Genetic Signature of a Past Anthropogenic Transportation of a Far-Eastern Endemic Cladoceran (Crustacea: Daphniidae) to the Volga Basin

Abstract

Most studies of water flea (Crustacea: Cladocera) invasions are concentrated on a few taxa with an obvious harmful influence on native ecosystems, while our knowledge of cases of anthropogenic introduction with not-so-obvious consequences, in most other taxa, is poor. We found in the Volga basin (European Russia) a population that contained D. curvirostris Eylmann, 1887 and its hybrids with D. korovchinskyi Kotov et al. 2021. The latter taxon is endemic to the Far East and it has appeared in the Volga basin as a result of past human-mediated transportation. The population from Bakhilovo is represented by two strongly different groups of the COI haplotypes belonging, respectively, to (1) D. curvirostris and (2) D. korovchinskyi. We detected SNPs in the position 60 of the HSP-90ex3 locus and in the 195 positions of 28S rRNA locus, which differentiate two species. Part of the specimens from Bakhilovo belonged to D. curvirostris s.str., demonstrating homozygote SNP sites in two loci, but two specimens had heterozygote SNP sites in both nuclear loci. They belong to D. curvirostris x korovchinskyi hybrids. Most morphological traits of the females were characteristic of D. curvirostris. We found in some specimens some characters which could suggest their hybrid status, but this opinion is a hypothesis only, which needs to be checked on more ample material. The exact hybrid system in this pond is not known. Moreover, we have no evidences of sexual reproduction of the hybrids; they could reproduce by parthenogenesis only as is known for hybrids of the D. pulex group, or continuously crossing with parents like some members of D. longispina group. However, poor parental D. korovchinskyi was not detected in the pond either morphologically or genetically. The exact vector of its past anthropogenic transportation to the Volga is unknown. Most probably, just ephippia of D. korovchinskyi were translocated replaced from the Khabarovsk Territory to the Samara Area somehow. This is the first report on hybrids within the D. curvirostris species complex. Here, we demonstrated that accurate studies with deep resolution increase the number of revealed cryptic invasions. We expect that the number of revealed cases of cryptic interspecific invasions will grow rapidly.

1. Introduction

Biological invasions and anthropogenically mediated mixing of faunas have been intensively studied recently due to their high practical importance. Water fleas (Crustacea: Cladocera) are among models for such studies, and several famous invasive species are known within them: Daphnia lumholtzi Sars, Bythotrephes cederstroemii Schödler, Cercopagis pengoi (Ostroumov). Their non-indigenous populations have appeared as a result of a long-distance human-mediated transportation; as a result, such invasion of newly colonized ecosystems led to destructive consequences for the latter [1,2,3]. Most studies of cladoceran invasions are concentrated on a few taxa with an obvious harmful influence on native ecosystems, while our knowledge of cases of anthropogenic introduction with not-so-obvious consequences, in most other taxa, is poor. Recent phylogeographic investigations reveal each time some non-indigenous haplotypes and clades of different taxa (mainly Daphnia O.F. Müller as the best studied genus of the cladocerans) in different regions of the world [4,5,6]. Such cryptic invasions in freshwaters are much more common as than currently acknowledged [7]; they are not so destructive for the ecosystems, but may potentially change them in certain regions [6,8,9].

Hebert et al. [10] proposed a simple approach for genetic identification of animals using the 5′-fragment of the first subunit of the mitochondrial cytochrome oxidase (COI) gene (“genetic barcoding”). Despite existence of well-known problems of its use, now it is obvious that this locus really could be used for a relatively accurate genetic identification of the animals [11], as resolution of the COI-based identification is better than of other standard mitochondrial and nuclear markers [12]. Just due to this, genetic barcoding is widely used for detection of non-indigenous haplotypes and clades [13,14]. Many cases of cladoceran anthropogenic transportation were detected in the course of genetic barcoding studies [4,5,14,15,16].

