Native distribution of Salvia species collected for this study.
Sample spectra of Salvia DNA content analyzed by flow cytometry. Each sample was stained with propidium iodide (PI) and analyzed with an internal standard. (Left) Sample spectrum shows Salvia hispanica as the left peak analyzed with the internal standard Solanum lycopersicum ‘Stupické’ on the right. (Right) Sample spectrum shows Salvia × ‘Jezebel’ as the left peak with the internal standard Pisum sativum ‘Ctirad’ on the right.
Distribution of mean holoploid 2C genome sizes of Salvia analyzed in this study compared with known ploidy at discrete values of 2x, 4x, 6x, or 8x.
Click on author name to view affiliation information
Salvia is a genetically diverse genus in the Lamiaceae family, with hundreds of species distributed globally. With base chromosome numbers ranging from 6 to 19 and ploidy levels ranging from diploid to octoploid, the genus has been proposed to be subdivided based on molecular data rather than morphology. However, little is known about total DNA content across the genus. The DNA content of 141 Salvia genotypes were analyzed using flow cytometry. Samples of Salvia were stained with propidium iodide and compared with the internal standards Pisum sativum ‘Ctirad’ and Solanum lycopersicum ‘Stupické’ to generate estimations of DNA content. Holoploid 2C genome sizes of the analyzed Salvia ranged from 0.63 pg to 6.12 pg. DNA content showed a wide distribution across chromosome number, ploidy, and clade. The wide distribution of DNA content across the genus further indicates the diversity of Salvia and may be useful for future breeding efforts.
Salvia is the largest genus in the Lamiaceae family and includes 1015 accepted species. It grows natively in Europe, Asia, Africa, Australia, and the Americas (Kew Science, 2019). The genus includes annual and perennial herbs and shrubs that are often highly aromatic and covered in hairs (Kew Science, 2019). Several species, including S. miltorrhiza and S. nemorosa, have been used for their medicinal properties (Božin et al., 2012; Li et al., 2013), whereas other species such as S. fruticosa, S. hispanica, and S. officinalis have been used culinarily (Cvetkovikj et al., 2015; Ullah et al., 2016). Most notably, Salvia is used as an ornamental plant to attract pollinators, including bees, butterflies, and birds, to the garden (Wester and Claßen-Bockhoff, 2011).
Great ornamental diversity exists within the genus, making Salvia useful for almost any garden. The inflorescences are vibrant and appear in whorls, with two-lipped corollas ranging from bright red to deep violet and including whites, yellows, and blues. Their visual impact is enhanced by pigmented calyces, formed by two cylindrically fused lips (Kew Science, 2019). The leaves are simple, frequently toothed, and display unique colors, such as S. argentea with silver foliage and S. elegans ‘Golden Delicious’ with chartreuse leaves (Whittlesey, 2014). As a result of their wide geographic spread, Salvias can be selected for a variety of climates. Species such as S. microphylla and S. patens do well in hot, humid regions, whereas others like S. nemorosa and S. verticillata are suited for colder climates. Plants including S. blepharophylla and S. miniata may be used in shade gardens, and others such as S. apiana and S. officinalis are useful in areas prone to drought (Clebsch, 2003).
In the last widely accepted classification of Salvia, the genus was subdivided into 12 sections based on differences in flower morphology (Bentham and Hooker, 1873). The distinguishing morphological characteristic that separates Salvia from other genera of the tribe Mentheae is the staminal structure. The flowers contain two stamens, with thecae separated by elongated connective tissue. Anterior thecae are often fertile, with aborted posterior thecae creating a unique lever mechanism that aids pollination (Walker et al., 2004). The taxonomic classification of the genus as a whole has not been altered since the arrangement by Bentham. However, based on genomic sequencing and phylogenic analysis, Salvia was confirmed to be polyphyletic, with highly variable staminal structure across the genus. A strong correlation was observed between genetic relatedness and geographic region, indicating a need to subdivide the genus into four clades (Walker et al., 2004).
Rather than characterizing the genus by morphological characteristics, molecular data provide a way to separate the genus by ancestry. Walker and Systma (2007) determined that the lever-like stamens developed multiple times in separate geographic regions and proposed dividing Salvia into three clades based on their lineage. Clade I and clade II are monophyletic and include other genera closely related to Salvia. The third clade was found to be distinct from the other two clades, but it was not confirmed to be monophyletic (Walker and Systma 2007). In a more recent study, Will and Claßen-Bockhoff (2017) conducted a robust analysis of the genus and identified four lineages. These four clades include other closely related genera and strongly correlate with a geographic region (Will and Claßen-Bockhoff, 2017). Despite various taxonomic revisions, comparatively little research has been performed to explore genome evolution.
