A occurring X-chromosome known as the sex-ratio (SR) chromosome (LAK-

Transcription

1 Copyright 1986 by the Genetics Society of America AN INVESTIGATION OF THE GENETICS OF MALE SEX-RATIO PHENOTYPE IN DROSOPHILA PSEUDOOBSCURA GARY COBBS Department of Biology, University of Louisville, Louisville, Kentucky 4292 Manuscript received May 27, 1985 Revised copy accepted February 28, 1986 ABSTRACT A laboratory strain of Drosophila pseudoobscura (L116) is studied that, when crossed to sex-ratio homozygous females, produces sons that exhibit varying levels of the male sex-ratio (msr) phenotype. The msr phenotype occurs only in sex-ratio males and is due to the production of a high frequency of nullo-xy sperm. The level of the msr phenotype is variable, and new variability is generated in one father-son transmission. Pedigree studies indicate the genes for msr reside on the Y chromosome or the autosomes of the L116 stock. number of species within the Drosophila obscura group contain a naturally A occurring X-chromosome known as the sex-ratio (SR) chromosome (LAK- OVAARA and SAURA 1982). The SR chromosomes, when present in males, causes them to produce almost entirely female progeny. Drosophila pseudoobscura SR/Y males usually produce progeny that are % female (POLI- CANSKY and DEMPSEY 1978); this is called the sex-ratio (sr) phenotype. The progeny of wild-type (ST/Y) males are composed of roughly one-half fertile males. The rare male progeny from SR/Y males are sterile and are thought to be X/O karyotypes (STURTEVANT and DOBZHANSKY 1936; HENAHAN and COBBS 1983). The mechanisms responsible for the sr phenotype are incompletely understood. POLICANSKY and ELLISON (197) report that, in D. pseudoobscura, mature sperm bundles of SR/Y males contain only half as many spermatids as those of ST/Y males, whereas the immature sperm bundles of SR/Y and ST/Y are not different. From this they infer that the sr phenotype is due to spermiogenic failure of Y-bearing sperm. Further support for this model is given by POLICANKSY (1974, 1979). The genetic basis for the sr phenotype also is incompletely known. Drosophila pseudoobscura SR chromosomes always differ from standard X-chromosomes by three nonoverlapping inversions in the right arm. WU and BECKENBACH (1983) performed crosses which suggest that the SR chromosome carries several genetic elements, with at least one in each inversion, that all must be present to produce the sr phenotype. Aberrant behavior of SRIY males is occasionally observed in which an unu- Genetics 113: June, 1986.