However, barcoding studies based on mitochondrial genes do not reveal hybridization between indigenous and non-indigenous taxa. It is known that species introductions promote secondary contacts between taxa with long histories of allopatric divergence, and gene flow can still occur after millions of years of divergence [17]. Moreover, interspecific and intergeneric hybridization is not so rare for aquatic animals: the hybrids could represent stable, although not so ample, components of water communities [18,19].

Hybridization of indigenous and non-indigenous species of Cladocera is mainly studied in a few species groups of the subgenus Daphnia (Daphnia) O.F. Müller, mainly the D. longispina group [20,21]. They are not known for the D. curvirostris complex, although cases of anthropogenic transportation of this taxon from Europe to North America were revealed several times [8,22,23]. In the course of our analysis of numerous samples collected in European Russia over many years, only a single taxon of the D. curvirostris complex sensu Kotov et al. [24] was found, namely D. curvirostris s.str. After two recent revisions of this complex [9,24], we have returned to the examination of “common” curvirostris-like populations in European Russia. The aim of this paper is to describe a specific population from the Volga River basin and confirm that it contained hybrids of D. curvirostris and D. korovchinskyi Kotov et al., which appeared as a result of a past human-mediated transportation of the latter from the Far East.

2. Materials and Methods

Many samples were collected in Russia by our team or our colleagues in the course of accomplishing the Federal Governmental Task AAAA-A18-118042490059-5 for the A.N. Severtsov Institute of Ecology and Evolution of the Russian Academy of Sciences and the State Assignment No. 121051100104-6 for the I.D. Papanin Institute of Inland Water Biology (see [24]). Field collection in public property in Russia does not require permissions. Samples from other countries were provided by our colleagues having permissions to collect them due to their scientific activity in the governmental institutes in the corresponding countries. In the case of routine analysis of the samples collected in Russia and some other European countries, we identified populations of D. curvirostris based on morphological analysis (Figure 1). Some of them were previously studied by genetic methods [24], but the hybrid specimens were found in a single sample among all European samples studied.

It was taken in a single pond in the village of Bakhilovo (N 53.4006°, E 49.6319°) by T.A. Chuzhekova in 20 June 2012 by a standard plankton net and fixed in 90% alcohol. The site is located on Samarskaya Luka—a huge bend of the Volga River belonging to the Samara Area of the Russian Federation.

The sample contained 16 specimens of Daphnia, among which only a single adult male, but no juvenile males and ephippial females. For morphological examination, all animals were picked from the sample under a dissecting microscope, placed on slides (in a drop of glycerol), and studied under a high-power optical microscope Olympus CX41 in toto. Then, five females and the adult male were dissected for analysis of fine morphological details, including the appendage structure. A system of numeration of setae on thoracic limbs was applied as in previous papers [24]. The sample with seven females was deposited at the Collection of Zoological Museum of M.V. Lomonosov Moscow State University, accession number MGU Ml 244.

Totally, we analyzed five specimens from the aforementioned population (specimens 12, 13, 14, 28, 30), and additional specimens of D. curvirostris, D. korovchinskyi and D. pulex from some other populations for comparative purposes (Table S1). Genomic DNA was extracted from single adult females using the Wizard Genomic DNA Purification Kit (Promega Corp., Madison, WI) and QuickExtract using manufacturer protocols, but volume of used reagents was reduced two times to increase the concentration of DNA. Amplification of the gene fragments was performed using specific primer pairs: Dcurv_COI F and R with internal primers Dcurv_COI iF + iR for COI locus; D1a and D1b2 for 28S locus; F and R for HSP-90 locus according to earlier described protocols [24,25,26]. PCR products were visualized in a 1.5% agarose gel stained with ethidium bromide and purified by QIAquick Spin Columns (Qiagen Inc., Valencia, CA, USA), according to the manufacturer’s protocol. Each PCR product was sequenced bi-directionally on the ABI 3730 DNA Analyzer (Applied Biosystems) using the ABI PRISM BigDye Terminator v.3.1 kit with subsequent analysis by the Syntol Co, Moscow. Initial analysis of the chromatograms, formation of contigs and their subsequent editing was made with the Sanger Reads Editor in the Unipro uGENE v.36 package [27]. The authenticity of the sequences was verified by BLAST comparisons in the NCBI GenBabk [28]. The sequences from this study were submitted to the NCBI GenBank database (accession numbers MW135336-40, MZ396552-54, MZ396619-20 for COI; MZ407638-50 and OK070772-3 for 28S; MZ441123-38 and OK073657-8 for HSP-90).

