Prevalence and genetic diversity of Echinorhynchus gymnocyprii (Acanthocephala: Echinorhynchidae) in schizothoracine fishes (Cyprinidae: Schizothoracinae) in Qinghai-Tibetan Plateau, China

Background The schizothoracine fishes, an excellent model for several studies, is a dominant fish group of the Qinghai-Tibet Plateau (QTP). However, species populations have rapidly declined due to various factors, and infection with Echinorhynchus gymnocyprii is cited as a possible factor. In the present study, the molecular characteristics of E. gymnocyprii in four species of schizothoracine fishes from the QTP were explored. Methods We investigated the infection status of E. gymnocyprii in 156 schizothoracine fishes from the upper Yangtze River, upper Yellow River, and Qinghai Lake in Qinghai Province, China. The complete internal transcribed spacer (ITS) of the ribosomal RNA (rRNA) gene and part of the mitochondrial cytochrome c oxidase subunit 1 (cox1) gene of 35 E. gymnocyprii isolates from these fishes were sequenced and their characteristics analyzed. In addition, we inferred phylogenetic relationships of the E. gymnocyprii populations based on the rRNA-ITS and cox1 sequences. Results The total prevalence of E. gymnocyprii in schizothoracine fishes was 57.69% (90/156). However, the prevalence among different species as well as that across the geographical locations of the schizothoracine fishes was significantly different. The results of sequence analysis showed that the four E. gymnocyprii populations from different hosts and regions of Qinghai Province were conspecific, exhibiting rich genetic diversity. Phylogenetic analysis based on rRNA-ITS and cox1 sequences supported the coalescence of branches within E. gymnocyprii; the cox1 gene of E. gymnocyprii populations inferred some geographical associations with water systems. In addition, three species of schizothoracine fishes were recorded as new definitive hosts for E. gymnocyprii. Conclusions To the best of our knowledge, this is the first molecular description of E. gymnocyprii populations in schizothoracine fishes from the Qinghai-Tibet Plateau that provides basic data for epidemiological surveillance and control of acanthocephaliasis to protect endemic fish stocks.


Background
Acanthocephalans are a group of widely distributed intestinal helminths, commonly found in fishes worldwide; they can cause malnutrition, tissue injury, and intestinal obstruction in hosts [1][2][3]. Echinorhynchus Zoega in Müller, 1776 (Acanthocephala: Echinorhynchidae) is a group of spiny-headed worms that is mainly parasitic in teleost fish and crustaceans found in various aquatic environments. It is the most widely distributed genus of acanthocephalans [4]; 52 species of Echinorhynchus have been recorded [5]. Because of its high morphological variability, there is a debate on the subdivision of this genus, and morphological identification of the species of Echinorhynchus is difficult [4,6]. Therefore, DNA sequence data from molecular markers have been used for classification, evolution, and phylogeny of the Echinorhynchus spp. [4,7,8]. Echinorhynchus gymnocyprii Liu, Wang & Yang, 1981, with an elongated and cylindrical proboscis, and 14-16 longitudinal rows of hooks (each row having 10-11 hooks of unequal length), was first described in Gymnocypris przewalskii (Kessler, 1876) in the Qinghai Lake of China [9]. Later, its presence was discovered in Triplophysa scleroptera (Herzenstein, 1888) in the Qinghai Lake and in T. stenura (Herzenstein, 1888) in the Lhasa River in Tibet [10,11]. However, to date, there has been no molecular data on E. gymnocyprii.
The schizothoracine fishes (Teleostei: Cyprinidae) are a dominant fish group of the Qinghai Tibet Plateau (QTP) ichthyofauna, and China has the largest distribution of schizothoracine species worldwide [12][13][14]. These fishes are endemic and adapted to the extreme environmental conditions (e.g. hypoxia, strong ultraviolet rays, food shortages and cold) of the QTP [12,[15][16][17]. Therefore, they are ideal models to investigate speciation, biological evolution, ecological effects, paleo-drainage changes, and biogeography of the QTP [13,[17][18][19][20]. However, a variety of factors have led to a sharp decline in the population of schizothoracine fishes in the last decades [21,22]. Parasites found in these fishes may contribute to this decline, as some studies suggest that parasites may have adverse effects on their host populations [23,24]. As an important part of QTP biodiversity, parasites of endemic fishes have gradually attracted the attention of researchers [11,25].
During the investigation of the helminth fauna of schizothoracine fishes in the QTP of China, it was found that these fishes were infected with acanthocephalans, which were initially identified as E. gymnocyprii based on morphological studies. In the present study, the infection status of E. gymnocyprii in four species of schizothoracine fishes from the upper Yangtze River, Qinghai Lake, and upper Yellow River in the QTP was evaluated. The molecular data on E. gymnocyprii populations from different hosts and geographical locations (including the E. gymnocyprii population from the type-host, G. przewalskii, and type-locality of Qinghai Lake) are described based on ribosomal RNA (rRNA), which includes the complete internal transcribed spacer (ITS) and the mitochondrial cytochrome c oxidase subunit 1 (cox1) gene. In addition, the phylogenetic relationships of the E. gymnocyprii populations were investigated based on the rRNA-ITS and cox1 sequences.

