Source of the fish

Three- to four-week old fry of the Tuxedo (TUX) guppy strain were obtained from Chin Lam Brothers Tropical Fish Farm in Singapore. The common name, Tuxedo, for this strain was given by guppy breeders. Wild-type (WT) feral guppies were collected from an isolated hill-stream near the Bukit Timah nature reserve in Singapore. Juvenile TUX were cultured in 180-liter fibreglass tanks (30 fish/tank) in the aquarium area of the Department of Biological Sciences, National University of Singapore, at temperatures of 25–28°C. WT fry were separated from the collected samples and raised in 30-liter clear plastic tanks (20 fish/tank). Under laboratory conditions, sexual maturation of WT fry usually occurs at 4–6 weeks of age. Juvenile WT were checked daily for developing males which are detected by gonopodial formation of the anal fin. Males, when spotted, were immediately removed and reared separately from females as virgin females were essential for the reciprocal crosses.

Description of the fish

Adult males and females of the TUX strain have a total length of 3–4 cm and 5–6 cm, respectively. Adult TUX males have black (melanic) or dark grey pigmentation on the caudal-peduncle region which masks normal wild-type male body coloration, and a caudal fin that ranges from blood-red to orange-red in color (Fig. 1A). Some TUX males may have a metallic blue or green sheen overlying the black caudal-peduncle. TUX females show drab wild-type olive-brown body coloration and grey caudal-peduncle with red tinges of varying intensity on an opaque greyish-white tail (Fig. 1B). Wild-type guppies are smaller than the domesticated TUX strain. Adult WT males are 2–2.5 cm in length and females are about 3–4 cm. As described earlier, WT males have highly polymorphic color patterns on the body and fins (Fig. 1C), while WT females are devoid of color patterns (Fig. 1D).

Fig. 1

(A) Adult male guppy of the Tuxedo (TUX) strain showing black caudal-peduncle and red tail color patterns. (B) Adult female guppy of the Tuxedo strain with grey caudal-peduncle and faint red tinges on an opaque greyish-white tail. (C) Adult male feral guppy with typical wild-type (WT) male body coloration. (D) Adult female wild-type guppy with a drab olive-brown body that is devoid of any bright color patterns. (E) Male progenies (F₁ and F₂) of the (a) TUX, (b) red tail, (c) black caudal-peduncle and (d) WT phenotypes. (F) Female progenies (F₁ and F₂) with typical TUX (top) and WT (bottom) colorations. (G) Recombinant females (F₁ and F₂) displaying red tail (top) and black caudal-peduncle (bottom) color patterns.

Reciprocal crosses

Inheritance of the black caudal-peduncle and red tail color patterns was elucidated by single-pair reciprocal crosses between the TUX strain and WT stock, using six-week old mature virgin fish. Each pair was kept in a 3.5-liter breeding tank. Broods were produced 4-6 weeks after mating. Single-pair full-sib F₁ males and F₁ females were mated to obtain the F₂ generation. The following notations were used: TUX♂♂×WT♀♀(Table 1A) and WT♂♂×TUX♀♀(Table 2A) for parental crosses, and F₁♂♂×F₁♀♀(Tables 1B, 2B) for full-sib F₁ crosses. Newly born fry were separated and raised to maturity in 3.5-liter clear plastic tanks (five fish/tank). All F₁ and F₂ progenies were segregated and scored according to phenotypes and sex. Progenies displaying color patterns such as Tuxedo, red tail and black caudalpeduncle were designated as the TUX, RT and BCP phenotypes, respectively, and those without such color patterns, WT phenotype. Tuxedo males of parental crosses were typed using Roman numerals (I, II, III, IV and V) according to their putative alleles following segregation of their F₁ and F₂ progenies. This was to facilitate description of the crosses.

