Chromosome Number, Ploidy Level, and Nuclear DNA Content in 23 Species of Echeveria (Crassulaceae)

Echeveria is a polyploid genus with a wide diversity of species and morphologies. The number of species registered for Echeveria is approximately 170; many of them are native to Mexico. This genus is of special interest in cytogenetic research because it has a variety of chromosome numbers and ploidy levels. Additionally, there are no studies concerning nuclear DNA content and the extent of endopolyploidy. This work aims to investigate the cytogenetic characteristics of 23 species of Echeveria collected in 9 states of Mexico, analyzing 2n chromosome numbers, ploidy level, nuclear DNA content, and endopolyploidy levels. Chromosome numbers were obtained from root tips. DNA content was obtained from the leaf parenchyma, which was processed according to the two-step protocol with Otto solutions and propidium iodide as fluorochrome, and then analyzed by flow cytometry. From the 23 species of Echeveria analyzed, 16 species lacked previous reports of 2n chromosome numbers. The 2n chromosome numbers found and analyzed in this research for Echeveria species ranged from 24 to 270. The range of 2C nuclear DNA amounts ranged from 1.26 pg in E. catorce to 7.70 pg in E. roseiflora, while the 1C values were 616 Mbp and 753 Mbp, respectively, for the same species. However, differences in the level of endopolyploidy nuclei were found, corresponding to 4 endocycles (8C, 16C, 32C and 64C) in E. olivacea, E. catorce, E. juarezensis and E. perezcalixii. In contrast, E. longiflora presented 3 endocycles (8C, 16C and 32C) and E. roseiflora presented 2 endocycles (8C and 16C). It has been suggested that polyploidization and diploidization processes, together with the presence of endopolyploidy, allowed Echeveria species to adapt and colonize new adverse environments.


Introduction
Echeveria is a genus in the Crassulaceae family consisting of perennial plants that grow naturally in America [1]. The distribution of the Echeveria genus spans from west Texas to Argentina, across Mexico, Guatemala, and Central and South America [2][3][4][5][6]. The Echeveria genus includes a vast diversity of species which are commonly found on rocky cliffs in different ecosystems, such as pine and oak forests, cloud forests, and some xerophytic scrublands, which are mainly limited to temperate zones of Mexico [2,3,5]. Diversification of Echeveria species has been influenced by the isolation of the populations by geographical barriers caused for the rugged orography of Mexico [3]. Hidalgo, Mexico, Oaxaca, and Puebla are the states with the highest richness and endemism of this genus in Mexico [3,7]. The number of species registered for Echeveria genus is approximately 170, Endopolyploidy is more frequent in plants with small genomes, such as Arabidopsis thaliana [54] and annual plants rather than perennial ones [33]. The occurrence and the degree of endopolyploidy is documented in several groups of plants, such as algae, bryophytes, ferns, gymnosperms, and angiosperms [50]. In species of Cactaceae, such as Mammillaria san-angelensis and several species of Opuntia and Nopalea, endopolyploidy has been found at different levels [24,34,55]. Nevertheless, regarding the genus Echeveria, there are no publications on its DNA content, nor the existence of endopolyploidy; however, this information could provide useful knowledge for taxonomic studies and biotechnology, conservation, and floriculture programs for these species.
The goals of this study are: (1) to assess the interspecific variation in chromosome counts 2n and ploidy levels of the Echeveria species; (2) to evaluate the variation of the nuclear DNA content in the genus, since there are no previous studies regarding nuclear DNA content for any Echeveria species; (3) to evaluate the presence of endopolyploidy in the taxa included in the study.

Plant Material
In total, 25 wild populations of 23 Echeveria species were collected in 9 states of Mexico: Durango, Guerrero, Jalisco, Michoacan, Oaxaca, Puebla, San Luis Potosi, and Veracruz (Table 1). In the case of Echeveria catorce and E. guerrerensis, two distinct populations from San Luis Potosi and Guerrero, respectively, were analyzed (Table 1). Individuals from each species (5)(6)(7)(8)(9)(10) were planted in pots containing organic soil, and a mix of inorganic substrate made of sand, tepojal (pumice stone), tezontle (red volcanic rock), and agrolite; plants were grown in greenhouse conditions at the Botanical Garden, Biology Institute, UNAM. Voucher specimens were deposited in the National Herbarium (MEXU) of the same Institution. Studied species correspond to 7 of the 17 series that comprise the Echeveria genus [6,7,56]. Echeveria catorce has not been formally described yet, however, it is widely recognized in the literature [19,56]. The species E. catorce and E. guerrensis have two collection numbers, but each one corresponds to a different population, so they were analyzed independently. The species E. longiflora, E. novogaliceana, E. olivacea, E. roseiflora, and E. uhlii have more than one collection number because they were collected on different dates or by different collectors, but belong to the same population, so they were analyzed together.

