Sex and speciation: genetic architecture and evolutionary potential of sexual versus nonsexual traits in the sibling species of the Drosophila melanogaster complex.
However, traits linked to mating components of reproduction are not the only ones that show species-specific characteristic. Among different animal groups, external male genitalia has been a useful taxonomic character at the species level. Eberhard (1985) surveyed major groups of animals, in which either primary or secondary external male genitalia were used to distinguish species (see tables 1.1, 1.2 in Eberhard 1985). Eberhard focused on morphological traits directly involved in copulation, excluding internal organs such as testes and accessory glands, and proposed sexual selection as the driving force responsible for the high interspecific divergence detected in male genitalia (Eberhard 1985). Recently, Eberhard and Cordero (1995) have tried to extend the role of sexual selection to seminal products that might influence female reproductive behavior and physiology as well as to female postcopulatory mechanisms for male discrimination (cryptic mate choice).
The high divergence of reproductive traits is certainly not restricted to external organs. There is extraordinary variation in sperm structure and size among different species of Drosophila (Joly et al. 1991). The extremely long sperm found in D. bifurca and D. hydei has become a challenging puzzle for evolutionary biologists. Explanations have included post-fertilization sperm contribution to the zygote (Bressac et al. 1994), sperm competition (Pitnick and Markow 1994; Joly et al. 1995), and species-specific sperm-egg interactions (Karr and Pitnick 1996).
Testis length also shows high divergence among Drosophila species. A pictorial representation of testis morphology by Patterson and Stone (1952; [ILLUSTRATION FOR FIGURES 1-18 OMITTED]) describes a range in length from species with highly coiled and long testes (such as D. virilis and D. funebris) to extremely short testes (such as D. pseudoobscura). However, most studies in Drosophila have focused on comparisons in sperm length with only some mention of testes length, and few studies have looked at sperm length divergence between closely related species in comparison to other nonsexual trait's divergence. One study that compared sperm, testes, and thorax length among the four species of the nanoptera group showed that these closely related species differ fourfold in testis and sperm length but are almost identical in thorax length (Pitnick and Markow 1994).
In this study, we have compared morphological variation within and between species of the melanogaster complex in sexual (testes length and area and the area of the posterior lobe of the genital arch) and nonsexual traits (wing length and width, tibia length, femur length, and length and area of the stalk connecting the malpighian tubule to the pyloric region of the ventriculus). We have addressed the question of whether sexual trait morphology is highly divergent between the four closely related species of the melanogaster complex. If sexual traits (i.e., those traits that are part of the genitalia but are not directly involved in reproduction) play a major role in the early stages of speciation, then these traits should show higher divergence than nonsexual traits between closely related species. To the extent that high levels of genetic divergence cause this morphological divergence, such higher genetic divergence is also expected to translate into a stronger disruption of the sexual trait's morphology in interspecific hybrids due to incompatible gene interactions. Interspecific hybrid breakdown and interspecific genetic interactions of sexual and nonsexual traits were analyzed by both patterns of dominance and comparisons of trait asymmetry between parental species and their hybrids.
Although Carson (1982, 1985) has previously proposed a dichotomy between a sexual and nonsexual gene pool, little is known about the nature of the interactions among genes in such pools. According to Carson's organization theory of speciation, the sex gene pool should be part of a closed polygenic system (Carson 1985). If speciation proceeds by major modifications in sexual traits, as proposed by Carson (1985), then it becomes necessary to understand the genetic basis of such traits. We have made an attempt to elucidate the genetic basis of the sexual and nonsexual traits included in this study by analyzing the patterns of inheritance of such traits in hybrids.
Our final question pertains to the role played by different evolutionary forces in the evolution of sexual and nonsexual traits. According to Paterson (1985) and Carson (1985), the high divergence observed among mating components of the reproductive system is the consequence of a direct disruption of a genetically buffered system. Such models predict that sexual traits should be almost invariant within species and show little asymmetry, an indicator of the trait's homeostasis (see Palmer and Strobeck 1986; Parson 1990), between the different sides of the character measured. Alternatively, the high divergence of sexual traits between species could be the result of directional selection during the early stages of speciation, with plenty of genetic variation remaining within species. Depending on the strength of the selective force, this predicts the same or higher asymmetry of sexual traits within species.
Our results extend the previous observations that sexual traits not directly involved in mating show high divergence between closely related species of Drosophila (Coulthart and Singh 1988; Thomas and Singh 1992; Civetta and Singh 1995). This high divergence, combined with the variation found within species and the pattern of traits asymmetry, suggest that the evolution of sexual traits may have been driven by directional sexual selection at the onset of species formation rather than by stabilizing selection and episodic allele fixation (Carson 1985; Paterson 1985). The latter scenario was also considered by Coulthart and Singh (1988), whose analysis of male reproductive tract proteins by twodimensional electrophoresis showed low variation within and high divergence between species. Finally, the different nature of the dominance relationship found for sexual and nonsexual traits and the higher interspecific hybrid asymmetry of sexual traits indicate that species-specific gene interactions may play an important role in the performance of sexual traits.