Table 1

Mating results of crosses between Tuxedo (TUX) males and wild-type (WT) females showing observed and expected numbers for each phenotypic class, expected segregation ratios, chi-square goodness-of-fit to the expected ratios and their corresponding adjusted values (χ²_adj) after application of Yates' correction for continuity, χ² test for homogeneity, probable genotypes and recombinants for (A) the F₁ generation of single-pair parental crosses, and (B) the F₂ generation of single-pair crosses between full-sib F₁ males and F₁ females. Recombinants (^§) due to crossing-over of the Rdt and Bcp genes were not considered in chi-square analyses. (Phenotypes: TUX♂♂=Tuxedo males with black caudal-peduncle and red tail; TUX♀♀=Tuxedo females with grey caudal-peduncle and faint red tinges on an opaque greyish-white tail; RT = red tail without black caudal-peduncle; BCP=black caudal-peduncle without red tail; WT = wild-type coloration without red tail and black caudal-peduncle. Genes: Bcp=black caudal-peduncle gene; Bcp⁺=absence of black caudal-peduncle gene; Rdt=red tail gene; Rdt⁺=absence of red tail gene). A. TUX♂♂×WT♀♀ (Parental Cross)

Table 2

Mating results of crosses between wild-type (WT) males and Tuxedo (TUX) females showing observed and expected numbers for each phenotypic class, expected segregation ratios, chi-square goodness-of-fit to the expected ratios and their corresponding adjusted values (χ²_adj) after application of Yates' correction for continuity, χ² test for homogeneity, probable genotypes and recombinants for (A) the F₁ generation of single-pair parental crosses, and (B) the F₂ generation of single-pair crosses between full-sib F₁ males and F₁ females. Recombinants (^#) due to crossing-over of the Rdt and Bcp genes were not considered in chi-square analyses. (Phenotypes: TUX♂♂= Tuxedo males with black caudal-peduncle and red tail; TUX♀♀= Tuxedo females with grey caudal-peduncle and faint red tinges on an opaque greyish-white tail; RT = red tail without black caudal-peduncle; BCP = black caudal-peduncle without red tail; WT = wild-type coloration without red tail and black caudal-peduncle. Genes: Bcp = black caudal-peduncle gene; Bcp⁺ = absence of black caudal-peduncle gene; Rdt = red tail gene; Rdt⁺ = absence of red tail gene).

Statistical analyses

Observed phenotypic distributions were tested for goodness-of-fit with predicted proportions using the chi-square (χ²) test (Sokal and Rohlf, 1981; Strickberger, 1990). Since observed and expected numbers in the phenotypic classes and sample sizes were small (n<200), Yates' (1934) correction for continuity was included in the calculation of χ² to improve the approximation to the χ² distribution, as shown by the χ²_adj values. The χ² test for homogeneity was used to determine whether there were significant differences among the phenotypic frequencies, and if the observations were sufficiently uniform and the population homogeneous after the data was pooled. The correction for continuity was not incorporated into the test for homogeneity because calculated χ² values had to be summed and χ²_adj values were not additive (Sokal and Rohlf, 1981; Strickberger, 1990). Following Winge (1922b, 1923, 1927, 1934), Nayudu (1979), Phang et al. (1989a, b, 1990), Phang and Fernando (1991), and Khoo et al. (1999), individuals with exceptional coloration due to crossing-over of the black caudal-peduncle (Bcp) and red tail (Rdt) genes between the X- and Y-chromosomes were not considered in chi-square analyses.

Recombination frequencies and map distances between Bcp and Rdt, and the sex-determining region (SdR) were estimated according to Strickberger (1990), Phang et al. (1990), Phang and Fernando (1991), Purdom (1993) and Khoo et al. (1999). Winge's (1922b, 1927, 1934) “zig-zag line diagram” method was applied to test all possible linkage combinations between Bcp, Rdt and SdR, and map these loci in sequential order. Due to the infrequency of single crossing-over in the guppy compared to the fruitfly, Drosophila (Winge, 1927, 1934; Purdom, 1993), double crossing-over was excluded from calculations of map distances. All crossovers in this study were thus regarded as single crossovers.

