Divergent subgenome evolution in the allotetraploid frog Xenopus calcaratus

Allopolyploid genomes are divided into compartments called subgenomes that are derived from lower ploidy ancestors. In African clawed frogs of the subgenus Xenopus (genus Xenopus ), allotetraploid species have two subgenomes (L and S) with morphologically distinct homoeologous chromosomes. In allotetraploid species of the sister subgenus Silurana , independently evolved subgenomes also exist, but their cytogenetics have not been investigated in detail. We used an allotetraploid species in Silurana — Xenopus calcaratus —to explore evolutionary dynamics of chromosome morphology and rearrangements. We ﬁnd that the subgenomes of X. calcaratus have distinctive characteristics, with a more conserved a-subgenome resembling the closely related genome of the diploid species X. tropicalis , and a more rapidly evolving b-subgenome having more pronounced changes in chromosome structure, including diverged heterochromatic blocks, repetitive sequences, and deletion of a nucleolar secondary constriction. Based on these cytogenetic diﬀerences, we propose a chromosome nomenclature for X. calcaratus that may apply to other allotetraploids in subgenus Silurana , depending on as yet unresolved details of their evolutionary origins. These ﬁndings highlight the potential for large-scale asymmetry in subgenome evolution following allopolyploidization.


Introduction
Whole genome duplication (polyploidization) and large-scale chromosomal rearrangements are important evolutionary driving forces that contribute to genome variability, for example by creating duplicate genes and affecting patterns of recombination (Wolfe, 2001). Polyploidization is a potential mode of sympatric speciation wherein reproductive isolation from progenitor species is achieved via inviability or sterility of offspring of a cross between parents of differing ploidy levels (Janko et al, 2018). Polyploidization occurs via genome duplication within a single species (autopolyploidization) or in association with hybridization between two or more distinct species (allopolyploidization). Divergence between homoeologous chromosomes could expedite "diploidization" of a polyploid genome, wherein chromosomes acquire disomic inheritance, with each having only one partner during cell division. This has the effect of creating separate genomic compartments called "subgenomes"; in allopolyploid genomes, each subgenome is derived mostly or entirely from a different ancestral species (Schiavinato et al, 2021).
Chromosomal changes such as translocations, inversions, insertions, and deletions are divided into small-and large-scale rearrangements. Small-scale rearrangements involve regions smaller than one megabase. Large-scale rearrangements affect regions larger than one megabase, may include fusions and fissions of different chromosomes, and may alter karyotype organization and gene expression. In any genome, the frequency of chromosomal rearrangements is governed by several factors including presence of retrotransposons in genome (Biémont and Vieira, 2006), spatial variation in DNA stability and fragility, and external factors such as radiation or environmental conditions (Feder et al, 2011). Within a polyploid genome, the frequency of rearrangements may differ between subgenomes (Brunet et al, 2006;Session et al, 2016;Li et al, 2021).

