First karyotype description and nuclear 2C value for Myrsine (Primulaceae): comparing three species

Abstract Cytogenetic studies in Primulaceae are mostly available for herbaceous species, and are focused on the chromosome number determination. An accurate karyotype characterization represents a starting point to know the morphometry and class of the chromosomes. Comparison among species within Myrsine, associating these data with the nuclear 2C value, can show changes that led the karyotype evolution. Here, we studied three Myrsine species [Myrsine coriacea (Swartz, 1788) Brown ex Roemer et Schultes, 1819, Myrsine umbellata Martius, 1841 and Myrsine parvifolia Candolle, 1841] that show different abilities to occupy the varied types of vegetation within the Brazilian Atlantic Forest. Cytogenetic characterization showed some individuals with 2n = 45 chromosomes for Myrsine parvifolia and Myrsine coriacea, with most individuals of the three species having 2n = 46. The first karyograms for Myrsine were assembled and presented morphologically identical and distinct chromosome pairs. In addition, differences in the mean 2C nuclear value and chromosome morphometry were found. Therefore, the first description of the Myrsine karyotype has been presented, as well as the nuclear 2C value. The procedures can be applied to other Myrsine species for future investigations in order to better understand its effects on the differential spatial occupation abilities shown by the species in Brazilian Atlantic Forest.

One interesting ecological aspect observed in Neotropical species of Myrsine that occur in Brazil is that some of them occur in more than one biome, as Cerrado, Atlantic Forest, and Amazonian Forest, while others are restricted of one of these biomes, as Atlantic Forest (BFG 2015). Among species that occur in Atlantic Forest, for example, some are able to occupy different types of vegetation within this biome, including Restinga Vegetation, High Altitude Campos, Rocky Outcrops, Ombrophyllous and Mixed Ombrophyllous Forests, while others are able to occupy just one type of vegeta-tion (Freitas and Kinoshita 2015). Considering the distinct ecological aspects, cytogenetic studies are relevant to show other differences between these species.
Studies combining cytogenetics and nuclear DNA content have offered data for understanding evolutionary processes in different species Carvalho 2008, Kolář et al. 2013). Measurement of the nuclear DNA content is complementary to cytogenetic information and is useful for detecting genome size variations between related species (Marhold et al. 2010, Kolář et al. 2013. Fine adjustments in cytogenetic procedures, combining advances in microscopy and image analysis systems, can provide accurate karyotype characterization for Myrsine species. Here, we study three species of Myrsine that occur in contrasting types of vegetation of the Brazilian Atlantic Forest, aiming to determine the chromosome number, describe the karyotype and measure the nuclear DNA.

Plant samples
Three species were selected for this study:

In vitro plantlet recovering
Fruit pericarp was manually removed and the seeds were desinfested according to Oliveira et al. (2013) and germinated in a medium composed of MS salts (Sigma) and vitamins (Murashige and Skoog 1962), 30 g L -1 sucrose (Sigma), 7 g L -1 agar and 2.685 µM naphthaleneacetic acid (NAA, Sigma). Solanum lycopersicum and P. sativum seeds were subjected to the same disinfestation procedure and inoculated in medium without NAA. Germination was done at 25 °C under a 16/8 hours (light/dark) regime.

Nuclear 2C value measurement
In order to adapt the FCM for Myrsine, the following procedures were done: (a) initially, from leaves collected in the field of male and female individuals (samples) and of the two standards; (b) afterward, replacing the dithiothreitol antioxidant by polyethylene glycol (PEG) in nuclei isolation buffer; and (c) from leaves of the samples and P. sativum plantlets in vitro cultivated.
Nuclei suspensions were obtained by co-chopping (Galbraith et al. 1983) leaf fragments (1 cm 2 ) cut from each sample (Myrsine species) and standard (S. lycopersicum or P. sativum). The suspensions were processed and stained following Otto (1990) and Praça-Fontes et al. (2011) and analyzed with the flow cytometer Partec PAS II/III (Partec GmbH). Myrsine genome size was measured by multiplying the 2C value of the internal standard using the fluorescence intensity corresponding to G 0 /G 1 nuclei peak. Mean 2C values were compared by the F test at 5% probability.
Slides were prepared and stained according to Carvalho et al. (2007) and analyzed on a Nikon eclipse Ci-S microscope (Nikon). Prometaphases and metaphases were captured using the 100× objective and a CCD camera (Nikon Evolution TM ) coupled to a Nikon microscope 80i (Nikon). About 100 slides were analyzed for each Myrsine species. Chromosome morphometry was characterized and the class was determined as proposed by Levan et al. (1964) and reviewed by Guerra (1986).
Using chromosome morphometric data (total, short and long arm length), the standardized Euclidean Distance and Unweighted Pair-Group Method Average (UP-GMA) was applied to each species. In addition, the value of the relative size (% size in relation to sum of the mean values of total length, Table 1) of each chromosome was compared among species by the Scott-Knott test at 5% probability. Analyses were made using the software R 3.2.4 (R Core Team 2016). Chrom = chromosomes; Total = total length; SD = standard deviation; Long/Short = arm length; r = arm ratio -long/short; M = metacentric; SM = submetacentric; A = acrocentric; Relative size = % size in relation to sum of the mean values of total length; Sum = sum of the mean values.