Segregation and recombination in TUX × WT F₁ and F₂ offspring

Three mating pairs TUX♂♂×WT♀♀(PT5, PT9 and PT10) produced 40 male and 34 female F₁ offspring in 8 broods (Table 1A). F₁ males exhibited the black caudal-peduncle and red tail color patterns of their TUX male parents (Fig. 1E), while F₁ females had a grey caudal-peduncle and an opaque greyish-white tail (Fig. 1F). F₁ males and females could have inherited the black caudal-peduncle and red tail color genes only from their TUX male parents (designated as type I). Table 1A also shows three other crosses in which the TUX male parents were heterozygous for black caudal-peduncle and red tail. To facilitate description of these crosses and their offspring, these TUX males were labelled as types III, IV and V. Type II TUX males were not observed in this study although they were found among crosses between the Tuxedo and Green Variegated guppy strains that we carried out in a later study (Khoo et al., submitted). For type III, five mating pairs gave 18 broods of 116 red tail (RT) males and 130 TUX females (Fig. 1E, F, Table 1A). Four F₁ broods of 33 WT males and 35 TUX females were produced by the cross between a type IV TUX male and a WT female (mating pair PT1), while 25 BCP males and 29 TUX females were obtained from mating pair PT6 (type V TUX male) (Fig. 1E, F, Table 1A). For all four types (I, III, IV and V) of TUX male parents, the F₁ male to female ratio was consistent with the expected ratio of 1:1 (Table 1A).

The F₂ generation of type I comprised 205 TUX males, 113 TUX females and 105 WT females with the observed numbers conforming to the expected phenotypic ratio of 2:1:1 (Fig. 1E, F, Table 1B). Four F₂ phenotypes that consisted of 108 TUX and 121 RT males, and 120 TUX and 113 WT females were obtained from 11 single-pair full-sib F₁ crosses of type III (Fig. 1E, F, Table 1B). These phenotypes fitted the hypothetical ratio of 1:1:1:1. Mating pair PT1 (type IV) also gave four F₂ phenotypes that were made up of 13 broods consisting of 61 TUX and 54 WT males, and 58 TUX and 48 WT females that agreed with the 1:1:1:1 ratio (Fig. 1E, F, Table 1B). For mating pair PT6 (type V), two full-sib F₁ crosses produced nine broods of 64 TUX males, 65 BCP males, 54 TUX females and 60 WT females that corresponded to the expected 1:1:1:1 ratio (Fig. 1E, F, Table 1B). Homogeneity χ² tests carried out for TUX males of types I and III showed that the F₁ and F₂ observations were uniform and did not form heterogeneous populations after being pooled (Table 1A, B).

F₁ and F₂ data for TUX♂♂×WT♀♀showed that the dominantly expressed black caudal-peduncle and red tail color patterns are due to single genes that are found at two different loci, Bcp and Rdt, respectively, on the sex chromosomes. Homozygous TUX male parents (type I) were elucidated to have the X_Bcp,RdtY_Bcp,Rdt genotype (Table 1, Fig. 2). When TUX males were heterozygous at the Bcp and Rdt loci, these individuals possessed the X_Bcp,RdtY_Bcp⁺,Rdt (type III), X_Bcp,RdtY_Bcp⁺,_Rdt⁺ (type IV) and X_Bcp,RdtY_Bcp,Rdt⁺ (type V) genotypes (Table 1, Fig. 2). The segregation of the Bcp and Rdt alleles in the TUX♂♂ ×WT♀♀cross is presented schematically by a genetic model in Fig. 2.

Fig. 2

Schematic diagram of the proposed genetic models showing segregation of the Bcp and Rdt color genes, and genotypes of the parents (P), F₁ and F₂ progenies, and recombinants (R) that occur in reciprocal crosses between the Tuxedo (TUX) and wild-type (WT) guppies.

Crossing-over between the sex-determining region (SdR), and the black caudal-peduncle (Bcp) and red tail (Rdt) loci at the parental and F₁ levels for TUX♂♂×WT♀♀produced exceptionally colored F₁ and F₂ recombinants (Table 1). These individuals which expressed TUX, RT, BCP and WT phenotypic colorations in males, and RT and BCP in females, deviated from the predicted phenotypes for types I, III, IV and V. Type I produced three RT and three WT recombinant F₂ males (Fig. 1E, Table 1B). The crossover frequency calculated from the percentage of crossover males out of the total number of F₂ males (3/211×100%) was 1.422% for SdR–Rdt and Rdt–Bcp, respectively. Occurrence of RT recombinant F₂ males indicated that the Y-chromosome of the guppy may possibly have a map order of SdR–Rdt–Bcp as these recombinants could not be produced using Winge's (1922b, 1927, 1934) “zig-zag line diagram” method if the gene order had been either SdR–Bcp–Rdt or Bcp–SdR–Rdt (Figs. 2, 3). A map distance of 1.802 map units was estimated for Rdt–Bcp from two RT and two BCP recombinant F₂ females of 222 F₂ females for type I (Fig. 1G, Table 1B). Recombinant F₁ TUX males and F₂ RT females of type III were not used to estimate map distances because crossovers could have taken place at SdR–Rdt and Rdt–Bcp, making it impossible to determine the actual region of crossover. Of 240 F₂ females of this cross, three BCP female recombinants gave a crossover frequency of 1.250% between Rdt and Bcp (Fig. 1G, Table 1B).

