|PD/A CRSP Eighteenth Annual Technical Report|
Table of Contents
Cite as: [Author(s), 2001. Title.] In: A. Gupta, K. McElwee, D. Burke, J. Burright, X. Cummings, and H. Egna (Editors), Eighteenth Annual Technical Report. Pond Dynamics/Aquaculture CRSP, Oregon State University, Corvallis, Oregon, [pp. ___.]
The variability in the factors affecting the sex ratios of Nile tilapia (Oreochromis niloticus) was studied using pair spawns from the Egypt, Ghana, and Ivory Coast strains. Sex ratios from 129 progeny groups were determined by microscopic inspection of the gonads. Of the 12,450 progeny sexed, 54.14% were males; this differed significantly from a 1:1 sex ratio (P < 0.001). Sex ratios from the 129 progeny groups produced a normal distribution (P > 0.45) but did not reflect a binomial distribution (P < 0.01) as would be expected from a simple monofactorial sex determination process. Therefore, other factors must account for the variation observed in sex ratios. Single spawn sex ratios ranged widely from 16 to 100% male. Chi-square tests revealed weak correlations between strain and progeny gender, individual parent and progeny gender, and male-female parent combinations and progeny gender. Sixty-five percent (11 of 17) of the repeat spawns by the same pairings produced sex ratios that extended beyond a range of 10%. The continuous range of sex ratios within the normal distribution suggests the presence of several minor sex-modifying factors in O. niloticus and raises questions as to whether sex inheritance in tilapia is predictable enough for a YY breeding program to be practical.
The nature of sex determination in tilapia is not clear. Clemens and Inslee (1968) and Chen (1969) proposed that in O. mossambicus, the female was the homogametic sex (XX) and the male heterogametic (XY). However, in both studies the authors obtained some sex ratios that could not be explained by a simple Mendelian XX:XY sex determination. Jalabert et al. (1971) studied the sex determination of O. niloticus and O. macrochirus by evaluating the sex ratios of hybrid progeny and concluded that O. niloticus had a basic XX:XY sex determination but that the sex ratios from the back cross of the male hybrid to a female O. niloticus did not conform to the expected ratios. Both species and their hybrid progeny had an identical karyotype (2N = 44) with no evident sex chromosome. The authors proposed that autosomes may play a role in sex determination.
The widely noted variation and general unpredictability of tilapia sex ratios has made the selection of true-breeding fish difficult. Shelton et al. (1983) reported sex ratios from mass spawnings of O. niloticus that ranged from 31 to 83% male. Tuan et al. (1999) reported a range of 15 to 100% male among 95 families of O. niloticus. These examples of variation in sex ratio and the uncertainty concerning the exact process behind sexual determination in tilapia present a challenge to the production of all-male populations through a YY breeding program.
YY breeding programs have been based on the premise that sex determination in O. niloticus is a simple Mendelian inheritance process where the female O. niloticus is homogametic (XX), the male heterogametic (XY), and animals having a Y chromosome develop as males. There are several approaches to a YY breeding program. Scott et al. (1989), working with O. niloticus, obtained a YY male in a gynogenesis study in which the parent female was thought to be of an XY genotype. Another approach is through hormone-induced sex inversion followed by selective breeding, as outlined by Shelton et al. (1978). Yongquan et al. (1979) applied this approach with O. mossambicus, another species in which the female is thought to be homogametic for sex determination, and successfully produced all-male progeny using YY males. Mair et al. (1993) successfully applied this approach to O. niloticus to produce all-male populations. Androgenesis, in which viable X or Y sperm are used to fertilize an irradiated oocyte, has also been used to produce YY male tilapia (Avila and Gonzalez, 1995; Myers et al., 1995; Marengoni and Onoue, 1998).
Numerous authors have proposed theories on the mechanism of tilapia sex determination, as reviewed by Wohlfarth and Wedekind (1991). It appears, however, that none has described the mechanism specifically enough to account for all the variation observed. This study was conducted to further examine the variation in sex ratios and to elucidate parts of the sex-determining process involved while trying to develop a line of fish suitable for use in a YY breeding program.
Where an XX:XY pattern of inheritance is assumed, the male would produce either an X or Y sperm, and an X-containing egg would be fertilized. If no autosomal factors influence sex ratios, the fertilization of the egg with a Y sperm should result in male progeny. This would be particularly evident in a highly inbred population such as one produced by androgenesis.
