Ten specimens of the Asian swamp eel were purchased from an animal pet shop in Bangkok, Thailand. Animal care and all experimental procedures were approved by the Animal Experiment Committee, Kasetsart University, Thailand (approval no. ACKU00958), and conducted according to the Regulations on Animal Experiments at Kasetsart University, Thailand. Mitotic chromosomes were obtained from gill and kidney cells using the air drying method. Briefly, after intraperitoneal injection of 0.01% colchicine (Sigma, St. Louis, Missouri, USA) in the proportion of 0.7 ml per 100 g of fish weight for 2 h, fishes were anesthetized in ice-cold water, and the anterior portion of the gill and kidney were removed and used for mitotic chromosome preparation. After hypotonic treatment of gill and kidney in 0.075 M KCl for 50 min at room temperature, the organs were minced and placed in the first fixative solution (3:1 methanol/acetic acid) for 5 min and in the second fixative solution (2:1 methanol/acetic acid) for 5 min on ice. The cells were collected by filtration using gauze, and then fixed with 3:1 methanol/acetic acid. The cells in suspension were dropped onto clean glass slides and air-dried. The slides were kept at -80°C until use. For karyotyping with conventional Giemsa staining, the chromosome slides were stained with 4% Giemsa solution (pH 7.2) for 10 min.

To examine the chromosomal distribution of constitutive heterochromatin, C-banding was performed using the standard barium hydroxide/saline/Giemsa method (Sumner 1972) with slight modification as follows: chromosome slides were treated with 0.2 N HCl at room temperature for 60 min and then with 5% Ba(OH)₂ at 50°C for 15 s, followed by 2× SSC at 65°C for 60 min.

Genomic DNA was extracted from liver and muscle tissue following the standard salting-out protocol as described previously (Supikamolseni et al. 2015), and used as templates for polymerase chain reaction (PCR). Partial DNA fragments of the 18S − 28S rRNA genes, and Rex retroelements (Rex1, Rex3, and Rex6) were amplified using following PCR primers (see Suppl. material 1). PCR amplification was performed using 20 μl of 1× ExTaq buffer containing 1.5 mM MgCl₂, 0.2 mM dNTPs, 5.0 μM the primers, and 0.25 U of TaKaRa Ex Taq (TaKaRa Bio, Otsu, Japan), and 25 ng of genomic DNA. PCR conditions were as follows: an initial denaturation at 94°C for 3 min, followed by 35 cycles of 94°C for 30 s, 53–59°C for 30 s, and 72°C for 45 s, and a final extension at 72°C for 10 min. The PCR products were cloned using the pTG19-T vector (Vivantis Technologies Sdn Bhd, Selangor Darul Ehsan, Malaysia), and nucleotide sequences of the DNA fragments were determined using DNA sequencing service (First BASE Laboratories Sdn Bhd, Seri Kembangan, Selangor, Malaysia). Nucleotide sequences of three to five DNA clones, and their consensus sequences were searched for homologies with annotated sequences in the National Center for Biotechnology Information (NCBI) database to identify the amplified DNA fragments, using the BLASTx and BLASTn programs (http://blast.ncbi.nlm.nih.gov/Blast.cgi). They were then deposited in the DNA Data Bank of Japan (DDBJ; http://www.ddbj.nig.ac.jp/index-e.html) (Suppl. material 2).

Multiple sequence alignments of the three data sets (Rex1, Rex3, and Rex6) were performed with those of other teleosts taken from the NCBI database (Suppl. material 2), using the default parameters of Molecular Evolutionary Genetics Analysis 6 (MEGA6) software (Center for Evolutionary Functional Genomics, The Biodesign Institute, Tempe, AZ, USA) (Tamura et al. 2013). Numbers of indels (insertions and deletions) for each data set of Rex retroelements were calculated using the multiallelic mode of DNAsp 5.0 (Librado and Rozas 2009). All unalignable and gap-containing sites were carefully removed from the data sets. Interspecific sequence divergence was estimated using uncorrected pairwise distances (p-distances), and for the Rex reverse transcriptase region, synonymous (K_s) and nonsynonymous (K_a) substitution rates (±standard error) were calculated using the Nei-Gojobori method (Nei and Gojobori 1986) with Jukes-Cantor correction (Jukes and Cantor 1969). Phylogenetic analyses were then performed using Bayesian Inference (BI) using MrBayes v3.0b4 (Huelsenbeck and Ronquist 2001) and the optimal model of DNA substitution was determined for each data set using Kakusan4 (Tanabe 2011). The Markov Chain Monte Carlo (MCMC) process was set to run four chains simultaneously for one million generations. After the log-likelihood value plateaued, a sampling procedure was performed every 100 generations to obtain 10,000 trees, and subsequently to provide a majority-rule consensus tree with average branch lengths. All sample points were discarded as burn-in prior to reaching convergence, and the Bayesian posterior probability in the sampled tree population was obtained in percentage terms. All phylogenetic trees were midpoint-rooted due to the absence of suitable outgroup in Rex3 data set. However, additional phylogenetic tree based on Rex1 and Rex6 sequences were constructed with using outgroup method from other Rex sequences.

