SELECTION FOR MALE RECOMBINATION IN DROSOPHILA MELANOGASTERl

Two-way selection for male recombination over seven intervals of the third chromosome in Drosophila melanogaster was practiced for nine generations followed by relaxed selection for five generations. Significant responses in both directions were observed but these mainly occurred in early generations in the low line and in later generations in the high line. Divergence of male recombination frequencies between the two selection lines was not restricted to any specific region but occurred in every measured interval of the chromosome. However, right-arm intervals showed a more pronounced response than either left-arm intervals or the centromeric region. Correlated responses in sterility and distodon of transmission ratios occurred as a result of selection for male recombination. Cluster distributions of male recombinants suggested a mixture of meiotic and late gonia1 events but relative map distances more closely resembled those of the salivary chromosome than standard meiotic or mitotic distances. Patterns of male recombination over time in both second and third chromosomes strongly suggested a major effect associated with the presence of third chromosomes from the Harwich strain. Evidence was also found for modifiers with relatively small effects located in other regions of the genome. The overall results are interpreted in terms of a two-component model of hybrid dysgenesis. HE complete absence of spontaneous crossing over in Drosophila melanoTgaster males can no longer be assumed. Since HIRAIZUMI’S (1971) initial observations, male recombination has been reported to be associated with chromosomes from numerous natural populations (e.g. VOELKER 1974; SVED 1974; WADDLE and OSTER 1974; KIDWELL and KIDWELL 1975a). The close correlation of male recombination with one o r more additional aberrant traits such as distortion of transmission frequencies, chromosomal aberration, lethal mutation and sterility have been observed frequently (e.g. HIRAIZUMI et al. 1973; SVED 1973 and and 1974; VOELKER 1974; YAMAGUCHI and MUKAI 1974). Large F, reciprocal differences in male recombination (KIDWELL and KIDWELL 1975a; WOODRUFF and THOMPSON, 1975) and recessive lethality (KIDWELL, KIDWELL and IVES 1976) have indicated that cytoplasm-chromosome interactions are involved in the spontaneous induction of these phenomena. This conclusion is strengthened by the direct demonstration that high frequencies of lethality. sterility and male rex This work was supported by NSF research grant GB-43820 and by USPHS research grant GM-20103 Genetics 84: 333-351 October, 1976. 334 M. G . KIDWELL A N D J. F. KIDWELL combination do not occur in crosses within strains but only in certain crosses between interacting strains (KIDWELL, KIDWELL and SVED, 1977, M. G. KIDWELL, unpublished). These phenomena are referred to collectively as ‘hybrid dysgenesis’ (SVED 1976; KIDWELL, KIDWELL and SVED 1977), which is defined as a syndrome of correlated genetic traits that is spontaneously induced in hybrids between certain mutually interacting strains, usually in one direction only. Many, but not all, of the observed traits are seemingly highly deleterious to their carriers; none has yet been found that is obviously advantageous with respect to fitness. In this respect it should be noted that distortion of transmission ratios is in the opposite direction to that of the well-known segregation distorter (SD) phenomenon ( ZIMMERING, SANDLER and NICOLETTI 1970). We postulate that hybrid dysgenesis is the result of interactions between two types of components, one that is maternally influenced and one that is chromosomally associated. Characterization of the maternal component is not yet complete but a strictly extrachromosomal factor seems to be ruled out (KIDWELL, KIDWELL and IVES 1976). Here we describe an examination of the chromosomally associated component under conditions in which the maternal component was held constant using the occurrence of male recombination as the method of assay. Specifically, we observed the response to selection for high and low male recombination in virtually the total length of the third chromosome. Correlated responses to selection in other dysgenic traits were measured. Changes of male recombination frequency over ten generations in both selected and unselected lines implicated genetically transmissable factors associated with wildderived third chromosomes and minor modifiers elsewhere in the genome. These, in the presence of an interactive maternal component, can induce male recombination in any interval of the second or third chromosomes. Patterns of male recombination and sterility in unselected second and third chromosome lines provided additional evidence on the relative effects of various chromosomes on dysgenic traits. MATERIALS AND METHODS The investigation consisted of two parts: (1) a preliminary experiment, in which second chromosome recombination was monitored over a period of ten generations in unselected males, (2) a selection experiment in which both selected and unselecfed lines were observed for male recombination in