It was the intent of this study to develop a clearer understanding of sex inheritance using androgenically produced males and females. However, it was not possible to produce a highly inbred population through androgenesis of adequate quantity for use in this study as initially proposed. Instead, three other inbred strains of tilapia were used to evaluate the contribution of the male and female genome to sex inheritance. The Ivory Coast strain of O. niloticus was introduced to Auburn University (AU) from Brazil in 1974 and was originally introduced into Brazil from the Ivory Coast in 1971. The Egypt strain was introduced from the Ismailia Canal near Cairo to AU in 1982. The Ghana strain was introduced to AU from Israel in 1982 but was originally collected in 1978 from Lake Volta near Accra, Ghana (Khater, 1985). These strains have been maintained at Auburn since their original introduction; no additional fish have been added to the respective populations from the outside.
Fish were allowed to spawn from June to October of 1997 and 1998 in outdoor 2-m3 net hapas with a mesh size of 1.5 mm. Four net hapas per tank were suspended in 20-m2 rectangular cement tanks at the Fisheries Research Unit, Alabama Experiment Station, Auburn University, Alabama. Each tank was maintained at an average water depth of 70 cm using water from a rain-fed reservoir with a mean alkalinity of 30 mg l-1. Tanks were not limed or fertilized during the spawning period. Reference water temperature and dissolved oxygen readings were taken four times per week prior to daily feeding using a YSI Model 54ABP dissolved oxygen meter.
All broodfish were tagged for individual identification using a PIT tag (Destron Fearing Corporation). One male was placed in each net hapa. Females were added at a ratio of three to five females per male. Each strain was held and spawned separately. Females were inspected for eggs eight to ten days after stocking. Brooding females were collected with a fine mesh dip-net and transported individually to the Alabama Experiment Station hatchery. Each female was placed into a 95-l aquarium for the duration of the egg incubation. When the eggs hatched and the larvae reached the swim-up phase, the female was returned to another spawning hapa containing a different male. Swim-up fry were collected from the aquarium and placed in a 57-l aquarium, reared for approximately 30 d, and then transferred to outdoor tanks and ponds to grow to a sexable size. Water temperatures in all aquaria ranged from 29 to 35°C. Dissolved oxygen concentration was maintained near saturation through aeration using one to two air stones in each aquarium.
When the experimental fish had reached a length of at least 5 to 6 cm, a random sample was taken from each progeny group and immediately preserved in a 10% formalin solution. The sex of each fish was determined by surgical extraction of the gonads. The gonads were placed on a microscope slide and stained using Fast Green. A cover slide was laid over the stained gonads and pressed firmly so as to rupture the gonads. The gonadal squash was examined under a microscope. Each gonad was designated as ovary, testis, or ovotestis.
Data on sex ratio of progeny were analyzed using SAS. S+ software was used to generate figures. Four Chi-square tests were performed relating chance of maleness of individual progeny members to strain, to individual sires, to individual dams, and to repeat pair spawns. A logistic regression analysis was used to distinguish differences in percent males among strains. A weighted least-squares regression analysis was used to compare average standard deviations between strains.
Sex ratios were analyzed for 129 progeny groups from pair spawnings of three strains of normal (non-hormone-treated) O.niloticus broodstock. Fifty to one hundred randomly selected offspring were sexed from each progeny group, depending upon the number of individuals present at the time of harvest. Of the total number of progeny sexed (12,450), the mean percent males was 54.1%, with a standard deviation of 0.14 and a range of sex ratios from single spawns of 16 to 100% male. The mean percent males was significantly different (P < 0.001) from a 1:1 population sex ratio due to the slight excess of males. Sex ratios reflected a normal distribution (P > 0.45), but not a binomial distribution (P < 0.01), due to the overdispersion of sex ratios. The normal distribution possessed a skewness value of 0.07 and a range of 0.838. The normal curve was fitted using the sample mean and standard deviation. The binomial curve was fitted using the sample mean and a conservative sample size of 90 fish per progeny group due to the slight variation in progeny sample sizes (Figure 1). The summation of spawns (N = 119) from all male parents with repeat spawns yielded an average progeny of 54.1% male, with a standard deviation of 0.16 (Figure 2). The summation of spawns (N = 81) from all female parents with repeat spawns yielded an average progeny of 55.0% male with a standard deviation of 0.16 (Figure 3). The male and female summations also produced a normal distribution (P > 0.42) without producing a binomial distribution (P < 0.01), likewise a result of the overdispersion of sex ratios. All normal curves were again fitted using the respective sample mean and standard deviation, and binomial curves with the respective mean and a conservative sample size of 90 fish per progeny group.