Because of the polyphyletic nature of Salvia and vast geographic spread, remarkable variations in chromosome number and ploidy exist across the genus. Ploidy levels range from diploid to octoploid, and chromosome base numbers include x = 6–11, 13, 14, 15, 17, and 19 (Delestaing, 1954; Ranjbar et al., 2015). A wide range of chromosome numbers appears to have developed over time through aneuploidy, dysploidy, and polyploidy. The majority of known polyploids come from section Calosphace, whereas the sections Salvia, Drymosphce, and Horminum contain only diploid species. Although polyploidy seems to be the driving factor for speciation in South America, species from China appear to have evolved through differences in karyotype structure. Cytomixis has been observed in S. atropatana and S. indica, resulting in gametes with different chromosome numbers (Ranjbar et al., 2015).
Although there is no clear correlation of base chromosome numbers with geographic regions, some chromosome numbers appear more frequently in certain regions. The most commonly occurring base chromosome numbers for species from Europe, Turkey, and the Mediterranean are x = 7, 8, and 11. Species from Iran typically have the base numbers x = 10 and 11, whereas species from China tend to have x = 8. South American and Central American species usually have x = 11, whereas North American species tend to have x = 15 and 16 (Ranjbar et al., 2015). Despite these trends, not all species within these regions follow the pattern exactly.
Although there are more than 1000 species of Salvia, there is little reported genetic information making the genus suitable for a genome study (Kew Science, 2019). Analyzing the DNA content across many species may lead to a greater understanding of species development and adaptation. Knowing the genome size also provides valuable information for future breeding efforts and can aid the process of genome sequencing (Doležel and Greilhuber, 2010). Genetic information, including nuclear DNA content, can be determined for plant samples through flow cytometry by comparing the fluorescent intensity of a plant sample to the fluorescent intensity of a known standard with a similar genome size (Doležel et al., 2007).
The accuracy of DNA content estimation also depends on the fluorophore used to tag the DNA strand. The fluorophore can be intercalating or bind preferentially to AT or GC bps (Doležel et al., 1992). Studies have shown that fluorophores such as 4′,6-diamidino-2-phenylindole (DAPI) and mithramycin, which preferentially bind to AT and GC bps, respectively, can yield inaccurate estimates of genome size based on differing AT% and GC% between the sample and standard (Coleman et al., 1981; Ortega-Ortega et al., 2019). Although DAPI has been used to estimate the genome size of various plant species, intercalating fluorophores such as propidium iodide (PI) should be preferentially selected to yield more accurate estimates (Contreras and Shearer, 2018; Ortega-Ortega et al., 2019).
As an alternative to counting chromosomes through direct observation, Hoshino et al. (2019) tested the ability of flow cytometry to estimate chromosome numbers for Lychnis senno accurately. The chromosome number of the unknown was estimated by comparing the relative fluorescent intensity of the sample to a known internal standard of the same species. These estimated values were confirmed through chromosome counts. Furthermore, a positive correlation was identified when fluorescent intensity was plotted against the chromosome number, suggesting that flow cytometry is a fast and reliable way to estimate taxa with unknown chromosome numbers (Hoshino et al., 2019).
The relationship between ploidy and fluorescent intensity was tested for several Vaccinium and Rubus species in separate studies. Species with known ploidy were analyzed through flow cytometry with a known standard. The relative fluorescence of each sample was plotted against the known ploidy level resulting in a positive, linear relationship between fluorescent intensity and ploidy. Using this method, researchers were able to estimate ploidy for several unknowns. Furthermore, it was determined that flow cytometry can accurately estimate the ploidy across species (Costich et al., 1993; Meng and Finn, 2002). The present study analyzed genome size across clades and tested the ability of flow cytometry to estimate the chromosome number and ploidy across species of Salvia.
