Stage-Specific Effects on Visual System Function in Drosophila mehnogaster

Of the many genes that are expressed in the visual system of Drosophila melanogasteradults, some affect larval vision. However, with the exception of one X-linked mutation, no genes that have larval-specific effects on visual system structure or function have previously been reported. We describe the isolation and characterization of two mutant alleles that define the lama1 photokinesis A ( Z p h A ) gene, one allele of which is associated with a P-element insertion at cytogenetic locus 8E1-10. Larvae that express lphA mutations are, like normal animals, negatively photokinetic, but they are less responsive to white light than ZphA+ controls. Larvae that are heterozygous in trans for a mutant lphA allele and a deficiency that uncovers the ZphA locus are blind, which indicates that the mutant allele is hypomorphic. lphA larvae respond normally to odorants and taste stimuli. Moreover, the lphA mutations do not affect adult flies' fast phototaxis or visually driven aspects of male sexual behavior, and electroretinograms recorded from the compound eyes of ZphA/deficiency heterozygotes and lphA '/ lphAz females are normal. These observations suggest that the lphA gene affects a larval-specific aspect of visual system function. A fundamental question of biology concerns the nature of genes that determine the manner in which systems develop and function. In holometabolous insects, in which many larval structures are replaced by their adult counterparts during metamorphosis, this question is of particular interest, since one can ask about the extent to which larval and adult homologues are regulated by the same genes. The visual systems of Drosophila melanogaster larvae and adult flies are ideal subjects for such analysis, since hundreds of mutations that affect the adult visual system have been identified ( e .g . , PAK 1979; HALL 1982), as have dozens of DNA sequences that are expressed there ( i . e . , SHIEH et al. 1989; HYDE et al. 1990). With regard to the question of whether these genes are also expressed in the larval visual system or affect its function, three mutations that affect various aspects of phototransduction in adult flies' eyes perturb larval responses to light, while five mutations that regulate other aspects of visual system development and function in adult flies have no detectable effects on larval vision ( HOTTA and KENC 1984). Three of the four opsin genes that are expressed in the adult visual system are also expressed in the larval photoreceptor organ (POLLACK and BENZER 1988), as is the chaoptic gene, which codes for an adhesion protein that is required for normal development of the rhabdomeres in the compound eye (POLLACK et al. 1990). Finally, mutations in the disconnected and imeguCorresponding author: Laurie Tompkins, Department of Biology ' Present address: Department of Biology, University of Tennessee, 015-00, Temple University, Philadelphia, PA 19122. Knoxville, TN 3799600810. Genetics 139: 1623-1629 (April, 1995) lar chiasm C genes, isolated because they affect axon projections in the optic lobes of adults, exert their primary effects on pioneering neurons in the larval visual system ( STELLER et al. 1987; BOSCHERT et al. 1990). On the basis of these observations, genes that affect vision can be classified in three groups, distinguished by their developmental specificity: genes that are expressed in the visual systems of larvae and adults; genes that are expressed in the larval visual system and, as an indirect consequence of their effects on larvae, also affect the adult visual system; and genes that are only expressed in the adult visual system. It would be reasonable to assume the existence of a fourth class of genes, those that exclusively affect the larval visual system. However, except for an unmapped X-linked mutation ( lnp) that affected larval responses to light without perturbing adult flies' behavior in a fast phototaxis assay ( HOTTA and KENG 1984), which may no longer be extant (see below), there is no evidence for this class. Accordingly, we have begun to isolate and characterize mutations that affect larval vision. This paper describes the first fruits of our labor: two serendipitously recovered mutations that define the @hA ( lama1 photokinesis A ) gene, which appears to have a larval-specific effect on visual system function. MATERIALS AND METHODS Stocks: The wild-type Canton-S stock, which is M cytotype ( BINCHAM et al. 1982) , was derived from a male-female pair in 1976 and has since been maintained in mass culture (TOMPKINS et al. 1980). P cytotype Harwich ( JSIDWELL et al. 1977), attached-X ( C(l)DX, y f / F M 7 ( P ) y snx2 B ) ( ENGELS 1985), and FM7a balancer stocks were obtained from J. CARL1624 B. Gordesky-Gold et al. SON. Two retinal degeneration A stocks (rdgA’ and rdgA3) were provided by W. STARK. cinnamon ( cin’) flies were collected from a stock obtained from M. WOLFNER. A stock designated as larval nonphototactic ( lnp) ( HOTTA and KENC 1984) was provided by Y. HOTTA. The Mid-American Droso hila Stock Center supplied three stocks: yellow ( y ) cut6 ( ct ) vermilion ( v ) forked ( f ) ; ocelliless ( o c ’ ) cut ( ct6)/FM6and M cytotype FM7a. A Of (1) C52/FM6 stock was obtained from the Drosophila Stock Center at Indiana University. Unless otherwise noted, all mutant alleles and special chromosomes are described in LINDSLEY and ZIMM (1992). All stocks were maintained on a cornmeal-corn syrup medium at 22 or 25”. Mutagenesis: The lphA mutations are