Rosato et al., 1998; Cheng and Lin, 2003; Lamb et al., 2007; Rosado et al., 2009). Supernumerary Bs are dispensable chromosomes lacking homology with any member of the normal complement, the A-chromosome (A) set. These accessory chromosomes represent a specific type of selfish genetic element with mechanisms of drive which allow them to increase their transmission rates through different processes of non-mendelian inheritance (Jones and Houben, 2003; Houben et al., 2014; Blavet et al., 2021). Although Bs follow their own species-specific evolutionary pathways, it is widely accepted that they are derived from their respective A complement (revisited in Houben et al., 2014).
Polymorphisms for presence/absence and different doses of Bs have been reported in annual Zea species (Longley, 1937; Ting, 1976; Kato, personal communication). Polymorphisms are well documented in maize, particularly for landraces from NWA and Bolivia (McClinctock et al., 1981; Quintela Fernández et al., 1995; Rosato et al., 1998; Fourastié et al., 2017; González and Poggio, 2021). In these landraces, a wide variation of population frequency of Bs was observed. The doses range between one and eight Bs per plant, with two Bs per plant being the most frequently reported dose (Table 2). Fourastié et al. (2017) proposed that the frequency of Bs depends not only on the cultivation altitude, but also on the genotypical and nucleotypical backgrounds of the landraces.
In NWA and Bolivian landraces, which grow along a wide altitudinal cline, the mean number and frequency of Bs are significantly and negatively correlated with the 2C-value of the A-chromosome complement (A-DNA) and positively correlated with altitude of cultivation (Figure 2) (Quintela Fernández et al., 1995; Poggio et al., 1998; Rosato et al., 1998; Fourastié et al., 2017; González and Poggio, 2021). These authors also found a significant negative correlation between the mean number of Bs and the percentage of knob heterochromatin. Moreover, they hypothesized that Bs are maintained at higher frequencies in populations with low percentage of heterochromatin to preserve an optimal nucleotype (sensuBennett, 1972). Such term defines the conditions of the nucleus affecting cell and developmental parameters such as cell volume, nuclear volume, chromosome size, mitotic cycle time, duration of meiosis and minimum generation time (Bennett, 1987; Poggio et al., 1998). The negative association observed between the frequency of Bs and the percentage of heterochromatin suggests that there is a maximum limit on the mass of nuclear DNA that allows the optimum nucleotype (Rosato et al., 1998; Fourastié et al., 2017). Based on the analysis of many landraces from NWA and Bolivia, González and Poggio (2021) proposed that the optimal nucleotype is the result of an intragenomic conflict between Bs and heterochromatin knobs, where genome adjustment may lead to an appropriate length of the vegetative cycle for maize landraces growing across altitudinal clines.
A better understanding of the causes accounting for genome size variation and their adaptive significance in maize landraces is essential for the development of successful breeding and conservation programs.