TABLE 1. A classification of the different traits measured. Traits Sexual Nonsexual External Genital arch (area) Tibia and femur (length) Wing (length and width) Internal Testis (length and area) Malpighian tubule (length and area)
MATERIALS AND METHODS
Drosophila Cultures and Stock Maintenance
All the stocks used in this study, except for Oregon R and D. mauritiana LG24 (provided by Dr. J. David), are currently available from the Bowling Green Species Resource Center and the stock numbers are given in parentheses. Three different strains of each of the four species of the melanogaster complex were studied. Drosophila melanogaster strains were Oregon R, Peru (0231.1), and Australia (0231.3); D. simulans Colombia (0251.2), California (0251.163), and South Africa (0251.164); D. mauritiana LG24, 72 (0241.3), and 207 (0241.5; all from Mauritius Island); D. sechellia 21 (0248.6), 22 (0248.2), and Robertson (0248.7; all from Seychelles Island). The strains were maintained at 22 [degrees] C under a 12:12 L: D cycle.
Flies used in our analysis were the progeny obtained from mating 10 virgin females to males in a 1:1 sex ratio in vials containing approximately 5 mL of banana medium. To avoid crowding, females were allowed to lay eggs and the adults were transferred to a new vial whenever the count of eggs per vial was around 30. Eggs were allowed to develop to adult stage at 22 [degrees] C under a 12:12 L:D cycle. Newly hatching adults were anesthetized with C[O.sub.2] and sexed. The males were transferred to vials containing fresh banana medium, where they remained for five days until collected for dissection. Between 10 and 15 individuals from each strain or cross were measured for the different traits analyzed.
The traits studied were chosen to compare sexual versus nonsexual characters that were internal as well as external (Table 1). The external, nonsexual traits were wing length and width, tibia length, and femur length. The external sexual trait was the area of the posterior lobe of the genital arch. The internal, nonsexual traits measured were the length and area of the stalk connecting the malpighian tubule to the pyloric region of the ventriculus (referred from now on as malpighian tubule). Testis length and area were used as the internal sexual counterpart. The fly organs were dissected from each individual fly on a slide containing a drop of Drosophila Ringer's solution (Ashburner 1989a). Wings, legs and genital arches were flattened by covering them with an 18-[mm.sup.2] coverslip and heating the slide until the solution boiled. Testes and malpighian tubules were transferred and stretched in a drop of liquid paraffin. A video camera (Hitachi VKC150) connected to a dissecting microscope and a Macintosh computer were used to capture images of the different preparations. Quick Image[TM] software package was used for image capturing, and NIH Image was used for measurement of the different morphological traits. Measurements were taken by using either a straight or free-hand, mouse-controlled line selection.
Wing length (WL), wing width (WW), tibia length (TiL), and femur length (FeL) were measured as in Long and Singh (1995), but the femur measurement did not include the trochanter. The posterior lobe of the genital arch (GA) was demarcared by a line across its base, and its area was assessed (Coyne 1983). Testes length (TL), testes area (TA), malpighian tubules length (MTL), and malpighian tubules area (MTA) were measured as shown in Figure 1. Data were obtained in pixels and then converted into millimeters by scaling with a micrometer. Measurements were taken from both sides of a trait and the mean was used as an individual score.
Analysis of Phenotypic Variation
Analysis of variance (ANOVA) of the data obtained from each trait measurement was used to examine the amount of variation associated with species and with strains within species. The analysis was done using JMP version 2.0.2 for the Macintosh. For malpighian tubules length and area, where only D. simulans (California and Colombia), D. mauritiana 72 and D. sechellia 21 were scored, measurements were analyzed by using a single-factor ANOVA model: [Y.sub.ij] = [Mu] + [Species.sub.i] + [[Epsilon].sub.ij]. The other character measurements were analyzed by a nested ANOVA model with species and strain within species effects: [Y.sub.ijk] = [Mu] + [Species.sub.i] + [Strain.sub.j[i]] + [[Epsilon].sub.ijk].
Both normal quantile plots and the Shapiro-Wilk statistic (Shapiro and Wilk 1965) were used to check for departures from normality. The genital arch area and malpighian tubules area data showed significant deviations from normality (W = 0.93, P = 0.0005 and W = 0.92, P = 0.0142, respectively). A normal distribution was recovered by logarithmic transformation of both GA and MTA data (W = 0.98, P = 0.633 and W = 0.96, P = 0.278, respectively).
The quantitative variables (trait measurements) showing the lowest correlation among themselves were standardized and retained as independent predictors of species membership. From this set of variables, linear combinations (canonical variables) that summarize between-species variation were obtained by a linear discriminant function analysis using SYSTAT 5.0 for the Macintosh.
Only D. simulans (California and Colombia), D. mauritiana 72, and D. sechellia 21 were used for interspecific crosses. The dominance relationship of the parents to their respective hybrids was scored as h:
[Mathematical Expression Omitted], (1)
where [Mathematical Expression Omitted] is the mean phenotypic value of the interspecific hybrid obtained from the cross between species, [Mathematical Expression Omitted] is the mean phenotype for species A, and [Mathematical Expression Omitted] is the mean phenotype for species B.
TABLE 2. Proportion of variance between strains within species and between species obtained from the analysis of variance of each trait among the species of the melanogaster complex. Variance components Between strains Trait within species Between species Sexual Testes length 0.0788 0.8885 Testes area 0.0974 0.8318 Genital arch area 0.0088 0.9867 Nonsexual Tibia length 0.1296 0.7232 Femur length 0.1619 0.6071 Wing length 0.1095 0.6920 Wing width 0.1620 0.6071 Malpighian tubules length n/a(1) 0.5475 Malpighian tubules area n/a 0.6424 1 n/a: not available.