2 356 G. COBBS sually high frequency of male progeny is produced. This has been termed the "male sex-ratio" (nisr) phenotype and was first described in D. affinis by NOV- ITSKI (1947). He also cites unpublished evidence for the existence of msr in D. pseudoobscura. The msr phenotype in D. afjnis was studied by VOELKER (1972) and was found to be due to a particular SR chromosome in a male lacking a Y chromosome. HENAHAN and COBBS (1983) report aberrant behavior of certain SR/Y males in D. pseudoobscura that produced 19.8% male progeny. These male progeny were all sterile and were thought to have an X/O karyotype. Following the terminology of NOVITSKI (1947), the aberrant behavior reported by HENAHAN and COBBS (1983) will be termed the msr phenotype of D. pseudoobscura. It must be kept in mind that the msr phenotypes of D. affinis and D. pseudoobscura, although similar, are not identical. In D. affinis the msr phenotype often produces 1% male progeny which are fertile X/O males. In D. pseudoobscura the msr phenotype involves variable amounts of completely sterile male progeny ranging from 1 to 95% of all progeny. Here, an investigation of the genetic basis for the msr phenotype described by HENAHAN and COBBS (1983) is reported. MATERIALS AND METHODS Strains and crosses: SRI and SR' are "sex-ratio" X-chromosomes isolated from natural populations at James Reserve on Mt. San Jacinto, California and San Diego, California, respectively. L116 is a wild-type strain isolated from Strawberry Canyon, California, in 1966 and obtained for these studies from R. C. LEWONTIN. The sex chromosomes from the L116 stock will be denoted as STL/YL. Stocks WWA-6 and WWA- 3 were obtained from the National Drosophila Species Resource Center at Bowling Green State University. The WWA-6 stock carries an X-linked dominant for pointed wings (Pt), and the sex chromosomes of this stock will be denoted as STP'/YW6; sex-ratio X-chromosomes are denoted as SR and standard sequence wild-type X-chromosomes by ST. The sex-ratio phenotype will be denoted by sr and the male sex-ratio phenotype by msr. The sex chromosomes of the WWA-3 stock will be denoted as STW3/Y3. The SR chromosomes were maintained by aunt-nephew crosses according to BECKENBACH (1978), using the L116 stock as the source of ST/Y males. The goi stock carries autosomal recessives orange () and incomplete (inc) on chromosomes IZ, ZZZ and ZV, respectively, and was obtained from TIMOTHY PROUT. The sex chromosomes from the goi stock will be denoted by Xg" and Yg"'. All crosses were performed on the instant potato medium described by HENAHAN and COBBS (1983) and maintained at 2". Chromosome preparations: Neural ganglia were removed from third-instar larvae in.7% NaCl with low4 M colchicine and were incubated in this medium for 1 hr at room temperature. The ganglia were then treated hypotonically according to HOLMQUIST (1975) and were stained in aceto-orcein (DARLINGTON and LACOUR 196) for 15 min. The ganglia were then squashed under a cover slip and were examined. Experiment 1: Here, the sex composition of progeny from several different parental types is studied. Each cross involved ten males mated en masse to ten females. The parents were transferred at 7-day intervals to fresh food vials, and all adults enclosing from the replicate were scored for sex. The fertility of both male and female progeny was also determined. Experiment 2: Here, a series of crosses are performed to determine if the msr trait maps to the SR chromosome. Ten STL/YL males were crossed en masse to ten SR'/SR' females, and 35 of the resulting FI SR'/YL males were randomly selected and each crossed singly to five STP'/STP' females, which shall be called harem 1. After 14 days, each FI SR'/YL male was mated to a second harem of five STPL/STPf females. Both

3 MALE SEX-RATIO IN DROSOPHILA 357 P F1 slj/* T SR~ /SRI SR1/+ T STPtlSTPt (harems 1 + 2) male i i =1,2,...,35 F2 count 86 i-?? F4 check ferti 1 i ty FIGURE 1.-Outline of experiment 2. The ST" X-chromosome carries the dominant Pt allele, whereas STL and SRI carry the fit+ allele. The Xp' chromosome may be a crossover product, but will carry either the SR genes or the ST genes. Examination of the F4 indicates the SR or ST genotype of the XP' chromosome in the Fs male. harems were transferred at 7-day intervals to fresh food vials. The progeny from harems 1 and 2, called broods 1 and 2, respectively, were analyzed for sex composition and fertility. Within each brood, each vial was scored separately, and all progeny eclosing from a vial during the 2 wk following the first eclosion were scored. Ten brood 1 females were chosen for each of the eight F1 males whose broods showed the highest proportion of males; then each was crossed en masse to ten ST"/Y"' males. Brood 1 females were also chosen from the seven Fl males giving the lowest proportion of males in brood 1, and each was crossed en masse to ten STPf/Y6 males. These 15 crosses each produced Fs males of both pt+ or Pt phenotypes. For each of the 15 broods, ten pt+ males were crossed singly to STP'/STP' females, and the resulting F4 progeny were scored for sex composition and fertility. Experiment 2 is outlined in Figure 1. Experiment 3: Here, a series of crosses are performed to determine if the variability of the msr trait is heritable. Eighteen males were selected from the L116 stock and each singly mated to ten SR'ISR' females creating 18 isogenic-y' clones. Ten F1 SRI/ YL males from the first cross were selected from each clone and were mated singly to a harem of ten SR1/SRP' females and transferred to fresh food vials at 7-day intervals. The FP progeny were then examined for fertility and for the numbers of males and females. Again, all adults eclosing in the 2-wk period following the first eclosion were included. After the first mating, each of the 18 parental males was mated a second time to a harem of ten L116 females to establish Y-chromosome clones for future work. Experiment 4: Here a series of crosses are performed to determine if the L 116 stock