Fig. 3

Genetic map of the Y-chromosome of the guppy, Poecilia reticulata, showing the positions of the black caudal-peduncle (Bcp) and red tail (Rdt) loci relative to the sex-determining region (SdR). Map distances of Bcp and Rdt from the SdR are based on recombination frequencies estimated from the TUX ♂♂×WT ♀♀ and WT ♂♂×TUX♀♀ crosses in Tables 1 and 2. Bcp and Rdt are also inferred to be located at similar positions on the X-chromosome of the guppy. The size of the SdR is not according to scale as the number of male-determining genes within that region is unknown.

For mating pair PT1 (type IV), 5.405 map units were estimated to separate the Rdt locus from SdR and Rdt from Bcp, respectively (Table 1A). Longer SdR–Rdt and Rdt–Bcp map distances for type IV compared to types I and III was due to frequent crossing-over which yielded two TUX and two BCP recombinant F₁ males out of only 37 F₁ males (Fig. 1E, Table 1A). The presence of F₁ BCP male recombinants in this cross also indicated a gene sequence of SdR–Rdt–Bcp (Figs. 2, 3). An F₁ BCP recombinant female of 30 F₁ females for mating pair PT6 (type V) gave a crossover rate of 3.333% between SdR and Rdt (Fig. 1G, Table 1A). Seven F₂ recombinant males that had WT coloration of 136 F₂ males of PT6 further supported an order of SdR–Rdt–Bcp, giving an approximate distance of 5.147 map units between SdR and Bcp (Fig. 1E, Table 1B). The mean genetic map distance between the SdR and Rdt, calculated from F₁ and F₂ recombinant data (Table 1), was 3.387±1.992 map units while the Bcp locus appeared to be about 5.147 map units from the SdR (Fig. 3).

Segregation and recombination in WT×TUX F₁ and F₂ offspring

Three mating pairs (PB2, PB5 and PB6) of the reciprocal cross, WT♂♂×TUX♀♀, gave nine F₁ broods of 69 males and 82 females, all of which had black caudal-peduncle and red tail color patterns typical of the Tuxedo (TUX) phenotype (Fig. 1E, F). The observed numbers of F₁ male to female offspring agreed with the expected ratio of 1:1 (Table 2A). With the exception of four males with BCP phenotype, the F₂ progeny of this cross segregated into 93 TUX males, 93 WT males and 183 TUX females according to the expected phenotypic ratio of 1:1:2 (Fig. 1E, F, Table 2B). Results for mating pair PB4 were not included with those of the above three mating pairs because it produced a very high number of F₂ recombinants (11 BCP males and 16 BCP females) (Table 2B). PB4, however, satisfied the expected F₁ male to female ratio of 1:1 with 26 TUX males and 29 TUX females (Table 2A). In addition, three full-sib single-pair F₁ crosses of PB4 generated 63 TUX males, 51 WT males and 105 TUX females that were consistent with the F₂ phenotypic ratio of 1:1:2 (Table 2B). Tuxedo females of mating pairs PB1 and PB3 were inferred to be heterozygous at the Bcp locus since they gave a total of 24 TUX males, 19 RT males, 18 TUX females and 19 RT females that conformed to the expected F₁ ratio of 1:1:1:1 (Table 2A). After pooling of the F₁ and F₂ data for WT♂♂× TUX♀♀, homogeneity χ² tests showed that each pooled population was homogeneous and uniform (Table 2A, B).