To enhance the genetic information of the genus Salvia, seeds and vegetative cuttings of 141 genotypes, including 92 individual species, were collected and grown for cytometric analysis. These species were obtained from germplasm collections worldwide to represent the wide natural distribution of species in the genus. Figure 1 shows the native distribution of species analyzed during this study. The nuclear DNA content was estimated using a flow cytometer (CytoFLEX S; Beckman Coulter, Hialeah, FL) and CytExpert software (Beckman Coulter). Two internal standards were selected for this study to represent the wide range of genome sizes existing in Salvia. These were Solanum lycopersicum ‘Stupické’, which has a holoploid 2C genome size of 1.96 pg, and Pisum sativum ‘Ctirad’, with a 2C value of 9.09 pg (Doležel and Bartoš, 2005). PI was chosen as the fluorochrome to yield an accurate estimation of the genome size (Doležel et al., 1992). Sample tissue of Salvia and the standards were collected for analysis by selecting young leaves at similar developmental stages.
Citation: Journal of the American Society for Horticultural Science 147, 3; 10.21273/JASHS05175-21
A PI stain kit (CyStain PI Absolute P; Sysmex America, Lincolnshire, IL) was used for this experiment. An internal standard was prepared with each sample. The chopped leaves were left in the extraction buffer for 60 s; then, the suspension was passed through a 50-μL filter (CellTrics; Sysmex America). The extracted cells were stained according to the recommended protocol and incubated for 20 min in a refrigerator before analysis. Three replicate samples were prepared for each species on separate days to confirm repeatability. The optical filter in the flow cytometer was set to select a wavelength range of 488 to 690 nm to include the excitation and emission maxima for PI bound to DNA. The suspended cells were drawn into the cytometer at a slow flow rate of 10 μL⋅s−1. A minimum of 1000 events was recorded within the gate of Salvia and 1000 events within the gate of the standard. A greater number of events was not required because of the low sd from replicate scans.
The 2C value of each species was estimated by comparing the mean fluorescent peak of Salvia to that of the internal standard using Eq. [1]. [1]
For each species, the 2C value was calculated by averaging three replicate scans. Most of the samples were analyzed with S. lycopersicum ‘Stupické’; however, many species had overlapping fluorescent peaks with S. lycopersicum, necessitating the use of P. sativum ‘Ctirad’. Figure 2 includes sample spectra of Salvia analyzed with each internal standard. For species with known ploidy, the DNA content of the nonreplicated genome (1Cx) was calculated by dividing the experimentally determined 2C value by ploidy (Contreras and Shearer, 2018).
Citation: Journal of the American Society for Horticultural Science 147, 3; 10.21273/JASHS05175-21
According to the genome size classifications described by Leitch et al. (1998), 45% of the analyzed Salvia samples are classified as having a very small genome (≤1.4 pg), and 48% are classified as having a small genome (≤3.5 pg) as shown in Table 1. A large genome (≥14.0 pg) was not identified for any species. Genome size does not correlate with the complexity of an organism (Doležel and Bartoš, 2005), but small genomes have been correlated with shorter reproductive cycles and the rapid establishment of seedlings. Therefore, the small genomes of Salvia may give them a competitive advantage over other plants (Leitch et al., 1998).
During this study, multiple accessions of S. bucharica, S. canariensis, S. coccinea, S. dorrii, S. elegans, S. greggii, S. guaranitica, S. hispanica, S. involucrata, S. leucantha, S. macrophylla, S. nemorosa, S. pratensis, S. prunelloides, S. sclarea, S. splendens, and S. verbenaca were analyzed for total genetic content. Minimal variation was observed within each species. For example, three accessions of S. coccinea were sourced from Argentina, California, and Florida. Despite having naturalization in different environments, their genomic contents were remarkably similar, with 0.85 ± 0.004, 0.87 ± 0.007, and 0.85 ± 0.008 pg, respectively. Therefore, genetic content was believed to be consistent within a species with the same ploidy. This finding is consistent with that of Castro et al. (2012) who reported that the chromosomal DNA content is constant across a species.
One example of the genome size not being consistent among cultivars was S. farinacea. The S. farinacea Duelberg series had genome sizes of 2.19 ± 0.051 pg and 2.15 ± 0.039 pg for ‘Augusta Duelberg’ and ‘Henry Duelberg’, respectively. In the S. farinacea Cathedral® series, the genome sizes were 2.17 ± 0.070 pg and 2.15 ± 0.016 pg for the lavender and purple cultivars, respectively. However, the genome size of S. farinacea Unplugged® So BlueTM was approximately double these values at 4.10 ± 0.068 pg. Although the species is reported to be 2n = 2x = 20, the doubled DNA content of S. farinacea Unplugged® So BlueTM may indicate tetraploidy (2n = 4x = 40) in this cultivar (Alberto et al., 2003). Chromosome counts could confirm polyploidy; however, genetic content is known to increase proportionally with ploidy.