Allopolyploid genomes of African clawed frogs (Xenopus)
African clawed frogs, genus Xenopus, (family Pipidae) are divided into subgenera Xenopus and Silurana, that include diploid and tetraploid species in Silurana, and tetraploid, octoploid, and dodecaploid species in Xenopus (Tymowska, 1991;Evans, 2008;Evans et al, 2015). This high species diversity (29 species in total) with a range of ploidy levels makes Xenopus a compelling focal group for studying cytogenetics and chromosomal rearrangements in the wake of polyploidization, with the potential to provide insights into whether and how genomic rearrangements are linked to speciation and adaptation. Available information from extant tetraploids, octoploids, and dodecaploids indicates that they are allopolyploids rather than autopolyploids (Evans et al, 2005(Evans et al, , 2015Session et al, 2016). In X. laevis (subgenus Xenopus) each subgenome contains distinct transposable element complements that originate from distinct ancestors (Session et al, 2016).
In subgenus Xenopus, several large-scale rearrangements have been identified. The largest involves a fusion between chromosomes 9 and 10 in a diploid ancestor of allopolyploids in subgenus Xenopus (Session et al, 2016). Comparison to a high-quality genome assembly of the diploid species X. tropicalis evidences several other rearrangements within the X. laevis S-subgenome and comparative genomic stability of the X. laevis L-subgenome (Session et al, 2016).
In Silurana tetraploids, divergence between the homoeologous a-and bsubgenomes is higher than between the a-subgenome and the genome of diploid X. tropicalis (Evans, 2008;Evans et al, 2015), as expected by allotetraploidy. In Silurana, one large-scale rearrangement has been identified-a nonreciprocal interchromosomal translocation of pericentromeric region between chromosomes 9b and 2a of the tetraploid species X. mellotropicalis (sensu Knytl et al, 2017Knytl et al, , 2018b. This rearrangement was identified using fluorescent in situ hybridization (FISH) in which whole chromosome painting (WCP) probes from X. tropicalis were hybridized to chromosomes of X. mellotropicalis (Zoo-FISH).
It remains unknown whether this rearrangement occurred in the ancestor of all three allotetraploid species in Silurana (X. mellotropicalis, X. epitropicalis, X. calcaratus) or more recently after the ancestor of X. mellotropicalis diverged from the ancestor(s) of the other species (Fig. 1). This is because evolutionary relationships among species in subgenus Silurana remain poorly resolved; very little genomic data has been collected from X. calcaratus and X. epitropicalis that could be leveraged for phylogenetic estimation along with genomic data from other Silurana species (Hellsten et al, 2010;Cauret et al, 2020). Available information from two tightly linked genes (RAG1 and RAG2 ) suggest that two separate allotetraploidization events occurred in Silurana: one generated X. calcaratus and the other resulted in the ancestor of X. mellotropicalis and X. epitropicalis (Evans et al, 2005;Evans, 2007) (Fig. 1b). However, a simplier scenario is also possible where allotetraploidization occurred only once 139  140  141  142  143  144  145  146  147  148  149  150  151  152  153  154  155  156  157  158  159  160  161  162  163  164  165  166  167  168  169  170  171  172  173  174  175  176  177  178  179  180  181  182  183  184 Springer Nature 2021 L A T E X template 4 Article Title in Silurana (Fig. 1a). Also lacking are cytogenetic characters that can clearly distinguish homoeologous chromosomes of the Silurana subgenomes. Fig. 1 Phylogenetic scenarios of the subgenus Silurana adapted from Evans et al (2015). The sister clade, subgenus Xenopus, is depicted as an outgroup. (a) A scenario involving one allotetraploidization via the fusion of two diploid ancestors, one of which was closely related to X. tropicalis and the other of which went extinct (dagger), to give rise to the most recent common ancestor of all three allotetraploid species in subgenus Silurana. (b) An alternative scenario involving two independent allotetraploidization events; one gave rise to the ancestor of X. calcaratus and the other to the most recent common ancestor of X. epitropicalis and X. mellotropicalis.
Large-scale genomic rearrangements occurred multiple times in genus Xenopus and may have been influential in the evolution of Xenopus. In this study, we examined cytogenetic evolution of X. calaratus with an aim of better understanding genome evolution in this species and also to further contextualize evolution of the interchromosomal translocation in X. mellotropicalis (Knytl et al, 2017). We used a suite of chromosome banding and FISH techniques and statistically processed chromosomal measurement information to provide unprecedented resolution of the karyotype of the allotetraploid Biafran clawed frog, X. calcaratus.

Establishment of primary cell cultures
Primary cell cultures were derived from the hind limb of tadpoles at stage NF55(±1) (Sinzelle et al, 2012) of X. calcaratus originated from Bakingili, Cameroon, and X. tropicalis, derived from the 'Ivory Coast' laboratory strain. Both species were bred at Charles University, Faculty of Science, Prague, Czech Republic. Briefly, tadpoles were anesthetized by 0.4% MS-222 (Sigma-Aldrich, St. Louis, MO, USA) and then washed with sterile MilliQ water following death. The hind limbs were removed and homogenized in cultivation medium which was prepared from components as described in Knytl et al (2017) and modified by addition of Gibco TM Antibiotic-Antimycotic (100X) and 0.1 mM Gibco TM 2-mercaptoethanol (both Thermo Fisher Scientific, Waltham, MA, USA). The explants were then cultivated at 29.5°C with 5.5% CO 2 for five days without disturbance. The medium was then changed every second day. The first and next passages were performed with trypsin-ethylenediaminetetraacetic acid according to Knytl et al (2017). For

Preparation of chromosomal suspension and metaphase spreads
Both X. tropicalis and X. calcaratus chromosomal suspensions were prepared according to Krylov et al (2010) and stored in fixative solution (methanol: acetic acid, 3:1, v/v) at −20°C. For laser microdissection, a fresh metaphase suspension was dropped onto a polyethylene-naphthalene membrane. For cytogenetic analysis, a chromosome suspension was dropped onto a microscopic slide according to Courtet et al (2001). Chromosome preparations were aged at −20°C for at least one week with the exception of that for the fluorescent in situ hybridization with tyramide signal amplification, FISH-TSA, in which the chromosome suspension was dropped and directly followed by experiment. For each experiment, mitotic metaphase spreads were counterstained with ProLong TM Diamond Antifade Mountant with the fluorescent 4',6-diamidino-2-phenylindole, DAPI stain (Invitrogen by Thermo Fisher Scientific). From ten to twenty metaphase spreads were analyzed per each banding technique and probe. Microscopy and processing of metaphase images using Leica Microsystem (Wetzlar, Germany) were conducted as detailed in Seroussi et al (2019).

