the formation of the polyploid hybrids from different subfamily … · 2007. 7. 25. ·...
TRANSCRIPT
Copyright � 2007 by the Genetics Society of AmericaDOI: 10.1534/genetics.107.071373
The Formation of the Polyploid Hybrids From Different Subfamily FishCrossings and Its Evolutionary Significance
Shaojun Liu,1 Qinbo Qin, Jun Xiao, Wenting Lu, Jiamin Shen, Wei Li, Jifang Liu,Wei Duan, Chun Zhang, Min Tao, Rurong Zhao, Jinpeng Yan and Yun Liu
Key Laboratory of Protein Chemistry and Fish Developmental Biology of the Education Ministry of China, College ofLife Sciences, Hunan Normal University, Changsha 410081, People’s Republic of China
Manuscript received January 26, 2007Accepted for publication March 17, 2007
ABSTRACT
This study provides genetic evidences at the chromosome, DNA content, DNA fragment and sequence,and morphological levels to support the successful establishment of the polyploid hybrids of red crucian carp3 blunt snout bream, which belonged to a different subfamily of fish (Cyprininae subfamily and Cultrinaesubfamily) in the catalog. We successfully obtained the sterile triploid hybrids and bisexual fertile tetraploidhybrids of red crucian carp (RCC) ($) 3 blunt snout bream (BSB) (#) as well as their pentaploid hybrids.The triploid hybrids possessed 124 chromosomes with two sets from RCC and one set from BSB; thetetraploid hybrids had 148 chromosomes with two sets from RCC and two sets from BSB. The females oftetraploid hybrids produced unreduced tetraploid eggs that were fertilized with the haploid sperm of BSB togenerate pentaploid hybrids with 172 chromosomes with three sets from BSB and two sets from RCC. Theploidy levels of triploid, tetraploid, and pentaploid hybrids were confirmed by counting chromosomalnumber, forming chromosomal karyotype, and measuring DNA content and erythrocyte nuclear volume.The similar and different DNA fragments were PCR amplified and sequenced in triploid, tetraploid hybrids,and their parents, indicating their molecular genetic relationship and genetic markers. In addition, thisstudy also presents results about the phenotypes and feeding habits of polyploid hybrids and discusses theformation mechanism of the polyploid hybrids. It is the first report on the formation of the triploid,tetraploid, and pentaploid hybrids by crossing parents with a different chromosome number in vertebrates.The formation of the polyploid hybrids is potentially interesting in both evolution and fish genetic breeding.
DISTANT crossing is defined as the interspecific orabove-specific hybridization. It is a useful strategy
for hybrid offspring to change genetic composition andphenotype. Interspecific hybridization normally resultsin genome-level alterations, including the occurrenceof triploid hybrids and tetraploids. For example, F1 hy-brids of grass carp (2n ¼ 48) 3 big head carp (2n ¼ 48)were triploids (3n ¼ 72) (Marian and Kraszai 1978;Beck et al. 1984). In reproduction, the triploid hybridswere sterile while the tetraploid hybrids were fertile.In general, the fertile tetraploid hybrids were used togenerate the sterile hybrids, which had some advan-tages such as faster growth and higher anti-disease ability(Liu et al. 2001). Until now, reports on the formation ofliving polyploidy hybrids produced by distant parentswith different chromosome number rarely existed. Tocreate a new type of polyploidy fish, in this study wemade the distant crosses by choosing different parentsthat belonged to a different subfamily and had a differ-ent chromosome number. In the catalog, the red cru-cian carp (RCC) with 100 chromosomes belonged to the
Cyprininae subfamily, and the blunt snout bream(Megalobrama amblycephala) (BSB) with 48 chromosomesbelonged to the Cultrinae subfamily (Yu 1989). In feed-ing habit, the red crucian carp is omnivorous whereas theblunt snout bream is herbivorous. In summary, in chro-mosome number, phenotype, and feeding habit, bothBSB and RCC are quite different. In this study, we suc-cessfully obtained bisexual fertile allotetraploid hybridsand the sterile triploid hybrids of RCC $ 3 BSB #, as wellas the pentaploid hybrids of the tetraploid hybrids $ 3
BSB #. In evolution, the fertile tetraploid hybrids arevery valuable because their fertility guarantees the for-mation of a new type of offspring with a new chromosomenumber. In fish genetic breeding, the sterile triploidhybrids have potential value because they will not disturbthe fish genetic resources in the natural environment.
In this study, we not only investigated the polyploidhybrids on the cytogenetic level, but also analyzed thegenetic relationship between the polyploid hybrids andtheir parents on the molecular level. We designed theprimers on the basis of the sequences of the conservedHMG box of Sox genes and conducted polymerase chainreaction (PCR) using the standard method. Sox genesare those genes that are characterized by a conservedDNA sequence encoding a domain of 80 amino acids.
1Corresponding author: College of Life Sciences, Hunan NormalUniversity, Changsha 410081, People’s Republic of China.E-mail: [email protected]
Genetics 176: 1023–1034 ( June 2007)
This domain, called the high mobility group (HMG), isresponsible for specific DNA sequence binding. Weamplified and sequenced the conserved HMG box ofSox genes in polyploidy hybrids and their parents andfound that the tetraploid and triploid hybrids, respec-tively, could be distinguished from their parents bypresenting different PCR products. Sequence homol-ogy among the fragments amplified from polyploidhybrids and their parents was compared to analyze theirgenetic relationship.
In summary, the formation of the triploid, tetraploid,and pentaploid hybrids is significant in both evolu-tion and fish genetic breeding. This study is the firstreport on the formation of the polyploid hybrids bycrossing parents with a different chromosome numberin vertebrates.
MATERIALS AND METHODS
Animals and crosses: BSB and RCC were obtained from theProtection Station of Polyploidy Fish in Hunan Normal Univer-sity. During the reproductive seasons (from April to June) in2004, 2005, and 2006, 15 mature females and 15 mature malesof both RCC and BSB were chosen as the maternals or pa-ternals, respectively. The crossings were performed by twogroups. In the first group, the RCC was used as the maternal,and the BSB was used as the paternal. In the second group, thematernals and paternals were reversed. The mature eggs andsperm of RCC and BSB were fertilized and the embryosdeveloped in the culture dishes at the water temperature of19�–20�. In each cross, 2000 embryos were taken at randomfor the examination of the fertilization rate (number ofembryos at the stage of gastrula/number of eggs 3 100%)and the hatching rate (number of hatched fry/number ofeggs 3 100%). The hatched fry were transferred to the pondfor further culture.
Because in the cross RCC $ 3 BSB # there existed theoffspring of triploid and tetraploid hybrids while in the reversecross BSB $ 3 RCC # there were no living progeny, in this studywe abbreviate triploid hybrids of RCC $ 3 BSB # as 3nRBhybrids and tetraploid hybrids of RCC $ 3 BSB # as 4nRBhybrids.