Chromosomal locations of the 18S – 28S rRNA genes, Rex retroelements (Rex1, Rex3, and Rex6), telomeric (TTAGGG)n sequences, and 19 microsatellite repeat motifs: (CA)₁₅, (GC)₁₅, (GA)₁₅, (AT)₁₅, (CAA)₁₀, (CAG)₁₀, (CAT)₁₀, (CGG)₁₀, (GAG)₁₀, (AAT)₁₀, (AAGG)₈, (AATC)₈, (AGAT)₈, (ACGC)₈, (AAAT)₈, (AAAC)₈, (AATG)₈, (AAATC)₆, and (AAAAT)₆ were determined using FISH, as described previously (Matsuda and Chapman 1995, Srikulnath et al. 2009). We used a 1,366-bp genomic DNA fragment of M. albus 18S – 28S rRNA genes (LC151290), a 533-bp genomic DNA fragment of M. albus Rex1 (LC110446), a 415-bp genomic DNA fragment of M. albus Rex3 (LC110447), a 471-bp genomic DNA fragment of M. albus Rex6 (LC110448), biotin-labeled 42-bp TTAGGG repeat, and 19 biotin-labeled oligonucleotide microsatellite repeat probes, respectively. We labeled 250 ng of DNA fragments with biotin-16-dUTP (Roche Diagnostics, Mannheim, Germany) by nick translation, according to the manufacturer’s protocol and ethanol-precipitated with salmon sperm DNA and Escherichia coli tRNA. After hybridization of biotin-labeled probes to M. albus chromosomes, the probes were stained with avidin labeled with fluorescein isothiocyanate (avidin-FITC; Invitrogen, CA, USA). Slides were subsequently stained with 0.75 μg/ml PI or 1 µg/ml DAPI. Fluorescence hybridization signals were captured using a cooled CCD camera mounted on a ZEISS Axioplan2 microscope and processed using MetaSystems ISIS v.5.2.8 software (MetaSystems, Alltlussheim, Germany).

For dual-color FISH, two probes differentially labeled with either biotin-16-dUTP or digoxigenin-11-dUTP (Roche Diagnostics) were mixed in hybridization buffer and co-hybridized to one slide. After hybridization, digoxigenin- and biotin-labeled probes were stained with anti-digoxigenin-rhodamine Fab fragments (Roche Diagnostics) and avidin labeled with fluorescein isothiocyanate (avidin-FITC; Invitrogen), respectively.

Over 10 Giemsa-stained metaphase spreads were examined for each M. albus individual. Diploid chromosome number is 24 (FN = 24) comprising twelve pairs of acrocentric chromosomes (Fig. 1a). The size difference of chromosome pairs was sequential, but most pairs were identified by size and banding pattern with DAPI and PI fluorescent staining. Large DAPI-positive bands were observed at the pericentromeric region of chromosome 9 (Fig. 1b), and large PI-positive bands were found at the pericentromeric region of chromosome 4 (Fig. 1c) coincident with a large C-positive heterochromatin bands (Fig. 1d).

Figure 1.

Giemsa-stained (a), DAPI-stained (b), PI-stained karyotype (c), and C-banded metaphase spread (d) of Monopterus albus. Arrowheads indicate the large DAPI-stained and large PI-stained regions. Arrows indicate C-positive heterochromatin blocks. Scale = 10 μm.

Fluorescence hybridization signals for the 18S – 28S rRNA genes were also detected at the pericentromeric region of chromosome 4 co-localizing with both PI-positive bands and large C-positive heterochromatin blocks (Fig. 2a, c, d, e). Hybridization signals of TTAGGG repeats were observed at telomeric ends of all chromosomes, but no interstitial signal was found (Fig. 2b, c).