Laser microdissection and whole chromosome painting
All ten individual chromosomes from X. tropicalis (20 copies of each chromosome) were separately isolated by laser microdissection as previously described in Kubickova et al (2002) using a PALM Microlaser system (Carl Zeiss MicroImaging GmbH, Munich, Germany). Whole chromosome painting probes (WCPs) were prepared according to Krylov et al (2010). During whole genome amplification, Digoxigenin-11-dUTP and Biotin-16-dUTP (both Jena Bioscience, Jena, Germany) were incorporated into the probes (the nucleotide ratio for dig probes: 10mM dATP, dGTP, dCTP: 6.5 mM dTTP, bio probes: 10mM dATP, dGTP, dCTP: 5 mM dTTP). Control double-colour intraspecies painting FISH on X. tropicalis and single-colour cross-species Zoo-FISH on X. calcaratus chromosomes were carried out as described by Krylov et al (2010) with minor modifications as detailed in Knytl et al (2017). Autoclaved X. tropicalis genomic DNA was used as a competitor (blocking DNA) according to Bi and Bogart (2006). In the control double-colour painting FISH, the digoxigenin and biotin labeled probe was detected by Anti-digoxigeninfluorescein (Roche, Basel, Switzerland) and CY TM 3-Streptavidin (Invitrogen, Camarillo, CA, USA), respectively, diluted with blocking reagents as used in double-colour rDNA FISH (Knytl et al, 2017). In the single-colour Zoo-FISH, fluorescent signal of digoxigenin labeled probes was visualized by Anti-digoxigenin-rhodamine (Roche).

Whole genome painting
Xenopus tropicalis genomic DNA (gDNA) was used as a probe for genomic in situ hybridization (GISH) experiments. Whole genome painting (WGP) probes were prepared using the GenomePlex Single Cell Whole Genome Amplification Kit (WGA4), Sigma-Aldrich, according to the manufacturer's whole genome amplification protocol with extracted gDNA. GenomePlex WGA Reamplification Kit (WGA3), Sigma-Aldrich, and labeling with Digoxigenin-11-dUTP (Jena Bioscience) was carried out as described in Krylov et al (2010). A combination of salmon sperm (Knytl et al, 2013b) and autoclaved X. tropicalis gDNA (Bi and Bogart, 2006) was used as a competitor DNA. Control GISH was performed on X. tropicalis chromosomes as detailed in painting FISH in Krylov et al (2010), and cross-species GISH was carried out on X. calcaratus chromosomes as detailed in Zoo-FISH in Krylov et al (2010) with minor changes described in Knytl et al (2017).

Ribosomal gene mapping and chromosome banding
Double-colour FISH was performed with 5S and 28S ribosomal DNA probes (rDNA FISH), followed by two chromosome banding techniques that were conducted sequentially on the same metaphase spread: C-banding, and then Chromomycin A 3 , CMA 3 (Sigma-Aldrich).
Xenopus calcaratus gDNA was used as a template for amplification of both 5S and 28S loci. Total gDNA was extracted from the tail tissue of a tadpole using the DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany) according to manufacturer's instructions. Both 5S and 28S primer sequences (Integrated DNA Technologies, Coralville, IA, USA) are listed in Table S1. Preparation of the 5S and 28S probes including modified PCR conditions (Sember et al, 2015) and labelling with Digoxigenin-11-dUTP and Biotin-16-dUTP (both Jena Bioscience) is detailed in Knytl and Fornaini (2021). The 5S and 28S probes were hybridized with chromosomal spreads of X. calcaratus. In total 22 µL of the hybridization mixture containing 100 ng of either 5S or 28S rDNA probe, and 14 µL master mix (10% dextran sulfate) was placed on a slide and covered with a 22 × 22 mm coverslip. Both probe and chromosomal DNA were denatured at 70°C for 5 min with subsequent overnight hybridization in a dark wet chamber. Post-hybridization washing and blocking reactions were performed as described for painting FISH in Krylov et al (2010). Probe signal was visualized as described in Knytl et al (2017).

Measurement and identification of X. calcaratus chromosomes
A total of 35 X. calcaratus individual metaphase figures were analyzed. Half of the metaphases selected for the identification of individual chromosomes was stained with 5% Giemsa/PBS solution (v/v). The rest were stained with DAPI during FISH experiments. Both arms of each chromatid were measured in pixels using ImageJ, V 1.53i (Schneider et al, 2012). The p and q arm lengths were quantified as described in Knytl and Fornaini (2021). To identify each chromosome, we analyzed chromosomal length (l), p/q arm ratio (r 1 ) (Tymowska, 1973), centromeric index (i), and q/p arm ratio (r 2 ) (Levan et al, 1964).
Chromosomal numbering was adopted from Knytl et al (2017) with proposed revisions detailed below. Data were analyzed in R software for statistical computing (V 4.1.0) (Team, 2020). R scripts were used and modified from Knytl and Fornaini (2021). One-way analysis of variance (one-way ANOVA) was performed to compare l and i values between each homoeologous pair of chromosomes. All steps outlining how the measured values were calculated and processed into tables and plots are shown on https://www.github.com/.