Alignment was carried out using MAFFT v.7 algorithm [29] in uGENE. The best-fitting models of the nucleotide substitutions for each nucleotide position in the codon (1st, 2nd 3rd) for COI were selected using ModelFinder v.1.6 [30] at the Center for Integrative Bioinformatics Vienna web-portal, Austria [31] based on minimal values of the Bayesian information criteria (BIC) [32]. Models of nucleotide substitutions [33] were identified as TNe + G4 for 1st position, F81 for 2nd position and GTR + G4 for 3rd position, which, in general, agrees with nucleotide variability in animals [34].

Phylogenetic reconstructions based on the maximum likelihood (ML) and Bayesian (BI) methods were made taking into consideration the codon position in a triplet. As the guide tree, we used the multilocus tree by Kotov et al. [24] with D. pulex as an outgroup.

For ML analysis, we used the IQ-TREE v.1.6.9 algorithm [35] using the web-portal CIBIV, Austria, with 10k replicas of the ultrafast bootstrap test, a UFboot [36] as the branch support test. Estimation of ML topology was based on a PhyML SH-like test [37] in the Building Phylogenetic Tree module in uGENE [27]. BI was conducted using BEAST2 v.2.6 [38]. We identified all the parameters of the substitution model using the BIC criterion from BEAUti v.2.6 [39]. In each analysis, we conducted four independent runs of MCMC (20M generations, with selection of each 10k generation). We used Tracer v.1.7 [40] to evaluate convergence of parameters (based on ESS > 200). Input files were composed in LogCombiner v. 2.6 and then the consensus tree based on the maximum clade credibility (MCC) was obtained in TreeAnnotator v.2.6 (as a part of the BEAST2 package). Only BI tree is represented below, as the ML tree has identical topology.

Nucleotide diversity analysis and genetic distance calculations were carried out using DnaSP v.6.12 [41] and MEGA-X v.10.1 [42]. We analyzed “simple” p-distances as most preferable for the barcoding purposes [43]. Comparison of the genetic distances and geographic distances were conducted by the Mantel test [44], according to the recommendations of Diniz-Filho et al. [45] in PASSaGE v.2.0 [46].

3. Results

3.1. Genetics Account

Figure 2 represents our new tree based on the COI sequences; we do not describe it as its topology is described in detail by [24].

All five PCRs for the specimens from Bakhilovo were successful. We found that the population from Bakhilovo is represented by two strongly different (p distance = 0.244) groups of the COI haplotypes belonging, respectively, to (1) D. curvirostris and (2) D. korovchinskyi according to our tree, while inter-group variability is less than 1%. Despite so strong a difference between two haplotype groups, there are only 12 mutations determining non-synonymous amino acid substitutions, but they do not change hydrophoby of the amino acid residues, and do not change the product conformation. The two groups are similar in nucleotide composition; they had identical G + C rate (0.365) and nucleotide diversity (pi 0.007). Due to high inter-group differences, the result of the neutrality test of Tajima [47] was negative (D = 1.755, not significant (P > 0.05) for both synonymous and nonsynonymous sites). The number of base differences per site from estimation of coefficient of evolutionary differentiation was 0.97, which is characteristic of geographically distant samples [48]. A high positive value of Fu’s statistic [49] (Fs = 5.21) confirms a high probability of the differentiation (or expansion) processes. However, the Mantel test did not reveal the relation between genetic and geographic distances (MA test NA). We hypothesized that our specimens identified as D. korovchinskyi based on the mitochondrial gene COI are hybrids.