4 358 G. CORRS TABLE 1 Sex composition, fertility and between-vial variability in crosses of SR/Y and ST/Y males in experiment 1 No. of vials -. Average proportion of males k SE 3.17" * " _ " f " f ' * ' f ' f.23 SD among P value ~- <.5 <.1 <o. 1 <.1 >.95 >.9 >O. 1 All males sterile and females fertile. ' Both males and females fertile. chromosomes have rearrangements such as Y-autosome translocation. The method involves testing fo~- independent assortment of the goi stock markers with themselves and with the sex chromosomes. Ten L116 males were crossed en masse to ten goi females. Ten of the resulting F, males were backcrossed to the goi stock, and the remaining F, were allowed to produce an FP. The F2 and the progeny from the backcross were scored for sex and all three mutant markers. Males were also scored for fertility by looking for motile sperm in squashed testes. RESULTS Experiment 1: The average proportion of male progeny over replicate vials for each of seven different crosses are given in Table 1. The msr phenotype is present only in crosses involving SR/YL males. When SR/Y males result from crossing males from either WWA-3 or WWA-6 to SRISR females, a normal sr phenotype is found. Also in Table 1, the L males are found to produce typical sex composition when crossed to L116 or WWA-6 females. Since the sex composition of a brood eclosing from a single vial is the basic observation reported here, it is important to know the amount of intrinsic variability of this measurement in relevant crosses. This is revealed by analyzing the replicate vials in experiment 1, as shown in Table 1. A xp test for homogeneity of sex composition among replicate vials was done for each cross, and the x2 distributed test statistic and its associated degrees of freedom are given in Table 1. The P values for the x2 statistics indicate that for several of the crosses there are significant differences between vials. A measure of the amount of variation is the standard deviation of vials within each cross; this also is reported in Table 1. BARTLETT'S test for homogeneity of variance indicated homogeneity of s2 for all seven crosses of Table 1, and the pooled standard deviation is.45. The cause of vial-to-vial variation within a cross is not revealed in these experiments, but is possibly due to an interaction of factors affecting growth rate, larval density and the difference in development time of males and females. The four crosses in Table 1 that produced sterile progeny were examined further, and in all cases it was found that males were sterile and females fertile.

5 MALE SEX-RATIO IN DROSOPHILA ) a. a a. a.) *a. a a a a...o a. a... O -. a. O O O D O a a.. O a. a. a.o a a. a. 2o I I IO 2?a RANK OF PARENTAL MALE FIGURE 2.-Vertical axis is the percentage of males in F2 vials., a single brood 1 vial;, a single brood 2 vial. The horizontal axis is the rank of the parental male according to the grand average percentage of male offspring. Experiment 2: The proportion of males was averaged over brood 1 and 2 vials for each of the 35 F1 males. The males were then ranked according to this overall average. The results for 143 brood 1 and 2 vials are presented in Figure 2. Figure 2 shows all 35 F1 males gave the msr phenotype and that there are dramatic differences among the parental males. The proportion of F2 males averaged over both broods ranged from a low of.34 to a high of.866, with a grand mean of.61. An additional 13 vials that produced fewer than 2 total F1 flies were not included because sampling variance would be excessively large for these vials. Among the vials presented in Figure 2 the number of F2 flies per vial ranged from a low of 28 to a high of 354, with a mean of and median of Summing over all vials, including those producing fewer than 2 flies, gives a total of 16,83 F1 flies, of which 9654 were sterile males (57.45%). Vials within the 59 different broods included in Figure 2 were analyzed by a homogeneity x2 procedure (total x2 = 27.27; d.f. = 87; P < Thus, there is a highly significant variation in sex composition between vials within broods. Figure 2 also indicates there is a brood effect, because for a given male, the vials of the first brood tend to produce a higher proportion of F2 males than do those of the second brood. In order to assess the significance of all these effects, a three-factor analysis

