ABSTRACT

The common carp (Cyprinus carpio) has a large variety of strains. The more popular are the koi (Japanese ornamental carp), which are still bred today to generate creative colors and patterns, giving rise to multiple phenotypes. Since koi are in great demand, there is a challenge to determine the genetics defining their quality. Two methods: 1) direct sequencing of five candidate gene regions, i.e., mitochondrial (cytochrome b, 12S gene and the D-loop) and nuclear (red sensitive opsin and Rag-1) loci, to detect single nucleotide polymorphisms (SNP)s and 2) random amplification of polymorphic DNA (RAPD), were used to differentiate among four koi strains (Kohaku, Sanke, Ghost and Ohgon) and the common carp. Novel SNPs, distinguishing between koi and the common carp, were revealed in cytochrome b, the D-loop and in the red sensitive opsin; one was a missense mutation in cytochrome b at position 15,860, in which threonine in the common carp became alanine in all koi strains examined. The Kohaku strain was found to have two alleles in the mitochondrial fragments, forming two different haplotypes (subpopulations). These novel SNPs distinguished between koi strains and the common carp, and the RAPD method enabled further differentiation among the four koi strains.

1. INTRODUCTION

The common carp (Cyprinus carpio) has a long history of domestication. Much artificial hybridization has been performed on this species over the years, leading to a large variety of strains in European and Asian countries [1,2]. Some of the carp strains generated over the years have been used as ornamental fish, because of the large variety in skin color and patterns. One of the more popular strains is the koi (Japanese ornamental carp). The koi fish are colored carp with much variation in color and patterns [3]. Some of the koi strains have been and are still being bred today in a specific manner, in order to generate creative colors and patterns, giving rise to multiple phenotypes among the strains existing today. The main demand today in Europe, the USA and East Asia, is for fish with an attractive and rare phenotype [4]. In a study on three main phenotypes, it was concluded that a gene complex, as opposed to a single gene, is responsible for the koi color [5,6]. These studies and others have found a correlation between the red phenotype and a microsatellite site, a specific allele correlated with red color [7]. Other groups have reported that at least five genes control the bright colors (white, yellow and orange) on the dark background of the wild type [8]. Genetic molecular markers are a useful tool in the identification of genetic polymorphisms and have contributed a lot to the research of intra and inter species polymorphisms, as well as to the building of phylogenetic trees at a high resolution [9]. The methods are divided into two main approaches: the use of random DNA-based genetic markers (such as random amplification of polymorphic DNA [RAPD], amplified fragment length polymorphism [AFLP] and microsatellites) and specific candidate gene single nucleotide polymorphism (SNP)s [10]. In this study, we used direct sequencing of candidate gene regions and RAPD. The candidate genes chosen were: two nuclear genes, red sensitive opsin and Rag-1, [11] and three mitochondrial genes, cytochrome b, 12S and the D-loop region of the mitochondrial DNA (mtDNA) [5]. The opsin protein is a photoreceptor with variable phenotypes, which are encoded by a few genes. The long-wave sensitive (LWS) gene was discovered to be polymorphic in guppy fish from different geographical areas and to have at least two copies in the genome [11]. Rag-1 encodes enzymes and is found in the VDJ recombination of immunoglobulins and in T cell receptors. An investigation on genetic intra-species polymorphism in the Mantella madagascariensis frog showed that 31 informative sites identified genetic changes among species, some of which are used today for the classification of the species of different strains [12].

MtDNA is known to be less conserved in evolution. Studies in vertebrates have shown that, since there are fewer mtDNA repair mechanisms, changes in mtDNA are more frequent. This is especially true in the D-loop, which is a noncoding sequence in mtDNA. However, this also occurs in the encoding genes, cytochrome b and 12S [13]. Another advantage of mtDNA is that it passes on maternal inheritance, and thus can trace complex heredity traits. MtDNA markers have been used a lot in research on sharks and eels. One study on mtDNA in the aquaculture field was carried out on guppy fish. A polymorphism of two nucleotides was found in cytochrome b which serves as a molecular marker of strains [14]. The different colors and patterns in ornamental carp are thought to be a result of mutations in the wild carp, followed by crossbreeding between different strains. The goal of this study was to identify molecular markers that are linked to different koi strains, and thus could specifically identify strains. Such markers would be valuable in aqua agriculture.

2. MATERIALS AND METHODS

The fish were obtained from the Dan Fish Farm. Eighty five fish from five strains of koi and the common carp were randomly collected: 15 Sanke, 15 Kohaku, 15 Ghost, 20 Ohgon, 10 Ochiba and 10 common carp (Cyprinus carpio). In the second phase of the study, seven males and females from the Sanke, Kohaku, Ghost and Ohgon strains were collected for crossbreeding to generate F1. Larvae remained on the farm, until they were six weeks of age, when it was possible to clearly identify the phenotype. From the growing pools, twenty F1 offspring were sampled randomly, five individuals from each species for the remainder the koi fish were as follows— Kohaku: red and white; Sanke: black, white and red; Ochiba: gray; Ghost: orange and black; and Ohgon: white and orange (Figure 1).