Unfortunately, the sample was not available for allozyme analysis (as the most reliable tool to identify the hybrids [50]), but we were able to analyze SNPs in the individual chromatograms of the capillary sequencing of nuclear genes in typical D. curvirostris from European Russia (specimens 089-090, 092-093, 098-099), typical D. korovchinskiy from its type locality in Khabarovsk (Far East of Russia) (018-022), and specimens from Bakhilovo (012-014, 028 and 030) (Figure 3). Sequences of nuclear fragments were obtained from four specimens from Bakhilovo only, while PCRs for both loci for the specimen 012 were not successful, probably due to a DNA fragmentation.

SNP in the position 60 of the HSP-90ex3 locus (20th amino acid of the exon 3) clearly differentiates typical D. curvirostris (cytosine, C) and D. korovchinskyi (thymine, T). Such substitutions are synonymous, both triplets code the serine in the translated amino acid sequence. All six specimens of “clean” D. curvirostris from European Russia are homozygous in this SNP, and all five D. korovchinskiy from Khabarovsk are also homozygous; therefore, the SNP should be considered species-diagnostic. Specimens 013 and 014 from Bakhilovo apparently belong to D. curvirostris s.str., demonstrating homozygote SNP sites. In contrast, the chromatogram of the specimens 028 and 030 demonstrates the overlapping of two peaks. The position was automatically marked by an ambiguity code (Y).

Other SNP is observed in the 195 position of 28S rRNA locus: substitution of cytosine (D. korovchinskyi) to thymine/uracil (D. curvirostris). Based on the RNAstructureWeb model [51], this substitution is located in the micro-loop and is not functionally significant for the RNA product. Again, all specimens of “clean” D. curvirostris are homozygous in this SNP, and all specimens of five D. korovchinskiy from Khabarovsk are also homozygous. If the specimens 013 and 014 from Bakhilovo belong to D. curvirostris s.str., the specimens 028 and 30 are hybrids having heterozygote SNP sites.

3.2. Morphological Account

Parthenogenetic female. Body subovoid in lateral view, maximum height approximately in the middle (Figure 4a–c and Figure 5a,b). Dorsal margin slightly convex, carapace elevated above head, a shallow depression between head and rest of body mostly absent. Postero-dorsal angle with a relatively short caudal spine. Head large, with a moderate rostrum in large adults (Figure 4d–g) and relatively short rostrum in juveniles (Figure 5c,d), its tip subdividing into two lobes by a “line” of pre-rostral fold. Posterior margin of head straight or slightly convex, but without a strong, heavily reticulated projection separated from the base of labrum by a very low depression. Ventral margin of head slightly concave. Compound eye large, ocellus minute. Body of antenna I almost fully reduced, aesthetascs protruding posterovetrally, their tips do reach rostrum tip. Abdomen short, consisting of four segments. The basalmost abdominal process especially long, the middle process well developed bent distally, distal most process globose, with rows of setules, and last segment lacking of process (5e,f). Postabdomen elongated, tapering distally, with long postabdominal seta, as long as preanal margin. Its preanal margin long, almost straight with series of minute setules, preanal and postanal angle not expressed. Postanal and anal portion with paired spines, their size continuously increasing distally. Postabdominal claw long, with proximal pecten consisting of about 10–12 thin teeth, medial pecten with 7–8 large teeth, distalmost pecten consisting of numerous fine setules (Figure 5g). Limb III with a flat exopodite bearing setae 2 and 4 the longest and with same size, seta 1 the shortest (Figure 6a,b). Inner-distal portion of limb III with endite 4 bearing a single anterior seta 1 and posterior seta (a), endite 3e with a single anterior seta 2 and posterior seta (b), endite 2 with a short anterior seta 3, endite 1 with large anterior seta 4. Size: 0.80–1.89 mm (n = 18).

