Cerrado is a biodiversity hotspot that has lost ca. 50% of its original vegetation cover and hosts ca. 11,000 species belonging to 1,423 genera of phanerogams. For a fraction of those species some cytogenetic characteristics like chromosome numbers and C-value were available in databases, while other valuable information such as karyotype formula and banding patterns are missing. In order to integrate and share all cytogenetic information published for Cerrado species, including frequency of cytogenetic attributes and scientometrics aspects, Cerrado plant species were searched in bibliographic sources, including the 50 richest genera (with more than 45 taxa) and 273 genera with only one species in Cerrado. Determination of frequencies and the database website (http://cyto.shinyapps.io/cerrado) were developed in R. Studies were pooled by employed technique and decade, showing a rise in non-conventional cytogenetics since 2000. However, C-value estimation, heterochromatin staining and molecular cytogenetics are still not common for any family. For the richest and best sampled families, the following modal 2n counts were observed: Oxalidaceae 2n = 12, Lythraceae 2n = 30, Sapindaceae 2n = 24, Solanaceae 2n = 24, Cyperaceae 2n = 10, Poaceae 2n = 20, Asteraceae 2n = 18 and Fabaceae 2n = 26. Chromosome number information is available for only 16.1% of species, while there are genome size data for only 1.25%, being lower than the global percentages. In general, genome sizes were small, ranging from 2C = ca. 1.5 to ca. 3.5 pg. Intra-specific 2n number variation and higher 2n counts were mainly related to polyploidy, which relates to the prevalence of even haploid numbers above the mode of 2n in most major plant clades. Several orphan genera with almost no cytogenetic studies for Cerrado were identified. This effort represents a complete diagnosis for cytogenetic attributes of plants of Cerrado.

Cerrado , cytogenetics, database, scientometrics, shiny-R

Cerrado, a phytogeographic domain from Brazil, is the third largest biodiversity hotspot in the world considering species endemism and degree of threat (Myers et al. 2000; Forzza et al. 2012). In fact, it has suffered more anthropic impact than the Amazonian tropical forest (Spera et al. 2016). With originally 2m km² of wilderness area, of which ca. 50% are currently cultivated, this region hosts ca. 11,000 species belonging to 1,423 genera of 171 families of phanerogams (Sano et al. 2008; Ferreira et al. 2016). Classical cytogenetic studies, comprehending somatic or meiotic chromosome counts are known for a considerable portion of species occurring in Cerrado but in many cases they were not performed on plants from this region. Meanwhile, studies of genome size, differential staining and molecular cytogenetics are very rare.

Since the chromosome number is the most basic karyotype feature and it can be observed by conventional staining of meristem cells or pollen mother cells, most cytogenetic databases deal only with this attribute (Peruzzi and Bedini 2014; Rice et al. 2015). Estimates of genome size, which are currently addressed mostly by comparing the relative fluorescence of propidium iodide stained nuclei measured in a flow cytometer to that of known patterns, are compiled in the Kew C-value database (Doležel and Greilhuber 2010; Garcia et al. 2014a). Other valuable information such as: karyotype formula, dependent on the relative position of centromere along the chromosome and chromosome length (Guerra 1986), silver staining of nucleolar organizer regions (AgNOR) (Vieira et al. 1990), and heterochromatin staining after treatment with acids, bases or denaturation (Schwarzacher et al. 1980; Guerra 2000; Barros e Silva and Guerra 2010) are not included in databases. In contrast, fluorescence in situ hybridization (FISH), in which a DNA probe produced by molecular biology methods anneals with chromosome preparations and is detected by fluorescence has gained more attention (Roa and Guerra 2012; Garcia et al. 2014b; Roa and Guerra 2015).

The aim of this study was to assess the current cytogenetic knowledge of Cerrado plants, aiming to provide a consistent database that didn’t exclude any attribute and diagnose recent progress in the field.

699 (16.1%) species of the searched list have known chromosome numbers with a general median of 2n = 28. Fig. 3 shows two ploidy related prevalent chromosome numbers; neopolyploids in Commelinids, which corresponds to Paspalum counts (2n = 20 and 40), and several paleopolyploid genera as Habenaria with 2n = 42 (Monocots) and Mikania with 2n = 36 (Eudicots-Asterids-Campanulids). An unusual conspicuous high frequency of 2n = 36 for the whole sample, corresponds mainly to genera Manihot Mill., 1754 (Euphorbiaceae), Tibouchina Aubl., 1775 (Melastomataceae) and Mikania Willd., 1803 (Asteraceae). Intraspecific chromosome number variation occurs in 14.3% of species, being polyploidy the cause in 88% of them. At the genus level, most high chromosome numbers are multiples of one of the lowest modal numbers of the genus, i.e. base numbers (database website). The even/odd ratio of n is higher for the groups with “n greater than the mode” in almost every clade, as expected, except for Asterids-Campanulids (ratio 1.5), while Asterids-Lamiids shows the greatest bias towards even numbers (ratio 6.5) in karyotypes with n greater than the mode (Table 2).