Twelve Egypt males had multiple partners (Table 1). The mean percent male among the 3,716 progeny was 52.7%, with an average standard deviation of 0.09 and a range of 26 to 64% among pair spawns. Forty-two percent (5 to 12) of the males produced sex ratios that ranged less than 10% or less than two standard deviations. A statistically significant association (P= 0.001) existed between individual male parent and progeny sex. The Pearson correlation coefficient (-0.037), however, demonstrated that this association was very weak.
Seven Ghana males had multiple partners (Table 1). The mean percent male among the 2,559 progeny was 57.2%, with an average standard deviation of 0.11 and a range of 16 to 78% among pair spawns. Twenty-nine percent (2 of 7) of the Ghana males produced sex ratios that ranged less than 10%. An association (P = 0.001) also existed here between individual male parent and progeny sex. The correlation coefficient (0.102) again displayed the relative weakness of the association.
Twelve Ivory Coast males had multiple partners (Table 1). The mean percent male among the 5,332 progeny was 54.1%, with an average standard deviation of 0.12 and a range of 16 to 100% among pair spawns. Eight percent (1 of 12) of the males produced sex ratios that ranged less than 10%. An association between the individual male parent and progeny sex did exist within this strain as well (P = 0.001) but again was shown to be weak by the correlation coefficient (0.018).
Ten Egypt females had multiple partners (Table 2). The mean percent male among the 2,632 progeny was 52.5%, with an average standard deviation of 0.11 and a range of 26 to 75% among pair spawns. Twenty percent (2 of 10) of the females produced sex ratios that ranged less than 10%. A Chi-square test showed an association (P = 0.002) between individual female parent and progeny sex. The association, as demonstrated by the correlation coefficient (0.029), was weak.
Six Ghana females had multiple partners (Table 2). The mean percent male among the 1,264 progeny was 58.5%, with an average standard deviation of 0.09 and a range of 16 to 78% among pair spawns. Thirty-three percent (2 of 6) of the females produced sex ratios that ranged less than 10%. The Chi-square test showed an association (P = 0.005) between individual female parent and progeny sex. The corresponding correlation coefficient (0.017) demonstrated the weakness of the association.
Twelve Ivory Coast females had multiple partners (Table 2). The mean percent male among the 4,015 progeny was 54.5%, with an average standard deviation of 0.14 and a range of 16 to 100% among pair spawns. Eight percent (1 of 12) of the females produced sex ratios that ranged less than 10%. The Chi-square test showed an association (P = 0.001) between individual female parent and progeny sex. The weakness of this association was reflected in its small correlation coefficient (0.054).
Four pairs of Egypt individuals spawned with one another at least twice (Table 3). The mean percent male among the 855progeny was 52.7%, with an average standard deviation of 0.11 and a range of 39 to 61%. None (0 of 4) of the Egypt pairs produced sex ratios that ranged less than 10%. The Chi-square test revealed no association between the combination of parents and progeny gender (P = 0.44).
Two pairs of Ghana individuals spawned with one another at least twice (Table 3). The mean percent male among the 370progeny was 69.7%, with an average standard deviation of 0.06 and a range of 49 to 78%. Fifty percent (1 of 2) of the Ghana pairs produced sex ratios that ranged less than 10%. A Chi-square test showed no association between the combination of parents and progeny gender (P = 0.16).
Eleven pairs of Ivory Coast individuals spawned with one another at least twice (Table 3). The mean percent male among the 2,451 progeny was 55.5%, with an average standard deviation of 0.10 and a range of 16 to 79%. Forty-five percent (5of 11) of the Ivory Coast pairs produced sex ratios that ranged less than 10%. The Chi-square test did reveal an association between the combination of parents and progeny gender (P = 0.001), which may have been influenced by the larger sample size. The correlation coefficient (0.022) again proved the association to be extremely weak.
There was no clear indication that sex determination in O. niloticus is sex-linked. The mean percentage male of 54.10% from all spawns of males that produced repeat spawns closely approximates the population mean. The mean percentage male of 55.04% from all spawns of females that produced repeat spawns was slightly higher than either the repeat-spawning males' or the population mean. Male and female averages possessed an identical standard deviation of 0.16. The data on the population level appear quite similar, with no apparent differences due to the parents' sex.
Individual males with multiple partners and individual females with multiple partners were investigated to determine the influence of individual parents on sex ratios. Although each individual parent was found to have a statistically significant effect on progeny sex ratio (P < 0.005), the correlation was weak, and the overall average ratio of males produced by male parents with repeat spawns versus males produced by female parents with repeat spawns was similar within each strain. Wohlfarth and Wedekind (1991) concluded that for O. niloticus the sex ratio was determined by the male, and skewed sex ratios could be selected for based on selection of the male parent. Meriweather (1980) suggested that in O. aureus the female parent had a greater influence on sex ratio of the progeny than did the male. Mair et al. (1991) found no link between male parent, female parent, and progeny sex ratios from diallele-type crosses of five male and five female O. niloticus.