Species identification
Species of Xenopus frogs are notoriously difficult to diagnose based on external anatomy. We therefore used Sanger sequencing of our experimental individual for species diagnosis. PCR amplifications of portions of the 5S and 28S nuclear   323  324  325  326  327  328  329  330  331  332  333  334  335  336  337  338  339  340  341  342  343  344  345  346  347  348  349  350  351  352  353  354  355  356  357  358  359  360  361  362  363  364  365  366  367  368 Springer Nature 2021 L A T E X template 8 Article Title rRNA genes, and the 16S mitochondrial rRNA gene resulted in approximately 200 and 300, and 900 bp long amplicons. Based on blastn results, the 5S amplicon showed 96% identity with the sequence of 5S rDNA of X. tropicalis (accession number X12624.1), and 28S amplicon showed 100% identity with 28S rRNA of X. tropicalis (accession number XR 004223802.1), which is consistent with membership in subgenus Silurana (at the time of this study, sequences of these genes for X. calcaratus are not present in the GenBank databese). The 16S amplicon showed 100% identity with 16S rRNA of X. calcaratus (accession number KT728037.1) from the same locality in Cameroon but 2.8-3.0% divergence from X. tropicalis (accession numbers KT728027.1 and KT728029.1), which is not known to occur in this locality. Together these results confirmed the species identity of the X. calcaratus cells. This species diagnosis is also consistent with chromosome counts as detailed below. Sequences are deposited to the NCBI/GenBank database (5S accession number XXX, 28S accession number OM910742, 16S accession number OM912675).
A key finding that emerged from the X. calcaratus Zoo-FISH experiments is that one homologous chromosome pair had consistently of higher fluorescent intensity than the other. Based on this, we defined X. calcaratus "a" homoeologs (or alpha, α, sensu Evans et al (2005)) to be the more intensely painted chromosomes and "b" homoeologs (or beta, β, sensu Evans et al (2005)) to be the less intensely painted ones. The observation of differential painting intensity is consistent with an allotetraploid origin of X. calcaratus from a diploid common ancestor of X. tropicalis and another divergent diploid ancestor whose diploid descendants are either extinct or have not yet been discovered (Fig.  1). That these experiments found no evidence of a chromosomal translocation between chromosomes homologous to XTR 9 and 2 in X. calcaratus suggests this translocation occurred in an ancestor of X. mellotropicalis after divergence from an ancestor of X. calcaratus.
Using the WCP probe from X. tropicalis chromosome 4, we found an additional signal on XTR 5 (Fig. 2b, red arrowheads) and XCA 5a (Fig. 3d

Genomic in situ hybridization with whole genome painting probes
Two GISH experiments were performed with X. tropicalis WGP probe. The first one was a control to evaluate the efficacy of the WGP probe against X. tropicalis chromosomes (Fig. 4a). All 20 chromosomes were consistently painted, indicating high efficacy. The second GISH experiment hybridized the WGP probe to X. calcaratus chromosomes with an aim of providing verification of a distinctive cytogenetic signature of each subgenome. Similar to the WCP experiments shown in Section 3.2, this technique recovered differences in the intensity of fluorescence between several homoeologous pairs (Fig. 4b).
In particular, chromosomes XCA 7a, 9a, and 10a were notably more intensively painted than XCA 7b, 9b, and 10b (labeled with arrowheads). These results are consistent with the analysis of p/q arm ratio and Zoo-FISH data discussed below, and with a closer evolutionary relationship between the XCA homoeologs "a" and XTR orthologs than between the XCA homoeologs "b" and XTR orthologs. Other homoeologous chromosomes within the X. calcaratus karyotype apart from XCA 7, 9, and 10 also were painted with slightly different but less distinguishable intensities. Fig. 4 Genomic in situ hybridization (GISH) using X. tropicalis whole genome painting (WGP) probes (in red) against (a) X. tropicalis and (b) X. calcaratus chromosomes. In (a), the X. tropicalis WGP probes hybridized to all 20 X. tropicalis chromosomes (XTR→XTR) with a consistent fluorescent signal. In (b), the X. tropicalis WGP probes hybridized to all 40 X. calcaratus chromosomes but with different intensities between homoeologs-especially between homoeologs XCA 7, 9, and 10 (white arrowheads). Each scale bar represents 10 µm.