Figure 3.

Dot-plot of observed and calculated 2n numbers for Cerrado plant species considering APG major clades. The three most common numbers (frequency ties separated by comma) and sample size in parentheses are indicated. Dot size varies depending on the maximum frequency of each group.

Only 88 C-values were found for the taxa of Cerrado, most of them ranging from 2C = ca. 1.5 pg to ca. 3.5 pg with a median value of 2.38 pg (Fig. 4). That corresponds to 1.25% of the addressed species list. For a sample of known TCL (or TCA) and C-value, a correlation analysis, although significant, resulted in low r² values (weak to moderate). Applying a linear model to predict 2C-value based on TCL, resulted in prediction intervals of ca. 2pg. TCL and TCA values together with some predicted 2C- values are shown in Suppl. material 5, 6 and the database website.

Figure 4.

Violin-plots for genome sizes in pg for Cerrado angiosperms. Y-axis contains clades, three highest peaks of the Gaussian kernel density and sample size. Bandwidth of the density estimator customizable in http://cyto.shinyapps.io/cerrado

There is strong interest for genomics and genetics of Cerrado species (Souza et al. 2016). Knowledge of basic data as chromosome number, ploidy level and genome size is critical for the selection of model species (Kelly et al. 2012). Though recent cytogenetic studies tend to incorporate phylogenetic frameworks and correlate cytogenetic with ecological variables, this kind of studies are scarce for Cerrado (Glick and Mayrose 2014; Suda et al. 2015; Silveira et al. 2016). As noted by Guerra (1990), previous to the 1990 decade, cytogenetic studies of authors from U.S.A. prevailed. After that period, mitotic analysis increased compared to meiotic studies, which relate to the establishment of cytogenetic laboratories allowing the treatment with antimitotics of root meristems. Still, the amount of chromosome number data for species occurring in Cerrado is lower than the world total, 16.1 vs. 20% (Rice et al. 2015) and several “orphan” genera were identified, such as Stachytarpheta (Verbenaceae), Microlicia (Melastomataceae) and Leiothrix (Eriocaulaceae). Unlike other techniques, C-banding and silver staining studies did not increase in recent decades, which might be related to low repetitiveness of methods (Guerra 1990). Despite recent efforts, the percentage of known CerradoC-values is still lower than the 2.1% for all angiosperms (Garcia et al. 2014a). Data of TCA, and TCL for a larger number of species were intended to be used as proxy for genome size after an analysis that resulted in weak to moderate significant correlation with 2C-value in pg. Though they can give a rough estimate of the magnitude of the genome, the prediction interval is high considering most genomes fall within a 2 pg range.

Discontinuities in chromosome numbers seen as multimodal distributions indicate that higher numbers are polyploids generated by genome duplication. A polyploid origin for most high n numbers is also evidenced by the high (> 1) even/odd ratios of n (Rice et al. 2015). In contrast, low even/odd value (for n greater than the mode) is rare and suggests increased dysploidy and/or low frequency of polyploidy, compared with groups that follow the opposite general trend.

The Database of Cytogenetics of Cerrado plants and its website is presented, making an important step in facilitating access to most known cytogenetic attributes (http://cyto.shinyapps.io/cerrado/). The amount of chromosome number data for Cerrado is lower than the world total. The chromosome number, is still lacking in several rich genera and families, and therefore, they can be considered “orphan”. This complete lack of information also happens for 70% of the single-species genera addressed. Analysis of chromosome numbers at several taxonomical levels revealed a straightforward relationship between polyploidy and high chromosome numbers. Regarding other techniques, like heterochromatin staining, molecular cytogenetics and C-value estimation, they have been applied to a very small percentage of species, however, those studies have been steadily increasing since 2000. Though cytogenetic data are the basis for some evolutionary and ecological studies, there is a lack of those kind of interdisciplinary studies for Cerrado.

FR wishes to thank CAPES for their financial support, through a PNPD fellowship. MPCT is supported by productivity grants CNPq/FAPEG/Pró-Centro-Oeste (563839/2010-4 and 201110267000125), Pronex – FAPEG/CNPq (Funding Announcement No. 07/2012) and CNPq (447754/2014-9). This publication was funded by the PPGS CAPES/FAPEG F. A. No. 08/2014. Current research is developed in the context of the National Institutes for Science and Technology (INCT) in Ecology, Evolution and Biodiversity Conservation, supported by MCTIC/CNpq (proc. 465610/2014-5) and FAPEG.