A multidimensional representation of the total traits measurement variation among species of the simulans clade (all strains) and the hybrids was summarized by a principal component analysis using JMP version 2.0.2 for the Macintosh. From this analysis, the minimum number of independent linear combinations (components) of the standardized measurements that explained the major part of the variation were retained. Only individuals from which GA, TL, TA, TiL, FeL, WL, and WW were measured, were included in the analysis. MT data were not included, because these measurements came from a different set of individual flies.
Analysis of Asymmetry
Drosophila simulans (California and Colombia), D. mauritiana 72, D. sechellia 21, and their respective interspecific hybrids were assessed for levels of asymmetry of sexual and nonsexual, internal and external morphological traits. Wing length, tibia length, the area of the posterior lobe of the genital arch, testes length and area, and malpighian tubules length and area were analyzed. The two sides of internal traits could not be distinguished as left or right, and all comparisons of trait asymmetry were made by using the absolute difference between sides of a trait. Three replicate measurements of each side of a trait were taken.
Phenotypic Variation within and between Species
Within each species sexual traits had less variation among strains than nonsexual traits (Table 2). The between-species comparisons revealed higher sexual than nonsexual trait variation, and not surprisingly, all traits exhibited significantly more variation among species than within species (Table 2).
Nonsexual traits were higher in mean value for D. melanogaster than the three species of the simulans clade, and D. sechellia appeared slightly smaller than the others. In contrast, sexual traits showed D. sechellia closer to D. melanogaster in mean value, although for testes area D. sechellia was closer to D. simulans (Table 3).
The product-moment correlations were significant and very close to perfect linearity for wing length and width (r = 0.977, P [less than] 0.01), tibia length and femur length (r = 0.953, P [less than] 0.01), malpighian tubules length and area (r = 0.949, P [less than] 0.05), and testes length and area (r = 0.843, P [less than] 0.01; [ILLUSTRATION FOR FIGURE 2 OMITTED]). However, D. sechellia seemed to fall off from a linear relationship between testes length and area, which suggests that testes differences among these closely related species involved a slight change in shape besides the change in size that is common to all traits [ILLUSTRATION FOR FIGURE 2 OMITTED]. When a linear function was fitted to the testes data, D. sechellia strains did not behave as outliers (i.e., standardized residuals were less than two) but the standardized residuals were consistently in the negative range [ILLUSTRATION FOR FIGURE 3A OMITTED]. Except for testes, the standardized residuals show a random distribution around the linear fit [ILLUSTRATION FOR FIGURE 3 OMITTED]. The unique pattern of positive deviations for D. melanogaster-D. simulans and negative for D. sechellia-D. mauritiana seen for testes measurements [ILLUSTRATION FOR FIGURE 3A OMITTED] supports a change in shape during the evolution of testes morphology in the species of the melanogaster complex.
The shape of the posterior lobe of the genital arch has been the only reliable morphological character used to classify the four species of the melanogaster complex (Ashburner 1989b), and its mean area was the measurement that most clearly distinguished the four species in this study (Table 3). The slightly higher divergence for sexual than nonsexual traits detected from the ANOVA suggested that perhaps sexual traits as a whole could be used as predictors of species membership. However, the efficiency of a particular trait to predict species membership could be influenced by correlation to other traits. Given the high correlation detected for measurements taken on the same character, the number of variables was reduced or combined to redefine variables with the lowest correlation among themselves. Malpighian tubules were excluded because data were not obtained from D. melanogaster, and only one strain of D. mauritiana and D. sechellia were measured. The multicollinearity among variables was tested based on estimates of the variance inflation factor (VIF), a function of the multiple correlation coefficient ([R.sub.i]) of an ith given variable and the remaining variables (Afifi and Clark 1990). Tibia length divided by femur length (TiL/FeL), wing length divided by wing width (WL/WW), testes length (TL), and the area of the posterior lobe of the genital arch (GA) gave the lowest VIF estimates, indicating independence among themselves (TiL/FeL = 1.03; WL/WW = 1.11; TL = 1.16; GA = 1.10).
Linear combinations of these four variables that summarize the variation between the grouping variables (species) were obtained by a discriminant function analysis. The first two discriminant functions ([V.sub.i]) showed the stronger correlation with the grouping factor (species) and they explained 99% of the between-species variation (Table 4). GA was the original variable with the highest loading associated to the first discriminant function, and TL had its highest discriminant coefficient associated with the second function. The highest coefficients for the third function were for WL/WW and TiL/FeL (Table 4). A plot of the first two discriminant function scores obtained for each individual gives the maximum possible [TABULAR DATA FOR TABLE 3 OMITTED] separation among the species. The first function scores separated all the species in the sample except for partial overlap between D. sechellia and D. melanogaster [ILLUSTRATION FOR FIGURE 4A OMITTED]. This result agrees with the separation obtained between the same species based on the size of the posterior lobe of the genital arch (Liu et al. 1996). Drosophila mauritiana/simulans and D. melanogaster/sechellia were separated by the second discriminant function, which had the highest coefficient associated with TL [ILLUSTRATION FOR FIGURE 4A OMITTED]. Figure 4b shows that the function with high coefficients for WL/WW and TiL/FeL ([V.sub.3]) did not discriminate between species.