Considerable variation was observed among repeat spawns of the same pair. Some pairs produced sex ratios skewed to male in one spawn and to female in the next spawn. The variation from repeat spawns of the same pair is supported by Tuan et al. (1999), who reported that progenies derived from repeat spawns of the pair were significantly heterogeneous (P < 0.01). Several pairs in this study did, however, reproduce with consistent sex ratios. Thirty-five percent (6 of 17) of pairs with repeat spawns produced sex ratios that ranged less than 10%.
Wohlfarth and Wedekind (1991) cited Wedekind (1987), who found sex ratio to be a stable, heritable trait, with repeat spawns of the same pair giving almost identical sex ratios. Specific males and females within the current study also gave consistent skews over multiple partners, indicating that selection for sex ratio may be productive, as stated by Shelton et al. (1983). The limited number of repeat spawns by the same pair in the current study suggests that a portion of the brood population may breed relatively true for sex ratio. However, this characteristic was not observed in all pairs; it is possible that numerous spawns from the same pair would yield a normal distribution of sex ratios.
Sex ratios of O. niloticus do not always conform to a simple Mendelian pattern of inheritance. Of the 129 spawns in the current study, only 57% conformed to a 1:1 sex ratio. The frequency distribution of sex ratios within this study was normal; therefore, no multiple modes existed that might suggest sex ratio inheritance is based on major sex-determining genes. Hammerman and Avtalion's (1979) prediction of eight distinct sex ratios (0:1, 1:3, 3:5, 1:1, 9:7, 5:3, 3:1, 1:0) might also be expected to display multiple modes. The frequency distribution did not reflect a binomial distribution as would be predicted by a simple XX/XY determination process; therefore, other factors such as autosomal genes must be present to account for the overdispersion of sex ratios.
Shelton et al. (1983) found a range of 31% to 83% males, with 21% differing from a 1:1 sex ratio, among 71 progeny groups of the Ivory Coast strain of O. niloticus. Many generations later, 42% of 57 Ivory Coast pairs produced progeny that did not conform to a 1:1 sex ratio. Inbreeding results in a more homozygous gene pool (Tave, 1986), and a highly inbred line such as the Ivory Coast strain should show reduced genetic variability. Tave and Smitherman (1980) and Teichert-Coddington and Smitherman (1988) concluded that there was little genetic variability in this Ivory Coast strain involving selection for faster growth. In the current study there is no evidence of a reduction in sex ratio variability as a result of inbreeding.
The results suggest that it will not be possible to have a successful YY breeding program without first breeding to minimize variation in sex determination among normal fish. Variation in sex ratio occurs at all levels of selection, in general, yielding a normal distribution of sex ratios. The variation occurs at the level of individual male-female mating because repeat spawns of the same pair are not always consistent in the sex ratios produced. Sex determination appears to be a product of individual parent contribution, but how this effect occurs and to what extent each parent contributes could not be determined here. No current theory on sex inheritance is specific enough to explain the variation observed in this study.
Monosex populations are an essential part of many tilapia production systems. These can be obtained by several methods, with direct inversion of sex through hormone administration the most common. Monosex populations of males can also be produced by genetic manipulation to obtain male brood fish of a YY genotype. This assumes that sex determination is based solely on the female being homo-gametic (XX) and the male being heterogametic male (XY) with no additional sex-modifying genes.
In this study it was evident that there are additional sex modifiers complicating a successful YY breeding program. There was no evidence that these modifiers were sex-linked or associated with a particular strain of fish. There was no evidence to justify a selection program for simple sex determination based on sex or strain of fish. However, the data suggest that it might be possible to build a line of fish acceptable for use in a YY breeding program based on single pair selections. Such an intensive selection will result in a highly inbred line of fish which may or may not have similar growth performance to other male tilapia.
Avila, M.C. and M.A.C. Gonzalez, 1995. Ensayo de androgenesis diploide en tilapia roja (Oreochromis sp.), Trewavas (1981), mediante manipulación cromosomal. Boletin Cientifico INPA, 3:6678.
Chen, F.Y., 1969. Preliminary studies on the sex determining mechanism of Tilapia mossambica Peters and T. hornorum Trewavas. Verh. Int. Ver. Limnol., 17:719724.