DNA samples, which were derived from 20 mg of tail fins, were extracted with a standard genomic DNA extraction kit. Mitochondrial DNA of four koi strains: Kohaku, Sanke, Ghost and Ohgon and the common carp, was amplified by polymerase chain reaction (PCR), followed by Sanger sequencing, using ABI 16 capillaries apparatus and primers, designed according to the Cyprinus carpio complete mitochondrial genome, GenBank accession number: ×61010. The fragments, which were amplified and sequenced, were from the mitochondrial genes: cytochrome b, 12S and D-loop and the nuclear genes: red sensitive opsin and Rag-1, GenBank accession numbers: AB055656 and AY787040, respectively.

RAPD analysis was performed on five koi strains: Kohaku, Sanke, Ghost, Ochiba and Ohgon. Eight primers were designed, and among them, two (opa-3, opa-9) were chosen; they showed the clearest patterns of the different haplotypes, enabling of the research. The color phenotypes of differentiation among the strains.

3. RESULTS

In order to evaluate the genetic distance between the common carp and koi, mtDNA fragments from cytochrome b, 12S and the D-loop (control region) genes were sequenced. Thirteen SNPs were found in a fragment of 356 base pairs (bp) from the cytochrome b gene; three of which were missense mutations (p.Thr190Ala c.A15860- G; p.Val192Ala c.T15867C; p.Ile218Val c.A15944G). Five of the SNPs, which included the Thr190Ala missense mutation, distinguished between the common carp and the four koi strains (Table 1 and Figure 1). The other eight SNPs were identical in the carp and the other strains, with the exception of the Kohaku strain, which comprised two subpopulations, one with the common allele (like in the other strains and the carp) and the second with a polymorphic allele. We named it the “rare allele”, since it was found only in 33% of the Kohaku strains and not in any other strain. These data enabled us to draw a phylogenetic tree based on the cytochrome b polymorphism, which implied that the koi “rare allele” strain branched before the separation between the common carp and the other koi strains (Figure 2). A cross between the Kohaku “rare allele” female and a Sanke male generated F1 offspring with both koi phenotypes and the “rare allele” genotype, indicating that this mtDNA sequence was maternally inherited.

The purpose of generating offspring (F1) from Sanke, Kohaku, Ghost and Ohgon strains was that the Kohaku strain is generally cross briding with the Sanke strain, and because we found a rare allele in the Kohaku strain only, we wanted to know if the origin of this rare allele is from the Sanke strain. The Sanke strain has a red, white and black colors and the Kohaku in red and white. So the cross was to see if the black color in the rare allele came from the Sanke. The cross between the Ghost to Ghost and between the Ohgon and Ohagon (generating F1 offspring’s), was as a control briding step only, and so we did not use the Ochiba and common carp as control as well.

However, this polymorphism did not distinguish between the two koi strain colors. Four SNPs in a 199 bp fragment from the mtDNA 12S gene were detected in all Kohaku strains. Thus, none of these SNPs were informative, since they were not able to differentiate between the common carp and the koi strains.

In a 380 bp fragment from the mtDNA D-loop region, 15 SNPs were found. Ten were only in the Kohaku strain, and five others demonstrated a significant difference between the koi and the common carp (Table 2 and Figure 3). These results were used to build a phylogenetic tree based on the D-loop polymorphism, which showed that the Kohaku with the “rare allele” is the most genetically distant, when compared to the other strains examined (Figure 4). A cross between the Kohaku D-loop “rare allele” female and a Sanke male generated F1 offspring with both koi phenotypes and the “rare allele” genotype, indicating that this mtDNA sequence was maternally inherited. However, this polymorphism did not distinguish between the two koi strain colors.

Two nuclear genes, the Rag-1 and red sensitive opsin, were also screened for SNPs. In the red sensitive opsin gene, a 458 bp fragment from the end of the coding region and the 3’ UTR were sequenced. The three SNPs (two in the coding region (p.Val328Val c.1076A > T; p.Ser342Ser c.1118T > C) and one in 3’ UTR, (c.1350G > A) (Table 3), were found to form five different haplotypes. The

different koi haplotypes were heterozygous in all three positions, 1076, 1118 and 1350, whereas the common carp was homozygous; A/A in position 1076, T/T in 1118 and G/G in 1350 (data not shown). In the Rag-1 gene, a fragment of 154 bp was sequenced and seven SNPs were identified, which generated nine different haplotypes in the four different koi strains and common carp. However, since there was no correlation between the haplotypes and the strains, these SNPs could not distinguish between inter or intra strain differences (data not shown). A phylogenetic tree based on the red sensitive opsin, cytochrome b and opsin polymorphisms distinguished the common carp from the koi strains (Figure 5).