Species Hybrids: Genetic Dominance in Sexual versus Nonsexual Traits
The size traits (wings and legs) showed a consistent pattern of overdominance as the hybrids had mean phenotypic values higher than both parents (see Appendix; [ILLUSTRATION FOR FIGURE 5 OMITTED]). Malpighian tubules measurements showed average underdominance (length) or overdominance (area; [ILLUSTRATION FOR FIGURE 5 OMITTED]), although the error associated with these measurements made the conclusions uncertain (Appendix). Sexual traits behaved additively in average value [ILLUSTRATION FOR FIGURE 5 OMITTED], but testes length and area showed a tendency toward dominance of the male parental species (Appendix). For testes length, this tendency was greater for the crosses between D. simulans and D. mauritiana, and D. sechellia female and D. simulans male. Crosses between D. sechellia female and D. mauritiana male and D. simulans female and D. sechellia male were closer to additivity (Appendix).
To represent the dominance relationship between parental species and hybrids, the original variables were transformed into uncorrelated variables by a principal component analysis. Only the first three components were used to represent dominance, because they explain most of the variance in the original data (Table 5). Wing and leg measurements had the largest coefficients, and hence the highest association, with the first component (PC1), testes with the second component (PC2), and genital arch with the third component (PC3; Table 5). The individual scores obtained from each component were plotted against each other and are shown in Figure 6. The first component (PC1), in which the bulk of the variance is due to nonsexual traits, showed overdominance as the hybrid scores laid on the positive range and beyond the distribution of the parental species. The hybrid scores for testes (PC2 axes) were additive in mean value, but closer to the male parental species. Genital arch area scores (PC3 axes) were additive.
Thus, the results showed that the phenotypic distributions of non sexual traits (malpighian tubules, legs, and wings) in the hybrids are beyond the parental range, whereas sexual traits were either additive or dominant.
Species Hybrids: Homeostasis of Traits
To compare asymmetry across traits, the absolute difference between sides was divided by the trait's size. The size of the trait was estimated as the average between sides. A logarithmic transformation of these measurements homogenized the variances across traits (O'Brien, Brown-Forsythe and Levene tests: [F.sub.6,716] [less than] 1.0, P [greater than] 0.1) but their distributions remained skewed.
Internal traits were consistently more asymmetric than external characters, and no pattern was detected based on the sexual versus nonsexual classification of traits [ILLUSTRATION FOR FIGURE 7 OMITTED]. It could be argued that a trait's asymmetry may be a consequence of measurement error. However, no association was observed between the degree of trait's asymmetry and differences among replicate measurements. For example, one of the least asymmetric traits (tibia length, see [ILLUSTRATION FOR FIGURE 7 OMITTED]) showed the largest difference between replicates (Kruskal-Wallis test: [Mathematical Expression Omitted], P = 0.518).
For each trait, differences in asymmetry between parental species and their hybrids were compared by using the logarithmic transformation of the absolute difference between sides. Sexual asymmetries were always significantly higher for interspecific hybrids than parental species, although the difference showed only borderline significance for testes area (Mann-Whitney test: Z = -1.90, [N.sub.H] = 65, [N.sub.P] = 38, P = 0.057; [ILLUSTRATION FOR FIGURE 8A OMITTED]). Except for malpighian tubules length, nonsexual traits showed a higher parental than hybrid asymmetry but the differences were not significant for wing length (Mann-Whitney test: Z = 1.19, [N.sub.H] = 63, [N.sub.P] = 49, P = 0.234) and malpighian tubules area (Mann-Whitney test: Z = -1.34, [N.sub.H] = 64, [N.sub.P] = 38, P = 0.179; [ILLUSTRATION FOR FIGURE 8B OMITTED]).
Sexual Traits: Interspecific Divergence and Sexual Selection
Our results showed higher divergence for sexual than nonsexual traits between species of the D. melanogaster complex and sexual traits were better predictors of species' distinctness than nonsexual traits. Our previous results on protein divergence among Drosophila species showed the same pattern: proteins expressed in testes and ovaries were more diverged among species than nonreproductive tract proteins (Civetta and Singh 1995). The emerging picture from DNA sequence analysis also shows that sex-related genes, that is, genes expressed in the male genitalia, involved in sex determination, or linked to mating behavior, have a larger number of amino acid replacements between species (Aguade et al. 1992; O'Neil and Belote 1992; Karotam et al. 1993). This pattern seems to be quite general in animal species (Tucker and Lundrigan 1993; Whitfield et al. 1993; Lee et al. 1995; Swanson and Vacquier 1995).
Why do sexual traits coincide in a common theme of high interspecific divergence? A few possible explanations follow.
According to the neutral model of evolution (Kimura 1983), the high divergence of male sexual traits may be the result of a higher mutation rate or a less constrained selection regime. based on such a model, variation within species should positively correlate with the variation detected between species. In our study, this trend is reversed: Sexual traits showed lower variation within but higher divergence between species than nonsexual traits. Patterson and Stone's (1952, see [ILLUSTRATION FOR FIGURES 1-18 OMITTED]) description of the male and female reproductive system of different Drosophila groups indicates that species with highly coiled and long testes also have long female seminal receptacles, whereas species with short testes have short seminal receptacles. A positive correlation between the length of the seminal receptacle and sperm length has been reported among species of the D. nanoptera group, and the seminal receptacle was the only female reproductive tract trait that showed significant divergence between species (Pitnick and Markow 1994). Similar results have been found for passerine birds, where sperm length and the length of female sperm storage tubules are positively correlated (Briskie and Montgomerie 1992). The coevolving nature of these male and female traits does not seem to support neutrality as a possible explanation for the divergence detected between species.