6 36 G. COBBS TABLE 2 Analysis of variance table resulting from a two-factor analysis on the angular transform of proportion of males in experiment 2 Source ss d.f. MS F P Parental males <.1 Broods <.1 Male-brood interaction.317? <.874 Residual The mean square of vials within broods and males were found to be homogeneous by BAR- TLETT S test. Abbreviations: SS, sum of squares; d.f., degrees of freedom; MS, mean square; F, FISHER S F-ratio; P, probability. of variance (ANOVA) was performed on the angular transform (SOKAL and ROHLF 1981) of the proportion of males, using male number (1-35), brood number (1-2) and vial number (1-3) as the factors. This analysis showed both male (P <.1) and brood (P <.1) effects to be highly significant, whereas the vial effect was not significant (P.2). The lack of a significant vial effect here indicates there is no consistent relationship between the proportion of F1 males and the time at which a vial of eggs is collected from a harem. The highly significant vial variation within broods detected by the x2 analysis will henceforth be regarded as being random variation. Viewing the vial-to-vial variation within broods as random uncontrolled variation, a two-factor analysis of variance was done using male (1-35) and brood (1-2) as the factors. The analysis was done on both the transformed and untransformed proportions, with no change in the conclusions. The resulting ANOVA table for the analysis of transformed proportions is given in Table 2. Both male and brood effects are highly significant with no significant interaction. The parameters of the ANOVA model applied to untransformed proportion indicate that the proportion of sterile males for first broods was, on the average,.1 1 higher than those of second broods, and the residual standard deviation was.949. Matings of FB pt males from the high and low broods produced 144 successful crosses. Each of these F4 progeny was examined to determine the phenotype of the FS male parent (wild type, sr or msr ). Seventy-two Fq progeny were found to contain fertile males and females and 72 were sterile with varying amounts of sterile males. The fact that half the FB pt males were wild type indicates no linkage of Pt with the SR genes. Twenty-five of the sterile F4 progeny were randomly chosen, and the proportion of males among those ranged from ( of 134) to.53 (11 of 197), with a mean standard error of.188 f.3. Experiment 3: The proportion of males was averaged over all F2 vials for each FI male. A grand mean over all FI males was computed for each of the 18 clones, and the clones were then ranked according to this grand mean. Figure 3 shows the distribution of the F1 male means for each of the 18 clones and is derived from counts of 355 F2 vials that produced 2 or more progeny. Forty-one F2 vials were not included in the analysis because they produced less

7 MALE SEX-RATIO IN DROSOPHILA gd o m o o m - Q 8 o m Qo o m Q)o QOD QD o m - Q) o a o o o c m Q) I II RANK OF CLONE FIGURE 3.-Vertical axis is the average percentage of male progeny in the F2 vials of experiment 2. The average is taken over all F2 vials for each F, male within each clone. The horizontal axis is the rank of the clone according to the average percentage of F2 males over all Ft males. than 2 total progeny. Most of the points in Figure 3 are averages of three or more F2 vials, although a few result from either two or one F2 vials. The number of F2 flies per vial ranged from a low of 2 to a high of 367, with a mean of and median of 18. The overall grand mean of the proportion of F2 males over all the vials is.647, with a standard error of.9. Summing all F2 vials gives a total of 41,23 F2 flies, of which 25,886 were sterile males (63.1%). Also, as found in experiment 1, the replicate vials were found to be heterogeneous by x2 analysis (P <.1). This vial-to-vial variation within F2 males is assumed to be random. Figure 3 reveals that there is extensive variability in proportion of sterile males among F1 males within clones, as well as between clones. To assess the

8 362 G. CORRS TABLE 3 Analysis of variance table resulting from a two-level nested analysis on the angular transform of the proportion of males in experiment 3. Soiircr SF d.f. MS F P Between clones <o.oo I FI males within clones co. I Vials within F, niales l he mean square of vials within FI males and the mean square of FI males within clones were both found to be homogeneous by BARTLETT S test. Abbreviations: SS. sun1 of squares: d.f., degrees of freedom; MS. mean square; F, FISHER S F-ratio; P. probability. c. rq FIGURE 4.-Neural phenotype. ganglia karyotypes. A. L116 male; B, male from a father showing the msr significance of this variability, a two-level nested ANOVA was performed on the entire data set. The ANOVA was first performed on the proportion of sterile males, and highly significant variation was found among FI males within clones, as well as between clones. BARTLE~ S test for homogeneity of variance indicated that the mean squares (MS) of vials within FI males were homogeneous over all FI males; however, the MS of FI males within clones were heterogeneous (.1 C P C.25). This latter effect can be seen in Figure 3, in that those clones with lower means over FI males have more variability among FI male means. The Spearman rank correlation of the 18 standard deviations among FI males with the corresponding means is -.778, which is highly significant (P C.5). This is a violation of the assumption of AN- OVA, and consequently the analysis was applied to the angular transform of the proportion. This analysis proved to satisfy homogeneity of variance at both levels and showed highly significant variation among FI males within clones and among clones (Table 3). Cytology: Ten male third-instar larvae from SR1/YL-SR1/STL parents were examined for neural ganglion karyotype. In all cases the karyotype was found to be lacking a Y chromosome, but otherwise was normal, as shown in Figure 4A. As a control group, ten male third-instar larvae from STL/YLSTL/STL parents were also examined and, invariably, were found to have normal kar-