Clemens, H.P. and T. Inslee, 1968. The production of unisexual brood of Tilapia mossambica sex reversed with methyltestosterone. Trans. Amer. Fish. Soc., 97:1821
Hammerman, I.S. and R.R. Avtalion, 1979. Sex determination in Sarotherodon (Tilapia). Part 2: The sex ratio as a tool for the determination of genotypeA model of autosomal and gonosomal influence. Theor. Appl. Genet., 55:177187.
Jalabert, B., P. Kammacher, and P. Lessent, 1971. Déterminisme du sexe chez les hybrides entre Tilapia macrochir et Tilapia nilotica. Etude de la sex-ratio dans les recroisements des hybrides de première génération par les espèces parentes. Ann. Biol. Anim. Biochim. Biophys., 11(1):155165.
Khater, A.A., 1985. Identification and comparison of three Tilapia nilotica strains for selected aquacultural traits. Ph.D. dissertation, Auburn University, Alabama, 67 pp.
Mair, G.C., A.G. Scott, D.J. Penman, J.A. Beardmore, and D.O.F.Skibinski, 1991. Sex determination in the genus Oreochromis. 1. Sex reversal, gynogenesis, and triploidy in O. niloticus. Theor. Appl. Genet., 82:144152.
Mair, G.C., J.B. Capili, J.A. Beardmore, and D.O.F. Skibinski, 1993. The YY-male technology for the production of monosex male tilapia, Oreochromis niloticus (L.). Proceedings of the AADCP International Workshop on Genetics in Aquaculture and Fisheries Management. University of Stirling, Scotland, pp. 9395.
Marengoni, N.G. and Y. Onoue, 1998. Ultraviolet-induced androgenesis in Nile tilapia Oreochromis niloticus (L.), and hybrid Nile x blue tilapia, O. aureus (Steindachner). Aquacult. Res., 29:359366.
Meriweather, F.H., 1980. Progeny sex ratios from pair spawning of Tilapia aurea (Steindachner). M.S. Thesis, Auburn University, Alabama, 61 pp.
Myers, J.M., D.J. Penman, Y. Basavaraju, S.F. Powell, P. Baoprasertkul, K.J. Rana, N. Bromage, and B.J. McAndrew, 1995. Induction of diploid androgenetic and mitotic gynogenetic Nile tilapia (Oreochromis niloticus L.). Theor. Appl. Genet., 90(2):205210.
Scott, A.G., D.J. Penman, J.A. Beardmore, and D. Skibinski, 1989. The "YY" supermale in Oreochromis niloticus (L.) and its potential in aquaculture. Aquaculture, 178(34):237251.
Shelton, W.L., K.D. Hopkins, and G.L. Jensen, 1978. Use of hormones to produce monosex tilapia. In: R.O. Smitherman, W.L. Shelton, and J.L. Grover (Editors), Culture of Exotic Fishes Symposium Proceedings. Fish Culture Section, American Fisheries Society, Auburn, Alabama, pp. 1033.
Shelton, W.L., F.H. Meriweather, K.J. Semmens, and E. Calhoun, 1983. Progeny sex ratios from intraspecific pair spawnings of Tilapia aurea and Tilapia niloticus. In: L. Fishelson and Z. Yaron (Compilers), International Symposium on Tilapia in Aquaculture, Nazareth, Israel. Tel Aviv University Press, Tel Aviv, pp. 270280.
Tave, D., 1986. Genetics for Fish Hatchery Managers. AVI Publishing Company, Inc., Westport, Connecticut, 299 pp.
Tave, D. and R.O. Smitherman, 1980. Predicted responses to selection for early growth in Tilapia nilotica. Trans. Am. Fish. Soc., 109:439445.
Teichert-Coddington, D.R. and R.O. Smitherman, 1988. Lack of response by Tilapia nilotica to mass selection for rapid early growth. Trans. Am. Fish. Soc., 117:297300.
Tuan, P.A., G.C. Mair, D.C. Little, and J.A. Beardmore, 1999. Sex determination and the feasibility of genetically male tilapia production in the Thai-Chitralada strain of Oreochromis niloticus. Aquaculture, 173:257269.
Wedekind, H., 1987. Untersuchungen zur Vererbung des Geschlechtes bei Afrikanischen Buntbarschen (Oreochromis niloticus). Diplomarbeit, Gottingen University, Gottingen, 110 pp.
Wohlfarth, G.W. and H. Wedekind, 1991. The heredity of sex determination in tilapias. Aquaculture, 92:143156.
Yongquan, Y., Z. Zhonging, L. Kehong, W. Yusheng, X. Zhen, H.Erchun, and G. Zhihui, 1979. A preliminary study on physiogenetic control of sexuality in Tilapia mossambica. Acta Genet. Sin., 6(3):1722.
Table of Contents