Alternatively, Paterson's recognition species concept (Paterson 1985) and Carson's organization theory of speciation (Carson 1985) suggest strong stabilizing selection on sexual characters within species. An episodic relaxation of such selective force is then required at the time of speciation. If stabilizing selection causes sexual traits to be well coadapted within species, then low variation and low asymmetry is expected within species, because any deviation from the norm becomes detrimental. Our results showed that testes length and area and the area of the posterior lobe of the genital arch are less variable within species than nonsexual traits. Although this result is in line with Carson's and Paterson's predictions, the variation within species was significant for sexual traits, suggesting that these traits may not be so tightly buffered by stabilizing selection. This conclusion is supported by the analysis of within-species trait asymmetry, because sexual traits were not less asymmetric than nonsexual traits. The level of asymmetry is dependent on how selection prevents a phenotype from randomly deviating from its norm. The phenotypes with the lowest asymmetry were wing and tibia length, two traits that are probably under strong stabilizing selective pressures because random variations between sides may affect locomotion and dispersal. Malpighian tubules length had the highest asymmetry, and there are no obvious reasons as to why fluctuations between sides in the length of the stalks connecting the malpighian tubule to the pyloric region of the ventriculus should affect the fitness of the organism.
TABLE 4. Results obtained from discriminant analysis. Each [V.sub.i] is a variable defined by the linear combination of the trait variables. The canonical squared correlations are between the [V.sub.i] variables and a corresponding linear function that results from the combination of the grouping variables (species). The eigenvalues show the variation among species explained by each linear function [V.sub.i]. For each linear combination ([V.sub.i]), a coefficient (or loading) is associated with each trait. Discriminant functions [V.sub.1] [V.sub.2] [V.sub.3] Squared correlation 0.986 0.894 0.257 Eigenvalue 75.34 8.57 0.35 Proportion of total variation 0.89 0.10 0.01 Variable Coefficients GA 1.07 -0.06 0.02 TL 0.20 1.03 0.01 WL/WW 0.30 0.21 0.86 TiL/FeL -0.16 0.03 0.69
Finally, directional sexual selection becomes an alternative explanation for the high between-species divergence observed in sexual traits. Models of speciation by sexual selection require within-species variation (Lande 1981; Kaneshiro 1989; Iwasa and Pomiankowski 1995), and empirical data show high variability for secondary sexual traits (Cade 1981; Carson and Lande 1984; Hedrick 1988; Houde 1992; Wilcokson et al. 1995) as well as mating behavior (Cade 1984) in a wide variety of organisms. Secondary sexual traits have also shown higher levels of asymmetry than other morphological characters, and it has been proposed that directional sexual selection on secondary sexual traits may act on modifier genes that reduce developmental control leading to rapid divergence and increased asymmetry (Moller 1993; Pomiankowski and Moller 1995). The intermediate level of asymmetry found for sex-related traits do not perfectly fit the expectations for a character under continuous directional sexual selection.
However, a model of speciation that relies on directional sexual selection on sex-related traits does not require a continuous selective pressure on such traits. This is contrary to models of speciation that rely on a random disruption of well coadapted traits that are continually buffered by stabilizing selective pressures within species (Carson 1982, 1985). Directional sexual selection at the time of species formation may have triggered the rapid divergence observed for sexual traits, but once speciation is completed, the selective pressure may become relaxed.
Species Hybrids and the Genetic Architecture of Sexual versus Nonsexual Traits
Except for malpighian tubule length, the mean phenotypic values for nonsexual traits of interspecific hybrids were constantly above the range of the parental species. This phenomenon was described by Dobzhansky (1952), who coined the term "luxuriance" to distinguish it from heterosis resulting from crosses between different varieties of the same species. Wright (1977) also refers to luxuriance and suggests that it may have more to do with the defective regulation of growth in the interspecific hybrids than a complementary action of the parental genes. A phenotype resulting from a misregulation of growth should be poorly buffered and thus more asymmetric than the parental forms. Our results for nonsexual traits showed not only a phenotypic distribution of the hybrids beyond the parental species, but also less asymmetry than the parental species. This is what has been observed for heterozygote hybrids obtained from crosses between different strains (for reviews, see Palmer and Strobeck 1986; Parson 1990). Both the overdominant effect and the lower asymmetry in the interspecific hybrids than the parental species suggest allele complementation in the hybrids.
Sexual traits did not manifest the phenomenon of luxuriance. The posterior lobe of the genital arch showed an intermediate phenotype in interspecific hybrids, whereas testes showed an average additive effect with a trend toward paternal dominance. Before attempting an explanation for the different phenotypic behavior of the two sex traits analyzed, it should be noted that similar differences in phenotypic behavior of sexual versus nonsexual traits can be found in previous studies that focus on other aspects of species morphological characteristics. Coyne (1983) and Liu et al. (1996) showed that the genital arch in male hybrids between D. simulans and D. mauritiana is intermediate in size between those of their parents. Coyne (1985) showed that the number of male sex comb teeth is intermediate in hybrids between D. simulans and D. mauritiana. Coyne et al. (1991) carried on an analysis of the genetic basis of morphological differences among the species of the melanogaster complex. Their measurements of tibia and femur length showed higher hybrid than parental species phenotypic values, whereas ovariole number and genital arch area behaved additively. The pattern looks constant throughout the literature and it is possible that the additivity versus overdominance behavior might pertain to sexual versus nonsexual traits.