Ribosomal gene mapping and chromosome banding
Nucleolar organizer regions (NORs) are regions in eukaryotic genomes that contain tandem arrays of three ribosomal genes (18S, 5.8S, 28S ). In eukaryotes  554  555  556  557  558  559  560  561  562  563  564  565  566  567  568  569  570  571  572  573  574  575  576  577  578  579  580  581  582  583  584  585  586  587  588  589  590  591  592  593  594  595  596  597  598 Springer Nature 2021 L A T E X template Article Title 13 the ribosome includes small and large subunits; the 18S rRNA gene is part of the small subunit and the 5.8S, 28S and 5S rRNAs are part of the large subunit, though the 5S rRNA gene is not contained within the NOR (Vierna et al, 2013). In diploid genomes there is generally only one pair of NORs (Schmid et al, 1987). Building on the findings from Zoo-FISH (Section 3.2), we expected a NOR pair to have been inherited from each of the diploid ancestral species of X. calcaratus. However, the FISH experiment with 28S rDNA probes identified only one pair of NORs situated on the nucleolar secondary constriction of XCA 7aq (Fig. 5a, red colour). The FISH with 5S probes revealed 8-10 positive loci located on telomeric regions of the XCA 2b, 4a, 5a, 6a, and 8b Fig.  5a, green colour. Xenopus calcaratus is closely related to X. mellotropicalis and for this reason we expected similar patterns and localization of the heterochromatic blocks in these two species. To test this expectation, we used C-and CMA 3banding, which are staining techniques that highlight GC rich chromosomal loci and blocks of constitutive heterochromatin. These staining methods can be helpful for distinguishing and assigning chromosomes based on banding patterns. C-and CMA 3 -banding showed positive signals on XCA 1bp ( Fig.  5b and Fig. 5c). C-banding also exhibited a faint signal in portions of stained regions on XCA 2b, 6b, 7b, and 8b (not shown by arrows). CMA 3 -banding also identified nucleolar secondary constriction of XCA 7aq (Fig. 5c). All X. calcaratus chromosomes bore weak CMA 3 -bands on telomeres that presumably are caused by repetitive sequences. (green) and 28S (red) ribosomal probes identifies XCA 2b, 4a, 5a, 6a, and 8b (green arrows) and the nucleolar secondary constriction on the q arm of XCA 7a (red arrowheads), respectively. (b) C-banding (brighter staining) highlights heterochromatic blocks on the p arm of XCA 1b (blue arrows). (c) CMA 3 banding in green (green arrows) shows heterochromatic blocks on the p arm of XCA 1b that co-localize with C-bands, and additional CMA 3 positive band on the q arm of XCA 7a (nucleolar secondary constriction) that co-localize with 5S rDNA loci. Scale bars represent 10 µm .   599  600  601  602  603  604  605  606  607  608  609  610  611  612  613  614  615  616  617  618  619  620  621  622  623  624  625  626  627  628  629  630  631  632  633  634  635  636  637  638  639  640  641  642  643  644 Springer Nature 2021 L A T E X template 14 Article Title

Mensural analysis of X. calcaratus chromosomes
As expected, across all X. calcaratus spreads (n = 35) we consistently counted 40 chromosomes, indicating that 2n = 4x = 40, where n refers to the haploid number of chromosomes of the extant species, and x refers to the haploid number of chromosomes in the most recent diploid ancestor of the extant species.
Each chromosome was measured from 35 metaphases spreads, with each chromosome identified based on comparison to the FISH results, and the median values of l, r 1 , and i were calculated. l was quantified as a percentage of the sum of l across all chromosomes in order to account for variation in resolution and pixel size of our images. Each individual chromosome was also assigned a chromosomal category based on the i value. If the i value was equal to or greater than 37.5, the chromosome was determined as metacentric. If the i value was equal to or higher than 25 and lower than 37.5, the chromosome was determined as submetacentric, and if the i value was equal to or 647 648  649  650  651  652  653  654  655  656  657  658  659  660  661  662  663  664  665  666  667  668  669  670  671  672  673  674  675  676  677  678  679  680  681  682  683  684  685  686  687  688  689  690 Springer Nature 2021 L A T E X template Article Title 15 greater than 12.5 and lower than 25, the chromosome was determined to be subtelocentric. The karyotype of X. calcaratus consists of 12 metacentric, four submetacentric, and four subtelocentric chromosomes. We found that pairs of homoeologs tend to have similar chromosomal morphology in terms of l and i values (Fig. 7a). For example, chromosomes 3a, 3b, 8a, and 8b have low centromeric indexes (between 12.5-25) and are therefore subtelocentric. Chromosomes 5a and 5b have centromeric indexes between 25-37.5 and are therefore both submetacentric. However, chromosomes 2a and 4a were in the submetacentric category whereas their homoeologous chromosomes 2b and 4b were in the metacentric category. The rest of chromosomes fell into the category of metacentric chromosomes (centromeric indexes of 37.5-50). Acrocentric chromosomes (centromeric indexes interval 0-12.5) and telocentric chromosomes without p arm (centromeric indexes = 0) are not present in the X. calcaratus karyotype.
The median values of l (%), r 1 and i are shown in Table 1, where these values are assigned to each chromosome.
Chromosome 7a and 7b of X. calcaratus can be distinguished because 7a has a secondary nucleolar constriction in the same position as XTR 7, whereas this feature is absent from XCA 7b. This inference is further supported by a higher intensity of Zoo-FISH and GISH signals with X. tropicalis probes on XCA 7a compared to 7b. The l values of some chromosomes were substantially different between "a" and "b" homoeologs (Table 1, significance codes). Based on the cytogenetic information presented here that was obtained from multiple methods, we defined a chromosome nomenclature for X. calcaratus that could extend to other tetraploid Silurana species depending on their phylogenetic relationships discussed below.

