Intrachromosomal karyotype asymmetry in Orchidaceae

Abstract The asymmetry indexes have helped cytotaxonomists to interpret and classify plant karyotypes for species delimitation efforts. However, there is no consensus about the best method to calculate the intrachromosomal asymmetry. The present study aimed to compare different intrachromosomal asymmetry indexes in order to indicate which are more efficient for the estimation of asymmetry in different groups of orchids. Besides, we aimed to compare our results with the Orchidaceae phylogenetic proposal to test the hypothesis of Stebbins (1971). Through a literature review, karyotypes were selected and analyzed comparatively with ideal karyotypes in a cluster analysis. All karyotypes showed some level of interchromosomal asymmetry, ranging from slightly asymmetric to moderately asymmetric. The five tested intrachromosomal asymmetry indexes indicated Sarcoglottis grandiflora as the species with the most symmetrical karyotype and Christensonella pachyphylla with the most asymmetrical karyotype. In the cluster analysis, the largest number of species were grouped with the intermediary ideal karyotypes B or C. Considering our results, we recommend the combined use of at least two indexes, especially Ask% or A1 with Syi, for cytotaxonomic analysis in groups of orchids. In an evolutionary perspective, our results support Stebbins’ hypothesis that asymmetric karyotypes derive from a symmetric karyotypes.