F₁ and F₂ results for WT♂♂×TUX♀♀confirmed the observations of the reciprocal cross (TUX♂♂×WT♀♀) that two single sex-linked genes, black caudal-peduncle (Bcp) and red tail (Rdt), are responsible for the dominant expression of the Tuxedo color pattern. TUX female parents (including PB4) used in this study were found to be homozygous for Bcp and Rdt, and had the X_Bcp,RdtX_Bcp,Rdt genotype (Table 2, Fig. 2). Conversely, the putative genotype for heterozygous TUX females of mating pairs PB1 and PB3 was X_Bcp,RdtX_Bcp⁺, ^Rdt (Table 2, Fig. 2). The segregation of Bcp and Rdt is illustrated by a genetic model in Fig. 2.

Four F₂ males of mating pair PB2 had a black caudal-peduncle and black tail instead of a red tail (Fig. 1E, Table 2B). In this instance, Bcp could have crossed over from the X- to the Y-chromosome in the F₁ male parent, and also from Y to X in the F₁ female parent to produce male BCP recombinants (Fig. 2). Since there were four BCP males of a total 190 F₂ male individuals, the crossover frequency between Bcp and Rdt was estimated to be 2.105% (2.105 map units) (Table 2B). As mentioned earlier, mating pair PB4 produced a very high number of F₂ BCP male (11) and female (16) recombinants out of 125 F₂ males and 121 F₂ females (Table 2B, Fig. 2). From these crossover data, the map distance between Bcp and Rdt was averaged to be 11.012±3.128 map units. Simultaneous occurrence of BCP males and females suggested a gene order of SdR–Bcp–Rdt. These results contradicted those of the reciprocal cross (TUX♂♂×WT♀♀) which pointed to a map order of SdR–Rdt–Bcp. Moreover, Bcp and Rdt seemed to be too far apart (11.012± 3.128 map units) for PB4 compared to the 2.397±1.714 map units estimated from mating pairs PB2, PB5 and PB6, and TUX♂♂×WT♀♀ (Tables 1, 2, Fig. 3).

Inheritance of the black caudal-peduncle and red tail color patterns

Observations for all parental (TUX♂♂×WT♀♀and WT ♂♂×TUX♀♀, Tables 1A, 2A) and full-sib (F₁♂♂×F₁♀♀, Tables 1B, 2B) crosses indicate that the black caudal-peduncle (Bcp) and red tail (Rdt) color pattern genes are responsible for the Tuxedo phenotype of the guppy. This study demonstrates that these two color patterns are simple sex-linked traits controlled by single genes: the Bcp allele is dominant for black caudal-peduncle over Bcp⁺, and Rdt is dominant for red tail over Rdt⁺. Wild-type or feral guppies do not display these color patterns as they have the recessive Bcp⁺ and Rdt⁺ genes (Fig. 1C, D). Using Tuxedo males that were heterozygous for Bcp and Rdt, we have also shown that the expression of Bcp and Rdt is dominant in both males and females (Tables 1, 2, Figs. 1, 2). This study confirms our preliminary findings that the black caudal-peduncle and red tail colour genes have a dominant mode of inheritance (Fernando and Phang, 1989, 1990, Phang et al., 1990). The typical genotypes for the Tuxedo guppy are thus inferred to be X_Bcp,RdtY_Bcp,Rdt for males (type I) and X_Bcp,RdtX_Bcp,Rdt for females. Tuxedo males that are heterozygous have the X_Bcp,RdtY_Bcp⁺,_Rdt (type III), X_Bcp,RdtY_Bcp⁺,_Rdt⁺ (type IV) and X_Bcp,RdtY_Bcp,Rdt ⁺ (type V) genotypes while the females are X_Bcp,RdtX_Bcp⁺,_Rdt. The presence of these heterozygous individuals suggests that the black caudal-peduncle and red tail color patterns have not yet become “fixed” traits among cultured stocks of the Tuxedo strain.