In vitro plantlet recovering
Approximately 60 days after in vitro inoculation, plantlets were obtained for the three Myrsine species. All plantlets exhibited sufficient and morphologically normal leaves and roots for FCM and cytogenetic analyses, respectively.

Nuclear 2C value measurement
FCM analysis performed on leaves collected in the field did not result in histograms showing profile G 0 /G 1 peaks. So, dithiothreitol antioxidant was replaced by PEG in the nuclei isolation buffer OTTO I. This change provided G 0 /G 1 peaks, exhibiting a coefficient of variation (CV) less than 5% for M. umbellata and the two internal standards. The channel of the P. sativum G 0 /G 1 peak however was closer to M. umbellata than S. lycopersicum. Thus, based on linearity international criteria for FCM, P. sativum was the standard chosen for the next measurements. The mean 2C value of the male (2C = 6.65 pg ± 0.02) and female (2C = 6.67 pg ± 0.11) M. umbellata individuals were statistically identical by the F test. Considering these previous results, the 2C value was measured from leaves of in vitro plantlets. The mean values were 2C = 4.81 pg ± 0.05 for M. parvifolia, 2C = 6.60 pg ± 0.14 for M. coriacea and 2C = 6.63 pg ± 0.13 for M. umbellata. The mean value of the M. umbellata in vitro plantlets was statistically identical to the males and females in the field. Therefore, the mean value adopted for this species was 2C = 6.65 pg, which was statistically equal to the M. coriacea.

Cytogenetic analysis
Roots exposed to a 12 h APM provided prometaphases, exhibiting chromosomes at a distinct chromatin compact level, and metaphases. Enzymatic maceration in 1:5 pectinase solution ensured the chromosomes remained inside the cell, allowing an accurate determination of 2n = 45 or 2n = 46. Chromosome number of 2n = 45 was found for 12.60% individuals of M. parvifolia and 8.45% of M. coriacea, with 2n = 46 for the three species. Based on these results, the next slides were made from roots of particular seedlings with 2n = 45 or 2n = 46. Root maceration with 1:20 enzymatic pool for 1h 30 min supplied chromosomes no damage to the chromatin structure, without overlapping, with well-defined centromeres and free of cytoplasm debris (Fig. 1). Karyotype characterization was possible only after carefully testing the time and concentration of the APM antitubulin and cell wall enzymes. Myrsine parvifolia presented a greater total sum of the length of the chromosomes despite having less nuclear DNA content. For this species only, we found prometaphase chromosomes showing low level of chromatin compaction (Fig. 2a), resulting in a higher sum of the total  (Table 1). Myrsine coriacea and M. umbellata did not show pronounced variation in chromatin compaction, but the quality of the chromosomes allowed us to characterize the karyotype and to assemble the karyogram (Fig. 2b-c, Table 1).
As the mean 2C values of M. coriacea (6.60 pg) and M. umbellata (6.65 pg) were statistically identical, the Scott-Knott test was used to compare the relative size (Table 1) of each chromosome of these species. Chromosomes 1, 2, 6, 7, 11, 14, 19 and 23 differed between the species, while the others were statistically identical (Fig. 3d, Table 2).