Although the interspecific hybrid testes morphology was additive on average, this trait showed a less consistent pattern than the others. The slight differences between hybrids for the phenotypic distribution of testes may relate to the complexity of its genetic basis. Joly et al. (1997) tried to elucidate the genetic basis of sperm and testes length through backcross analysis between D. simulans and D. sechellia and the use of phenotypic markers. They obtained the same result of testes length additivity we observed for the [F.sub.1] hybrid between D. simulans female and D. sechellia male but the backcross to D. simulans males showed a complicated range of phenotypes that, with the exception of one particular backcross type, ranged from the average distribution between the parental species to the average phenotype of the original D. sechellia male parent. The study suggested the existence of autosomal factors on both arms of chromosome 3 of D. sechellia interacting with D. sechellia chromosome 2 as responsible for the enhanced length of the testes in the backcrossed male progeny (Joly et al. 1997). The dominance effect seen in our study may be the result of either similar autosomal effects or Y chromosome effects. Interestingly, only in crosses involving D. sechellia males, the hybrids remained closer to additivity suggesting that the Y chromosome of D. mauritiana and D. simulans, but not D. sechellia, may have an effect on the length of testes. The genetic basis of testes morphology seem quite complex, and it will require further genetic analysis to properly elucidate it.
The lack of overdominant interactions among sex genes in the hybrids together with the higher sex trait's asymmetry in hybrids than in the parental species suggest that sexual trait genes have rapidly diverged between the closely related species of the simulans clade to the point of having lost their ability to properly interact in the hybrids. This phenotypic disruption seems to be enhanced in hybrids between more distant species such as D. melanogaster and D. sirnulans, in which testes become atrophied (Pontecorvo 1943). It is possible that the complex phenotypic behavior of testes in the different hybrids between the three species of the simulans clade represents a first step toward phenotypic breakdown.
Conclusion: The Evolutionary Potential of Sexual Traits in Speciation
It could be argued that postmating barriers, such as hybrid sterility, might not be the first step in speciation. For example, D. heteroneura and D. silvestris, two well-recognized species of the Hawaiian group, are interfertile in the lab but are strongly isolated in mate-choice experiments (Kaneshiro 1976; Carson 1978). The two species differ in two male secondary sexual characters, one involved in male-male competition (the head; Spieth 1981) and the other whose function relates to the stimulation of the female's abdomen during courtship (the male tibial cilia; Carson et al. 1982). These kind of observations have generated speciation models that emphasize the role of the behavioral components of the reproductive system as key traits responsible for species identity (Carson 1985; Paterson 1985).
TABLE 5. Results obtained from the principal component analysis on the species and hybrids data. The eigenvalues show the variance in the data explained by each component. The value [a.sub.ij] is the coefficient of the jth variable for the ith principal component. Principal components PC1 PC2 PC3 Eigenvalues 3.18 1.96 0.98 Percent of variation 45.4 28.0 14.0 Cumulative percent 45.4 73.4 87.4 Traits [a.sub.ij] Testes length -0.121 0.658 -0.117 Testes area 0.159 0.630 0.167 Genital arch 0.294 0.168 0.781 Wing length 0.474 -0.281 0.115 Wing width 0.517 -0.122 0.059 Tibia length 0.460 0.097 -0.401 Femur length 0.411 0.197 -0.413
However, Haldane's observation that in the [F.sub.1] offspring of two different species the heterogametic sex is rare, absent, or sterile (Haldane 1922) has become a keystone for those attempting to resolve the speciation problem. Wu and Davis (1993) have shown that in groups such as Drosophila and mammals, crosses between species almost always result in hybrid male sterility, whereas hybrid females are either inviable or sterile among Lepidoptera and birds. The results suggest that the high proportion of hybrid male sterility in Drosophila and mammals is not only a condition brought about by the heterozygote nature of the sex chromosomes in the males. The diminished hybrid male fertility may also be the result of an easier disruption of the male senitalia during the early stages of speciation.
Based on these previous observations, the high betweenspecies divergence of sexual morphological traits not directly involved in copulation, their higher asymmetry in hybrids than parental species, their [F.sub.1] hybrid dominance relationship, and our previous result showing high between-species divergence of proteins expressed in the male reproductive tract that are not necessarily involved in reproduction (Coulthart and Singh 1988; Thomas and Singh 1992; Civetta and Singh 1995), we suggest that all aspects of sexuality undergo rapid changes during speciation. Regardless of whether pre- or postmating barriers establish isolation between species, the common theme seems to be the major role played by sexual traits in the establishment of such barriers.
Finally, the high divergence between species coupled with the within-species variation found for the sexual traits analyzed in this study, and the fact that sexual traits were not less asymmetric than nonsexual traits, argue in favor of their role in speciation through directional sexual selection coupled with population isolation rather than strong stabilizing selection and episodic random perturbations of the genome though founder events. Sexual selection has been traditionally linked to behavioral or morphological characters directly involved in male-male competition, courtship, or mating (Andersson 1994). We propose a distinction between this original narrow-sense sexual selection definition and a broader selective process affecting all aspects of sexuality (molecular and morphological) not directly linked to courtship or mating (i.e., broad-sense sexual selection).
We would like to thank R. A. Morton and A. G. Clark, whose valuable comments have contributed to the improvement of earlier versions of the manuscript. This work was supported by a research grant from the Natural Sciences and Engineering Research Council of Canada to RSS.