the products of a hybrid dysgenesis cross, that is, mobilization of paternally derived Pelements in the germline of the progeny of P cytotype males and M cytotype females ( ENCELS 1983). To initiate the mutagenesis, Harwich males were mated to Cantons females. The F, male progeny of this cross, whose X chromosomes were inherited from their Canton-S mothers, were then mated to C(l)DX, y f / YP cytotype females to stabilize any Pelement insertions that had occurred and ensure that mutagenized X chromosomes would be inherited patroclinously. Since we were also interested in mutations that affected larval taste, the F2 larvae were screened in a taste assay ( GORDESKY-GOLD et al. 1988) to see whether they responded normally to 1 M NaCl, which repels wild-type larvae. Larvae that responded mistactically in two consecutive tests were transferred to vials to complete development and then sexed. Each male fly was individually mated to P cytotype FM7a females to establish, after two generations, a line in which all of the flies were homozygous or hemizygous for a mutagenized X chromosome. Larvae from each line were then tested in the GORDESKY-GOLD et al. taste assay. Since larvae from two of the lines responded abnormally to 1 M NaCl when the lines were first tested, those lines were retained for further analysis. In subsequent tests, larvae from the two lines behaved normally when they were tested with 1 M NaCl, but their responses to light were consistently abnormal in comparison with those of control larvae with unmutagenized Canton-S chromosomes. Accordingly, the lines were retained for further analysis of their visual phenotypes. Genetic analysis: Since the lphA mutations are X-linked (see below) , it was necessary to analyze the behavior of female larvae to assess the mutations’ dominance and their behavior in complementation tests. This analysis was facilitated by constructing X chromosomes bearing the lphA mutations and a mutant cinnamon allele ( cin’) by meiotic recombination. For complementation tests and determination of dominance, females homozygous for these chromosomes were crossed to cin+ males, generating cin’/ cin+ females and c i n ’ / Y males. Since homozygous or hemizygous cin’ progeny of cin’/ cin’ mothers die as embryos (BAKER 1973), this cross yields allfemale populations of larvae that are heterozygous or homozygous for a l@hA mutation, depending on the genotype of the male parent with respect to the lphA gene. To map the lphA gene by meiotic recombination, @ h A 2 females were crossed toy c t6 v fmales, and F2 males representing single recombinant classes (e .g . , y+ ct6 v f ) were isolated. The males were mated to females from an FM7a stock ( P cytotype ) to generate stocks in which all of the flies were hemizygous or homozygous for the recombinant chromosomes. Larvae from each stock were then tested in the photokinesis assay (see below) . To correlate excision of a P element in the lphA ’ stock with reversion of the mutant phenotype, we isolated two &hA+ revertant lines. To do this, lphA’/Ymaleswere crossed to C(l)DX, y f / Y females from an M cytotype stock to initiate hybrid B dysgenesis ( ENCELS 1983). After eight generations, -20 individual males were crossed to females from a P cytotype FM7a stock to generate, in two generations, stocks in which all flies were homozygous or hemizygous for the mutagenized Xchromosome that was derived from the lphA’/Ymale progenitor. Larvae from each stock were then tested in the photokinesis assay. Two lines whose larvae responded normally to light were classified as revertant and retained for further analysis. To confirm the map osition of the lphA gene and characterize the mutant lphA allele, we analyzed the behavior of female larvae that were heterozygous in trans for the @ h A Z allele and Of (1) C52. Since the deletion is embryonic lethal when homozygous, generation of lphAz/ Of (1) C52 larvae was facilitated by construction of a chromosome bearing the @hA’ allele and a mutant allele of shihre (shi‘”’) , which has heatsensitive effects on larval and adult mobility ( POODRY et al. 1973), by meiotic recombination. To generate &Az/ Of (1) C52 larvae, shi‘”/ Y males were mated to Of ( 1 ) C52/ FM6 females, and the Df (1) C52/ shifs’ female progeny were crossed to lphAz shi”’/ Ymales. The progeny of this cross were reared at 22“, a temperature at which larvae that express the shi*’ mutation are mobile (POODRY et al. 1973). Since the response of wild-type larvae to light is temperature sensitive (J. M. WARRICK, D. P. KUTZLER and L. TOMPKINS, unpublished data) , the progeny of this cross were tested in the larval photokinesis assay at 22”, after which the assay plates were transferred to 29” for 810 min to paralyze the shi‘”’/ Y and shi‘”’/ lphA2 shi‘”’ larvae ( POODRY et al. 1973; L. TOMPKINS, unpublished data). Immediately after the 29” temperature shift, the mobile larvae, which were Df (1) C52/ lphA2 shi””, were observed to determine their distribution on the assay plate. To confirm that the deficiency-bearing chromosome did not have dominant effects on larval photokinesis, Df (1) C52/FM6 females were crossed to shi“’/Y males, and the Of (1) C52/ shit”’ female progeny were crossed to shi“”/Y males. Progeny of this cross were raised at 22”, tested in the larval photokinesis assay at 22” and then shifted to 29” as described above to paralyze the shifS1/ shi”’ and shi“”/ Y larvae. The Of (1) C52/ shi”’ larvae, which were not paralyzed, were immediately observed to determine their distribution on the assay plates. Chromosome in