Phenotypic map of Bcp, Rdt and the SdR

Our results prove that the Bcp and Rdt alleles are able to cross over from the X- to the Y-chromosome and vice versa since male and female recombinants of the TUX, RT, BCP and WT phenotypes were obtained from both F₁ and F₂ offspring of TUX♂♂×WT♀♀(types I, III, IV and V) and WT♂ ♂×TUX♀♀(Tables 1, 2, Figs. 1, 2). The phenomenon of alleles migrating between the X- and Y-chromosomes as a result of crossing-over was first documented in the guppy by Winge (1922a, b, 1923). Subsequent analyses by Winge (1927, 1934), Winge and Ditlevsen (1938), Dzwillo (1959), Nayudu (1975, 1979) and Kirpichnikov (1981) showed that the X- and Y-chromosomes of the guppy are cytologically indistinguishable and possibly share undifferentiated homologous regions along the lengths of their chromatids. Therefore, crossing-over occurs between the sex chromosomes, and recombination rates of up to 10% have been recorded between the Vitellinus (Vi) and Elongatus (El), and Doppelschwert (Ds) and Pigmentierte caudalis (Cp) genes (Winge, 1927, 1934; Winge and Ditlevsen, 1947; Dzwillo, 1959; Kirpichnikov, 1981; Purdom 1993). In our study, the occurrence of recombination shows that Bcp and Rdt are located within homologous regions on the X- and Y-chromosomes, and are approximately 5.147 and 3.387±1.992 map units, respectively, away from the sex-determining region (SdR) (Tables 1, 2, Fig. 3).

From the F₁ and F₂ recombinant data (Tables 1, 2), genetic map distances of 3.387±1.992, 5.147 and 2.397±1.714 map units were obtained for SdR–Rdt, SdR–Bcp and Rdt–Bcp, respectively (Fig. 3). As expected, the map distance between SdR and Bcp (5.147 map units) is close to the sum of distances (3.387+2.397=5.784 map units) for SdR–Rdt and Rdt–Bcp (Fig. 3). Estimates for crossing-over between two loci that are far apart are, however, never exactly the sum of the estimates for smaller regions amidst them (Purdom, 1993). This is because a crossover between two loci usually inhibits a second crossover from occurring in an adjacent region (Strickberger, 1990). Crossing-over at SdR–Bcp, SdR–Rdt and Rdt–Bcp will thus influence each other, thereby affecting their frequency of occurrence. Despite this, our results show that Rdt is closer to SdR than Bcp because lower crossover frequencies (shorter map distances) were obtained between Rdt and SdR than from Bcp to SdR for all crosses (Tables 1, 2). Recombinant data for Tuxedo male parents of types I, IV and V also provides evidence that Rdt lies between the SdR and Bcp, hence indicating a gene map order of SdR–Rdt–Bcp (Fig. 3). The exception is mating pair PB4 which suggests a sequence of SdR–Bcp–Rdt. The latter may possibly result from translocation of a segment of the sex chromosomes that contain at least one of these color genes onto another, or pericentric inversion of a region that has two of these genes (Purdom, 1993).

In conclusion, the black caudal-peduncle (Bcp) and red tail (Rdt) genes of the domesticated Tuxedo guppy strain (1) occur as single genes at two different loci, and are (2) dominantly expressed, (3) X- and Y-linked, and (4) fully capable of crossing-over from the Y- to the X-chromosome and vice versa since they are situated in an undifferentiated homologous region on these chromosomes. Genetic map distances for the sex-determining region (SdR)–Rdt, SdR–Bcp and Rdt–Bcp are estimated to be 3.4, 5.1 and 2.4 map units, respectively. We therefore propose a phenotypic map of SdR–Rdt–Bcp (Fig. 3) for the Y-chromosome of the guppy. The Bcp and Rdt loci are also inferred to be at similar positions on the X-chromosome that was postulated by Winge (1927, 1934) to have a corresponding feminine segment of the SdR (Fig. 3). Our findings for the black caudal-peduncle and red tail genes, and the sex-determining region should prove valuable for the mapping of color pattern genes onto the sex chromosomes of the guppy as initiated by Winge (1927, 1934), and carried on by Winge and Ditlevsen (1947), Dzwillo (1959), Nayudu (1975, 1979), Kirpichnikov (1981), Fernando and Phang (1989, 1990), Phang et al. (1989a, b, 1990), Phang and Fernando (1991), Purdom (1993) and Khoo et al. (1999).