AFIFI, A. A., AND V. CLARK, 1990. Computer-aided multivariate analysis. 2d. ed. Van Nostrand Reinhold Co., New York.
AGUADE, M., N. MIYASHITA, AND C. H. LANGLEY. 1992. Poly- morphism and divergence in the Mst26A male accessory gland gene region in Drosophila. Genetics 132:755-770.
ANDERSSON, M. 1994. Sexual selection. Princeton Univ. Press, Princeton, NJ.
ASHBURNER, M. 1989a. Drosophila a laboratory manual. Cold Spring Harbor Laboratory Press, New York.
-----. 1989b. Drosophila a laboratory handbook. Cold Spring Harbor Laboratory Press, New York.
BRESSAC, C., A. FLEURY, AND D. LACHAISE. 1994. Another way of being anisogamous in Drosophila subgenus species: giant sperm, one-to-one gamete ratio, and high zygote provisioning. Proc. Nat. Acad. Sci. USA 91:10399-10402.
BRISKIE, J. V., AND R. MONTGOMERIE. 1992. Sperm size and sperm competition in birds. Proc. R. Soc. Lond. B Biol. Sci. 247:89-95.
CADE, W. H. 1981. Alternative male strategies, genetic differences in crickets Gryllus integer. Science 212:563-564.
-----. 1984. Genetic variation underlying sexual behaviour and reproduction. Am. Zool. 24:355-366.
CARSON, H. L. 1975. The genetics of speciation at the diploid level. Am. Nat. 109:83-92.
-----. 1978. Speciation and sexual selection in Hawaiian Drosophila. Pp. 93-107 in P. F. Brussard, ed. Ecological genetics: the interface. Springer-Verlag, New York.
-----. 1982. Speciation as a major reorganization of polygenic balances. Pp. 421-433 in C. Barigozzi, ed. Mechanisms of speciation. Liss, New York.
-----. 1985. Unification of speciation theory in plants and animals. Syst. Bot. 10:380-390.
CARSON, H. L., AND R. LANDE. 1984. Inheritance of a secondary sexual character in Drosophila silvestris. Proc. Nat. Acad. Sci. USA 81:6904-6907.
CARSON, H, L., F. C. VAL, C. M. SIMON, AND J. W. ARCHIE. 1982. Morphometric evidence for incipient speciation in Drosophila silvestris from the island of Hawaii. Evolution 36:132-140.
CIVETTA, A., AND R. S. SINGH. 1995. High divergence of reproductive tract proteins and their association with postzygotic reproductive isolation in Drosophila melanogaster and Drosophila virilis group species. J. Mol. Evol. 41:1085-1095.
COULTHART, M. B., AND R. S. SINGH. 1988. High level of divergence of male-reproductive tract proteins, between Drosophila melanogaster and its sibling species, D. simulans. Mol. Biol. Evol. 5:182-191.
COYNE, J. A. 1983. Genetic basis of differences in genital morphology among three sibling species. Evolution 37:1101-1118.
-----. 1985. Genetic studies of three sibling species of Drosophila with relationship to theories of speciation. Genet. Res. Camb. 46:169-192.
COYNE, J. A., J. RUX, AND J. R. DAVID. 1991. Genetics of morphological differences and hybrid sterility between Drosophila sechellia and its relatives. Genet. Res. Camb. 57:113-122.
DEMEREC, M. 1950. Biology of Drosophila. Hafner, New York.
DOBZHANSKY, TH. 1952. Nature and origin of heterosis. Pp. 218-223 in J. W. Gowen, ed. Heterosis. Iowa State College Press, Ames, IA.
-----. 1970. Genetics of the evolutionary process. Columbia Univ. Press, New York.
EBERHARD, W. G. 1985. Sexual selection and animal genitalia. Harvard Univ. Press, Cambridge, MA.
EBERHARD, W. G., AND C. CORDERO. 1995. Sexual selection by cryptic female choice on male seminal products a new bridge between sexual selection and reproductive physiology. Trends Ecol. Evol. 10:493-496.
HALDANE, J. B. S. 1922. Sex ratio and unisexual sterility in hybrid animals. J. Genetics 12:101-109.
HEDRICK, A. V. 1988. Female choice and the heritability of attractive male traits: an empirical study. Am. Nat. 132:267-276.
HOUDE, A. E. 1992. Sex-linked heritability of a sexually selected character in a natural population of Poecilia reticulata (Pisces: Poeciliidae) (guppies). Heredity 69:229-235.
IWASA, Y., AND A. POMIANKOWSKI. 1995. Continual change in mate preferences. Nature 377:420-422.
JOLY, D., C. BRESSAC, J. DEVAUX, AND D. LACHAISE. 1991. Sperm length diversity in Drosophilidae. Drosophilia Inf. Serv. 70:104-108.
JOLY, D., C. BRESSAC, AND D. LACHAISE. 1995. Disentangling giant sperm. Nature 377:202.
JOLY, D., C. BAZIN, L.-W. ZENG, AND R. S. SINGH. 1997. Genetic basis of sperm and testis length differences and epistatic effect on hybrid inviability and sperm motility between Drosophila simulans and D. sechellia. Heredity 78:354-362.
KANESHIRO, K. Y. 1976. Ethological isolation and phylogeny in the planitibia subgroup of Hawaiian Drosophila. Evolution 30: 740-745.
-----. 1989. The dynamics of sexual selection and founder effects in species formation. Pp. 279-296 in L. V. Giddings, K. Y. Kaneshiro, and W. W. Anderson, eds. Genetics, speciation, and the founder principle. Oxford Univ. Press, New York.