At the age of 1 year, no 3nRB and 4nRB hybrids were foundto be fertile. At the age of 2 years, the 3nRB hybrids were stillfound to be sterile. However, the 2-year-old female 4nRBhybrids produced mature eggs whereas the 2-year-old maletetraploid hybrids produced only water-like semen. The female4nRB hybrids produced two sizes of eggs. The diameters of200 larger eggs and 200 smaller eggs were measured. The eggsof 4nRB hybrids were fertilized with the semen of BSB to pro-
duce pentaploid hybrids. We abbreviate pentaploid hybrids of4nRB $ 3 BSB # as 5nRB hybrids. On the other hand, the eggsand semen of 4nRB hybrids were mated to each other to pro-duce the next generation. All crossing procedures and for-mation of the polyploid hybrids are illustrated in Figure 1.
Preparation of chromosome spreads: To determine ploidy,chromosome counts were done on kidney tissue for 10 RCC,10 BSB, 10 3nRB, 10 4nRB, and 10 5nRB hybrids at 1 year ofage. After culture for 1–3 days at the water temperature of 18�–22�, the samples were injected with concanavalin one to threetimes at a dose of 2–8 mg/g body weight. The interval time ofinjection was 12–24 hr. Six hours prior to dissecting, eachsample was injected with colchicine at a dose of 2–4 mg/g bodyweight. The kidney tissue was ground in 0.9% NaCl, followedby hypotonic treatment with 0.075 m KCl at 37� for 40–60 minand then fixed in 3:1 methanol–acetic acid for three changes.Cells were dropped on cold, wet slides and stained for 30 minin 4% Giemsa. The shape and number of chromosomes wereanalyzed under microscope. For each type of fish, 200 meta-phase spreads (20 metaphase spreads in each sample) of chro-mosomes were analyzed. Preparations were examined underan oil lens at a magnification of 3330. Good-quality meta-phase spreads were photographed and used for analysis ofkaryotypes. Lengths of entire chromosomes, long and shortarms, were measured. Chromosomes were classified on thebasis of their long-arm to short-arm ratios according to thereported standards (Levan et al. 1964): values of 1.0–1.7were classified as metacentric (m), 1.7–3.0 as submetacentric(sm), 3.1–7.0 as subtelocentric (st), and 7.1 as telocentric (t)chromosome.
Measurement of DNA content: To measure the DNA con-tent of erythrocytes of 3nRB, 4nRB, and 5nRB hybrids andRCC and BSB, the 1–2 ml of red blood cells were collectedfrom the caudal vein of the above fish into syringes containing�200–400 units of sodium heparin. The blood samples weretreated with the nuclei extraction and DAPI DNA stainingsolution, cystain DNA 1 step (Partec). Then all the sampleswere filtered. A flow cytometer (cell counter analyzer, Partec)was used to measure the DNA content. Under the sameconditions, the DNA content of each sample was measured.To calculate the probabilities of the ratios of the DNA contentof the polyploid hybrids to the sum of that of RCC and BSB, thex2 test with Yate’s correction was used for testing deviationfrom expected ratio values.
Measurement of nuclear volume and appearance of eryth-rocytes: Nuclear micromeasurements of erythrocytes were alsoused to further ascertain ploidy. A 1- to 2-ml blood sample wasdrawn from 3nRB, 4nRB, and 5nRB hybrids and from RCCand BSB. Blood smears were fixed in ethanol and stained withGiemsa solution. To measure nuclear diameters, 20 erythro-cytes from each fish were observed under oil at 3330 using anocular micrometer. Nuclear volume was calculated by (4
3)Q
ab2,where a is the major semi-axis and b is the minor semi-axis of aperfect ellipsoid. Volume may have been slightly overstimulated
Figure 1.—Crossing procedure and formationof the polyploid hybrids.
1024 S. Liu et al.
due to cell flattening. To calculate the probabilities of the ratiosof nuclear volume of the polyploid hybrids to the sum of that ofRCC and BSB, the x2 test with Yate’s correction was used fortesting deviation from expected ratio values. The appearance ofthe erythrocytes’ nuclei was also observed under microscope.
Morphological traits and feeding habit: At 1 year of age, 20RCC, 20 BSB, 20 3nRB, 20 4nRB, and 20 5nRB hybrids, re-spectively, were morphologically examined. The examinedmeasurable traits included the average values of the wholelength, the body length and width, the head length and width,and the tail length and width. The average ratios of wholelength to body length (WL/BL), body length to body width(BL/BW), body length to head length (BL/HL), head lengthto head width (HL/HW), tail length to tail width (TL/TW),and body width to head width (BW/HW) were calculated. Theexamined countable traits included the number of dorsal fins,abdominal fins, anal fins, lateral scales, upper and lower lateralscales.
For both measurable and countable data, we used the soft-ware of SPSS to analyze the covariance of the data betweentwo kinds of fish in 3nRB, 4nRB, and 5nRB hybrids and in RCCand BSB. The feeding habit of 3nRB, 4nRB, and 5nRB hybridswas also investigated.
PCR: The HMG box of Sox genes was amplified by PCRusing the degenerate primers P (1) [59-TGAAGCGACCCATGAA(C/T)G-39] and P (�) [59-AGGTCG(A/G)TACTT(A/G)TA(A/G)T-39]. The genomic DNAs extracted from theblood cells of RCC, BSB, 3nRB, and 4nRB hybrids by routineapproaches (Sambrook et al. 1989) were used as templates.The PCR was performed in a volume of 25 ml with �80 nggenomic DNA, 1.5 mm of MgCl2, 200 mm of each dNTP, 0.3 mm
of each primer, and 0.9 unit of Taq polymerase (Takara). Thecycling program included 35 cycles of 94� for 30 sec, 52� for30 sec, and 72� for 1 min, with a final extension of 7 min at 72�.All PCR products, including three DNA fragments in RCC, twoDNA fragments in BSB, three DNA fragments in 3nRB hybrids,and four DNA fragments in 4nRB hybrids were separated in1.5% agarose gels.
Sequencing: After electrophoresis, all fragments of PCRproducts in 3nRB and 4nRB hybrids and their parents werepurified using a gel extraction kit (Sangon) and ligated intothe pMD-18T vector. The plasmids were amplified in DH5a.The inserted DNA fragments in the pMD-18T vector weresequenced by an automated DNA sequencer (ABI PRISM3730). The sequences were aligned with the correspondingsequences of Sox genes in zebrafish, rainbow trout, medaka,rice field eel, mouse, and the common carp, etc., derived fromthe NCBI database using Jellyfish (2.1) software and namedaccordingly. Sequence homology and variation among thefragments amplified from RCC, BSB, 3nRB, and 4nRB hybridswere analyzed using Clustal W software (http://www.ebi.ac.uk/clustalw/intex.html).