Mitotic Chromosome Counts
Based on 5-10 individual plant samples of each population of Echeveria, 9 mitotic cells at the metaphase stage were observed for the determination of chromosome numbers (2n). Elongated secondary root tips were collected in the morning and placed in 8-hydroxyquinoline 0.002 M solution for 6 hours at 18 • C in the dark. After wards, the root tips were fixed in fresh Farmer s solution (three-parts absolute ethanol: one-part glacial acetic acid) for 24 h at 18 • C. The root tips were hydrolyzed in hydrochloric acid 1 N for 11 min at 60 • C and transferred to Schiff reagent for 1 h, and then to 1.8% propionic orcein to stain the chromosomes for 25 min following the procedure of García [57]. An additional treatment was required in some species with very hardened tissue-10% pepsin (Sigma) at 37 • C for 90 minutes; this treatment was used after staining with Schiff s reagent. Slides were prepared and frozen with dry ice [58], dehydrated in absolute alcohol, and mounted in Canada balsam. Nine of the best cells from each plant were photographed using an Axio Vision Rel. 4.7 camera in a Zeiss photomicroscope III (Gottingen, Germany).

Estimation of Nuclear DNA Content
Of the 23 studied Echeveria species. three to five young plants collected from the wild and kept in a greenhouse while they were analyzed and used to estimate nuclear DNA content, utilizing flow cytometry. Three replicates of each individual plant were analyzed. Internal standards for the genome size estimates consisted of Solanum lycopersicum cv. Stupické polni rané, 2C DNA = 1.96 pg [37]; Zea mays cv. CE-777, 2C DNA = 5.43 pg and Pisum sativum L. cv. Ctirad, 2C DNA = 9.09 pg [37]. Samples were prepared according to the two-step protocol with Otto solutions as detailed in Dolezel et al. [37], with some modifications.
Leaf tissue from Echeveria plants was used for the analysis of DNA content, from which the waxy cuticle had been previously removed to prevent the nuclei from adhering to each other or to the surface of the containers (due to their wax content). The waxy cuticle was removed using a razor blade; a shallow cut was made at the edge of the leaf and then the cuticle was pulled and separated with the help of the same blade and fine-tipped pins. Due to the difficulty of obtaining isolated nuclei of Echeveria and the great differences in the content of nuclei between the Echeveria leaves and the internal standards, it was almost impossible to obtain a good relative proportion of nuclei. Therefore, it was decided to use the pseudo-standardization technique, where the nuclei solutions are prepared separately, check Temsch et al. [59]. Between 120 and 250 mg of Echeveria leaf tissue was chopped with a razor blade in a Petri dish containing 1.5 to 2.00 mL of Otto 1 solution (1.5 mL of 0.1 M citric acid and 0.5% Tween 20), and then filtered through a 50 µm nylon mesh. Separately, 20-42 mg of internal standard plant was chopped into another Petri dish containing 1 mL of Otto1 solution and filtered in the same way [59]. Both solutions were incubated for 15 min at room temperature. Then, samples were pelleted by centrifugation (90 g for 3 min), and each was suspended in 500 µL of Otto 1 solution.
Subsequently, 250 µL of each nuclei solution (Echeveria and Internal plant) was poured into a sample tube and 2 mL of 0.4 M Na 2 HPO 4 (Otto 2 solution) was added to the suspension along with stock solutions of propidium iodide and RNase, both at 50 µg mL −1 .
The definitive sample was filtered through a 50 µm nylon mesh (to avoid the passage of possible crystals formed in the Otto 2 solution, given its high salt concentration, which could cause obstruction in the cytometer conduits) and then analyzed using a Partec CyFlow SL Cytometer (equipped with a 488 nm solid state laser). The instrument gain was adjusted so that the peak representing either the G 1 nuclei of Echeveria or the G 1 nuclei of the internal standard was placed in channel 50 of a 250-channel linear scale.
At least 10,000 nuclei were analyzed for each sample. Peak means areas and coefficient of variation (CV%) were obtained for each peak of interest (sample and standard) using the gating function in the FloMax software for cytometry (Partec). The CV accepted for the samples was less than 5.00%; for the case of the internal standards, values between 2.80% and 3.72% were obtained, while for Echeveria the range was between 3.83% and 4.98%. Nuclear genome size was calculated according to Dolezel et al. [37] using the formula: where A = Echeveria 2C nuclear DNA content (pg); B = Echeveria G 0 /G 1 peak mean; C = internal standard G 0 /G 1 peak mean; and D = 2C DNA content internal standard. The 1Cx-value was calculated for all the studied species by dividing nuclear DNA content by the ploidy level, as suggested by Greilhuber et al. [39] and multiplying it later by 978 to convert it to Mbp [37].

