Ethiopian material was obtained from tributaries of the Sobat River, a tributary of the White Nile, in southwestern Ethiopia (Table 1). Fish were collected by the Joint Ethio-Russian Biological Expedition (JERBE) with the permissions of the National Fisheries and Aquatic Life Research Center under the Ethiopian Institute of Agricultural Research (EIAR) and the Ethiopian Ministry of Science and Technology. Two individuals – male and female – of Nannaethiops bleheri Géry et Zarske, 2003, collected from the roadside ditch in the interfluve of the Alvero and Gilo rivers (between towns of Abobo and Funido, 7°45.307'N, 34°15.639'E) were karyotyped. The rest of karyotyped Ethiopian material was obtained from the two localities: (1) Alvero River just downstream of the Abobo Dam (7°52.503'N, 34°29.960'E) and (2) Baro River at the City of Gambela (8°14.878'N, 34°34.044'E). Two males and a female of Distichodus engycephalus Günther, 1864, as well as a female of Ichthyborus besse (Joannis, 1835), were collected at locality 1. Two males of I. besse and a female of Nannocharax niloticus (Joannis, 1835), were collected at locality 2.

Four specimens (a female, two males and one unsexed) of an unidentified species representing the genus Nannaethiops and seven specimens (five females and two males) of Neolebias unifasciatus Steindachner, 1894 were purchased from the Nigerian aquarium fish dealers through the mediation of the company Aqua Logo Engineering (https://www.aqualogo-engineering.ru).

After colchicine treatment, fish were euthanized with an overdose of tricaine methanesulfonate (MS-222), identified, measured with an accuracy of 1 mm, dissected for gonad examination and tissue sampling, and preserved in 10% formaldehyde or 70% ethanol. Species identification was done based on morphological characters (Gosse and Coenen 1990; Golubtsov et al. 1995). The experiments were carried out in accordance with the rules of the Severtsov Institute of Ecology and Evolution (IEE) and approved by IEE’s Ethics Committee. Vouchers are deposited at the Severtsov Institute of Ecology and Evolution (Moscow), under provisional labels of JERBE.

In order to clarify the phylogenetic position of Nannaethiops and Neolebias specimens, two genetic markers – Cytochrome oxidase subunit I (COI) and 16S ribosomal RNA (16S rRNA) – were studied in 13 karyotyped fish and one additional specimen of N. bleheri from an another location in Ethiopia (Suppl. material 1: table S1). We extracted total genomic DNA from the ethanol-preserved tissues using the DiatomDNA Prep 100 (Izogen, Moscow) extraction kit. The PCR mixture contained 5 pmol of each primer and the precast PCR mixture from DIALAT Ltd (Russia). The primers used for COI amplification were designed by Ward et al. (2005): FishF1-5′TCAACCAACCACAAAGACATTGGCAC3′ and FishR1-5′TAGACTTCTGGGTGGCCAAAGAATCA3′. The PCR cycle profiles were as follows: 5 min initial denaturation at 94 °C, followed by 35 cycles of 1 min at 94 °C, annealing for 45 sec at 55 °C, extension for 1 min at 72 °C; final extension for 7 min at 72 °C. The primers 8f-5′AGAGTTTGATCCTGGCTCAG3′ (Edwards et al. 1989) and 1492r-5′GGTTACCTTGTTACGACTT3′ (Stackebrandt and Liesack 1993) were employed for the 16S rRNA amplification. The PCR cycle profiles were as follow: 3 min initial denaturation at 94 °C, followed by 30 cycles of 30 sec at 94 °C, annealing for 30 s at 50 °C, extension for 30 sec at 72 °C; final extension for 7 min at 72 °C. PCR products were visualized by electrophoresis in 1.5% agarose gel in TBE buffer with addition of ethidiumbromide. DNA sequencing was performed using an Applied Biosystems 3500 genetic analyzer. All new DNA sequencies were deposited in GeneBank (Suppl. material 1: table S1).