9 MALE SEX-RATIO IN DROSOPHILA 363 yotype, Figure 4B. The Y chromosome present in the L116 stock appeared to be normal for D. pseudoobscura and is a type V Y-chromosome described by DOBZHANSKY (1937). Fertility of 142 L 116 males was checked by crossing each singly to ten virgin females. Those males not producing larvae in the first mating were mated to a second and third harem of virgin females. Eight males (5.33%) proved to be sterile, and four of these had enlarged abdomens, indicating perhaps an infectious cause of sterility. The fertile males were observed to have much lower fecundity than other strains of D. pseudoobscura, although no progeny counts were done. Experiment 4: The phenotype proportions occurring in the F2 and in the progeny from the backcross were in good agreement with the independent assortment of all three mutant markers with each other, as well as with the sex chromosomes. Fertility test on F2 males gave a sterility rate of % based on a sample of 44 males. DISCUSSION The msr phenotype is a property of the L116 strain (Table 1). Only when SR/Y males result from L116 fathers does the msr phenotype occur. Both SR1 and SR2 showed msr when crossed to males of the L116 strain. Also it is apparent that msr varies in degree of phenotype, as the SR/P6 and SR/P3 males do produce some sterile male progeny but at a much lower level than that of the SR/YL males. In fact, the average for SR/P6 males in Table 1 (7.16%) is significantly different from the average for SR/P3 males (1.29%), (P <.1). Thus, even though WWA-6 and WWA-3 strains are considered normal when crossed to SR, they actually differ with respect to how much msr trait they show. Experiment 2 shows clearly that within the L 1 16 strain there is extensive male-to-male variability in the degree of the msr phenotype present in their SR/YL sons and that the amount of msr of a given SR/YL male tends to lessen somewhat as the male ages. The absence of significant interaction between the male and brood effects indicates that the age effect is consistent, as can also be seen in Figure 2, where the proportion of males in first broods are consistently greater than the second broods. The F3 and Fq vials of experiment 2 show a complete disappearance of the msr phenotype. The highest proportion of sterile males in Fg vials was 5.3%, substantially less than the low of 3.4% found in the F2 vials and consistent with the value for the WWA-6 strain of 7.2% (see Table 1). This indicates that the determinants of msr are not transmitted through the F2 females as would be expected if loci for msr were X-linked or autosomal dominant. The pattern of transmission seen in these experiments may be explained by either a Y-linked allele or by an autosomal recessive allele. The heritable variability in degree of msr suggests that the genetic basis may not be simple. Efforts to locate the msr genes are currently in progress. The independent assortment of the three autosomal markers in the goi stock shown in experiment 4 indicates that the L116 stock does not carry interchromosomal rearrangements, such as Y-autosome translocation.