Discussion
We used a combination of conventional (C-and CMA 3 -banding) and more advanced molecular cytogenetic techniques (rDNA FISH, intra-and crossspecies painting FISH, FISH-TSA, and GISH) to study the allotetraploid karyotype architecture of X. calcaratus. Our results, together with previous cytogenetic findings (Tymowska and Fischberg, 1982;Schmid and Steinlein, 2015;Knytl et al, 2017), enabled us to distinguish a-and b-subgenomes of X. calcaratus. We identified several cytogenetic changes discussed below in the bsubgenome that were not present in the a-subgenome of X. calcaratus or in X. tropicalis, which indicate that the b-subgenome has undergone more rapid (and asymmetric) evolution compared to the a-subgenome. Table 1 Measured values used for identification of Xenopus calcaratus chromosomes (XCA). Median of chromosomal length (l (%)), median of p/q arm ratio (r 1 ), and median of centromeric index (i). Chromosomal categories correspond to m = metacentric, sm = submetacentric, and st = subtelocentric chromosomes. Significance codes following l and i values define whether pairs of homoeologous chromosomes are significantly different based on ANOVA test. Significantly different homoeologs are depicted by significance codes "***", "**", and "•" showing the p-value p < 0.001, p < 0.01, and p < 0.1, respectively. The allotetraploid species X. laevis has one NOR on chromosome 3Lp (Tymowska and Kobel, 1972;Schmid et al, 1987;Roco et al, 2021), with a homologous NOR on chromosome 3S thought to have been lost after allotetraploidization in subgenus Xenopus (Session et al, 2016). In the diploid X. tropicalis the NOR is on chromosome 7q (Tymowska, 1973;Tymowska and Fischberg, 1982;Uehara et al, 2002;Roco et al, 2021), the allotetraploid X. epitropicalis and X. mellotropicalis also have one NOR on chromosome 7aq (Tymowska and Fischberg, 1982;Tymowska, 1991), with the other on chromosome 7b hypothesized to have been lost after allotetraploidization (Knytl et al, 2017). In X. calcaratus, 28S rDNA FISH identified nucleolar (secondary) constriction on XCA 7aq (Figs. 5a). There are at least two possible explanations for the conserved location and number of NORs in X. mellotropicalis and X. calcaratus. If these two species each evolved via independent allotetraploidization events (Fig. 1b), loss of one pair of NORs from the same ancestral species could have occurred twice independently after each allotetraploidization event.
C-and CMA 3 -banding identified a prominent heterochromatic block in the form of non-nucleolar secondary constriction, but some of these signals may colocalize with NORs (e.g. 28S -positive signal, Fig. 5a, and CMA 3 -positive signal on XCA 7a, Fig. 5c). Several Silurana tetraploids have one intensively labeled non-nucleolar secondary constriction in the form of heterochromatic block: for X. tropicalis it is on chromosome 9q (Tymowska and Fischberg, 1982;Knytl et al, 2017), for X. mellotropicalis it is on 2ap (re-designated below to be 2bp) (Knytl et al, 2017), for X. epitropicalis it is on the pericentromeric region of 2a (re-designated below to be 2b) involving both p and q arms (Tymowska and Fischberg, 1982), and for X. calcaratus it is on 1bp. Different size and position of the non-nucleolar secondary constrictions within Silurana species indicate that these constrictions are very dynamic structures with independent evolution but the most similar position in X. epitropicalis and X. mellotropicalis (chromosome 2b) which is consistent with their close evolutionary relationship (Evans et al, 2005).
The number and position of NORs has been been extensively explored in both diploid and polyploid animals. Amphibians and fish are good groups with which to explore the effect of whole genome duplication on the evolution of NOR and other ribosomal structures (Knytl et al, 2018a) because both include a diversity of diploid and polyploid species. In most diploid and polyploid amphibians, only one chromosome pair bears a NOR (Schmid et al, 1987;Inafuku et al, 2000;Alves et al, 2012;Schmid, 1982). However, in fish there is more extensive variation in the number of NORs, including observations ranging from one (Alves et al, 2012) to seven pairs (Gromicho et al, 2006). For instance, some tetraploid fish possess two pairs of NORs located on one homoeologous set of chromosomes (Diniz et al, 2009;Knytl et al, 2013a). One NOR pair may therefore reflect functional diploidization, wherein bivalents rather than multivalents form during meiosis, which appears to be the case for all Xenopus (Tymowska, 1991).
Another possible explanation for among species variation in the number of nucleolar loci is that some have been transcriptionally inactive in the preceding interphase. Therefore, these loci have not been transcribed, and might not be detected by ribosomal probes (Schmid, 1982). Number of nucleolar loci could be also changed by the translocation of individual rDNA copies from clusters via viruses or transposons (Gromicho et al, 2006).