situ hybridization: A pUCPT probe was biotinylated as specified in the protocol accompanying the nick translation kit (Enzo Biochemicals) , utilizing bio-16 dUTP as the biotinylated nucleotide. The biotinylated probe was hybridized to salivary gland polytene chromosomes, utilizing a standard protocol (ASHBURNER 1989) with, in some cases, the following modifications: the chromosomes were pretreated with I X PBS rather than 2X SSC, they were not acetylated, and the predenaturing dehydration step was omitted. Peroxidase staining was performed according to the protocol accompanying the Detek HRP kit (Enzo Biochemicals), utilizing a modified AEGsubstrate mixture ( 7 ml water, 1 ml of 8X reaction buffer, 160 pl of 0.5 M EDTA, and 80 p1 of AEC) . Chromosomes were counterstained with Giemsa. Behavioral assays: The larval photokinesis assay has been described ( LILLY and CARLsON 1990). Briefly, two opposing quadrants of a partitioned X-Plate Petri dish (Falcon 1009) are filled with clear agarose, while the remaining quadrants are filled with an agarose solution that is dyed black with commercial food coloring. Approximately 100 larvae are transferred to the center of the dish, which is then placed on a light box to illuminate it from below. After 5 min, a stimulus index, which we define as the percent of larvae on the two quadrants that are filled with clear agarose, is calculated. In this assay, the larvae are responding to light, rather than a chemical stimulus associated with the food coloring in the black agarose, since wild-type (Canton$) larvae distribute themselves randomly on the blackand clear-agarose-filled F Mutations Affecting Larval Vision 1625 quadrants if they are tested on photokinesis assay plates in the dark (mean 5 SEM stimulus index = 50 5 2; n = 4 ) . Evidence that the response of wild-type larvae is a photokinesis ( a change in behavior effected by light) rather than a negative phototaxis ( a directed movement away from light) is provided by the observation that individually tested CantonS larvae, transferred to the center of a clear-agarose-filled quadrant, move quickly in the direction in which they were oriented when they were placed on the agarose; if they reach the edge of the Petri dish, they move rapidly along the periphery. Once they have crossed into a black-agarose-filled quadrant, the larvae immediately begin to move more slowly, describing randomly oriented circles. Thus, light changes the rate at which wild-type larvae move and their pattern of movement but does not elicit directed movement away from the stimulus. The LILLY and CARL~ON larval taste assay ( LILLY and CARLSON 1990) is identical to the photokinesis assay except that 0.025 M or 1 M NaCl is substituted for the food coloring in one of the agarose solutions. For the taste assay, we define a stimulus index as the percent of larvae on the two quadrants that are filled with the agarose-NaCl solution. In the GORDESKY-GOLD et al. assay for larval taste ( GORD ESKY-GOLD et al. 1988), -100 larvae are transferred to the interface between two 21.5 X 7 cm blocks of 0.4% agar, one containing 1 M NaCI. After 30 min, a stimulus index, which is the percent of larvae in the stimulus-containing agar block, is determined. The larval olfaction assay is described in MONTE et al. (1989). A filter paper disc on one side of a Petri dish is saturated with 25 p1 of an odorant solution, diluted with dis tilled water, and a disc on the other side of the dish is saturated with 25 pl of distilled water. Approximately 100 larvae are transferred to the center of the dish. After 5 min, a stimulus index, which we define as the percent of larvae on the half of the dish with the odorant-saturated filter paper, is calculated. Adult flies' fast walking phototactic behavior was assayed in a Y-maze ( QUINN et al. 1974), in which one arm is illuminated with white fluorescent light. Flies are shaken into a start tube and then agitated; after 30 sec, the flies in the illuminated arm, the dark arm, and the start tube and body of the apparatus are counted. Adult flies' sexual behavior was assayed by transferring a virgin male and female to a clear plastic cylindrical observation chamber (volume ca. 0.2 cm') and then observing the flies under 7 X magnification for 10 min. At the end of the observation period, the male's courtship index (the percent of time spent by the male performing any courtship behaviors) (TOMPKINS et al. 1980) is calculated. If the flies mate during the observation period, this is noted, and a courtship index is calculated for the time elapsing before copulation ensues. Flies that do not mate during the 10-min observation period are observed for an additional 20 min; if they mate during the second observation period, this is noted. Electroretinograms: Electroretinograms were recorded from the compound eyes of adult flies as described in WOODARD et al. (1989). Statistical analysis: The significance of differences between means was determined by calculating t (Student's t test, twotailed) after arcsin transformation of the data, if appropriate. The significance of differences in the distribution of individuals into classes was determined by calculating contingency chi square with Yates correction factor. The significance of deviations from a 1:l ratio of the distribution of individuals into classes was determined by calculating chi square. Statgraphics statistical software was used for all statistical analyses. TABLE 1 Dominance and complementation of ZphA mutations' effects on larval photokinesis Genotype Stimulus index n