Chromosome counts have been performed for S. discolor, S. elegans, S. repens, and S. sessilifolia; however, ploidy was not reported for these species. Because base chromosome numbers for Salvia are limited to x = 6–11, 13, 14, 15, 17, or 19, ploidy can be inferred for each of these species by their chromosome counts (Ranjbar et al., 2015). S. elegans and S. repens are reported as 2n = 20 (Chromosome Counts Database, 2021; Kumar and Subramaniam, 1987). Using known base chromosome numbers, S. elegans and S. repens must be diploid with a base chromosome number x = 10. By the same logic, S. discolor, reported as 2n = 24, must be a tetraploid with a base chromosome number x = 6 (Kumar and Subramaniam, 1987). Finally, S. sessilifolia, with 2n = 44, must be tetraploid with 11 chromosomes (Chromosome Counts Database, 2021).
Available hybrids were analyzed by comparing the 2C value of the hybrid to the 2C value of the parental species. The interspecific hybrid S. mexicana ×gesneriiflora, with an estimated 2C genome size of 1.59 ± 0.002 pg, was halfway between the parental species S. mexicana (2C = 1.21 ± 0.043 pg) and S. gesneriiflora (2C = 1.96 ± 0.055 pg). The same trend was observed for the hybrids S. fruticosa ×officinalis, S. ×jamensis ‘Elk Cranberry Red’, and S. splendens ×guaranitica ‘Purple & Bloom’. Interspecific hybrids with a genome size between the genome sizes of their parental species have also been observed in other genera, such as Cirsium (Bureš et al., 2004), Diphasiastrum (Hanušová et al., 2014), Dryopteris (Ekrt et al., 2010), Ficaria (Popelka et al., 2019), and Sarcococca (Denaeghel et al., 2017).
Contrary to these examples, however, some of the studied hybrids did not have an intermediary genome size compared with those of their parents. Rather, S. ×jamensis ‘Red Velvet’, a hybrid of S. greggii and S. microphylla, has a larger DNA content (3.27 ± 0.077 pg) than that of either parent. S. ‘Jezebel’ is also a hybrid of S. microphylla, but it has an unknown pollen source. Like S. ×jamensis ‘Red Velvet’, ‘Jezebel’ also has a larger genome size (2.48 ± 0.031 pg) than S. microphylla. A similar observation was made by Baack et al. (2005) during the study of Helianthus hybrids. The interspecific hybrids of H. annuus and H. petiolaris had a greater total DNA content than either parental species (Baack et al., 2005). Therefore, the genetic content is not always reliable evidence of the hybridization of species.
Our data was analyzed according to the arrangement of the genus by Will and Claßen-Bockhoff (2017); species are divided into four clades based on their phylogenetic relationships. Estimated 2C values and calculated 1Cx values are summarized in Table 2 with their assigned clades. Section classification based on the arrangement by Bentham and Hooker (1873) are included for reference. Most plants obtained from this study were from clades I and II, a few species were from clade IV, and none were obtained from clade III (Will and Claßen-Bockhoff, 2017). Each clade showed a significant range of genome sizes, from a 1Cx value of 0.20 pg in the tetraploid S. sessilifolia to a 1Cx value of 3.04 pg in the diploid S. carduacea. The distribution of 2C and 1Cx values within each clade is summarized in Table 3.
Salvia chromosomes are small and difficult to distinguish, resulting in many species lacking chromosome counts (Haque, 1981). Unfortunately, flow cytometry cannot be used to predict chromosome numbers in Salvia; however, they can be shown by comparing known chromosome numbers to calculated genome sizes summarized in Table 1. For example, S. macrophylla ‘Purple Leaf’ (2n = 2x = 18) had a genome size of 0.93 ± 0.008 pg (Bolkhovskikh, 1969). In comparison, S. hispanica (2n = 2x = 12) had the same genome size (0.93 ± 0.016 pg) despite having a different base chromosome number (Estilai et al., 1990). Contrary to these findings, flow cytometry has been used to estimate chromosome number within the species Lychnis senno and among species in the genera Scrophularia, Verbascum, and Veronica (Castro et al., 2012; Hoshino et al., 2019).