Although the relative size of homoeologous pairs corresponds to subgenomes in X. laevis (Session et al, 2016), this is not necessarily the case in other groups. Because the relative intensities of fluorescence signal in FISH corresponds with phylogenetic distance between the probe and the genome being interrogated, evolutionary affinities of chromosomes within each subgenome can be inferred based on the intensity of the probe fluorescence (Markova et al, 2007;Liu et al, 2015). Here, we distinguished homoeologous chromosome in X. calcaratus subgenome "a" and "b" (1a-10b) based on measurements of chromosomal p and q arm length, and comparative cytogenetic findings from Zoo-FISH and GISH using probes from the diploid species X. tropicalis. This approach allowed for an evolutionary-based classification system built on divergence from X. tropicalis orthologous chromosomes. Chromosomes from the X. calcaratus a-subgenome are more similar and closely related to X. tropicalis orthologs than chromosomes from the X. calcaratus b-subgenome. With this approach, we identified one homoeologous pair -chromosome 2 -where the larger chromosome was actually from b-subgenome, warranting revision on the chromosomal nomenclature (Table 2). Consistent with this reappraisal, the intensity of the Zoo-FISH signal on the longer X. mellotropicalis chromosome 2, (2a sensu Knytl et al (2017)) is lower than intensity of the smaller X. mellotropicalis chromosome 2 (2b sensu Knytl et al (2017); see Fig. 4B in Knytl et al (2017)). Based on results presented here for X. calcaratus, X. mellotropicalis 2a (XME 2a) and 2b should thus be re-classified as XME 2b and 2a, respectively. Additionally, our findings indicate that chromosomes XCA 1b and 2b had higher p/q arm ratio, centromeric index and chromosomal length than XCA 1a and 2a, respectively. This is also inconsistent with previous designations based on conventional staining such as Giemsa, C-banding, or replication banding (Tymowska, 1991;Schmid and Steinlein, 2015).
Xenopus calcaratus and X. mellotropicalis a-subgenomes have no translocations, although the b-subgenome of X. mellotropicalis contains a large-scale   875  876  877  878  879  880  881  882  883  884  885  886  887  888  889  890  891  892  893  894  895  896  897  898  899  900  901  902  903  904  905  906  907  908  909  910  911  912  913  914  915  916  917  918  919  920 Springer Nature 2021 L A T E X template 20 Article Title Table 2 List of r 1 values (p/q arm ratio) for each chromosome in X. tropicalis (Tymowska, 1973), X. calcaratus (this study), and X. mellotropicalis (Knytl et al, 2017). Xenopus tropicalis chromosome nomenclature according Khokha et al (2009) was used. Each X. tropicalis ortholog is divided into the "a" and "b" homoeologous chromosomes in tetraploid Silurana karyotypes ("homoeolog" column rearrangements between XME 2b and 9b (Knytl et al, 2017). Heterochromatin is evident on XCA 1b (Figs. 5b and 5c) but not on XTR 1 or XME 1b (Knytl et al, 2017) hypothesized the heterochromatin transpositions within b-subgenome. The pattern that the nucleolar 28S rDNA loci (a part of 45S rRNA) were detected only on XCA 7aq but not on 7bq indicates gene loss within X. calcaratus b-subgenome. Overall, these results suggest asymmetric subgenome evolution in Silurana, with a-subgenome being more stable and resembling the X. tropicalis diploid genome, and b-subgenome having experienced more chromosomal rearrangements. As discussed above, asymmetric subgenome evolution has also been reported in the allotetraploid X. laevis which is in subgenus Xenopus (Session et al, 2016), which argues for the plausibility of this phenomenon in subgenus Silurana. The p/q arm ratio of XTR 5 is more similar to the p/q arm ratio of XCA 4a, 4b, XME 4a, and 4b than to the p/q arm ratio of XCA 5a, 5b, XME 5a, and 5b (Table 2). It is not clear whether X. tropicalis chromosomes were identified correctly by Tymowska (1973) and Khokha et al (2009), or whether both X. calcaratus and X. mellotropicalis homoeologs 5a and 5b have undergone changes undetectable by our experimental approaches. It is also possible that intraspecies variability in chromosome structure exists among geographically Article Title with repetitive sequences (Knytl et al, 2017). Extensive rearrangements and GC-rich secondary constrictions were not identified on the XCA 2b (CMA 3 or C-banding positive, Figs. 5b and 5c) which suggests the repetitive regions associated with the t(9b;2b) of X. mellotropicalis also accumulated after divergence from the ancestor of X. calcaratus. The effects on recombination of a rearrangement may not be bounded by the breakpoints of the rearrangement (Xia et al, 2020). It is thus conceivable that this translocation influenced speciation through the suppression of chromosomal pairing and recombination (Fishman et al, 2013).  Fig. 1. (a) One allotetraploidization event gave rise to all three species (X. calcaratus, X. epitropicalis, and X. mellotropicalis). (b) Two independent allotetraploidizations, the first gave rise to X. calcaratus. The second tetraploidization gave rise to a common ancestor of X. epitropicalis and X. mellotropicalis. Blue spots indicate potential translocation between ancestral chromosomes 9 and 2 (t(9;2)) in a diploid, and between chromosomes 9b and 2b (t(9b;2b)) in a tetraploid common ancestor of X. epitropicalis and X. mellotropicalis. Red spots indicate translocation occurred separately in X. mellotropicalis lineage. All scenarios exclude the hypothesis of one shared translocation event in a common ancestor of all Silurana tetraploids.