Sampling
A total of 156 schizothoracine fishes (20 Ptychobarbus kaznakovi Nikolskii, 1903, 73 Gymnocypris eckloni Herzenstein, 1891, 11 Gymnodiptychus pachycheilus Herzenstein, 1892, and 52 Gymnocypris przewalskii), from four localities and three water systems in the northeastern QTP of Qinghai Province, China, were examined for parasites (Fig. 1). The sampling locations were in Zhiduo County (ZD) of the upper Yangtze River, Maduo County (MD) and Dari County (DR) of the upper Yellow River, and Qinghai Lake (QHL); the altitudes of ZD, MD, DR, and QHL are 4285 m, 4300 m, 3970 m and 3196 m, respectively. The Yangtze and Yellow rivers are both freshwater rivers, and the Qinghai Lake is a closed-basin brackish lake. The acanthocephalans were washed immediately in saline after collection from the intestines of fishes, preserved in 70% ethanol, and stored at 4 °C. The preliminary identification of E. gymnocyprii was mainly based on morphological characteristics (elongated and cylindrical proboscis, armed with 14-16 longitudinal rows of hooks, each row having 10-11 hooks of unequal length) [9,26]. Subsequently, 35 acanthocephalans were selected for DNA analysis. The E. gymnocyprii found in G. przewalskii of Qinghai Lake was the type-species.

DNA extraction, amplification, cloning and sequencing
Before performing genomic DNA analysis, the anterior extremities of the acanthocephalans were cut off to avoid contaminating the DNA with fish tissues attached to the proboscis. Next, the genomic DNA of individual acanthocephalans was extracted using the QIAamp ® DNA Mini kit (Qiagen, Hilden, Germany) according to the manufacturer's recommendations.

Sequence analysis
The sequence obtained from each acanthocephalan was assembled using DNAStar software [28] and then aligned with Clustul X 2.0 [29] and edited manually. Subsequently, the data obtained from these corrected sequences were analyzed with MEGA 6.06; both the pairwise distance and the mean distance within/ between the populations were estimated using the Kimura 2-parameter model [30,31]. In addition, the sequences were compared with sequences from Gen-Bank by BLAST analysis. The sequences of ITS1 and ITS2 were defined in a previous study [32]. Based on the ITS1, ITS2 and cox1 sequences of each population, haplotype analysis was carried out using DNASP 5.10 [33].