Discussion
The first step in FCM was to define the best antioxidant and internal standard. The presence of secondary metabolites in the Myrsine leaves, such as tannins, saponins, flavonoids and steroids (Abbi et al. 2011) made this challenging. These compounds probably prevented us from measuring the 2C value in individuals from the field when the OTTO I buffer (Otto 1990) was supplemented with dithiothreitol. Cytosolic compounds can reduce or inhibit the interaction of the fluorochromes and DNA during the nuclei staining step (Noirot et al. 2003). Antioxidants inhibit this interference, preserving the chromatin structure (Shapiro 2003). Nevertheless, the dithiothreitol was not efficient at providing nuclei suspensions suitable for FCM. Thus, this compound, which is more specific for molecules that possess free sulfhydryl groups, was replaced by PEG because of its wide spectrum for antioxidant activities, an effect called PEGylation (Term Fisher Scientific 2016). Due to this effect, PEG was more efficient at inhibiting the action of cytosolic compounds, resulting in G 0 /G 1 peaks for M. umbellata and P. sativum with CV below 5%. Owing to the linearity parameter, P. sativum was a more adequate standard relative to S. lycopersicum, which reduced measurement errors. Table 2. Chromosome groups of the Myrsine karyotype suggested from karyogram evaluation ( Fig. 2 and Table 1) and confirmed by UPGMA clustering (Fig. 3a-c).
Secondary metabolite interference was completely resolved for other Myrsine species using in vitro plantlets propagated in a controlled environment. FCM measurements from leaves collected in the field may have been influenced by environmental conditions. Secondary metabolite production is influenced by humidity, temperature, light intensity and the availability of water and nutrients (Akula and Ravishankar 2011). Thus, the conditions at each elevation gradient can be associated with the FCM result, suggesting a differentiated production of secondary metabolic compounds for Myrsine at distinct altitudes.
Genome size in Myrsine had only been reported for M. africana as 2C = 2.46 pg (Hanson et al. 2003), which was measured by Feulgen microdensitometry using Vigna sp. as standard. Levels of endoreduplication in cells of V. radiata, varying from 2C to 64C, were reported by Pal et al. (2004). Thus, the differences, which were about 200% between the values found for Myrsine species in this study and the value observed for M. africana, can be related to the C value of V. radiata used as reference.
As well as for FCM, karyotype data about Myrsine species in the literature are very limited, with only the chromosome number reported (Beuzenberg and Hair 1983, Dawson 1995, Dawson 2000, Molero et al. 2002, Molero et al. 2006, Rice et al. 2015. In vitro Myrsine plantlets were fundamental for providing sufficient quantities of roots for the cytogenetic study independent of the reproductive period. Meticulous standardization of the antimitotic agent and enzymatic maceration were also essential for accurate chromosomal characterization.
Chromosome number 2n = 46 (Beuzenberg and Hair 1983, Dawson 1995, Dawson 2000, Molero et al. 2002, Rice et al. 2015 and 2n = 48 (Molero et al. 2006) had been reported, but this was the first record of 2n = 45. The odd chromosome number 2n = 45 was well-marked by absence of the homologue pair of the chromosome 23 (Fig. 2). So, other cytogenetic approaches should be performed from Myrsine individuals separately to know the cause of this aneuploidy. Some chromosome groups determined by statistical analysis are morphologically distinct, such as chromosomes 22 and 23 of M. coriacea. Although clustered (Fig. 3b), these chromosomes are cytogenetically distinct, with 22 being acrocentric and 23 submetacentric (Fig. 2b, Table 2). Likewise, distinct chromosomes clustered in M. parvifolia (Fig. 3a, Table 2) and M. umbellata (Fig. 3c, Table 2). Chromosome 1 of M. coriacea and 1 and 2 of M. parvifolia and M. umbellata presented the highest contrast, considering the morphology and Euclidean distances (Fig. 2,  Fig. 3). Similarities and differences regarding relative size (% size in relation to sum of the mean values of total length, Table 1) were shown between M. coriacea and M. umbellata through the Scott-Knott test. The similarities, which were shown for some chromosomes, imply that these species could have originated from a common ancestor. The distinct chromosomes are likely to be attributed to karyotype changes that happened throughout their evolution, altering the chromosome relative size and contributing to taxa diversification. Comparative investigations of the karyotypes of related species have usually been applied to infer the evolutionary role of karyotypic modifications in different taxa and to describe the pattern and directions of chromosomal evolution within a group (Stebbins 1971, Soltis and Soltis 2012, Amaral-Silva et al. 2016.
Based on the constant chromosome number displayed by Myrsine species, interspecific variation of the nuclear 2C value between M. parvifolia compared to M. coriacea and M. umbellata was also caused by karyotype alterations. The changes to the nuclear DNA content have also been attributed to structural rearrangements and/or heterochromatin polymorphisms (Pellicer et al. 2014, Amaral-Silva et al. 2016.
In conclusion, the first karyotype description and data about nuclear 2C value were shown for three Myrsine species. Besides of the comparison between them, these data represent the basis to understand karyotype evolution in Myrsine.