Conclusions
Allopolyploid genomes are interesting subjects for cytogenetic investigation because evolutionary phenomena in each subgenome can be distinguished and compared. In this study, we identified two subgenomes in allotetraploid frog X. calcaratus, and found them to be distinguished by cytogenetic characteristics that are consistent with asymmetric evolution. Results are consistent with the proposed allopolyploid origin of X. calcaratus (X. new tetraploid 2 in (Evans et al, 2005)), with one progenitor species also being an ancestor of X. tropicalis and another ancestral species either extinct or undiscovered. To better understand genome evolution in subgenus Silurana it is necessary to perform advanced cytogenetic investigation of X. epitropicalis, the remaining member of a group of allotetraploid species, similar in scope with this study.
Supplementary information. Supplementary results: Box plot statistics of chromosomal length (l ); Box plot statistics of centromeric index (i ); Supplementary table: Table S1: The list of genes used for our study with appropriate gene symbol and name, GenBank accession number, length of the PCR amplicon, and PCR primer sequence.; Supplementary figures: Fig. S1: FISH-TSA with positive red signals on XCA metaphase spreads. The gyg2β and α were Article Title 23 localized on the p arm of XCA 2a and 2b, respectively. Chromosomes were counterstained with DAPI (blue-green). Each scale bar represents 10 µm.; Fig. S2: FISH-TSA with positive red signals on XCA metaphase spreads. The cept1β and α were localized on the p arm of XCA 2a and 2b, respectively. Chromosomes were counterstained with DAPI (blue-green). Each scale bar represents 10 µm.; Fig. S3: FISH-TSA with positive red signals on XCA metaphase spreads. The sf3b1α and β loci were mapped on the q arm of XCA 9a and 9b, respectively. Chromosomes were counterstained with DAPI (bluegreen). Each scale bar represents 10 µm.; Fig. S4: FISH-TSA with positive red signals on XCA metaphase spreads. The ndufs1α and β loci were mapped on the q arm of XCA 9a and 9b, respectively. Chromosomes were counterstained with DAPI (blue-green). Each scale bar represents 10 µm.; Fig. S5: FISH-TSA with positive red signals on XCA metaphase spreads. The fn1α and β loci were mapped on the q arm of XCA 9a and 9b, respectively. Chromosomes were counterstained with DAPI (blue-green). Each scale bar represents 10 µm. Article Title github.com/). Raw data of chromosomal arm measurements can be provided in the form of RData file as R workspace, also upon request. • Code availability Not applicable • Authors' contributions MK conceived and designed research. MK, NRF, BB, VG, and HČ conducted experiments. MK analyzed and validated data and performed statistical analyses. VG performed the fieldwork. MK wrote the initial manuscript and MK, BJE and VK then edited the manuscript. All authors read and approved the manuscript and agreed to the published version of the manuscript. BJE and VK are joint last authors .  1107  1108  1109  1110  1111  1112  1113  1114  1115  1116  1117  1118  1119  1120  1121  1122  1123  1124  1125  1126  1127  1128  1129  1130  1131  1132  1133  1134  1135  1136  1137  1138  1139  1140  1141  1142  1143  1144  1145  1146  1147  1148  1149  1150  1199  1200  1201  1202  1203  1204  1205  1206  1207  1208  1209  1210  1211  1212  1213  1214  1215  1216  1217  1218  1219  1220  1221  1222  1223  1224  1225  1226  1227  1228  1229  1230  1231  1232  1233  1234  1235  1236  1237  1238  1239  1240  1241  1242