RESULTS
Formation of triploid, tetraploid, and pentaploidhybrids: In the crossings of RCC $ 3 BSB #, we observeda higher fertilization rate (.60%) and hatching rate(.50%) and obtained �5000, 20,000, and 100,000 liv-ing hybrids in 2004, 2005, and 2006, respectively. In thereverse crossing, RCC # 3 BSB $, there was no survival.In 2006, we obtained 25,000 living hybrids with a 67%fertilization rate and a 51% hatching rate in the cross ofthe mature eggs from tetraploid hybrids of RCC $ 3 BSB #
and the white semen from BSB, and only 8 living hybridsgenerated by fertilizing the mature eggs with the water-like semen of the tetraploid (4nRB) hybrids of RCC $ 3
BSB #. All the fish grew well, except the 8 hybrids pro-duced by male–female mating of the 4nRB hybrids, whichwere dead at the age of 4 months due to accidental oxy-gen loss in the water.
Examination of chromosome number and formationof karyotypes: Table 1 presents the distribution of chro-mosome number in BSB, RCC, 3nRB, 4nRB, and 5nRBhybrids. Of all examined samples in BSB, 93% of chro-mosomal metaphases possessed 48 chromosomes (Table1), indicating that they were diploids with 48 chromo-somes (2n¼ 48) (Figure 2A) with the karyotype of 18 m1 22 sm 1 8 st. A pair of the largest submetacentricchromosomes in BSB was observed, which could be usedas the chromosomal marker to identify this species (Fig-ure 2A). Of all examined examples in RCC, 96% ofchromosomal metaphases had 100 chromosomes (Ta-ble 1), indicating that they were diploids with 100 chro-mosomes with the karyotype of 22 m 134 sm 122 st 1 22 t,which was the same as described in our previous study(Liu et al. 2001). In the chromosomes of RCC, no speciallarger submetacentric chromosome was found. In theoffspring of RCC $ 3 BSB # without the barbells in thephenotype, 77% of chromosomal metaphases had 124chromosomes (Table 1, Figure 2B), showing that theywere triploids (3nRB). In these metaphases, only onelarge submetacentric chromosome, similar to the twofound in BSB, was observed, suggesting that the 3nRBhybrids possessed 24 chromosomes from BSB and 100chromosomes from RCC. In the offspring of RCC $ 3
BSB #, which bore the barbells in appearance, 73%of chromosomal metaphases had 148 chromosomes
TABLE 1
Examination of chromosome number in BSB, RCC, 3nRB, 4nRB, and 5nRB hybrids
Distribution of chromosome number
Fish type No. in metaphase ,48 48 ,100 100 ,124 124 ,148 148 ,172 172
BSB 200 14 186RCC 200 8 1923nRB 200 46 1544nRB 200 54 1465nRB 200 18 182
Formation of the Polyploid Hybrids and Its Evolutionary Significance 1025
(Table 1, Figure 2C), indicating that they were tetra-ploids (4nRB). In these metaphases, the two largestsubmetacentric chromosomes, similar to the two inBSB, were found, suggesting that the 4nRB hybridspossessed 48 chromosomes from BSB and 100 chromo-somes from RCC. The karyotype (Figure 3A) of 40 m 1
56 sm 1 30 st 1 22 t in 4nRB hybrids indicated that their40 m chromosomes contained 22 m chromosomes fromRCC and 18 m chromosomes from BSB; their 56 smchromosomes contained 34 sm chromosomes fromRCC and 22 sm chromosomes from BSB; their 30 stchromosomes contained 22 st chromosomes from RCCand 8 st chromosomes from BSB; and their 22t chro-mosomes contained 22 t chromosomes from RCC. Inthe offspring of 4nRB $ 3 BSB #, 91% of chromosomalmetaphases had 172 chromosomes (Table 1, Figure2D), suggesting that they were pentaploids (5nRB). Inthese metaphases, the three largest submetacentric chro-mosomes, similar to those found in BSB (Figure 2A),were observed, indicating that the 5nRB hybrids hadthree sets of chromosomes of BSB and two sets ofchromosomes of RCC. The karyotype of 5nRB hybrids(Figure 3B) showed that they possessed 49 m chromo-
somes (22 m from RCC and 27 m from BSB), 67 smchromosomes (34 sm from RCC and 33 sm from BSB),34 st chromosomes (22 st from RCC and 12 st fromBSB), and 22 t chromosomes (22 t from RCC).
Measurement of DNA content: We used the sum ofthe DNA content of RCC and BSB as the controls.The results of the distribution of DNA content of all thesamples are shown in Table 2 and Figure 4, A–E. Themean DNA content of 3nRB hybrids was equal (P .
0.01) to the sum of that of RCC and half of BSB,suggesting that 3nRB hybrids had two sets of chromo-somes from RCC and one set of chromosomes from
Figure 3.—Karyotypes of 4nRB and 5nRB hybrids. (A) Thekaryotype of 4nRB hybrids, 40 m 1 56 sm 1 30 st 1 22t, whichconsists of two sets of chromosomes from RCC and two setsof chromosomes from BSB. The arrows indicate a pair of thelargest submetacentric chromosomes similar to the two inBSB. (B) The karyotype of 5nRB hybrids contained two setsof chromosomes from RCC and three sets of chromosomesfrom BSB. The arrows indicate the three largest submetacen-tric chromosomes similar to the two in BSB. Bar in A and B,4 mm.
Figure 2.—Chromosome spreads at metaphase in BSB,3nRB, 4nRB, and 5nRB hybrids. (A) The 48 chromosomesof BSB, in which a pair of the largest submetacentric chromo-somes (arrows) are indicated. (B) The 124 chromosomes of3nRB hybrids of RCC 3 BSB in which a piece of the largestsubmetacentric chromosomes (arrow) is indicated. (C) The148 chromosomes of 4nRB hybrids of RCC 3 BSB in whicha pair of the largest submetacentric chromosomes (arrows)are indicated. (D) The 172 chromosomes of 5nRB hybridsproduced by crossing the females of 4nRB hybrids with themales of BSB, in which the three largest submetacentric chro-mosomes (arrows) are indicated. Bar in A–D, 4 mm.
1026 S. Liu et al.
BSB. The mean DNA content of 4nRB hybrids was equal(P . 0.01) to the sum of that of RCC and BSB, indicat-ing that 4nRB hybrids had two sets of chromosomesfrom RCC and two sets of chromosomes from BSB. Themean DNA content of 5nRB hybrids was equal (P . 0.01)to the sum of that of RCC and one and half of BSB, show-ing that 5nRB hybrids had two sets of chromosomes fromRCC and three sets of chromosomes from BSB.