Endopolyploidy Determination
Leaf tissue without the waxy cuticle was also used for endopolyploidy determination and the same procedure was carried out for the determination of the nuclear DNA content, but without the addition of an internal standard. The gain of the Partec CyFlow SL Cytometer was adjusted so that all the peaks could be seen on a logarithmic scale. The nuclei number and coefficient of variation (CV) were obtained for each peak using the gating function in the FloMax Software by Partec.
The cycle value or endoreduplication index was calculated according to procedure described by Barow, 2003 [60], using the formula:

Statistical Analyses
A one-way variance analysis (ANOVA) was conducted to compare the chromosome numbers of the series Gibbiflorae, Angulatae, and Racemosae. Data were transformed to reach normality and variance homogeneity by applying a Box-Cox transformation calculated by the JMP program. The algorithm calculated to produce the best transformation was: , where x = data with transformation, and N = data without transformation.
Also, a Tuckey-Kramer test comparison was performed among groups in this series. However, the series Echeveria, Nudae, Pruinosae, and Urbinae contains only one species and it was not possible include in the ANOVA analyses.
Differences between the DNA content values (pg) of Echeveria populations were evaluated through a bi-factorial nested variance analysis (Nested ANOVA model). Random effects of repeated measurements on the same individual were considered. Natural logarithm transformed values of the response variable in the ANOVA model was significant, as well as each of its components. Multiple mean comparison test analyses of the DNA 2C content result values were done to compare the 25 populations of the 23 species of Echeveria based on a Tukey-Kramer test, a post-hoc test conducted after ANOVA. Mean and standard errors are reported in the original scale.
Pearson correlation and regression analyses were applied to evaluate the correlation among polyploidy levels and the 2n number, and between the ploidy level and the 1Cx-value.
The comparison of the cycle value between the study species was carried out using a one-way variance analysis (ANOVA). The normality of the studentized residuals was verified and the homogeneity of the variances was verified. A post-hoc Tuckey-Kramer test was applied after the one-way ANOVA.
All statistical analyses were done with JMP 8.0 and JMP 10 (SAS Institute, Cary, NC, USA) and using the software R (R: A language and environment for statistics computing. R Foundation for Statistical Computing, Vienna, Austria, URI, https://www.R-project.org (accessed on 17 July 2021) [62].

Endemism, Chromosome Numbers, and Ploidy Level
Chromosome numbers of 23 Echeveria species were analyzed in this research. Of them, sixteen were endemic, corresponding to 69.6% of the total species analyzed here ( Table 2). The largest diversity and endemism of this genus were observed in Oaxaca. There were different values of 2n and x in the 23 species of Echeveria, as expected in a genus known to be polybasic. For the first time in this research, the 2n for 16 species of Echeveria was reported ( Table 2; Figures 1-4), which ranged from 24 to 270. On the other hand, the most frequent x value was x = 27, present in 56.5% of the species was also the higher basic chromosome number ( Table 2). The lowest value of x and 2n was observed in two diploid populations of E. catorce and of E. schaffneri with 2n = 2x = 24, and both species with x = 12.
The differences among the series concerning chromosome numbers were significant among Gibbiflorae, Racemosae, and Angulatae (ANOVA: F 2,18 = 20.28, p < 0.0001). The Tukey-Kramer test comparisons among the series showed that the mean of the chromosome numbers of the Gibbiflorae series (x = 104.3 ± 15.6) was significantly higher than the mean of chromosome numbers of the Racemosae series (x = 40 ± 29.2). Similarly, the mean of the chromosome numbers of the Racemosae series was significantly higher than Angulatae series (x = 24 ± 33.7) ( Table 3).  Figure 5).
Based on the 2C value (pg) of the 23 species of Echeveria analyzed by flow cytometry, one-way ANOVA revealed significant differences among species (α = 0.05, p < 0.0001). Ten significantly different groups were observed in the Tukey-Kramer test. The first group included Echeveria roseiflora, the second group was represented by E. novogaliciana, and the third group corresponded to E. gibbiflora and E. altamirae. Each group had statistically different means (Table 2). There were no significant differences among the next five groups (see central part in Table 2

Endopolyploidy
Differences in the percentage of endopolyploidy nuclei and number of endocycles were observed in the leaf parenchyma of 23 species of Echeveria analyzed (25 populations  ; peaks corresponding to nuclei 2C, 4C, 8C, and 16C from E. juarezensis are shown, peaks SG1 and SG2 represent nuclei from Solanum lycopersicum, used as the Internal standard. (B) Isolated nuclei analysis of E. altamirae (2n = 4x = 108); peaks corresponding to nuclei 2C, 4C, 8C and 16C from E. altamirae are shown; peak SG1 represents nuclei from Zea mays, used as the internal standard. (C) Isolated nuclei analysis of E. novogaliciana (2n = 6x = 176); corresponding to nuclei 2C, 4C, and 8C from E. novogaliciana are shown; peak SG1 represents nuclei from Pisum sativum used as the internal standard.