Introduction
The karyotype is the first phenotypic expression of the genotype and provides an overview of the organization of the genetic material in the chromosome (Guerra, 2008).Among the information that can be extracted from karyotypes, i.e. number and morphology of the chromosomes, diversity of heterochromatic bands, gene location, etc., a very peculiar characteristic stands out: the karyotype asymmetry, which is the subject of long debates.Changes in karyotype symmetry often involve modifications in chromosome size and morphology usually caused by DNA sequence expansions or deletions or by centric fusion/fissions (accompanied by disploidy) (Weiss-Schneeweiss and Schneeweiss, 2013).
The search for an index that reflected the karyotype asymmetry started with Lewitsky (1931) and was followed by Huziwara (1962), Arano (1963), Stebbins (1971), and many other authors (for a detailed discussion see Peruzzi and Erolu, 2013).For a long time, these indexes were employed by many cytogeneticists and cytotaxonomists to discuss the taxonomic relationships among related species (Dematteis, 1998;D'Emerico et al., 1999;Selvi et al., 2006;Felix et al., 2007;Peruzzi et al., 2009;Souza et al., 2010).Stebbins (1971) suggested that asymmetric karyotypes were originated from symmetrical ones, which has not been properly tested until now.
The existing indexes are separated into two groups: interchromosomal asymmetry indexes, which quantify the heterogeneity in chromosome size, and intrachromosomal asymmetry indexes, which quantify the relative differences in the centromere position among chromosomes of a complement (Stebbins, 1971;Peruzzi and Erolu, 2013).Among the interchromosomal asymmetry indexes, A 2 (Romero-Zarco, 1986) and CV CL (Paszko, 2006) are the most used due to their accuracy in the evaluation of chromosome dissimilarities (Chiarini and Barboza, 2008;Souza et al., 2010;Pierozzi, 2011;Alves et al., 2011;Assis et al., 2013).However, there is no consensus about the best method for calculating the intrachromosomal asymmetry (Romero-Zarco, 1986;Peruzzi and Erolu, 2013).
Among the proposed intrachromosomal asymmetry indexes, the following stand out: The four categories of Stebbins (1971): from A to D according to the proportion of acrocentric and/or telocentric chromosomes in a karyotype, i.e. proportion of chromosomes with a ratio between chromosome arms < 2:1.The four categories have subtypes 1 to 3 according to the ratio between the large/small chromosome arms, giving a total of 12 categories (Table 1).
The total form percentage (TF%; Huziwara, 1962): ratio between the sum of the short arms (p) length and the sum of the total chromosomes length: The karyotype asymmetry index percentage (Ask%; Arano, 1963): ratio between the length of the long arms (q) of the chromosome set and the total length of the chromosome set: The symmetric index (Syi; Greilhuber and Speta, 1976): ratio between the average length of the short arms (p) and the average length of the long arms, multiplied by 100: Syi = (Smean of p length/Smean of q length) ´100 The intrachromosomal asymmetry index A 1 (Romero-Zarco 1986): sum of the ratio between the average length of the short arms in each homologous pairs (b i ) and the average length of the long arms in each homologous pair (B i ) divided by the number of homologous chromosome pairs (n): This proposal was later modified by Watanabe et al. (1999), who created the asymmetry index A, and followed by Peruzzi and Erolu (2013), who called the same index M CA.This index results from the sum of the ratio between the differences in the long arm length (B i ) and the short arm length (b i ) of each chromosome and the sum of the lengths of the long and short arms of each chromosome (B i + b i ).The sum is divided by the haploid chromosome number (n): The coefficient of variation of the centromeric index CV CI (Paszko, 2006): based on the index of interchromosomal asymmetry A2 (standard deviation/total average of the chromosome length): With so many ways to calculate the intrachromosomal asymmetry, Zuo and Yuan (2011) developed six models of ideal karyotypes to test the accuracy of these methods.These authors observed that the CV CI does not reflect the intrachromosomal asymmetry in karyotypes when it is composed by telocentric and/or acrocentric chromosomes.Since the standard deviation decreases in acrocentric or telocentric chromosomes, the CV CI is more suitable to indicate the heterogeneity of the centromeric index, i.e. how different is the position of the centromeres among the chromosomes of the complement (see Zuo and Yuan, 2011).Peruzzi and Erolu (2013) discouraged the use of any intrachromosomal asymmetry index for karyotypes with small chromosomes (£ 1 mm), due to the inaccuracy in the arms' measurement.However, the mean chromosome size in plants is 1.5-2.0mm.Moreover, if the suggestion of Peruzzi and Erolu (2013) is followed, the chromosome symmetry analysis will be prohibitive in a large number of plant species, including a great part of Bromeliaceae, Fabaceae and Orchidaceae species.
Considered one of the most diverse and taxonomically complex plant families among the angiosperms, Orchidaceae comprises 25,971 species with global distribution (Pridgeon et al., 1999;Joppa et al., 2011).The Orchidaceae present a large karyotype variation, with all types of chromosome morphology distributed in species with chromosome numbers varying from 2n = 12 in Erycina pusilla (L.) N.H.Williams & M.W.Chase (Felix and Guerra, 1999) to 2n = 240 in Epidendrum cinnabarinum Salzm.ex Lindl.(Guerra, 2000;Felix and Guerra, 2010;Assis et al., 2013).Except for the subfamily Cypripedioideae, Orchidaceae species are characterized by small chromosomes (Larsen, 1968;Okada, 1988).Table 1 -Intrachromosomal asymmetry indexes.The total of 12 indexes is composted of the four categories of Stebbins (1971) -A to D according to the proportion of acrocentric and/or telocentric chromosomes in a karyotype -and subtypes 1 to 3 according to the ratio between the large/small chromosome arms in each of these.
Ratio: largest/smallest chromosomes Proportion of chromosomes with arm ratio < 2:1 0.0 0.01 -0.5 0.51 -0.99 1.0 The wide karyotype diversity observed in Orchidaceae makes this plant family an excellent group for evaluating the applicability of karyotype asymmetry indexes, especially the intrachromosomal index.Thus, the present study aimed to compare different intrachromosomal asymmetry indexes, in order to indicate the most efficient for the estimation of karyotype asymmetry in orchids, including species with small chromosomes.Besides, we aimed to compare our results with the Orchidaceae phylogenetic proposal to test the hypothesis of Stebbins (1971) that asymmetric karyotypes derived from symmetric ones.

Chromosome measurements
A literature search was performed to select informative photographic records of metaphases quality -clear identification of centromere and secondary constrictionsand the available voucher (Table 2).
The arm ratio (r = length of the long arm/length of the short arm) was used to classify the chromosomes as metacentric (M: r = 1.00 to 1.49), submetacentric (S: r = 1.50 to 2.99), acrocentric (A: r ³ 3.00) and telocentric (T: r = ¥), according to Guerra (1986).We did not consider differences between acrocentric and telocentric chromosomes.For chromosome measurements we used Imagetool ® software version 3.0 (available at http://compdent.uthscsa.edu/dig/itdesc.html)calibrated with scales available on the selected images.