Measurement of the nuclear volume of erythrocytes:The results of the measurements of the mean erythro-cyte nuclear volume for BSB, RCC, 3nRB, 4nRB, and5nRB hybrids are shown in Table 3 and Figure 5, A–F.With the increase of the polyploid level, the meanerythrocyte nuclear volume regularly increased among3nRB, 4nRB, and 5nRB hybrids. The mean erythrocytenuclear volume ratio of the 3nRB hybrid to the sumof that of RCC and half of BSB was not significantlydifferent (P . 0.05) from the 1:1 ratio, suggesting that3nRB hybrids were triploids. The mean erythrocyte nu-clear volume ratio of the 4nRB hybrid to the sum of thatof RCC and BSB was not significantly different (P .
0.05) from the 1:1 ratio, indicating that 4nRB hybridswere tetraploids. The mean erythrocyte nuclear volumeratio of the 5nRB hybrid to the sum of that of RCC andone and one-half of BSB was not significantly different(P . 0.05) from the 1:1 ratio, indicating that 5nRBhybrids were pentaploids.
Nuclear traits of erythrocytes: Figure 5 showed thenuclear traits of erythrocytes in BSB, RCC, 3nRB, 4nRB,and 5nRB hybrids. The size of the nuclei of erythrocytesin 3nRB, 4nRB, and 5nRB hybrids was larger than thatof BSB or RCC. With the increase of the ploidy level,the size of the nuclei of erythrocytes regularly enlarged
TABLE 2
Mean DNA content in RCC, BSB, 3nRB, 4nRB, and5nRB hybrids
Fishtype
MeanDNA
content
Ratio
Observed Expected
BSB 42.92RCC 52.193nRB 72.63 3nRB/(RCC 1 0.5 BSB) ¼ 0.99a 14nRB 97.05 4nRB/(RCC 1 BSB) ¼ 1.02a 15nRB 117.46 5nRB/(RCC 1 1.5 BSB) ¼ 1.01a 1
a The observed ratio was not significantly different (P .0.05) from the expected ratio.
Figure 4.—Cytometric histograms of DNA fluorescence forBSB, RCC, and polyploid hybrids. (A) The mean DNA con-tent of RCC (peak 1: 52.19). (B) The mean DNA contentof RCC (peak 1: 42.92). (C) The mean DNA content of3nRB hybrid (peak 3: 72.63). (D) The mean DNA contentof 4nRB hybrid (peak 2: 97.05). (E) The mean DNA contentof 5nRB hybrid (peak 1: 117.46).
Formation of the Polyploid Hybrids and Its Evolutionary Significance 1027
among 3nRB, 4nRB, and 5nRB hybrids. In addition tothe difference in nuclear size, there also existed thedifference in nuclear appearance between the poly-ploidy hybrids and their parents. For example, only thenormal erythrocytes with one nucleus, but no unusualerythrocytes with two nuclei, were observed in both RCC
and BSB (Figure 5, A and B). However, unusual eryth-rocytes with two nuclei were found in 3nRB, 4nRB, and5nRB hybrids (Figure 5, C–F). The unusual erythrocyteswith three nuclei were also observed in 5nRB hybrids(Figure 5F). The two-nucleus erythrocytes accountedfor 3.4, 9.4, and 21.8% in the 3nRB, 4nRB, and 5nRBhybrids, respectively, indicating that, with the increaseof ploidy level, the percentage of unusual erythrocyteswith two nuclei increased. The unusual erythrocyteswith three nuclei in 5nRB hybrids accounted for 0.7%.
Morphological traits and feeding habit: Tables 4 and5 present the examined measurable traits and countabletraits in 3nRB, 4nRB, 5nRB hybrids and in RCC andBSB. For the measurable traits between 3nRB hybridsand BSB, all ratios were significantly different. Between3nRB hybrids and RCC, apart from the ratio of BL/HL,which was not significantly different (P . 0.01), otherratios were significantly different. Between 4nRB hybridsand BSS, all ratios were significantly different. Between4nRB hybrids and RCC, apart from the ratios of BL/BWand HL/HW, which were not significantly different (P .
0.01), other ratios were significantly different. Between5nRB hybrids and BSB, apart from the ratio of BL/BW,which was not significantly different (P . 0.01), otherratios were significantly different. Between 5nRB hybridsand RCC, all ratios were significantly different. On theother hand, between 3nRB and 4nRB hybrids, apart fromthe ratio of HL/HW, which was not significantly different(P . 0.01), other ratios were significantly different.Between 3nRB and 5nRB hybrids, all ratios were signif-icantly different. Between 4nRB and 5nRB hybrids, allratios were significantly different.
For the countable traits between 3nRB hybrids andBSB, all data were significantly different. Between 3nRBhybrids and RCC, apart from the number of lateralscales, dorsal fins, and abdominal fins, which were notsignificantly different (P . 0.01), other data were sig-nificantly different. Between 4nRB hybrids and BSS, alldata were significantly different. Between 4nRB hybridsand RCC, apart from the number of dorsal fins andabdominal fins, which were not significantly different(P . 0.01), other data were significantly different. Be-tween 5nRB hybrids and BSB, all data were significantly
TABLE 3
Mean erythrocyte nuclear volume measurements for BSB, RCC, 3nRB, 4nRB, and 5nRB hybrids
Volume ratio
Fish type Major axis (mm) Minor axis (mm) Volume (mm3) Observed Expected
BSB 5.07 6 0.47 2.44 6 0.80 15.76 6 1.92RCC 4.99 6 0.27 2.95 6 0.20 22.71 6 2.423nRB 6.96 6 0.70 3.22 6 0.33 37.64 6 6.32 3nRB/(RCC 1 0.5 BSB) ¼ 1.23a 14nRB 7.46 6 0.56 3.27 6 0.34 42.01 6 8.31 4nRB/(RCC 1 BSB) ¼ 1.09a 15nRB 7.71 6 0.81 3.35 6 0.23 45.56 6 8.18 5nRB/(RCC 1 1.5 BSB) ¼ 0.98a 1
a The observed ratio was not significantly different (P . 0.05) from the expected ratio.
Figure 5.—Erythrocytes of the RCC, BSB, and polyploid hy-brids. (A) Normal erythrocytes with one nucleus in RCC. (B)Normal erythrocytes with one nucleus in BSB. (C) Normal er-ythrocytes with one nucleus and unusual erythrocytes with twonuclei (arrows) in a 3nRB hybrid. (D) Normal erythrocyteswith one nucleus and unusual erythrocytes with two nuclei(arrows) in a 4nRB hybrid. (E) Normal erythrocytes withone nucleus and unusual erythrocytes with two nuclei (ar-rows) in a 5nRB hybrid. (F) Normal erythrocytes with one nu-cleus and unusual erythrocytes with two nuclei or three nuclei(arrow) in a 5nRB hybrid. Bar in A–F, 0.01 mm.