Endopolyploidy
Differences in the percentage of endopolyploidy nuclei and number of endocycles were observed in the leaf parenchyma of 23 species of Echeveria analyzed (25 populations Figure 6).  Figure 6). Among the species analyzed, there were some that presented two endopolyploidy levels. The variation was detected between organisms of the same species and population. From these, E. olivacea, E. juarezensis, E. perezcalixii, and one of the populations of E. catorce (Accession number 5469), presented nuclei with 3 (8C, 16C, and 32C) or 4 (8C, 16C, 32C, and 64C) endocycles, while E. longiflora presented 2 (8C and 16C) or 3 (8C, 16C, and 32C)  Among the species analyzed, there were some that presented two endopolyploidy levels. The variation was detected between organisms of the same species and population. From these, E. olivacea, E. juarezensis, E. perezcalixii, and one of the populations of E. catorce (Accession number 5469), presented nuclei with 3 (8C, 16C, and 32C) or 4 (8C, 16C, 32C, and 64C) endocycles, while E. longiflora presented 2 (8C and 16C) or 3 (8C, 16C, and 32C) endocycles, and E. roseiflora showed 1 (8C) or 2 (8C and 16C) endocycles ( Table 5). The distribution of the relative percentage of endopolyploid nuclei populations was variable, even among species with the same endopolyploidy level. (Table 4.). The variation in the number of endocycles and in the relative nuclei distribution in each C-level was reflected in the cycle value, which show values from 0.690 to 2.562, which is an adequate parameter to compare the degree of endopolyploidy [64] (Tables 4 and 5). Table 5. Distribution of the percentage of endopolyploidy nuclei, number of endocycles, and cycle value in 6 Echeveria species that present 2 endopolyploidy levels among individuals of the same population.

Correlation Polyploidy, Chromosome Number, and 2C DNA Content
There was a positive and high correlation between the ploidy level and the chromosome number (2n) (r = 0.93, p < 0.001, Figure 7). On the other hand, a negative correlation between polyploidy and the 1Cx-value (r = −0.43, p = 0.03. Figure 8) was observed, which implies that as polyploidy level increases, the chromosome number also increases, but the DNA content of one monoploid genome decreased, suggesting a reduction of the DNA content of chromosomes in plants with the highest ploidy level.
There was a positive and high correlation between the ploidy level and th some number (2n) (r = 0.93, p < 0.001, Figure 7). On the other hand, a negative c between polyploidy and the 1Cx-value (r = −0.43, p = 0.03. Figure 8) was observ implies that as polyploidy level increases, the chromosome number also increas DNA content of one monoploid genome decreased, suggesting a reduction of content of chromosomes in plants with the highest ploidy level. Correlation results were statistically significant (p < 0.001). Numbers in the graphic correspond to species numbers in Table 2 and  Table 2 and Figures 1-4.

Endemism, Chromosome Numbers, and Polyploidy in Echeveria
The species of the genus Echeveria included in this study come from 9 States of Mex ico; 16 of the 23 species were endemic which corresponds to 69.6% (Tables 1 and 2). Of th total endemic species, 11 were collected in the Oaxaca State, 2 in Guerrero, 1 in Puebla, in Michoacán and 1 more in Veracruz. The high number of endemic species in the state o Oaxaca is consistent with the observations of other authors, who consider Oaxaca as th state with the highest diversity and endemism of this genus in Mexico [7]. In fact, Mexic presents a high percentage of endemism for this genus because, of the 170 species de scribed, approximately 140 are endemic, which corresponds to 85% of endemism distrib uted in different localities in Mexico [3,8].
The 23 studied species of Echeveria belonged to 7 of the 17 series of the genus (Tabl 1). A great diversity of chromosome numbers (2n) was observed within each one of th series and also when those values were compared between the series (Tables 1 and 2). Th