Adult male. Body elongated, dorsal margin almost straight, shallow depression between head and valves, postero-dorsal angle distinct, with relatively short caudal spine (Figure 7a). Head with a well-developed rostrum, region of antenna I joint with a distinct depression (Figure 7b). Antenna I slightly curved, without rows of setules, antennular sensory seta thin, its tip reaching the distal part of antenna I (Figure 7c). Aesthetascs of different size, largest aesthetasc longer than antenna I diameter. Male seta relatively short, its length same as antenna I length. Anterior margin of valves (Figure 7d) slightly convex, with short setules, antero-ventral angle with row of long setules, ventral margin almost straight, bearing row of setules. Postero-ventral angle rounded, with row of small spines (Figure 7e), caudal spine bearing spines on ventral and dorsal portion. Inner portion of postero-ventral part bearing row of setules and small setules between them (Figure 7f–h), a bunch of long setae in middle of ventral margin (Figure 7i). Abdomen without processes on distal segments. Postabdomen elongated, tapering distally, with long postabdominal seta, as long as preanal margin (Figure 7j). Preanal margin slightly convex, preanal angle expressed, spines on postanal and anal portion differenced in size distally. Gonopore opens subdistally (Figure 7). Postabdominal claw bearing three pectens, first pecten with 7 setules, second with five strong spines, third pecten with row of fine setules (Figure 7k). Outer distal lobe of limb I (Figure 7l–n) large, bearing a very long seta, its tip with row of small setules (Figure 7l); inner distal lobe with a bent copulatory hook and two seta of different size. Distal most endite of limb II with a modified hook-like seta (Figure 7o). Size: 0.82 mm (n = 1).

Variability and morphological identification of the hybrid specimens. According to the key of Kotov et al. [24] (p. 799), Daphnia curvirostris group differs from D. korovchinskyi species groups in: (1) rostrum of female notably long (as a result, tips of aesthetascs are located far from rostrum tip); and (2) sensory seta on male antenna I reaches tip of post-aesthetasc projection. In addition, D. curvirostris s.str., in contrast to D. korovchinskyi; (3) lacks a strong medial keel on the posterior head margin, separated from the labrum by a deep incision (while the posterior head margin is sometimes slightly convex in the former); (4) has seta 4 on exopodite III markedly longer than seta 3; (5) bears seta 3 on inner-distal lobe of the limb III relatively long. In general, all studied specimens correspond morphologically to D. curvirostris, while the characteristic traits of D. korovchinskyi (especially strong medial keel) were not found in any studied specimens. We were not able to find differences between D. curvirostris s.str. and its hybrids with D. korovchinskyi based on our detailed morphological analysis of all specimens from Bakhilovo.

The difficulties of morphological differentiation of D. curvirostris s.str. and its hybrids with D. korovchinskyi in our case are related with a size structure in the studied sample: the maximum specimen size was 1.89 mm, while females of D. curvirostris reach 2.88 mm [24]. The studied population was represented by relatively small-sized and relatively young females with taxonomically valuable characters which are not fully expressed yet, i.e., the rostrum is shorter than in large females [24,25]. A single male in our population was not fully typical of D. curvirostris having a relatively short flagellum on antenna I. It could be a hybrid specimen, but an adult male of D. korovchinskyi is not described yet, and we cannot discuss an “intermediate” morphology of the male from Bakhilovo. Moreover, we do not know, was this male fertile? Therefore, we found in some specimens the characters which could suggest their hybrid status, but this opinion is a hypothesis only.

4. Discussion

Our genetic analysis revealed both D. curvirostris s.str. and its hybrids with the Far-Eastern endemic D. korovchinskyi in the sample from Bakhilovo. The latter taxon was known to date only from two ponds in the Amur River valley (Khararovsk Area, Far East of Russia) being a locally distributed pre-Pleistocene relict. Our previous analysis of all possible sequences of D. curvirostris reveals a moderate relation between genetic and geographic distances (Mantel test p = 0.89), the Volga sample analysis led to an opposite conclusion. Just anthropogenic transportation (when a species is momentarily replaced through thousands of kilometers) explains a contradiction between conclusions on the differentiation or expansion as most possible biogeographic scenario (neutrality test) and absence of genetic differences between distant populations. It is impossible to assume a natural way to form such a genetic pattern.