1028 S. Liu et al.
different. Between 5nRB hybrids and RCC, apart fromthe number of dorsal fins, which was not significantlydifferent (P . 0.01), other data were significantly dif-ferent. On the other hand, between 3nRB and 4nRBhybrids, apart from the number of lateral scales andlower lateral scales, which were significantly different,other data were not significantly different (P . 0.01).Between 3nRB and 5nRB hybrids, apart from the numberof lower lateral scales and buttock fins, which were notsignificantly different (P . 0.01), other data were sig-nificantly different. Between 4nRB and 5nRB hybrids,apart from the number of dorsal fins and anal fins,which were not significantly different (P . 0.01), otherdata were significantly different.
In morphological traits, the 3nRB (Figure 6A), 4nRB(Figure 6B), and 5nRB hybrids (Figure 6C) showedobvious differences from RCC and BSB. Tables 4 and 5indicated that most of the morphological data in 3nRB,4nRB, and 5nRB hybrids were significantly differentfrom those in both RCC and BSB, suggesting that thevariation traits occurred in the polyploid hybrid off-spring. For example, the ratio of the body length towidth length in 3nRB hybrids was 1.67 whereas it was2.18 in RCC and 2.37 in BSB. The most interestingvariation trait, the presence of the barbells, occurred in4nRB hybrids while there was no barbell in theirparents, RCC and BSB. On the other hand, there weresome significantly different traits between 3nRB and4nRB hybrids. For example, the 4nRB hybrids had a pairof barbells on the mouth, and their lateral scale num-ber was 31–33 whereas the 3nRB hybrids did not presentany barbell on the mouth, and their lateral scale num-ber was 28–30. In the offspring of RCC $ 3 BSB #, 4nRBhybrids accounted for 77% and 3nRB hybrids accountedfor 23%. Furthermore, some significantly different traitsbetween 4nRB and 5nRB hybrids also existed. For exam-ple, the 5nRB hybrids did not possess any barbell on themouth, and their lateral scale number was 34–36. For thefeeding habit, like BSB, 3nRB, 4nRB, and 5nRB hybridswere herbivorous.
Sizes of eggs produced by 4nRB hybrids: The femalesof 4nRB hybrids produced two sizes of eggs. The averagediameter of the larger eggs was 0.20 cm, which accountedfor 95%, and the average diameter of the smaller eggswas 0.17 cm, which accounted for 5% (Figure 7).
Sterility of 3nRB hybrids and fertility of 4nRBhybrids: During the reproductive season, no egg orsemen could be squeezed out from 3nRB hybrids at theages of 1 and 2 years. No egg or semen could be strippedout from 4nRB hybrids at the age of 1 year. However,the mature eggs (Figure 7) or water-like semen weresqueezed out from the females and males of the 4nRBhybrids, respectively, at the age of 2 years. The eggs of4nRB hybrids were mated with the semen of the BSB togenerate �25,000 living pentaploid hybrids. However,after the water-like semen of 4nRB hybrids was fertilizedwith the eggs of 4nRB hybrids, only 8 living hybrids wereproduced, suggesting that, at the age of 2 years, thefemale 4nRB hybrids had stronger fertility than that ofthe male 4nRB hybrids.
DNA fragments based on the primers of the HMGbox of Sox genes: The PCR results based on the primersof HMG of Sox genes and the sequencing results showedthat there were three DNA fragments (215, 617, and1958 bp) in RCC, two DNA fragments (215 and 712 bp)in BSB, three fragments (215, 616, and 1955 bp) in3nRB hybrids, and four fragments (213, 616, 918, and1959 bp) in 4nRB hybrids (Figure 8). All the sequencesof the PCR products have been submitted to GenBankand their accession numbers are in Table 6. By compar-ing the sequences, we confirmed that a 215-bp DNAfragment that existed in RCC, BSB, and 3nBB hybridsbelonged to Sox 11 whereas a 213-bp DNA fragment in4nRB hybrids represented the Sox 1 gene. The 616- and617-bp DNA fragments existed in RCC, 4nRB, and 3nRBhybrids and the 712-bp fragment in BSB represented Sox9a gene. The 918-bp fragment in 4nRB hybrids was newlyformed and represented Sox 9b gene. The 1955-, 1958-,and 1959-bp DNA fragments in the 3nRB hybrid, RCC,and the 4nRB hybrid were from the Sox 4 gene (Table 7).
Table 7 indicates the percentage of nucleotide simi-larities of separate regions of the DNA fragments pro-duced by PCR with the primers of the HMG box ofSox genes in 3nRB and 4nRB and their parents, RCCand BSB. In the sequences of the 215-bp DNA frag-ments in RCC, BSB, and 3nRB hybrids and the 213-bpDNA fragment in 4nRB hybrids, 84% identity betweenRCC and BSB, 86% identity between BSB and 3nRB hy-brids, and 96% identity between RCC and 3nRB hybridswere found, indicating that the sequence of this DNA
TABLE 4
Comparison of the measurable traits between the hybrid offspring and RCC and BSB
Fish type WL/BL BL/BW BL/HL HL/HW TL/TW BW/HW
3nRB 1.31 6 0.03 1.67 6 0.03 3.70 6 0.03 1.11 6 0.04 0.71 6 0.04 2.31 6 0.044nRB 1.18 6 0.02 2.18 6 0.02 3.83 6 0.03 1.08 6 0.04 0.75 6 0.04 1.92 6 0.025nBRB 1.25 6 0.03 2.40 6 0.06 3.35 6 0.03 1.25 6 0.03 1.11 6 0.03 1.66 6 0.03RCC 1.22 6 0.02 2.18 6 0.02 3.72 6 0.03 1.07 6 0.03 0.82 6 0.03 1.84 6 0.03
BSB 1.19 6 0.03 2.37 6 0.03 4.75 6 0.04 1.14 6 0.03 1.08 6 0.04 2.09 6 0.04
Formation of the Polyploid Hybrids and Its Evolutionary Significance 1029
fragment in 3nRB hybrids was highly homologous tothat of RCC (Table 7). The sequence identity between4nRB and BSB (69%) is higher than that between 4nRBand RCC (65%), indicating that the sequence of thisDNA fragment in 4nRB hybrids was similar to that ofBSB (Table 7).