Endemism, Chromosome Numbers, and Polyploidy in Echeveria
The species of the genus Echeveria included in this study come from 9 States of Mexico; 16 of the 23 species were endemic which corresponds to 69.6% (Tables 1 and 2). Of the total endemic species, 11 were collected in the Oaxaca State, 2 in Guerrero, 1 in Puebla, 1 in Michoacán and 1 more in Veracruz. The high number of endemic species in the state of Oaxaca is consistent with the observations of other authors, who consider Oaxaca as the state with the highest diversity and endemism of this genus in Mexico [7]. In fact, Mexico presents a high percentage of endemism for this genus because, of the 170 species described, approximately 140 are endemic, which corresponds to 85% of endemism distributed in different localities in Mexico [3,8].
The 23 studied species of Echeveria belonged to 7 of the 17 series of the genus (Table 1). A great diversity of chromosome numbers (2n) was observed within each one of the series and also when those values were compared between the series (Tables 1 and 2). The chromosome number (2n) of 16 Echeveria species are reported for the first time, which corresponds to 69.6% of the total species studied in the present investigation ( Table 2, Figures 1-4).
From the Gibbiflorae series, 13 species were analyzed: 7 from the state of Oaxaca, 2 from Guerrero, 1 from Zacatecas, 1 from Jalisco, 1 from Michoacán, and 1 from Durango. This series presents high diversity in terms of the chromosome numbers; 6 of the 13 species studied in this series are diploids, five of which (E. perezcalixii, E. guerrerensis, E. pallida, E. magnifica, and E. juarezensis) had 2n = 2x = 54 with x = 27 while E. triquiana had 2n = 2x = 32 with x = 16. Regarding this species, Uhl [13] reported three diploid populations of E. juarezensis with n = 27, similar to the one studied here; the first one was found in Sierra de Juarez, the second one came from Sola de Vega locality, and the third population belonged to Zoquiapan; all these localities are in Oaxaca, Mexico. Uhl [13] informed gametic numbers about a tetraploid population of E. juarezensis with n = 54, from San Felipe, Oaxaca. Similarly, Uhl [13] analyzed other triploid populations of E. juarezensis from Sierra de Juarez, Oaxaca, and another hybrid of E. juarezensis, with irregular meiosis from Sierra de Juárez, Oaxaca. Uhl [13] also mentioned one population of E. pallida collected in Oaxaca, also a diploid with n = 27 and two other tetraploid-cultivated populations with n = 54 of E. pallida, one from Oaxaca, and another from Mexico City.
The three tetraploid species analyzed in this research (E. altamirae, E. dactyliffera, and E. cupreata) belong to the Gibbiflorae series, and the three of them have 2n = 4x = 108 and x = 27. Uhl [13] pointed out that one population, also a tetraploid of E. dactylifera, from Durango state in Mexico, n = 54, like the one reported in this investigation. On the other hand, from this same series, we analyzed a hexaploid species (E. longiflora) with 2n = 6x = 162 and x = 27; this species is like a hexaploid population of Echeveria scopolorum (2n = 6x = 162) reported by Uhl [13] which was collected on the road from Michoacán to Veracruz, Mexico. The highest chromosome number of all the species analyzed in this study was found in E. roseiflora, which had 2n = 10x = 270 with x = 27 and belongs to the series Gibbiflorae.
From the Angulatae series, two species were analyzed, E: shaffneri and E. catorce (two localities were studied of the latter). The three populations were collected in San Luis Potosí and were diploids with 2n = 2x = 24 with x = 12. These numbers were coincident with results reported by Uhl for these species from the same locality of the state of San Luis Potosi, Mexico [13]. Regarding the Echeveria series, only E. zorzaniana was analyzed, which was collected in Oaxaca and was a diploid; this was the only species with 2n = 2x = 40 and x = 20 in the present study and corresponds to the first count informed for this species (Table 2, Figure 4g). Nevertheless, Uhl [19] reported n = 20 for Echeveria secunda, from the central area of Mexico.
From the Nudae series, E. multicaulis from Guerrero was the only species analyzed, being a diploid with 2n = 2x = 32. For this same species, Uhl [12,16,17] found one diploid population, and one triploid population (2n = 3x + 6 = 48) in Merida, Venezuela (South America). Both E. caamanoi from the Urbinae series (collected in Puebla), and E. cuicatecana from the Pruinosae series (collected in Oaxaca), turned out to be 5x with 2n = 5x = 60. It is important to mention that in this study that we confirmed the 2n = 60 (5x) for E. cuicatecana, which was previously reported by Reyes et al. [63]. The basic chromosome number x = 12 is frequent within the genus Echeveria. In fact, Uhl has also reported this number for other species of Echeveria, such as E. lutea, E. secunda, E. strictiflora, and E. tenuifolia, from some localities of San Luis Potosi [11,13,14,16,17].
In this investigation, we found 13 different chromosome numbers: 2n = 24, 28, 32, 36, 40, 42, 54, 60, 108, 162, 172, 176, and 270, and 2 species that show different levels of polyploidy-aneuploidy. In fact, evidence from the literature indicates a wide diversity of chromosome numbers within the genus Echeveria. Nevertheless, some chromosome numbers as those of E. carmenae [11,13] with chromosome number 2n = 65 has not been reported in this study; in addition to this, there are some polyploids-aneuploids showing chromosome numbers different to those reported here, for example, n = 320 in E. bakery, n = 119 + 2 in E. chiclensis from Ecuador to Argentina [15], which suggests that the genus Echeveria has experienced a strong chromosomic evolution. With respect to the ploidy level, it is well known that the whole genome duplication (WGD) has led to an increase in species richness [30]. Polyploidy as a mechanism can generate individuals capable for the adaptation and colonization of new ecological niches, favoring the survival and reproduction of individuals more capable to adapt to new environments, with respect to diploid individuals [47,65]. In this investigation, a positive and strong correlation between the polyploidy and the chromosome number (r = 0.93, p < 0.0001) in the 25 populations of the 23 species of Echeveria was analyzed ( Figure 5). Although this correlation could be expected, its verification in this genus is interesting, especially when Uhl has pointed out the importance that this phenomenon represents in the evolution and adaptation of this genus [5].
Although polyploidization is an important process in the evolution of angiosperms, it has as a consequence during meiosis: there is a high frequency of non-disjunction of sister chromatids, which is due to the fact that sister chromatids associate in multivalents, rather than in bivalents during meiotic prophase I, resulting in the formation of aneuploid gametes [66]. On the other hand, the cellular machinery and the entire organism become unviable with indefinite increases in DNA and chromosomes, so one of the processes that polyploidization causes is so-called diploidization, which occurs thanks to a series of massive chromosomal rearrangements, including reductions in the number of chromosomes and a significant loss of repetitive sequences and duplicated genes [65,67]. Thanks to these processes, there is a reduction in the size of the genome, which generates enormous variation (for example in Asteraceae) [65,68]. Therefore, the diploidization phenomenon involves mechanically diverse processes, which operate together and in the long term result in the generation of descendants that behave like normal diploids during meiosis, but that reflect vestiges of past polyploidy events in their genomes [67]. Diploidization has been evident in autopolyploid plants, as was demonstrated in Zea mays [69] and has been observed in autotetraploid plants of Zea perennis [70], Festuca sp. [71], and autotetraploid cytotypes of Gibasis schiedeana [72].
In the present investigation, three tetraploid species were found with n = 54 (E. altamirae, E cupreata and E. dactilifera), two pentaploids with n = 30 (E. caamanoi and E. cuicatecana) and a 6x species with n = 81 (E longiflora), in addition to two polyploid aneuploids, one 6x + 10 with n = 86 (E. gibbiflora) and another 6x + 14 (E. novogaliciana), while in diploid species values of n = 12, 14, 16, 18, 20, 21 and 27 were observed. On a larger scale, Uhl [14,16,17] proportionated values of n = 12-100, 119, 135-162 and polyploid up to 13x-, 20x, and 42x, mainly in species of Echeveria from South America, where he confirmed that polyploidy was the base for the evolution of the chromosome number in the species of Echeveria, through diploidization of their genomes. Moreover, Uhl [15] argues that genetic recombination processes in the genomes of polyploid-aneuploid species found in localities across Ecuador and Argentina are due to adaptation to different new environments, when compared to E. gibbiflora and E. novogaliciana from México. Diploidization processes probably allowed these species to survive in other localities distinct from their original environments Uhl [15].
Because the species belonging to the Gibbiflorae series are the most represented in this study, it is important to point out that high diversity was found in the levels of ploidy and in the number of chromosomes and DNA content, and in fact, the two species that had the highest number of chromosomes and DNA content in this study belong to this series ( Table 2). These results coincide with the fact that this series has been reported as highly diversified and widely distributed [7]. The increase in the chromosome number has an advantage in colonizing new environments, as the distribution of this series confirms in Mexico, where it could be hypothesized that the diploidization process was the main mechanism of series diversification. It is relevant to mention that data on the chromosome number and DNA content can help in the phylogenetic resolution of this genus, because even though important molecular studies in the Gibbiflorae series show a monophyletic origin, there are still some phylogenetic relationships that need to be resolved [73].