The second approach was to screen differences and polymorphisms randomly, instead of according to special candidate genes. This was carried out by the random amplification of polymorphic DNA (RAPD) assay with two

primers, opa-3 and opa-9. Amplification with the opa-3 primer revealed nine fragments, producing 23 different haplotypes (Figure 6). With the opa-9 primer, the nine resulting fragments gave 28 different haplotypes. When these 18 fragments were analyzed by means of a phylogenetic tree, two main branches were obtained, one with the Ghost and Ohgon strains and another, which subdivided into two more branches, one including Kohaku and Sanke and the other with the Ochiba strain (Figure 7).

4. DISCUSSION

In this study, an attempt was made to identify genetic differences among four strains of koi and the common carp, using SNPs on three mitochondrial and two nuclear genes and basic RAPD technology. Genetic relationships among six populations of domesticated carp (Cyprinus carpio), a species with a partially sequenced genome, has been carried out using genetic markers, such as 18S and RAPD, as well as SNPs [3]. From the five different koi strains, Kohaku and Sanke are known to be more crossbred because of their color demand, and as a result, harbor more genetic polymorphisms and are more genetically related [3]. In our study, we also found multiple genetic polymorphisms in the mtDNA of the Kohaku strain. According to the RAPD method, these two strains are genetically closer.

The assumption that koi fish are descended from the common carp has existed for a long time [19]. Several studies over the years have engaged in detecting genetic differences among strains of different carp and koi species, revealing phylogenetic relationships [3,6,14,20,21]. The present project has contributed more information towards its completion, by focusing on genetic regions, which indicate clear, but close, phylogenetic distances between the common carp and koi strains.

Since mtDNA is less stable than nuclear DNA in acquiring new mutations [22], we chose these two types of DNA regions. In this study, differences in two mitochondrial genes (cytochrome b and D-loop) and in one nuclear one, the red sensitive opsin, were discovered. These novel SNPs distinguished the common carp from the koi, but did not differentiate among the koi strains. In the red sensitive opsin nuclear gene, three SNPs were found to be heterozygous (or rarely homozygous) only in the koi strains and not in the common carp (Table 3).

Genetic polymorphism has been detected in the mtDNA of guppy strains [5]. In the present study, we found several novel SNPs in the three regions tested in the mtDNA, five novel SNPs in the cytochrome b gene (one of which is a missense mutation; Thr190Ala) and five other novel SNPs in the D-loop region, which clearly segregate the koi strains from the common carp. Thus, these are good genetic markers, which can be utilized for distinguishing between the koi and the common carp, even in very early stages of fish development.

We used the three genomic regions (cytochrome b, D-loop and red sensitive opsin) that were found to harbor genetic polymorphisms, as a basis for genetic variance between the koi strains and the common carp. The Anova test for genetic variance was executed twice, once on the koi strain data alone, and a second time, when the data on the common carp was included. Omission of the common carp from the Anova analysis decreased the genetic variance and the Fst value, which was 0.221 (21%, P < 0.05), in comparison to a Fst value of 0.342 (34%, P < 0.01), when the common carp was included. When this test was applied to the RAPD data, based on the genetic polymorphism obtained from the opa-3 and opa-9 primers “Figure 6”, the Fst value was 0.183 (18%, P = 0.01). According to this type of analysis, the SNPs revealed higher genetic variations between the koi strains and the common carp, in comparison to the RAPD, but this finding is probably affected more by the polymorphism found in the red sensitive opsin gene and less by the mtDNA regions, where we found less genetic variance and which was mostly due to the Kohaku strain. The RAPD analysis showed more genetic similarity between the Kohaku and the Sanke strains, in comparison to the other koi strains. This is in agreement with previous findings by Jackson et al. [6].

In the Kohaku strain, two different alleles were identified in the three mitochondrial fragments; thus, forming two different haplotypes, which differed in two amino acids in the cytochrome b gene: V192A and I218V. Apparently, this is a subpopulation of the Kohaku. This is based on the phylogenetic tree of the cytochrome b polymorphism, which shows that the Kohaku “rare allele” strain probably branched before the separation between the common carp and the other koi strains “Figure 2”.

In summary, by means of candidate gene sequencing, we identified novel SNPs, which can distinguish between the four koi strains and the common carp, but not among the different koi strains. However, the RAPD method was able to distinguish among the koi strains studied here, as well as between them and the common carp.

REFERENCES