Phylogenetic analysis
The rRNA-ITS and cox1 sequences were independently analyzed and aligned with sequences from GenBank. The rRNA-ITS sequences were as follows: E. gadi Zoega in Müller, 1776 (EF107647, EF107648), Pomphorhynchus laevis Zoega in Müller, 1776 (KJ756500), P. bosniacus Kistaroly & Cankovic, 1969 (MH319900), and P. zhoushanensis (KY472823); the cox1 sequences were: E. gadi (AY218095), E. salmonis Müller, 1784 (KP261017), E. truttae Schrank, 1788 (DQ089710), and Acanthocephalus lucii Müller, 1776 (KP261016), according to the software Clustal X 2.0 [29]. After amendment, the rRNA-ITS and cox1 datasets included 633 characters and 585 characters, respectively. Phylogenetic trees were constructed by the neighbor-joining (NJ) method using MEGA 6.06 [30] and Maximum Likelihood (ML) with IQ tree 1.6.10 [34]. For NJ, the Kimura 2-parameter model was used. The TPM3 + F + I for the rRNA-ITS dataset, and the HKY + F + I for the cox1 dataset, were suggested as the best-fit models by ModelFinder [35] of IQ tree 1.6.10 based on the Bayesian information criterion (BIC). Pomphorhynchus zhoushanensis and A. lucii were selected as the outgroups of the rRNA-ITS and cox1 datasets, respectively. The branch reliability of the phylogenetic tree constructed by the two methods was tested by bootstrapping with 1000 replicates. A haplotype network was constructed from the cox1 DNA sequence data with the software TCS using the probabilistic method of statistical parsimony [36,37].

Statistical analysis
The differences in E. gymnocyprii infection rates among different fishes and locations were analyzed using the chi-square test in SPSS 25.0 for Windows (IBM Corp., New York, USA). P < 0.05 was considered as statistically significant.
Regarding the cox1 gene, 35 newly generated sequences were 702 bp long; the sequences of QH1 and QH5 were identical, and the sequences of ZD4 and ZD5 were identical. The overall cox1 GC content was 37.3%, showing a high AT bias; the percent identity between these sequences was 95.4-100.0%. The overall average of pairwise distances was 0.02182, and the percent identity/mean distance within each population was 98.   Table 3, and the values of h, k, and π indicate that the populations of E. gymnocyprii are rich in genetic diversity.

Phylogenetic analyses
The topological structures of the phylogenetic trees based on the rRNA-ITS and cox1 sequences of the E. gymnocyprii populations were similar; all the E. gymnocyprii populations clustered in a single well-supported clade, with only slight differences in the bootstrap values for some nodes (Figs. 2, 3, 4). The phylogenetic NJ tree of E. gymnocyprii populations based on the cox1 sequences, inferred 4 subclades (subclades MD, DR, QHL and ZD; Fig. 4), corresponding to 4 geographical locations belonging to 3 water systems. The MD and DR groups belong to the Yellow River system, the QHL group belongs to the Qinghai Lake system, and the ZD group belongs to the Yangtze River system. However, the DR and ZD groups appear to be paraphyletic in the ML tree based on the cox1 sequences (Fig. 3).
The 35 specimens comprised 33 unique haplotypes, and the statistical parsimony analysis revealed 3 distinct networks (QHL, ZD and MD-DR; Fig 5), corresponding to three water systems. Network QHL contained 4 haplotypes from 5 individuals, QHL1 and QHL5 were considered the central haplotype; network ZD contained 4 haplotypes from 5 individuals, ZD4 and ZD5 were considered the central haplotype; network MD-DR contained 25 haplotypes from 25 individuals, with MD13 as the ancestral haplotype.