Cluster analysis
The intrachromosomal index values were separately used for cluster analysis.The obtained values were categorized and used to define the limits of each category (Table S1).For the cluster analysis we used the UPGMA algorithm (unweighted pair-group method with arithmetic means) implemented in the software Mesquite ® (Maddison  and Maddison, 2015).Ten trees were generated for each index using the distances from the data matrix by majority consensus.Subsequently, only one consensus tree was stored.The software Dendroscope ® (Huson and Scornavacca, 2012) was used to root the tree with the most ideal symmetrical karyotype (karyotype A) as an outgroup, following the hypothesis of Stebbins (1971).

Statistical analysis
To test the hypothesis of Stebbins (1971), the mean values of the interchromosomal index A2 and the intrachromosomal asymmetry indexes were separately used, in order to compare the karyotype asymmetry levels for each subfamily.The variation between the mean values of asymmetry indexes for the subfamilies was compared statistically by ANOVA followed by Tukey's test using BioEstat v.5.3 (Ayres et al., 2007).

Cluster analysis for intrachromosomal indexes
Cluster analysis using the values found for TF% grouped most species with the ideal karyotype C, the Sarcoglottis grandiflora with the ideal karyotype B and Christensonella pachyphylla with the ideal karyotype D (Figure S1).Ask% and A 1 indexes presented identical trees (Figure S2) in the cluster analysis, with most species grouped with ideal karyotype B, plus a clade, separated into two groups: (1) a polytomy with the ideal karyotypes D, E and F and (2) a group with Christensonella pachyphylla, C. subulata (Lindl.)Szlach.and the ideal karyotype C. The cluster analysis for A and Syi also formed identical trees (Figure S3), similar to trees obtained with Ask% and A1 (Figure S2).A difference was found with Christensonella subulata, which was grouped with most species and the ideal karyotype B (instead of C).The indexes Ask% and A1 (Figure S2), and Syi and A (Figure S3) grouped Sarcoglottis grandiflora as a sister group of karyotype A. The indexes Ask%, A1, Syi and A provided more consistent groups (Figure 2), reflecting the species karyotype composition in the ideal karyotypes, as proposed by Zuo and Yuan (2011).
Comparing the four most congruent intrachromosomal indexes, Ask%, A1, Syi and A, with the current proposed Orchidaceae phylogeny (Chase et al., 2015), all indexes presented similar mean and mode values for Orchidoideae and Cypripedioideae (Figure 3).Index A did not detected a difference among subfamilies, after Tuckey's test (despite the F = 4.1420, p = 0.0203).The indexes Ask% and Syi indicated that karyotypes from Epidendroideae and Orchidoideae are more asymmetrical than Cypripedioideae -the most basal subfamily among the three (F = 4.4915 and 4.7008, respectively; p = 0.01 for both indexes).The A1 index suggested Epidendroideae as the most asymmetric karyotype among subfamilies (F = 5.77, p = 0.0054).

Discussion
The inter and intrachromosomal asymmetry values observed here corroborate previous studies, with a slight variation for some species, such as Epidendrum   For each subfamily the mean value (dot), the amplitude of variation (bar), the number of species analyzed and mode value (last two data in the parenthesis, respectively) are presented.
paniculatum Ruiz & Pav., E. fulgens Brongn.(Assis et al., 2013), Cyclopogon calophyllus (Barb.Rodr.)Barb.Rodr.and C. elatus (Sw.)Schltr.(Grabiele et al., 2013).Therefore, we can observe that the relationship between the two kinds of asymmetry (intra and interchromosomal) is not always unidirectional, but it is a result of complex rearrangements that modify both the centromere position and the chromosome size in a karyotype.