Figure 2.

Chromosomal locations of the 18S – 28S rRNA genes and (TTAGGG)n sequences in Monopterus albus. Hybridization pattern of FITC-labeled 18S – 28S rRNA genes (green) (a) and rhodamine-labeled TTAGGG repeats (red) (b) on DAPI-stained chromosomes, and their co-hybridization pattern (c). Hybridization pattern of FITC-labeled 18S – 28S rRNA genes (green) (d) on PI-stained chromosomes. PI-stained patterns of the same metaphase spreads of (d) is shown in (e). Arrowheads indicate FISH signals of the 18S – 28S rRNA genes. Arrows indicate the large PI-stained region. Scale =10 μm.

Eight of the 19 microsatellite repeat motifs were dispersedly mapped onto most chromosomes (Fig. 3). Notably, strong hybridization signals of trinucleotide (CGG)₁₀ were localized to chromosomes 2, 4, and 6, tetranucleotide (AAAT)₈ to chromosomes 3 and 5, (AGAT)₈ to chromosomes 5 and 9, (ACGC)₈ to chromosomes 1, 2, 4, 7, 8 and 9, and pentanucleotide (AAATC)₆ to chromosomes 1 and 8. No signal was observed from the other 11 microsatellite repeat motifs ((CA)₁₅, (GC)₁₅, (GA)₁₅, (AT)₁₅, (CAA)₁₀, (CAG)₁₀, (CAT)₁₀, (GAG)₁₀, (AAT)₁₀, (AAGG)₈, (AATC)₈, (AAAC)₈, (AATG)₈, and (AAAAT)₆).

Figure 3.

Chromosomal locations of microsatellite repeat motifs in Monopterus albus. Hybridization pattern of FITC-labeled (CAA)₁₀ (a), (CAG)₁₀ (b), (CGG)₁₀ (c), (GAG)₁₀ (d), (AGAT)₈ (e), (ACGC)₈ (f), (AAAT)₈ (g), and (AAATC)₆ (h) on PI-stained chromosomes.

M. albus Rex1 (MALRex1) obtained from a single M. albus individual was localized to the pericentromeric region and interstitial sites of all chromosomes, except for chromosomes 4 and 9 where MALRex1 was found only at interstitial sites (Fig. 4a). MALRex3 was located scattered in all chromosomes with strong hybridization signals observed on chromosomes 1–4 and 8 and weak signals on chromosomes 5–7 and 9–12 (Figs 4b, 5b, d). FISH signals of MALRex6 were found on chromosomes 1, 2, 5, 6, 8, and 10 as dispersion along the chromosomes (Figs 4c, 5c, d).

Figure 4.

Chromosomal locations of Rex1, Rex3, and Rex6 in Monopterus albus. Hybridization pattern of FITC-labeled Rex1 (green) (a) on PI-stained chromosomes, and rhodamine-labeled Rex3 (red) (b) and Rex6 (red) (c) on DAPI-stained chromosomes. Scale =10 μm.

Figure 5.

Chromosomal locations of Rex3 and Rex6 in Monopterus albus. Hybridization pattern of FITC-labeled Rex3 (green) (b) and rhodamine-labeled Rex6 (red) (c) on DAPI-stained chromosomes, and their co-hybridization pattern (d). DAPI-stained patterns of the same metaphase spreads of (b, c, and d) is shown in (a). Scale =10 μm.