Preprocessing and alignment of the obtained sequences was carried out using SeqMan Pro 7.1.0 and BioEdit 5.0.9. For phylogenetic reconstruction all sequences of the two markers (COI and 16S rRNA) available in GenBank for Nannoethiops and Neolebias specimens were used. These sequences are listed below. The distichodontid species Belonophago hutsebouti Giltay, 1929, Distichodus nefasch (Bonnaterre, 1788) and D. sexfasciatus Boulenger, 1897, as well as citharinid Citharinus citharus (Geoffroy Saint-Hilaire, 1809), were selected as outgroups. The GenBank accession numbers for outgroups are given in Suppl. material 1: table S1.

Comparative material included the GenBank sequences of six species representing the genera Nannoethiops and Neolebias for CO1 and seven such species for 16S rRNA (Fig. 1, Suppl. material 1: table S1). For CO1, these were Nannaethiops bleheri from Ethiopia (the GenBank accession number KF541848, Arroyave et al. 2013), Nannaethiops gracilis (Matthes, 1964) (KF541851, KF541852, Arroyave et al. 2013), Nannaethiops unitaeniatus Günther, 1872 (KF541849, KF541850, Arroyave et al. 2013), Neolebias ansorgii Boulenger, 1912 (KF541858, KF541859, KF541860, Arroyave et al. 2013; HM418212, HM418213, Sonet et al. 2019), Neolebias trewavasae Poll et Gosse, 1963 (KF541853, KF541857, Arroyave et al. 2013) and Neolebias trilineatus Boulenger, 1899 (KF541854, KF541855, KF541856, Arroyave et al. 2013; KT193336, Decru et al. 2016; HM418214, HM418215, MK074510, MK074511, Sonet et al. 2019), all from West Africa. For 16S rRNA, these were Nannaethiops bleheri from Ethiopia (JX985104, Lavoué et al. 2017), Nannaethiops unitaeniatus (JX985105, Lavoué et al. 2017), Neolebias ansorgii (AY788058, Calcagnotto et al., 2005; JX985107, Lavoué et al. 2017), Neolebias powelli Teugels et Roberts, 1990 (AY788061, Calcagnotto et al. 2005), Neolebias trewavasae (JX985132, Lavoue et al. 2017), Neolebias trilineatus (AY788063, Calcagnotto et al. 2005) and Neolebias unifasciatus Steindachner, 1894 (JX985103, Lavoué et al. 2017), all from West Africa.

Figure 1.

Maximum Likelihood (ML) trees with compressed subtrees based on (A) 615-bp COI fragment and (B) 387-bp 16S rRNA fragment. Length of branches is proportional to the genetic distances between haplotypes; bootstrap support (Felsenstein, 1985) is indicated next to the branching nodes and calculated with ML/Maximum Parsimony/Bayesian Inference methods from 1000 replicas (“*” - bootstrap support is equal to 100% or 1, “-- ” or not specified - bootstrap support is less than 50%).

For phylogenetic reconstruction, we used Maximum Likelihood (ML), Maximum Parsimony (MP) (Nei and Kumar 2000) and Bayesian Inference (BI) methods. For ML, the chosen models of molecular evolution were as follows: Hasegawa-Kishino-Yano (HKY +G+I; parameter +G = 1.77; +I = 0.6) (Hasegawa et al. 1985) for COI and Tamura-Nei (TN93+G; parameter +G = 0.13) (Tamura and Nei 1993) for 16S rRNA. For ML and MP, the bootstrap support for branch nodes was calculated with 1,000 replicates (Felsenstein 1985). Genetic distances and other parameters for phylogenetic ML and MP analysis were calculated using the MEGA X software package (Kumar et al. 2018). The nucleotide substitution model for BI was selected by means of the Bayesian Information Criterion (BIC) as implemented in jModelTest (Posada 2008). BI was carried out in MrBayes version 3.1.2 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003) and implemented using Markov Chain Monte Carlo algorithm for 10,000 generations with a sampling period of 1,000 generations.