As for the sequences of the DNA fragments of 617 bpin RCC, 712 bp in BSB, and 616 bp in 3nRB and 4nRBhybrids, a 72% similarity between RCC and BSB, 73%similarity between BSB and 3nRB hybrids, and 99%similarity between RCC and 3nRB hybrids existed,indicating that the sequence of this DNA fragmentin 3nRB hybrids was similar to that of RCC (Table 7).A 72% similarity between BSB and 4nRB hybridsand a 99% similarity between RCC and 4nRB existed,
TA
BL
E5
Co
mp
aris
on
of
the
cou
nta
ble
trai
tsb
etw
een
the
po
lyp
loid
san
dth
eir
par
ents
Fis
hty
pe
No
.o
fla
tera
lsc
ales
No
.o
fu
pp
erla
tera
lsc
ales
No
.o
flo
wer
late
ral
scal
esN
o.
of
do
rsal
fin
sN
o.
of
abd
om
inal
fin
sN
o.
of
anal
fin
s
3nR
B29
.45
60.
51(2
9–30
)6.
606
0.50
(6–7
)7.
706
0.47
(7–8
)II
I1
18.1
06
0.79
(III
117
–19)
8.35
60.
49(8
–9)
III
16.
306
0.47
(III
16–
7)4n
RB
31.6
56
0.49
(31–
32)
6.55
60.
51(6
–7)
6.45
60.
51(6
–7)
III
118
.70
60.
98(I
II1
17–2
0)8.
606
0.50
(8–9
)II
I1
6.40
60.
68(I
II1
5–7)
5nR
B35
.20
60.
77(3
4–36
)7.
606
0.50
(7–8
)7.
706
0.47
(7–8
)II
I1
19.1
56
1.23
(III
117
–21)
10.7
56
1.07
(9–1
2)II
I1
6.55
60.
51(I
II1
6–7)
RC
C29
.20
60.
70(2
8–30
)5.
606
0.50
(5–6
)5.
706
0.47
(5–6
)II
I1
18.6
56
0.49
(III
118
–19)
8.55
60.
51(8
–9)
III
15.
656
0.49
(III
15–
6)B
SB50
.90
60.
91(5
2–49
)9.
656
0.49
(9–1
0)10
.05
60.
69(9
–11)
III
18.
656
0.49
(III
18–
9)9.
106
0.55
(8–1
0)II
I1
25.8
56
0.59
(III
125
–27)
Figure 6.—Appearance of 3nRB, 4nRB, and 5nRB hybrids.(A) The 3nRB hybrid did not present any barbell on themouth. Its ratio of body length to body width was 1.67, andits lateral scale number was 28–30. (B) The 4nRB hybridhad a pair of barbells on the mouth. Its ratio of body lengthto body width was 2.19, and its lateral scale number was 31–33.(C) The 5nRB hybrid did not possess any barbell on themouth. Its ratio of body length to body width was 2.42, andits lateral scale number was 34–36. Bar in A–C, 1 cm.
1030 S. Liu et al.
suggesting that the sequence of this DNA fragment in4nRB hybrids was similar to that of RCC (Table 7).
As for the sequences of the DNA fragment of 1958 bpin RCC, 1955 bp in 3nRB hybrids, and 1959 bp in 4nRBhybrids, 99% similarity between RCC and 3nRB hybridsand 100% similarity between RCC and 4nRB hy-brids existed, suggesting that the sequence in this DNAregion of both 3nRB and 4nRB hybrids was very closeto that of RCC. In the 4nRB hybrids, the 918-bp DNAfragment was newly formed, and the sequence wasdeposited in GenBank under accession no. EF370033(Table 6).
DISCUSSION
The distant crossing is an important and effectivemeans of increasing the possibility of variation in thehybrid progeny. With this method, it is possible that thewhole haploid genome of one species could be trans-lated into another species. The distant genome trans-lation not only will result in changes of the ploidy levelsin chromosomes, but also will lead to changes in phe-notypes. Furthermore, it will lead to changes in re-productive behavior, for example, the formation of thesterile triploid hybrids.
In this study, the ploidy levels of the crossing offspr-ing were confirmed by counting chromosomal number(Figure 2), forming chromosomal karyotype (Figure 3),and measuring DNA content (Figure 4, Table 2) and theerythrocyte nuclear volume (Figure 5, Table 3). All ofthe above results were in agreement that 3nRB, 4nRB,and 5nRB hybrids were triploids, tetraploids, and pen-taploids, respectively. Counting the chromosome num-ber is a direct and accurate method of determining the
ploidy of the samples. Using the instrument of the flowcytometer to examine the DNA content of samples is arapid and simple as well as accurate method. Theerythrocyte nuclear measurement is also a simple anduseful method for determining the ploidy.
The 3nRB, 4nRB, and 5nRB hybrids were differentfrom RCC and BSB in chromosome number, largestchromosome marker, DNA content, nuclear volume oferythrocytes, and morphological traits, indicating thatthey were from hybridization rather than from gyno-genesis or androgenesis. The feeding habit in 3nRB,4nRB, and 5nRB hybrids, like that of BSB, was herbiv-orous, which also suggested that the polyploid offspringhad chromosomes from BSB and that they were notfrom gynogenesis. Their herbivorous feeding habitindicated that they had wider feeding resources, whichwas very beneficial in aquaculture.
At the chromosome level, the chromosomal numberand the special chromosome(s) can be used as themarker(s) for analyzing the changes of the ploidy levelsand the origin of the chromosomes in the hybrid prog-eny. In this study, the diploid BSB had 48 chromosomeswith one pair of the largest submetacentric chromo-somes (Figure 2A), which could be used as markers foridentifying BSB from RRC possessing 100 chromosomeswithout an obviously large submetacentric chromosome.With 124 chromosomes and the presence of one sub-metacentric largest chromosome similar to the two in BSB,the 3nRB hybrids were considered to have two setsof chromosomes from RCC and one set of chro-mosomes from BSB (Figure 2B). On the basis of thechromosomal number (148) and one pair of the largestchromosomes similar to those in BSB, we concluded
Figure 8.—Amplified DNA fragments resulting from PCRbased on the primers of HMG of Sox genes in BSB, RCC,3nRB, and 4nRB hybrids. M, DNA ladder markers with an in-crease of 200 bp. In lane1, there are three DNA fragments(215, 617, and 1958 bp) in RCC; in lane 2, two DNA fragments(215 and 712 bp) in BSB; in lane 3, four fragments (213, 616,918, and 1959 bp) in 4nRB hybrids; in lane 4, three fragments(215, 616, and 1955 bp) in 3nRB hybrids.
Figure 7.—Eggs produced by 4nRB hybrids. There are twosizes of eggs produced by 4nRB hybrids. The larger eggs, eachwith a diameter of 0.2 cm, were considered 4n eggs, and thesmaller eggs (arrows), each with the diameter of 0.17 cm, wereconsidered 2n eggs because their size was the same as that ofthe diploid eggs produced by the tetraploid hybrids of redcrucian carp $ 3 common carp # as described in our previousstudy (Liu et al. 2001). Bar, 0.4 cm.