The nucleotide sequence of a 533 bp-fragment of MALRex1 was used in multiple sequence alignment with 28 other teleosts, evidencing 32 indel sites. Sequence divergence among species varied from 0 to 50.13% with an average of 29.56±1.13% (Suppl. material 3). MALRex1 sequences in M. albus showed the minimum interspecific sequence divergence of 1.88% from nototheniids Dissostichus mawsoni Norman, 1937 and Notothenia coriiceps Richardson, 1844 (Perciformes) and the maximum divergence of 41.95% to Poeciliopsis gracilis Heckel, 1848 (Cyprinodontiformes); the average is 24.51±8.14%. The phylogenetic placement of Rex1 sequences showed that most species were grouped in their respective orders (Fig. 6, Suppl. material 6). The average K_s/K_a value of Rex1 sequences was 2.19±0.08 (Table 1). The nucleotide sequence of a 415 bp-fragment of MALRex3 was used in multiple sequence alignment with 24 other teleosts, showing 23 indels. The average sequence divergence among species was 33.94±17.24%, ranging from 2.65% to 69.54% (Suppl. material 4). MALRex3 sequences showed the minimum interspecific sequence divergence of M. albus, 18.54%, from Esox lucius Linnaeus, 1758 (Esociformes) and the maximum divergence, 66.65%, from Astyanax fasciatus Cuvier, 1819 (Characiformes); average 31.84±12.74%. The phylogenetic placement of Rex3 sequences showed a clade for each order except for Perciformes fishes (Fig. 7). The average K_s/K_a value of Rex3 sequences was 1.05±0.05 (Table 2). The nucleotide sequences of a 471 bp-fragment of MALRex6 was used in multiple sequence alignment with 17 other teleosts showing 15 indels. The sequence divergences among species varied from 3.13 to 65.546% (average 27.94±19.53%). MALRex6 sequences showed the minimum interspecific sequence divergence of M. albus, 60.31%, from Geophagus proximus Castelnau, 1855 (Perciformes) and the maximum divergence, 65.54%, from Oreochromis niloticus Cuvier, 1832 (Perciformes,); average 62.60±1.14% (Suppl. material 5). The phylogenetic placement of Rex6 sequences showed a clade for each order (Fig. 8, Suppl. material 7). The average K_s/K_a value of Rex6 sequences was 0.85±0.04 (Table 3).

Figure 6.

Phylogenetic placements of partial nucleotide sequences of Rex1 from 28 teleosts. Support values at each node are Bayesian posterior probability.

Figure 7.

Phylogenetic placements of partial nucleotide sequences of Rex3 from 24 teleosts. Support values at each node are Bayesian posterior probability.

Figure 8.

Phylogenetic placements of partial nucleotide sequences of Rex6 from 17 teleosts. Support values at each node are Bayesian posterior probability.

The karyotype of M. albus (2n = 24, FN = 24) composed of 12 acrocentric chromosome pairs was found to be similar to that reported by Yu et al. (1989) and Ji et al. (2003). The chromosome number of M. albus is the lowest among synbranchids, e.g., M. cuchia (2n = 42, FN = 46) (Rishi and Haobam 1984), Synbranchus marmoratus Bloch, 1795 (2n = 42–46, FN = 46–54) (Carvalho et al. 2012; Utsunomia et al. 2014), Ophisternon aenigmaticum Rosen & Greenwood, 1976 (2n = 46, FN = 52) (Nirchio et al. 2011), and O. bengalense McClelland, 1844 (2n = 46, FN = 52) (Carvalho et al. 2012), as well as the species of family Mastacembelidae of the same order (2n = 48, FN = 58–88) (Khuda-Bukhsh and Barat 1987). The fundamental numbers of M. albus is reduced to 50% of norm in synbranchid fishes and teleosts, which suggests that the acrocentric chromosomes of M. albus may have been formed by repeated tandem fusion of the ancestral acrocentric chromosomes contained in the ancestral karyotype of Synbranchidae. However, the hybridization signal of (TTAGGG)n at interstitial telomeric sites (ITSs) that appears to be remnants of fusion or inversion (Srikulnath et al. 2009, 2011, 2015) was not found in any chromosomes of M. albus in this study (Fig. 2). Comparative chromosome mapping of Asian swamp eel with zebrafish (Danio rerio Hamilton, 1822) using human bacterial artificial chromosome (BAC) probes revealed the Asian swamp eel retains a number of gene copies found in tetrapods, while other teleosts underwent the third genome duplication (GD), leading to multiple copies of the genes (Yi et al. 2001, Zhou et al. 2002). This suggests that Asian swamp eel retained the genome composition before the event of the third GD that occurred in teleosts (Zhou et al. 2002). Molecular structure of the pericentromeric regions of chromosome 4 which were high GC-rich have evolved in a concerted manner with amplification of the 18S – 28S rRNA genes. However, the chromosomal locations of the 18S – 28S rRNA genes varied in M. albus individuals (Fig. 2d, e), a phenomanon also observed in Chinese population on pair of chromosome 3 and/or chromosome 7 (Ji et al. 2003). In other synbranchid fishes, the 18S – 28S rRNA genes are generally located on a pair of chromosome 1 and on a pair of medium-sized acrocentric chromosomes in O. aenigmaticum (Nirchio et al. 2011), as well as on several other chromosome pairs in various pattern of S. marmoratus (Utsunomia et al. 2014). These results suggest that chromosomal locations of the 18S – 28S rRNA genes considerably differ in Synbranchidae.