Inconsistent chromosome reports further complicate chromosome estimation. Different chromosome numbers have been assigned to the same species, as shown in Table 1. For example, S. nemorosa has been reported to have 12 or 14 base chromosomes, whereas S. pratensis has been observed to have 16, 18, or 20 (Haque, 1981). Salvia with the same number of chromosomes also vary in genome size. For example, species with 16 base chromosomes (2n = 2x = 32) range from 0.92 ± 0.003 pg in S. leucantha ‘Midnight’ to 6.07 ± 0.129 pg in S. carduacea (Epling et al., 1962; Haque, 1981). Therefore, flow cytometry is an unreliable method for estimating chromosome counts in Salvia. These findings are consistent with the analysis of the genus Helleborus. All species in Helleborus have the same base chromosome number, but their DNA contents have a significant range (Zonneveld, 2001). In the case of Helleborus and Salvia, the inability to predict chromosome numbers based on genome size is caused by each species having a different amount of DNA in their chromosomes.
Measured holoploid 2C values of Salvia were plotted against known ploidy. Genome size was variable across ploidy, as shown in Fig. 3. Few species in our collection had a ploidy level of 4x, 6x, or 8x. Therefore, regression analysis results were inconclusive when determining a relationship between genome size and ploidy. However, the variations in chromosome number and size throughout the genus make it unlikely for these two factors to be correlated. Previously, flow cytometry to estimate ploidy has been applied for genera with constant base chromosome numbers, including Rubus (Meng and Finn, 2002) and Vaccinium (Costich et al., 1993). Although flow cytometry cannot be used to estimate ploidy in Salvia, it may be a valuable way to confirm induced polyploidy within a species (Bose et al., 1989).
Citation: Journal of the American Society for Horticultural Science 147, 3; 10.21273/JASHS05175-21
Although genome sizes could not predict chromosome number or ploidy, data collected in this study may benefit future breeding efforts. For example, interspecific hybridization is difficult to achieve in Salvia; however, genome size data may be helpful to predict crosses (Akbarzadeh et al., 2021; Tychonievich and Warner, 2011). In a study conducted by Bureš et al. (2004), the cross-compatibility of Cirsium was compared with the total DNA content. A negative correlation was observed between the total genetic content and the ability to hybridize between species. For Cirsium, smaller genomes were more likely to form interspecific hybrids (Bureš et al., 2004). Genome size data can also increase the understanding of species development and adaptation. For example, larger plant genomes have been related to an increased ability to withstand humid environments and increased atmospheric CO2 levels (Jasienski and Bazzaz, 1995; Vesleý et al., 2012). However, smaller plant genomes are less susceptible to damage by ionizing radiation (Sparrow and Miksche, 1961). Of the species analyzed in this study, none were found to have a large genome according to the classifications by Leitch et al. (1998). However, the trends of adaptability relating to genome size may be useful for future breeding efforts.
The 141 genotypes of Salvia analyzed in this study were intentionally selected from many geographic regions to provide a representative subset of the genus. In addition to having wide geographic distribution, multiple base chromosome numbers, and variation in ploidy across the genus, significant variations in total DNA contents were observed across the species screened in this study. This study showed that flow cytometry is not a reliable way of estimating the chromosome number or ploidy across species in Salvia. However, the wide distribution of genome size among clades provides further evidence of species diversity within the genus and may aid in future breeding work.
Native distribution of Salvia species collected for this study.
Sample spectra of Salvia DNA content analyzed by flow cytometry. Each sample was stained with propidium iodide (PI) and analyzed with an internal standard. (Left) Sample spectrum shows Salvia hispanica as the left peak analyzed with the internal standard Solanum lycopersicum ‘Stupické’ on the right. (Right) Sample spectrum shows Salvia × ‘Jezebel’ as the left peak with the internal standard Pisum sativum ‘Ctirad’ on the right.
Distribution of mean holoploid 2C genome sizes of Salvia analyzed in this study compared with known ploidy at discrete values of 2x, 4x, 6x, or 8x.
Contributor Notes
J.M.R. is the corresponding author. E-mail: ruter@uga.edu.
Native distribution of Salvia species collected for this study.
Sample spectra of Salvia DNA content analyzed by flow cytometry. Each sample was stained with propidium iodide (PI) and analyzed with an internal standard. (Left) Sample spectrum shows Salvia hispanica as the left peak analyzed with the internal standard Solanum lycopersicum ‘Stupické’ on the right. (Right) Sample spectrum shows Salvia × ‘Jezebel’ as the left peak with the internal standard Pisum sativum ‘Ctirad’ on the right.
Distribution of mean holoploid 2C genome sizes of Salvia analyzed in this study compared with known ploidy at discrete values of 2x, 4x, 6x, or 8x.