Formation of the Polyploid Hybrids and Its Evolutionary Significance 1031
that 4nRB hybrids contained two sets of chromosomesfrom BSB and two sets of chromosomes from RCC(Figure 2C; Figure 3A). With 172 chromosomes and thethree largest submetacentric chromosomes similar tothose in BSB (Figure 2D), the 5nRB hybrids proved tohave two sets of chromosomes from RCC and three setsof chromosomes from BSB (Figure 3B).
At the DNA fragment level, the RCC presented threedifferent DNA fragments whereas the BSB presentedonly two different DNA fragments (Figure 8), indicatingthat these two different species possessed different DNAcharacteristics. As the hybrid offspring of RCC and BSB,the 4nRB hybrids not only shared all the DNA fragmentsfound in both RCC and BSB, but also possessed thenewly formed 918-bp DNA fragment, indicating that the4nRB hybrids not only changed in chromosomal ploidylevels, but also changed in DNA sequences. The 4nRBhybrids shared the common and different fragmentsfound in RCC and BSB (Figure 8), providing the geneticevidence that the 4nRB hybrids came from RCC andBSB, but were different from RCC and BSB. On theother hand, the specific 712-bp DNA fragment (Figure8) of BSB was absent in 3nRB and 4nRB hybrids,providing further genetic evidence that the 3nRB and4nRB hybrids were different from BSB (Figure 8). Table7 indicated that most of the sequences in the amplifiedDNA fragments were similar to those of RCC, consistentwith the result that 3nRB and 4nRB hybrids had morechromosomes from RCC than from BSB.
The presence of the unusual erythrocytes with twonuclei could be used as one of the cytological markers todistinguish the polyploid hybrids from their diploid
parents (Figure 5). With the increase of the ploidy level,the percentage of the unusual erythrocytes with twonuclei increased, suggesting that the higher the ploidylevel, the higher the possibility of the presence of theunusual erythrocytes in polyploid hybrids. In our pre-vious study (Liu et al. 2003), 32.4% of unusual eryth-rocytes with two nuclei were also found in the tetraploidhybrids of red crucian carp $ 3 common carp #. It wasconcluded that the higher DNA content made it easierfor the nuclei to divide.
In this study, only 3nRB and 4nRB hybrids, but nodiploid hybrids, were found to survive in the cross ofRCC $ 3 BSB #. However, we still observed diploidembryos with 74 chromosomes by examining the chro-mosomal number (data not shown), but no survivaldiploid hybrid. This means that the diploid hybridscould not survive probably due to the larger differencein chromosomal number between the maternal RCC(100) and the paternal BSB (48), which inhibits theembryos from developing into living diploid hybrids.The tetraploidization and triploidization were helpfulto the hybrids in overcoming the obstacle to developinginto living hybrids. The 3nRB hybrids resulted from theretention of the second polar body of the fertilized eggsand the 4nRB hybrids resulted from the inhibition ofthe first cleavage of the fertilized eggs.
In general, animals generate the half-reduced gam-etes through meiosis. For example, most of diploid fishproduce haploid gametes. However, there were a fewreports about the production of diploid gametes gen-erated by diploid hybrids. For example, the diploidfemale hybrids of crucian carp (Carassius auratus gibelio)
TABLE 6
GenBank accession numbers of the sequences in 3nRB and 4nRB and of their parents, RCC and BSB
GenBank accession no. of the sequences
DNA fragments RCC BSB 3nRB 4nRB
213 or 215 bp EF219273 EF219276 EF370039 EF370036616, 617, or 712 bp EF219274 EF219277 EF370038 EF370035918 bp Absent Absent Absent EF3700341955, 1958, or 1959 bp EF219275 Absent EF370037 EF370033
TABLE 7
Nucleotide similarities of separate regions of the DNA fragments produced by PCR with the primers of the HMG box ofSox genes in 3nRB, 4nRB, and their parents, RCC and BSB
DNA fragments RCC and BSB RCC and 3nRB BSB and 3nRB RCC and 4nRB BSB and 4nRB
213 or 215 bp 84 96 86 65 69616, 617, or 712 bp 72 99 73 99 72918 bp Absent in both Absent in both Absent in both Absent in RCC Absent in BSB1955, 1958, or 1959 bp Absent in BSB 99 Absent in BSB 100
Numbers are percentages.
1032 S. Liu et al.
($) 3 common carp (Cyprinus carpio L.) (#) (interspecificcrossing) produced unreduced diploid eggs (Cherfas
et al. 1994). The diploid female hybrid of medaka be-tween Oryzias latipes and O. curvinotus also spawneddiploid eggs (Shimizu et al. 2000). Some diploid malehybrids of Rutilus alburnoides (Teleostei, Cyprinidae)produced unreduced diploid sperm (Alves et al. 1999).In our previous study (Liu et al. 2001), we made thedistant crossing of RCC $ 3 common carp # and foundthat both F1 and F2 hybrids were diploids with 100chromosomes. However, some females and males of F2
hybrids were able to generate diploid eggs and diploidsperm, respectively, which fertilized each other to formthe bisexual tetraploid hybrids in F3. Until now thepopulation of F3–F16 was proved to be tetraploidhybrids, indicating that their tetraploidy could be stablyinherited from one generation to another (Liu et al.2001, 2003, 2004; Sun et al. 2003; Guo et al. 2006). Inthat study, the formation of unreduced diploid gameteswas due to the premeiotic endoreduplication of oogo-nia or spermatogonia of females or males of the F2 (Liu
et al. 2001). The fertile males and females of F2 hybridsreached maturity when they were 2 years old. In thisstudy, the 2-year-old female 4nRB hybrids reachedmaturity and produced many mature eggs whereas themale 4nRB hybrids produced only the water-like semenrather than the normal white semen as the male RCC orBSB did, indicating that the females of 4nRB hybridsmatured early and better than the males of 4nRB hybrids.In contrast, no egg and semen were squeezed fromeither 1- or 2-year-old 3nRB hybrids. Interestingly, the fe-males of 4nRB hybrids generated two sizes of eggs. Thelarger eggs, each with a diameter of 0.2 cm (Figure 7),were considered as tetraploid eggs because the pentaploidhybrids were found in 4nRB $ 3 BSB #. The smallereggs, each with a diameter of 0.17 cm, were consideredas diploid eggs because their size was similar to that ofthe diploid eggs of the tetraploid hybrids of RCC $ 3
common carp # (Liu et al. 2001). In the progeny of4nRB $ 3 BSB #, we found only the pentaploid hybrids,but no triploid hybrids. It is probably because 95% of theeggs produced by 4nRB hybrids were tetraploid eggs andonly 5% were diploid eggs. We thought the formationof the unreduced tetraploid eggs was also due to thepremeiotic endoreduplication of oogonia of 4nRB hy-brids. The similarity between this study and our previousstudy was the formation of unreduced gametes producedby the distant crossing hybrids. The difference betweenthis study and our previous study is that, in our previousstudy, the unreduced diploid gametes resulted fromthe diploid hybrids and, in this study, the unreducedtetraploid eggs resulted from the tetraploid hybrids. Onthe other hand, in our previous study, the tetraploids’maternal and paternal had the same chromosomalnumber (2n¼ 100) and belonged to the same subfamilyand different genus in the catalog. In this study, the RCCand BSB had different chromosomal numbers (2n ¼
100 and 2n¼ 48) and belonged to different subfamiliesin the catalog. The formation of 4nRB hybrids resultedfrom the inhibition of the first cleavage of the fertilizedeggs, but not from the fertilization of the diploid eggsand sperm produced by F2 hybrids of RCC $ 3 commoncarp #. Both the fertile males and females of the 4nRBhybrids were found when they were 2 years old andmany pentaploid hybrids were found in the progeny of4nRB $ 3 BSB #, suggesting that it is very possible to ob-tain the octoploid hybrids with 296 chromosomes if weobtain more fertile male 4nRB hybrids in 2007.