KAROTAM, J., A. C. DELVES, AND J. G. OAKESHOTT. 1993. Conservation and change in structural and 5[prime] flanking sequences of esterase 6 in sibling Drosophila species. Genetica 88:11-28.
KARR, T. L., AND S. PITNICK. 1996. The ins and outs of fertilization. Nature 379:405-406.
KIMURA, M. 1983. The neutral theory of molecular evolution. Cambridge Univ. Press, Cambridge, U.K.
LANDE, R. 1981. Models of speciation by sexual selection on polygenic traits. Proc. Nat. Acad. Sci. USA 78:3721-3725.
LEE, Y.-H., T. OTA, AND V. D. VACQUIER. 1995. Positive selection is a general phenomenon in the evolution of abalone sperm lysin. Mol. Biol. Evol. 12:231-238.
LIU, J., J. M. MERCER, L. F. SWAM, G. C. GIBSON, Z.-B. ZENG, AND C. C. LAURIE. 1996. Genetic analysis of a morphological shape difference in the male genitalia of Drosophila simulans and D. mauritiana. Genetics 142:1129-1145.
LONG, A. D., AND R. S. SINGH. 1995. Molecules versus morphology: the detection of selection acting on morphological characters along a cline in Drosophila melanogaster. Heredity 74: 569-581.
MAYR, E. 1963. Animal species and evolution. Harvard Univ. Press, Cambridge, MA.
MOLLER, A. P. 1993. Developmental stability, sexual selection and speciation. J. Evol. Biol. 6:493-509.
O'NEIL, M. T., AND J. M. BELOTE. 1992. Interspecific comparison of the transformer gene of Drosophila reveals an unusually high degree of evolutionary divergence. Genetics 131:113-128.
PALMER, A. R, AND C. STROBECK. 1986. Fluctuating asymmetry: measurement, analysis, patterns. Ann. Rev. Ecol. Syst. 17:391-421.
PARSON, P. A. 1990. Fluctuating asymmetry: an epigenetic measure of stress. Biol. Rev. 65:131-145.
PATERSON, H. E. H. 1985. The recognition concept of species. Pp. 21-29 in E. S. Vrba, ed. Species and speciation. Transvaal Museum Monograph no. 4. Pretoria, South Africa.
PATTERSON, J. T., AND W. S. STONE. 1952. Evolution in the genus Drosophila. Macmillan, New York.
PITNICK, S., AND T. A. MARCOW. 1994. Male gametic strategies: sperm size, testes size, and the allocation of ejaculate among successive mates by the sperm-limited fly Drosophila pachea and its relatives. Am. Nat. 143:785-819.
POMIANKOWSKI, A., AND A. P. MOLLER. 1995. A resolution of the lek paradox. Proc. R. Soc. Lond. B Biol. Sci. 260:21-29.
PONTECORVO, G. 1943. Hybrid sterility in artificially produced recombinants between Drosophila melanogaster and D. simulans. Proc. R. Soc. Edinburgh B 41:385-397.
SHAPIRO, S. S., AND M. B. WILK. 1965. An analysis of variance test for normality (complete samples). Biometrika 52:591-611.
SPIETH, H. T. 1981. Drosophila heteroneura and Drosophila silvestris: head shapes, behavior and evolution. Evolution 35:921-930.
SWANSON, W. J., AND V. D. VACQUIER. 1995. Extraordinary divergence and positive Darwinian selection in a fusagenic protein coating the acrosomal process of abalone spermatozoa. Proc. Nat. Acad. Sci. USA 92:4957-4961.
TEMPLETON, A. R. 1980. The theory of speciation via the founder principle. Genetics 94:1011-1038.
-----. 1981. Mechanisms of speciation-a population genetics approach. Annu. Rev. Ecol. Syst. 12:23-48.
THOMAS, S., AND R. S. SINGH. 1992. A comprehensive study of genic variation in natural populations of Drosophila melanogaster. VII. Varying rates of genic divergence as revealed by two-dimensional electrophoresis. Mol. Biol. Evol. 9:507-525.
TUCKER, P. K., AND B. L. LUNDRIGAN. 1993. Rapid evolution of the sex determining locus in old world mice and rats. Nature 364:715-717.
WHITFIELD, L. S., R. LOVELL-BADGE, AND P. N. GOODFELLOW. 1993. Rapid sequence evolution of the mammalian sex-determining gene SRY. Nature 364:713-715.
WILCOCKSON, R. W., C. S. CREAN, AND T. H. DAY. 1995. Heritability of a sexually selected character expressed in both sexes. Nature 374:158-159.
WRIGHT, S. 1977. Evolution and the genetics of populations. Vol. 3. Experimental results and evolutionary deductions. Univ. of Chicago Press, Chicago.
WU, C.-I, AND A. W. DAVIS. 1993. Evolution of postmating reproductive isolation: the composite nature of Haldane's rule and its genetic basis. Am. Nat. 142:187-212.
|Printer friendly Cite/link Email Feedback|
|Author:||Civetta, Alberto; Singh, Rama S.|
|Date:||Aug 1, 1998|
|Previous Article:||Cytological characterization of premeiotic versus postmeiotic defects producing hybrid male sterility among sibling species of the Drosophila...|
|Next Article:||A molecular phylogeny of the Drosophila willistoni group: conflicts between species concepts?|