Moreover, Kotov et al. [24] found that a differentiation of the D. curvirostris group and the D. sinevi-korovchinskyi group happened in the Mesozoic era, confirming the idea of a very old age of the cladoceran taxa of different ranks, including species groups [52,53,54]. Apparently, the revealed pattern could not be explained by an ancient mitochondrial introgression. Sequences of HSP90 and 18S loci of D. curvirostris s.str. from Bakhilovo are identical to those of other populations of this species from European Russia, but, if this pattern is natural, we can expect some genetic differentiation since the Mesozoic. The sequence of one DNA strain in the hybrids is identical to that in D. curvirostris, while the sequence of the other strain is identical to D. korovchinskyi. This is a signature of a very recent hybridization.

It is important that poor D. korovchinskyi was not detected in our sample, both morphologically and genetically. The exact hybrid system in our pond is not known. Moreover, we have no evidences of sexual reproduction of the hybrids, they could reproduce by parthenogenesis only, as is known for the hybrids of the D. pulex group [21], or continuously crossing with parents, as do some members of the D. longispina group, making also backcrosses [55,56,57].

We have found genetic evidence of a past (but apparently in the 20th-21st century) anthropogenic transportation of a cladoceran from the D. curvirostris group from a distant region to the Volga basin. Such discovery became possible only now, after two global revisions of this group and development of the genetic method for accurate species identification [9,24]. Keeping in mind previously detected human-mediated introduction of D. curvirostris s.str. to North America, we can conclude that all curvirostris-complex has a significant potential for invasions, probably because its ephippia are easily dispersed and strongly adapted to survive drying, freezing, etc. due to which D. curvirostris is very common in semiarid temporary pools [9,58]. This is the first report on the hybrids within the D. curvirostris species complex.

The Volga basin, the largest in Europe, provides a good example of the fauna mixing, i.e., of Cladocera. Range expansions of the haplopods of the Pontocaspian origin (Cercopagis pengoi, Pseudevadne tergestina) towards the south are well known there [59,60]. Few ctenopods and anomopods are also regarded as expanding their distribution ranges north [61,62]. Yet, no one case of cladoceran transportation from the Far East to the Volga basin was known to date, although East-Asian invaders to the Volga Basin are well-known among fishes [63,64] and their parasites [65,66], bivalve mollusks [67], freshwater crabs [68], aquatic plants [69] and other aquatic organisms [70]. However, we do not expect serious changes in the water bodies of the Volga basin due to D. korovchinskyi appearance there. At least our studies revealed only a single case of its presence in the Volga basin, without a further expansion of its distribution range.

The exact vector of this invasion to the Volga is unknown. Most probably, just ephippia of D. korovchinskyi were relocated from Khabarovsk Territory to Samara area somehow. Samarskaya Luka is located in between two large industrial centers, Samara and Tolyatty (8th and 19th cities of Russia according to their population size [71]), both with a strong industry and intensive transportation to/from other regions of Russia. In the region, there are many factories using accessories from different regions of Russia, including the Far East. Moreover, the Far East of Russia is an important transitional region for the transfer of second-hand Japanese and Korean made cars. Ephippia could be occasionally attached to the car wheels, or occasionally taken with some accessories. A chance of transportation of any specimens in active phase instead of dormant stages is minimal in our case.

Morphological characters of hybrid specimens are studied mainly in the Daphnia longispina species group, where they usually are morphology intermediate between characters of two parents [57,72,73]. In contrast, in the D. pulex group, no specific traits were revealed in the hybrid females, but it could be explained by an insufficient level of morphology-based taxonomy of the group [74]. Moreover, some specimens of D. logispina group having a hybrid origin were previously described as independent biological species [72], which is a bad practice in terms of the International Code of Zoological Nomenclature [75]. However, as the D. curvirostis complex is among the best-studied cladoceran groups [9,24,25], we are sure that there were no hybrid forms among the formal taxa from this complex described to date; see the WORMS checklist [76]. Hybrids of Daphnia are prospective objects of morphological studies, which need to be conducted in combination with detailed molecular studies.

5. Conclusions

Here, we demonstrated that accurate studies with deep resolution increase the number of revealed cryptic invasions. At the same time, many cladoceran taxa are studied inadequately, and analogous discoveries are impossible for them. We expect that the number of revealed cases of cryptic interspecific invasions will grow rapidly.