Before preparation, fish were treated intraperitoneally with 0.1% colchicine (0.01 ml / 1 g of their weight; for Ethiopian material, under field conditions) or 0.025% colchicine (0.01 ml / 1 g of their weight; for Nigerian material, under laboratory conditions) for 3–5 hours. After euthanasia, chromosome preparations were obtained from kidney tissue following Kligerman and Bloom (1977) for Ethiopian and Nigerian material or from kidney, spleen, intestine and liver following Bertollo et al. (2015) for Nigerian material with some modifications for both protocols, as described in Simanovsky et al. (2022). The chromosome spreads were stained conventionally with 4% Giemsa solution in a phosphate buffer solution at pH 6.8 for 8 min and then analysed using an Axioplan 2 Imaging microscope (Carl Zeiss, Germany) equipped with a CV-M4⁺CL camera (JAI, Japan) and Ikaros software (MetaSystems, Germany). Final images were processed using Photoshop software (Adobe, USA). Karyotypes were arranged according to the centromere position following the nomenclature of Levan et al. (1964), but modified as metacentric (m), submetacentric (sm) and subtelocentric/acrocentric (st/a). Chromosome pairs were arranged according to their size in each chromosome category. To determine the chromosomal arm number per karyotype (fundamental number, FN), metacentrics and submetacentrics were considered as biarmed, and subtelocentrics/acrocentrics as monoarmed. The total numbers of complete metaphases studied for each species is presented in Table 1.

An analysis of 615 bp of the mitochondrial CO1 in 13 individuals representing the genera Nannoethiops and Neolebias and 387 bp of the mitochondrial 16S rRNA in seven individuals representing the same genera included the Ethiopian samples of Nannaethiops bleheri, as well as the West African samples (from the Nigerian aquarium fish dealers) of the genera Nannoethiops and Neolebias. The alignment used for phylogenetic reconstructions included 47 CO1 sequences and 18 16S rRNA sequences in total.

The thirteen newly obtained COI sequences were collapsed in six haplotypes deposited in GenBank with accession numbers OQ891056–OQ891061. Two of them made an independent cluster corresponding to Neolebias unifasciatus (Fig. 1). Genetic distance (p-distance) was 0.002 between haplotypes. Two more cluster together with a sequence of Nannaethiops bleheri deposited earlier by Arroyave et al. (2013) (p-d 0.002–0.003). The remaining two new haplotypes formed an independent cluster recognized by us as Nannaethiops sp. that is a sister to Nannaethiops bleheri (Fig. 1). In general, haplotypes of the genera Nannoethiops and Neolebias comprise a monophyletic group without a clear division into two genera (Fig. 1). This is fully consistent with the conclusion of Géry and Zarske (2003) – supported by Arroyave et al. (2013) and Lavoué et al. (2017) – who considered Neolebias as a junior synonym of Nannaethiops.

The seven newly obtained 16S rRNA sequences were collapsed in three haplotypes. One of them appeared to be identical to the sequence (JX985103) earlier deposited in GenBank for Neolebias unifasciatus by Lavoue et al. (2017). Two other haplotypes we deposited in GenBank with the accession numbers OQ911366 and OQ911367. The former cluster together with the haplotype deposited for Nannaethiops bleheri by Lavoué et al. (2017) (p-d 0.003); the latter belongs to the Nannaethiops sp. clade.

In summary, both the COI and 16S rRNA analyses support: (1) our identification of Nannaethiops bleheri; (2) the distinctiviness of Nannaethiops sp.; and (3) the 16S rRNA analysis supports our identification of Neolebias unifasciatus.

The karyotype of Distichodus engycephalus has 2n = 52 and consists of 30 metacentrics and 22 submetacentrics, FN = 104 (Fig. 2). It differs substantially from the karyotype of D. affinis (2n = 48, 32m + 16sm, FN = 96) reported by Rab et al. (1998) (Table 2). No distinguishable sex chromosomes were observed in complements of D. engycephalus, similar to the finding by Rab et al. (1998) in D. affinis. This is true for all distichodontids studied by us.

Figure 2.