The 4nRB hybrids could be identified from 3nRBhybrids by traits such as the presence of the barbells, adifferent chromosomal number, a different DNA con-tent, and different molecular markers of DNA fragmentsof Sox genes. The 3nRB hybrids presented a similarnumber and size of DNA fragments to their maternalRCC, suggesting that two sets of chromosomes of RCCplayed a major role in inheritance material, whichresulted in the 3nRB hybrid phenotype being similar tothat of RCC. The appearance and the DNA markers alsosuggested that 4nRB hybrids showed more hybrid traitsthan 3nRB hybrids displayed. Furthermore, the 213-bpDNA fragment of 4nRB hybrids belonged to Sox 1, ratherthan to Sox 11, as happened in RCC, BSB, and 3nRBhybrids, suggesting that more DNA variation occurred in4nRB hybrids than in 3nRB hybrids.
In this study, compared with their parents, obviousdifferences occurred in 3nRB, 4nRB, and 5nRB hybrids inthe measurableand countable data, indicating the distanthybridizing effect. The most obvious variation trait wasthe presence of the barbells in 4nRB hybrids, which wereabsent in their parents, RCC and BSB. The changes in thephenotype traits resulted from the genomic changes ofthe chromosomes, DNA fragments, or sequences.
Also, in this study, we found that both females andmales of the 4nRB hybrid were fertile. Therefore, thechanges in chromosomes and sequences could be in-herited from one generation to another. In evolution,the fertility of the 4nRB hybrid makes it available tocreate a new type of offspring with a changed chromosomenumber. For example, we have created the 5n hybrids,and it is very possible to create the octoploid hybrids bycrossing the females and males of the 4nRB hybrids. Intheory, the octoploid hybrids should also be fertile. Inaquaculture, the 3nRB hybrids combined the biologicalcharacteristics between RCC and BSB and they will be-come a new fish product in the market. Their sterilityensures that they will not disturb the natural fish geneticresources. In a word, the bisexual fertile 4nRB hybridsand sterile 3nRB hybrids are significant in both evolu-tion and aquaculture.
This research was supported by grants from the National NaturalScience Foundation of China (nos. 30330480 and 30571444), theProgram for Changjiang Scholars and Innovative Research Team inUniversity (no. IRT0445), the Key Item of Science and Technology ofHunan Province (no. 2006NK2008), and the Training Project of
Formation of the Polyploid Hybrids and Its Evolutionary Significance 1033
Excellent Young Researchers of the State Education Ministry of China(no. 200248).
LITERATURE CITED
Alves, M. J., M. M. Coelho, M. I. Prospero and M. J. Collares-Pereira, 1999 Production of fertile unreduced sperm by hybridmales of the Rutilus alburoides complex (Teleost, Cyprinidae):an alternative route to genome tetraploidization in unisexuals.Genetics 151: 227–283.
Beck, M. L., C. J. Biggers and C. J. Barker, 1984 Chromosomal andelectrophoretic analyses of hybrids between grass carp and big-head carp. Copeia 81: 337–342.
Cherfas, N. B., B. L. Gomelsky, O. V. Emelyanova and A. V.Recoubratsky, 1994 Induced diploid gynogenesis and poly-ploidy in crucian carp, Carassius auratus gibelio (Bloch) 3 com-mon carp, Cyprinus carpio L., hybrids. Aquac. Fish. Manage. 25:943–945.
Guo, X. H., S. J. Liu and Y. Liu, 2006 Evidence for recombination of mi-tochondrial DNA in triploid crucian carp. Genetics 172: 1745–1749.
Levan, A., K. Fredga and A. Sandburg, 1964 Nomenclature forcentromeric positions on chromosomes. Hereditas 52: 201–220.
Liu, S. J., Y. Liu, G. J. Zhou, X. J. Zhang, C. Luo et al., 2001 Theformation of tetraploid stocks of red crucian carp 3 common
carp hybrids as an effect of interspecific hybridization. Aquacul-ture 192: 172–186.
Liu, S. J, Y. D. Sun, G. J. Zhou, X. J. Zhang and Y. Liu, 2003 Theultra-microstructure of the mature gonads and erythrocytes in al-lotetraploids. Prog. Nat. Sci. 13: 194–197 (in Chinese).
Liu, S. J, Y. D. Sun, C. Zhang, K. K. Luo and Y. Liu, 2004 Productionof gynogenetic progeny from allotetraploid hybrids red cruciancarp 3 common carp. Aquaculture 236: 193–200.
Marian, T., and Z. Kraszai, 1978 Karyological investigation onCtenopharyngodon idella and Hypophthalmichthys nobilis and theircrossing-breeding. Aquac. Hungarica 1: 44–50.
Sambrook, J., and D. W. Russell, 1989 Molecular Cloning: A Labora-tory Manual, Ed. 3. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, NY.
Shimizu, Y., N. Shibata, M. Sakaizumi and M. Yamashite, 2000 Pro-duction of diploid eggs through premeiotic endomitosis in thehybrid medaka between Oryzias latipes and O. ourvinatus. Zool. Sci.17: 951–958.
Sun, Y. D., S. J. Liu, C. Zhang, J. Z. Li, W. R. Wang et al., 2003 Thechromosome number and gonadal structure of F9–F11 allotetra-ploid crucian carp. Chin. J. Genet. 30: 37–41 (in Chinese).
Yu, X. J., 1989 China Freshwater Fisheries Chromosome, pp. 44–75. Sci-ence Publishing House, Beijing (in Chinese).
Communicating editor: J. A. Birchler
1034 S. Liu et al.