Nuclear DNA Content and Ploidy Levels in Echeveria
It is important to mention that there are no reports in the literature for the genome size (nuclear DNA content) for any species of Echeveria. Therefore, results of this study are the first reported for the studied species of this genus.
We obtained the DNA content of parenchyma tissue from 25 populations of 23 Echeveria species (Table 2); the highest value of 2C that we found was in E. roseiflora with 7.70 pg, and 1Cx of 0.77pg, while the lowest value was found in 1 of the 2 populations of E. catoce with 2C = 1.26 pg, and 1Cx = 0.63 pg. Although there are no data in the literature on the size of the genome in the genus Echeveria, there are some records of Mexican species of other genera of the Crassulaceae family; one of them is that of Sedum suaveolens from Durango with 2C = 18.20 pg [40] which, as mentioned before, is the highest chromosome number recorded in angiosperms. On the other hand, Sedum burrito, which is cultivated in Guadalajara and Veracruz, Mexico, but is not known in the wild, was 2C = 1.3 pg [40,74]. Other species of the genus Graptopetalum, which is phylogenetically closely related to the genus Echeveria [7], are G. macdougallii from Oaxaca with 2C = 6.70 pg [40] and G. bellum from Chihuahua with 2C = 8.40 pg [40,75]. As can be seen, there is a wide variation in the size of the genome in these genera and in both, some 2C values are even higher than those found in this investigation.
It is also relevant to compare the size of the genome with the level of ploidy and the number of chromosomes, but of the Mexican genera mentioned, these data are only available for S. suaveolens, which is 20x with 2n = 640 and 2C = 18.20 pg [12,40,75]. Meanwhile, E. roseiflora, which was the species with the highest level of ploidy in this study, was 10x with 2n = 270 and 2C = 7.70 pg. However, with respect to European species, Sedum forsteriaum was reported as a diploid species with 2n = 24 and 2C = 0.92 pg [40,76] while we observed 2C = 1.26 pg and 1.29 pg for E catorce and 2C = 1.50 pg for E. shaffneri, both diploid species with 2n = 24. Another European species, S. sediforme, is also 2x but with 2n = 32 and 2C = 1.16 pg [40,76], it can be compared with E. triquiana and E. multicaulis, both diploids with 2n = 32 but with 2C = 2.07 pg and 2C = 1.40 pg, respectively.
In general, within the species analyzed by us, the 2C values in diploid species varied between 2.96 pg in E. perezcalixii and 1.26 pg in E.catorce, while in the tetraploid species, values between 2C = 3.54 pg in E. altamirae and 2C = 2.50 pg in E. cupreata were observed. In E. cuicatecana and E. caamanoi, both pentaploids, values 2C = 2.44 pg and 2C = 1.96 pg were observed, respectively, and in E. longiflora, which is hexaploidy, it was 2C = 2.54 pg. Of the two polyploid-aneuploid species, E. gibbiflora (6x + 10) presents 2C = 3.68 pg and E. novogaliciana presents (6x + 14) 2C = 5.81 pg. These data reflect a certain tendency to increase in the 2C value as the ploidy level increases, but it is highly variable because finally, it also depends on the basic chromosome number and chromosome size for each species.
Although a positive significant correlation was observed between ploidy level and chromosome number, a negative significant correlation between ploidy level and the 1Cx-value was observed. The decrease in the monoploid genome size as polyploidy level increases is a process known as genome downsizing [77] and can be explained as a strategy for the reduction in the number of chromosomes and a significant loss of both, repetitive sequences and duplicated genes related to the diploidization process [65,67]. Wang et al. [77] propose that genome downsizing may be a byproduct of various processes that give rise to smaller genomes, which could offer a selective advantage as an emergent property. As Wendel [67] has mention, angiosperms history includes many multiple events of polyploidy and reduction in chromosome numbers through massive rearrangement, which cause the reduction in genome size [65,78].