In this study, eight microsatellite repeat motifs [(CAG)₁₀, (CAA)₁₀, (CGG)₁₀, (GAG)₁₀, (AGAT)₈, (ACGC)₈, (AAAT)₈, and (AAATC)₆] were dispersedly mapped on different chromosomes (Fig. 3). This suggests that the amplification of several microsatellite repeat motifs has occurred independently in the genome of M. albus. Interestingly, the dispersion of the microsatellite repeat motifs signals was co-localized to M. albus chromosomes with Rex retroelements. A similar case was found in cichlid species Cichla monoculus Agassiz, 1831, Pterophyllum scalare Schultze, 1823, and Symphysodon discus Heckel, 1840 (Schneider et al. 2015). This suggests that both Rex retroelements and microsatellite repeat motifs have co-amplified in the evolutionary process of the genome of M. albus.

The diversity of chromosomal distribution for Rex retroelements (Rex1, Rex3, and Rex6) was found in teleosts (Table 4). Two major distinctive patterns were observed: (1) compartmentalization as found in pericentromeric, centromeric, or telomeric regions, and (2) uniform dispersion throughout the genome or along the chromosomes (Ozouf-Costaz et al. 2004). Chromosomal distribution of Rex1, Rex3, and Rex6 were generally located in the specific region together as compartmentalization within each family/order (Table 4). In this study, although MALRex1 was dispersed throughout the genome, this element was predominantly localized to pericentromeric regions of all chromosomes except for chromosomes 4 and 9. By contrast, strong hybridization signals of MALRex3 were dispersed on five chromosome pairs, with weak signals on seven chromosome pairs, which implies that MALRex3 were specifically amplified in chromosomal regions of M. albus.

The differences in the copy number and chromosomal distribution of MALRex1, MALRex3, and MALRex6 suggest that these retroelements were independently amplified or lost in the lineage of M. albus, where MALRex3 is prone to retain a copy number higher than MALRex1 and MALRex6. A similar case of copy number variation in Rex retroelements was also found in several Antarctic nototheniid species (Ozouf-Costaz et al. 2004).

Three Rex retroelements were identified in the genome of M. albus, and the degree of sequence divergence for the three retroelements was high (14–67%) from other species in comparison. MALRex1 and MALRex3 showed high interspecific sequence divergences from Cyprinodontiformes and Characiformes, respectively, but low interspecific sequence divergences from Perciformes fishes for Rex1 and Esociformes for Rex3 (Suppl. materials 3 and 4). This suggests that M. albus and Perciformes or Esociformes shared relatively recent activity of Rex1 or Rex3, respectively. The average K_s/K_a value of Rex1 was higher than 1 between all compared species and between M. albus and other species (Table 1). These results suggest that Rex1 evolved under purifying selection and that retrotranspositions occurred during the evolution of teleosts. By contrast, the average K_s/K_a value of Rex3 was closer to 1, which suggests that after retrotransposition, Rex3 was influenced by pseudogene-like evolution (Table 2) (McAllister et al. 1997).

Only few data of Rex6 sequences were available because specific PCR primers were not feasibly effective to detect this element in the genome of teleosts (Volff et al. 2001, Ozouf-Costaz et al. 2004, Schneider et al. 2013). The absence of Rex6 was observed in several Antarctic nototheniid species, but Rex6 exists in some other species of the same order Perciformes (Volff et al. 2001, Ozouf-Costaz et al. 2004, Schneider et al. 2013). This suggests that Rex6 might have rapidly diverged in teleosts. MALRex6 showed high interspecific sequence divergences (approximately 60%) of M. albus from other teleosts (Suppl. material 5). This may indicate that the divergence of Rex6 sequences of M. albus (or Synbranchidae in general) and other teleosts was rather ancestral. The average K_s/K_a value of Rex6 was less than 1 (Table 3). This suggests that Rex6 has a more diverse function in teleosts.

The present results of chromosomal distribution and molecular diversity of four repetitive element groups (the 18S – 28S rRNA gene, telomeric sequences, microsatellite repeat motifs, and Rex retroelements) revealed the chromosome constitution and genome organization of Asian swamp eels. This enabled us to learn more about the chromosome constitution in synbranchid fishes and teleosts as a whole. Further work is required to investigate and compare synbranchid fishes, including M. cuchia, to better understand the process of karyotype and genome evolution in this lineage.