the third chromosome over a similar period. The Harwich strain of Drosophila melanogaster was used in both experiments. This is a wildtype strain, founded by DR. MARTIN L. TRACEY JR. from two females collected in Harwich, Massachusetts in 1967. A chromosome ZZ multiply-marked stock, a1 cl b c sp2 (KIDWELL and KIDWELL 1975b) was crossed with Harwich in the first experiment and rucuca, a chromosome ZZZ multiply marked stock (LINDSLEY and GRELL 1968) was crossed with Harwich in the second experiment. These two marker stocks were previously found to interact strongly in specific combinations with Harwich (KIDWELL and KIDWELL 1975a). A general outline of the mating scheme is shown in Figure 1. The first experiment was restricted to only the unselected A and A’ lines. The second experiment included all four lines. In both experiments, the unselected A lines were derived from the initial cross A mating. The unselected A’ lines were derived from the reciprocal cross B matings but subsequent to the initial generation they were maintained in an identical manner to the A lines. Within each experiment, the unselected A and A’ lines were expected to differ from each other only in the SELECTION FOR MALE RECOMBINATION 335 source of the Y chromosomes; the A line carried Harwich Y chromosomes and the A' line the marker stock Y chromosomes. In the selection experiment, the high and low lines were both derived from the original cross A mating. Nine generations of selection were completed. Every generation within all lines, each of about 40 nonrecombinant males, heterozygous for Harwich and the marker stock, were back-crossed to four virgin marker stock females which were allowed to deposit eggs for nine days. Frequencies of recombination in all marked intervals were observed. In the unselected lines of both experiments, each of the 40 original F, chromosome lines was maintained by a single randomly chosen male, in order to ensure a minimum of selection. In the high and low selection lines of the second experiment, the eight male progeny groups with the highest and lowest frequencies of recombination in all seven intervals combined were identified. Five nonrecombinant males from each of these eight progeny groups constituted the next generation. Reserves were substituted for sterile males whenever possible. Matings to measure male recombination were made in 8-dm shell vials. Parents were removed on the 9th day after introduction and all progeny emerging up to the 19th day were counted and classified. A standard cornmeal-molasses-agar medium, seeded with live yeast, was employed. Proprionic acid was used as a mold-inhibitor. The temperature was maintained at 25 f 1 '. Three different, but related estimates of male recombination were computed. For a discussion of their relative merits see KIDWELL and KIDWELL (1975b). 1. Mean frequency of male recombination, expressed as a percentage of total progeny: (Total number of recombinants x 100) /(Total number of progeny). 2. Mean minimum frequency of independent events, expressed as a percentage of total progeny: (Estimated number of recombination events x 100)/(Total number of progeny). 3. Percentage of males with recombination: (Number of males with one or more recombinant progeny x 100)/(Total number of males). For purposes of calculation of the second estimate, a recombination event is judged to have occurred when one or more recombinant progeny of a single male parent are observed in any given interval. Cluster size is defined as the number of identical and complementary recomhinants within a given interval which were observed in the progeny of a single male. A sterile individual is defined as one that produced no progeny when mated with three members of the opposite sex from a standard fertile stock. EXPERIMENT SECOND CHROMOSOME-NO SELECTION Male recombination: Mean frequencies of male recombination events, for all four second chromosome intervals combined, are plotted against generation number in Figure 2. As expected, reciprocal differences between A and B crosses were very large in the first (F,) generation. Thereafter, the B line became the A' line (when homozygous marker females were necessarily employed) and this change was accompanied by a male recombination increase of similar magnitude to that of the F, reciprocal differences. From the second generation onwards both lines had very similar frequencies of male recombination events and these were characterized by a steep decline in the early generations and a levelling off at around the 0.5 per cent level in later generations. The size of the decrease was roughly proportional in all four intervals. Table 1 provides a comparison of the A and A' lines for two measures of male recombination within each of the four intervals, averaged over all ten generations. Male recombination values in the A line exceeded those in the A' line for both measures and in every interval. However, differences between the lines in the frequency of male recombination tended to be much larger than differences in the minimum frequency of male recombination events. These large differences are reflected in large cluster sizes