Endopolyploidy
As there are not previous reports in the literature of the endopolyploidy in any species of Echeveria, results on endopolyploidy in this investigation are the first records for the studied species of this genus. However, Zonneveld [79] mentions the presence of an endopolyploidy in some genera belonging to the Crassulaceae family, such as: Sedum, Crassula, Sempervivum, and Graptopetalum, which presented endoploidy levels between 8C and 32C.
In this investigation, we observed the existence of endopolyploidy patterns in all the analyzed species, but with differences in the number of endocycles and percentage of nuclei in each endocycle, and observed species that presented from one to four endocycles, which represents endopolyploidy levels from 8C to 64 C (Table 3) in the different species. Also, the presence of two levels of endopolyploidy was observed within the same species and population in E. roseiflora, E. longiflora, E.olivacea, E. juarezensis, E. perezcalixii, and E catorce (Accession number 5469) ( Table 4). Variations in endopolyploidy levels and in the number of nuclei in each endocycle were reflected in the cycle value that showed a range between 0.682 and 2.562 (Tables 3 and 4; Figure 1d, Figure 3a,b,f and Figure 4b,c).
Endopolyploidy represents a metabolic adaptation of plants that favors the survival and reproduction of individuals living in arid environments [80,81] and it is considered an emergent response in species that live in these environments, as it has been observed in species of Mammillaria [34] and Opuntia [24]. Moreover, it has been suggested that the biological significance of endopolyploidy is to provide a high DNA content to sustain the demand of DNA synthesis in the cells of species with small genomes and with specialized functions [82], such as in the endosperm cells of Arum maculatum specialized in nutrition, where the presence of an endopolyploidy pattern with levels up to 24,576C (13 endocycles) has been reported [33]. However, at least for certain tissues, this assumption seems unlikely because the proportionality between the ploidy level and genome size has been inconsistent in some plant families and they do not have a relationship that indicates the need for a particular amount of DNA to maintain cell function [33,60].
The genus Echeveria has a wide distribution and presents different mechanisms that allow them to adapt to different environments. Endopolyploidy has been suggested to play an important role in how plants cope with situations that constitute some form of stress for them. Recently, information has been generated that shows that stress itself, whether biotic or abiotic, can trigger endopolyploidy as part of a plant response to stress, helping to mitigate its effect on the plant. For example, it allows continued growth by endopolyploidy induced cell expansion when growth via cell division is inhibited by a low temperature, or by inducing the formation of larger cells via endopolyploidy which can reduce the number of guard cells and consequently the loss of water, or providing a layer of cells that absorb potentially damaging light and thus protecting the plant against possible damage caused by light [50,53]. Endopolyploidy could also be related to the production of mucus (polysaccharides) to retain moisture, as Zonnevelt [79] suggests. Furthermore, endopolyploidy level is variable to some degree and may even present differences between individuals or populations of the same species in response to different environmental conditions [50,61], as seen in six of the species analyzed in this investigation (E. catorce, E. olivacea, E. juarezensis, E. perez-calixi, E. longiflora, and E. roseiflora). In fact, the cycle values obtained reflect this variability within the same species and population, and can be deduced from the results of the Tukey-Kramer test, where it was observed that several species were grouped within the same block, despite the fact that the range of variation of the cycle values was wide.