10 364 G. COBBS The significant clone effect found in experiment 3 indicates that variability in msr is heritable. The significant variation of F1 males suggests that new variability in msr is generated in one father-son transmission. Such a mechanism for generating variation in msr is consistent with the extensive variation for msr found within the L116 stock. The degree of the msr trait found within the L116 stock has changed dramatically during the past several years while being maintained in my laboratory. The study of HENAHAN and COBBS (1983), which was very similar to experiment 2 reported here, gave an overall average of 19.8% sterile males, whereas experiment 2 gave a corresponding value of 61.%. The two experiments are separated by roughly 3 yr, during which the L116 stock has apparently evolved to produce a much more pronounced msr phenotype. It is not clear why a more extreme msr phenotype would evolve in the L116 stock. It is possible that the mechanism that generates the variability in msr also causes an increase in its intensity. An indication of selection against the msr trait is found by comparing the mean level of experiment 1 (61.%) to the levels found in the SR /YL-SR /STL (34.49%) and SR2/YL-SR2/YL (36.3%) crosses reported in Table 1. These experiments were separated by only a few months, and one would expect roughly comparable results. One explanation for the difference is sampling error due to the variable L116 stock. The close agreement of the two crosses in Table 1 suggest that this is not the case. Also, it is possible to show that a sample of ten males from an L116 stock with a mean of 61% is very unlikely to produce a mean of 34-36%. A more plausible explanation is selection. The 35 males in experiment 1 were mated singly to harems, whereas the crosses in Table 1 were performed en masse, and thus the ten SR/YL males could contribute unequally to progeny. If SR/YL males that show high msr phenotype are less fecund than those with lower msr phenotype, a lower overall frequency would result. The msr phenotype appears to be the result of the combination of genes from the L116 stock with the SR chromosome, because there is only a low rate of male sterility within the L116 stock, and F1 males from L116 males crossed to goz females were completely fertile. The observation of XI karyotype in the sterile male progeny suggests the msr phenotype is due to the production of a high proportion of null-xy sperm. The cytological mechanism for this phenomenon is as yet unknown, but if revealed, may provide insight into the role of the sex chromosomes during spermiogenesis. I wish to thank MIKE GOLDMAN for sharing his unpublished data on msr and HELEN MILLER ALEXANDER for critical reading of the manuscript. LITERATURE CITED BECKENBACH, ANDREW, 1978 The sex-ratio trait in Drosophila pseudoobscura: fertility relations of males and meiotic drive. Am. Nat. 112: DARLINGTON, C. D. and L. F. LACOUR, 196 York. The Handling of Chromosomes. Macmillan, New

11 MALE SEX-RATIO IN DROSOPHILA 365 DOBZHANSKY, T., 1937 Further data on the variation of the Y chromosome in Drosophila pseudoobscura. Genetics 22: HENAHAN, J. and G. COBBS, 1983 Origin of XI progeny from crosses of sex-ratio trait males of Drosophila pseudoobscura. J. Hered. 74(3): HOLMQUIST, GERALD, 1975 Hoechst fluorescent staining of Drosophila chromosomes. Chromosoma. 49: LAKOVAARA, S. and A. SAURA, 1982 Evolution and speciation in the Drosophila obscura group, pp In: The Genetics and Biology of Drosophila. Vol. 3b, Edited by M. ASHBURNER, H. L. CARSON and J. N. THOMPSON. Academic Press, New York. NOVITSKI, E., 1947 Genetic analysis of an anomalous sex ratio condition in Drosophila affinis. Genetics 32: POLICANKSY, D., 1974 Sex-ratio, meiotic drive, and group selection in Drosophila pseudoobscura. Am. Nat. 18: POLICANSKY, D., 1979 Fertility differences as a factor in maintenance of the sex-ratio polymorphism in Drosophila pseudoobscura. Am. Nat POLICANSKY, D. and B. B. DEMPSEY, 1978 Evolution 32: POLICANSKY, D. and J. ELLISON, 197 ure. Science 169: SOKAL, ROBERT R. and ROHLF, JAMES F., 1981 Modifiers and sex-ratio in Drosophila pseudoobscura. Sex-ratio in Drosophila pseudoobscura: spermiogenic fail- Biometry, Ed. 2. W. H. FREEMAN, San Francisco. STURTEVANT, A. H. and T. DOBZHANSKY, 1936 Geographical distribution and cytology of sexratio in Drosophila pseudoobscura and related species. Genetics 21: VOELKER, ROBERT, 1972 Preliminary characterization of sex-ratio and rediscovery and reinterpretation of male sex ratio in Drosophila afinis. Genetics 71: Wu, CHUNG-I and ANDREW T. BECKENBACH, 1983 Evidence for extensive genetic differentiation between the sex-ratio and the standard arrangement of Drosophila pseudoobscura and D. persimilis and identification of hybrid sterility factors. Genetics 15: Communicating editor: W. W. ANDERSON