Karyotypes of six representatives of the family Distichodontidae. Scale bar: 10 μm.

The karyotype of Ichthyborus besse has 2n = 46 and consists of 40 metacentrics and 6 submetacentrics, FN = 92. The karyotype of Nannocharax niloticus has 2n = 54 and consists of 46 metacentrics, 6 submetacentrics, and 2 subtelocentrics/acrocentrics, FN = 106. The latter species exhibits the highest numbers of chromosomes and chromosome arms among all distichodontids studied (Table 2).

The karyotypes of Nannaethiops bleheri, Nannaethiops sp. and Neolebias unifasciatus appeared to be similar. These karyotypes have 2n = 50 and consists of 38 metacentric, 10 submetacentric, and 2 subtelocentrics/acrocentrics, FN = 96. These taxa, along with Nannocharax niloticus, possess the only pair of monoarmed chromosomes; the other distichodontids studied have exclusively biarmed chromosomes in their compliments.

The molecular phylogeny of the order Cithariniformes as it is reconstructed by Arroyave et al. (2013) and Lavoué et al. (2017) is as follows. The family Citharinidae is a sister group to the family Distichodontidae. Xenocharax Günther, 1867 comprises a sister group to all other distichodontids. Nannaethiops + Neolebias represent a sister group to other distichodontids excluding Xenocharax. Monostichodus Vaillant in Rivière, 1886 + Ichthyborus comprise a sister group to all remaining distichodontids. Branching of the remaining three clades (Distichodus + Paradistichodus Pellegrin, 1922, Nannocharax, Belonophago Giltay, 1929 + Phago Günther, 1865 with the related genera) is not well supported and different in Arroyave et al. (2013) and Lavoué et al. (2017). Nevertheless the monophyly of the each of three groups is well supported. Thus, we analysed the representatives of four clades out of six excluding Xenocharax and Belonophago + Phago with the related genera.

There is an apparent correspondence between molecular phylogeneetic and cytogenetic data. There are differences in cytogenetic characteristics between Distichodus (2n = 48–52), Ichthyborus (2n = 46), Nannocharax (2n = 54) and Nannaethiops + Neolebias (2n = 50) representing the four different clades revealed by phylogenetic analyses. Moreover, there are differences in cytogenetic characteristics between all these distichodontids and the two species of Citharinus (2n = 40–44) (Table 2). These data clearly suggest a substantial role of chromosome fusions/fissions in the evolution of Cithariniformes karyotypes.

Regarding variation within the clades, we see two opposing trends. Two species of Distichodus, D. affinis and D. engycephalus, differ both in diploid chomosome numbers and karyotypic formulae. On the contrary, no differences were found between karyotypes of Nannaethiops bleheri, Nannaethiops sp. and Neolebias unifasciatus representing another clade. The latter point corroborates the position of authors who considered Neolebias as a junior synonym of Nannaethiops (Géry and Zarske 2003, Arroyave et al. 2013, Lavoué et al. 2017). Variability of karyotype structure in the genus Distichodus makes it possible to use the cytogenetic data in its taxonomy when a sufficient array of such data is accumulated. The same is true for the family Distichodontidae as a whole.

Due to the lack of data on the diversity of karyotypes in both the families Citharinidae and Distichodontidae it might be premature to make assumptions about the trend of karyotype evolution in the order Cithariniformes. The great prevalence of biarmed chromosomes (the karyotypes of most species contain exclusively biarmed chromosomes) is a distinctive characteristic of Cithariniformes compared to Characiformes and Siluriformes, sister groups to Cithariniformes. Characiformes and Siluriformes are characterized by karyotypes with various proportions of biarmed and monoarmed chromosomes (Arai 2011; Simanovsky et al. 2022). There is reason to suggest that the ancestral karyotype of Cithariniformes consisted exclusively/predominantly of biarmed chromosomes. However, the karyotypes of representatives of the basal group of Distichodontidae – genus Xenocharax – have yet to be determined. Thus, the cytogenetic information about this genus and other unexamined taxa of Cithariniformes would be of great interest.