Conclusions
A high variability of the number of chromosomes (2n and x) could be observed among the species of the genus Echeveria analyzed; 13 different chromosome numbers were found, and 2n = 54 and x = 27 were the most frequent. This variability was particularly observed in the diploid species that presented 7 different basic chromosome numbers that were x = 12, 14, 16, 18, 20, 21 and 27, the last one being the most frequent. Meanwhile, in the pentaploid species x = 12, and in tetraploid, hexaploid, decaploid, and hexaploidy-aneuploid species x = 27 was observed. The chromosome number (2n) of 16 species of the genus is reported for the first time in this research, which coincided with some of those reported by Uhl for other species.
Results of genome size were reported for the first time in 25 populations of 23 species of Echeveria in this research. The lowest value was observed in E. catorce where 2n = 2x = 24 with 2C DNA = 1.26 pg; the highest value of 2C DNA = 7.70 pg was observed in E. roseiflora 2n = 10x = 270. Moreover, the presence of endopolyploidy pattern in the leaf parenchyma of species of Echeveria was also observed for the first time in this genus, corresponding to cells with endopolyploidy levels from 8C up to 64C and cycle values between 0.690 and 2.562 in the different species of Echeveria analyzed.
The variability observed in the basic number of chromosomes, as well as the presence of different ploidy levels and two polyploid-aneuploid species in the analyzed species, coincide with several of Uhl's [13][14][15] observations in various species of this genus from Mexico and from Central and South America. These data, together with the positive correlation found between ploidy level and the number of chromosomes, and the negative correlation between polyploidy levels and 1Cx-value, could provide evidence of the existence of different polyploidization-diploidization cycles in each species, and as was suggested by Wendel [67], generate changes in the phenotypes on which natural selection acts, resulting in the diversification of the species. Additionally, we found that the species analyzed in this research were polysomatic, and it has been suggested that endopolyploidy allows plants to adapt to various types of stresses and survive in different environments [50,60,61]. Thus, all these processes acting simultaneously may be the cause of the high degree of endemism as well as the wide diversification and distribution of this genus. Although no clear differences could be found between the different series of the genus included in this study, the obtained results are relevant in studies about systematics, phylogeny, karyotype evolution [67,83], and improvement programs of the genus Echeveria and ornamental hybrid plant production [84].