The interchromosomal index
The A 2 index employed here yielded values close to zero for some species, mainly in the subfamilies Epidendroideae and Orchidoideae.In such cases, the index reflects a conservation among chromosome size in the karyotype; other species, however, presented high A 2 values.
The highly asymmetric karyotypes could be the result of chromosome rearrangements, what could also cause bimodality, as observed in Cephalanthera damasonium (Epidendroideae; Moscone et al., 2007) and Pteroglossa lurida (Orchidoideae; Martinez 1984), both with A 2 = 0.60.The origin of bimodal karyotypes could be due to the loss of chromosome segments after polyploidy, resulting in the formation of smaller chromosomes (Weiss-Schneeweiss and Schneeweiss, 2013), or due to unequal translocations (Stebbins, 1971), differential amplification of heterochromatic regions (de la Herrán et al., 2001), or even in the hybridization between species with chromosome sizes.All these events increase the interchromosomal asymmetry by increasing the morphological discontinuities between chromosomes in a karyotype.

The intrachromosomal indexes
Regarding the intrachromosomal asymmetry, we showed that Orchidaceae karyotypes ranged from slightly asymmetric to moderately asymmetric.The intrachromosomal asymmetry is defined by the presence of a greater number of acrocentric/telocentric chromosomes in relation to the metacentric and submetacentric ones, a consequence of changes in centromere position (Stebbins, 1971) -in which case the chromosome rearrangement could affect all  chromosomes in the same way and even increase the karyotype asymmetry.Therefore, the efficacy of the intrachromosomal asymmetry indexes is dependent on the precise identification of the centromere and a well-defined chromosome morphology, and not on chromosome size.The indexes Ask% and A 1 proved to be more useful in determining the intrachromosomal asymmetry, even in species with small chromosomes, like Campylocentrum neglectum.
The extreme symmetry (ideal karyotype A) or the extreme asymmetry (ideal karyotype F) karyotypes are hardly found in nature.However, in the present analysis, an extreme of symmetric karyotype was found in Sarcoglottis grandiflora, grouped with the ideal karyotype A. Christensonella pachyphylla showed the most asymmetric karyotype, but this species was grouped with ideal karyotypes C and D and not with the extreme ideal karyotype F.

The Stebbins' hypothesis
The relationship between karyotype asymmetry and species evolution could be discussed based on intrachromosomal indexes, since the interchromosomal index does not differ among subfamilies.The intrachromosomal asymmetry indexes indicated the karyotypes of some representatives of the subfamily Epidendroideae as the most asymmetric -in agreement with the hypothesis of Stebbins (1971) that asymmetric karyotypes had been originated from symmetrical ones.Based on the statistical results and cluster analysis the congruent indexes Ask%, A1 and Syi indicated Epidendroideae as the most derivate subfamily, presenting the most asymmetrical karyotype, while the rep-resentatives of the subfamily Cypripedioideae have more symmetrical karyotypes.

Conclusions
Considering our results, the indexes Ask% (Arano, 1963), A 1 (Romero-Zarco, 1986) and Syi are recommended for the estimation of intrachromosomal asymmetry in cytotaxonomic studies, especially in a combined fashion.We showed that the critical point for the efficacy of an asymmetric index is the well-preserved chromosome morphology and precise definition of the centromere position -and not the size of chromosomes.Moreover, the higher karyotype asymmetry associated with the derivative subfamily Epidendroideae supports Stebbins' hypothesis that asymmetric karyotypes tend to derive from symmetric karyotypes.

Figure 1 -
Figure 1 -Interchromosomal index A2 values for Orchidaceae subfamily.For each subfamily the mean value (dot), the amplitude of variation (bar), the number of species analyzed and mode value (last two data in the parenthesis, respectively) are presented.

Figure 2 -
Figure 2 -Ideograms of the ideal karyotypes A, B and C, as well as the most similar species, grouped by UPGMA, equally obtained by the indexes Ask%, A1, Syi and A. The numeric scale at the right side of the ideogram is given in micrometers (mm).

Figure 3 -
Figure 3 -Intrachromosomal asymmetry values obtained by Ask% (blue), A1 (red), A (green) and Syi (orange) indexes for the Orchidaceae subfamily.The numeric scale at the right side indicates the mean value for the four intrachromosomal indexes.The Syi value was divided by 100.Subfamilies indicated by the same letters are not significantly different (Tukey test, p < 0.05).