Discussion
In this study, we found that the prevalence and intensity of E. gymnocyprii from G. przewalskii in Qinghai Lake had similarity to those reported in previous studies [10]. Previously, a range of 33.9-100% for prevalence and 14.9-41.4 for mean intensity of E. gymnocyprii in G. eckloni, Gymnodiptychus pachycheilus, and P. kaznakovi was reported. In addition, the prevalence of E. gymnocyprii in schizothoracine fishes differed significantly depending on fish species and location. There are several possible explanations for these differences, such as the characteristics (e.g. feeding behavior, age, physiology and abundance) of different definitive hosts, different numbers of samples, and the ecological environment (e.g. density of intermediate hosts, aquatic organisms, water conditions and altitude) [38][39][40][41][42][43][44]; these differences result in different survival conditions for both the schizothoracine fishes and acanthocephalans. Our results suggest that E. gymnocyprii is widely distributed in native fishes of the QTP, a finding that should be considered when working to protect the schizothoracine fish population. Further studies of E. gymnocyprii may serve as a model for studying the co-evolution of native fishes and parasites in the QTP.
However, the host species, geographical distribution, age, and other factors may cause changes in the morphology of Echinorhynchus spp. These changes sometimes exceed the suggested generic boundaries, thereby leading to inconvenience and confusion in the classification and identification of Echinorhynchus spp. [2, 4, 6, [45][46][47]. Therefore, DNA sequences from the ribosomal gene cluster (e.g. 18S, ITS and 28S) or mitochondria (e.g. cox1) provide a complementary methodology for analyzing the taxonomy and phylogeny of Echinorhynchus spp. [4,7,47]. Taken together, these results support the use of rRNA (containing ITS) and the cox1 gene as genetic markers for the identification of E. gymnocyprii. In our study, the values of π, k, and Hd based on the cox1 gene indicated that the E. gymnocyprii populations exhibit rich genetic diversity [33], which may explain the existence of these populations in a variety of endemic  Fig. 2 Phylogenetic relationships of E. gymnocyprii. The tree is inferred from the rRNA-ITS sequences using neighbor-joining (NJ) method. The maximum likelihood (ML) method produced phylogenetic tree with the same branch topologies. Bootstrap support from ML/NJ analysis are shown above the nodes. The scale-bar indicates the number of substitutions per site. Pomphorhynchus zhoushanensis was used as the outgroup fish in the QTP and their adaptation to the harsh environment of the plateau. The overall π (0.02122) and k (14.95966) were significantly higher than those of π (0.00397-0.00851) and k (2.800-6.000) of single populations from MD, DR, QHL and ZD, which indicates that geographical factors may have a positive impact on the evolution of the cox1 gene in the E. gymnocyprii populations.
The phylogenetic analysis based on the rRNA-ITS and cox1 gene supports the coalescence of branches within the E. gymnocyprii populations, and provides evidence that these populations belong to the same species. Interestingly, the NJ tree based on the cox1 gene showed four subclades (Fig. 4), and the statistical parsimony analysis revealed three distinct networks (Fig. 5), both reflecting geographical associations with water systems. The E. gymnocyprii populations from the Yellow River system (MD, DR and QHL) were more closely related to each other than to those of the Yangtze River system (ZD). This result might be explained by geological research findings, which imply that the Qinghai Lake was connected to the ancient Yellow River approximately 0.15 mega-annum before the present (Ma BP) [48]. The distribution and evolution of schizothoracine fishes, as hosts of E. gymnocyprii, in the QTP are strictly due to differences in water systems and habitats [15,[49][50][51][52]. The cox1 gene of E. gymnocyprii populations showed some geographical associations with water systems, indicating that adaptive changes in the cox1 gene enabled the species to survive in different water environments and preserve genetic differentiation. Consequently, the cox1 gene is potentially useful for investigating the zoogeography of E. gymnocyprii.
In the present study, to the best of our knowledge, two DNA markers of E. gymnocyprii were analyzed for the first time, and three new host records were added, which  Figure S1 may be the first step in further understanding the speciation and evolution of this species in the QTP. This study also provides important molecular data for revision of the genus Echinorhynchus [4] in the future.

Conclusions
Echinorhynchus gymnocyprii is widely distributed in native fish of QTP; three species of schizothoracine fishes (G. eckloni, Gymnodiptychus pachycheilus and P. kaznakovi) were identified as new definitive hosts of E. gymnocyprii. Our study is the first molecular characterization of E. gymnocyprii populations in schizothoracine fishes from the QTP; it provides basic data for epidemiological surveillance and control of acanthocephaliasis to protect endemic fish stocks in the QTP. Statistical parsimony network of E. gymnocyprii haplotype based on cox1 gene sequences. The connection limit excluding homoplasic changes was set to 95%. Each oval represents a haplotype, and the ancestral haplotype (with the highest outgroup probability) is indicated by a square, the size of the square or oval corresponds the haplotype frequency. Each line equates to one mutational step, and the small circles are hypothetical haplotypes