Toward a Molecular Genetic Analysis of Spermatogenesis in Drosophila Melanogastec Characterization of Male-Sterile Mutants Generated by Single P Element Mutagenesis

Toward a Molecular Genetic Analysis of Spermatogenesis in Drosophila melanogastec Characterization of Male-sterile Mutants Generated by Single P Element Mutagenesis

Diego H. Castrillon,* Pierre Gonczy? Sherry Alexander,* Robert Rawson,* Charles G. Eberhart,* Sridhar Viswanathan?’ Stephen DiNardot and Steven A. Wasserman* *Department of Biochemistry, University of Texas Southwestern Medical Center, Dallas, Texas 75235-9038,and tThe Rockefeller University, New York, New York 10021-6399 Manuscript received February 23, 1993 Accepted for publication June 18, 1993

ABSTRACT We describe 83 recessive autosomal male-sterile mutations, generated by single P element mutagenesis in Drosophila melanogaster. Each mutation has been localized to a lettered subdivision of the polytene map. Reversion analyses, as well as complementation tests using available chromosomal deficiencies, indicate that the insertionsare responsible for the mutant phenotypes. These mutations represent 63 complementation groups,58 of which are required for spermatogenesis. Phenotypesof the spermatogenesis mutants were analyzedby light microscopy. Mutations in 12 loci affect germline proliferation, spermatocyte growth,or meiosis. Mutations in 46 other loci disrupt differentiation and maturation of spermatids into motile sperm.This collection of male-sterile mutants provides the basis for a molecular genetic analysis of spermatogenesis.

PERMATOGENESIS in Drosophila begins with been generated by chemical mutagenesis in screens S the formation of a spermatogonium from a pop- conducted by several laboratories (see forexample ulation of germlinestem cells. Afixed number of LIFSCHYTZ1987; HACKSTEINet al. 1990; HACKSTEIN mitotic division ensues, followed successively by cel- 1991). However, most of the mutants generated in lular growth, meiosis, and an elaborate pathway of this manner have not been mapped, and the majority cellular differentiation. Throughout this process the are no longeravailable. developing germ cells are accompanied by a pair of We have chosen a different approach, thatof single somatic cells. The entireprogram from stem cell P element insertional mutagenesis (COOLEY,KELLEY division to functional sperm requires 10 days, and is and SPRADLINC1988), to conduct a genetic analysis initiated every 10 hr. Thus,all stages of spermatogen- of spermatogenesis in Drosophila melanogaster. With esis are present in a single adult testis, with earlier this method, each mutation is genetically tagged by stages at thetip and later ones at thebase (for reviews an eye color marker andcan be readily located on the see LINDSLEYand TOKAYASU1980; FULLER1993). physical map by in situ hybridization to polytene chro- The opportunityto explore the maintenance of mosomes. Furthermore, new mutations,including stem cell identity, the regulation of an invariant num- null alleles, can be efficiently generated by remobili- ber of mitotic divisions, the execution of meiosis, the zation of the P element with a transposase source. initiation of a program of morphogenesis, and the Lastly, DNA flanking an insertion site can be readily interaction between germ line and soma make sper- cloned by bacterial transformation or by polymerase matogenesis an ideal targetfor investigations into chainreaction (PCR), greatly facilitating molecular differentiation and cell-cell interaction. analysis of the locus. Despite a detailed understanding of spermatogene- We present here theisolation and phenotypic char- sis at themorphological level, relatively little is known acterization of 83 male-sterile autosomalmutations about the genes controlling this differentiation pro- generated by single P element mutagenesis. The vast gram. Exceptions include the testis-specific ,&-tubulin majority affect spermatogenesis, with defects ranging and the cdc25 homolog twine, bothof which were from nonproliferation of germ cells to the production identified as homologs of gene products studied in of defective sperm. The results of complementation other organisms (KEMPHUESet al. 1979; ALPHEYet al. tests with available chromosomal deficiencies, to- 1992; COURTOTet al. 1992). Male-sterile mutants gether with reversion analyses, indicatethat the P defective for different stages of spermatogenesis have element insertions are responsible for the observed ’ Current address: Department of Developmental Biology, Stanford Uni- male sterilephenotypes. These mutationsshould versity School of Medicine, Stanford, California 94305. therefore provide a starting point for the dissection

Genetics 135: 489-505 (October, 1993) 490 D. H. Castrillon et al. of mechanisms controlling thecourse of spermatogen- Reversionanalysis: Insertion-bearing chromosomes esis. were brought together with a transposase source by crossing flies from the balanced mutant stocks to flies carrying the MATERIALS AND METHODS P[ry+ A 2-31 transposase source on either the second or third chromosome (ROBERTSONet al. 1988; M. SANICOLA Genetic strains:All marker mutations and balancer chro- and W. GELBART,personal communication). In the next mosomes are described and referenced by LINDSLEYand generation, the P[ry+ A 2-31 chromosome was crossed out ZIMM (1992). Crosses were carried out at 25" in vials con- and ry derivatives of the P[Z] or P[lac, ry+]A chromosome taining freshly yeasted cornmeal molasses agar. Ore on R, were selected and used to establish lines. The ry derivatives Sevelen (EDGARand SCHUBIGER1986), st e or w" Y* flies were then made homozygous or put in trans to the original were used as wild type. insertion for tests of male and female fertility. Generation of single P element insertion lines: The Characterization of male-sterile phenotypes:For gross

P[lacZ, ry +]element, hereafter called P[Z], contains a drug examination of reproductive organs, five males,one toseven resistance gene, plasmid replication sequences, the Drosoph- days old, homozygous forthe P element insertion were ila rosy+ gene, and afusion of the lacZ gene to thefirst exon dissected in modified Drosophila Ringer's solution (5 mM of the P transposase gene (MLODZIKand HIROMI1992). The PIPES, pH 6.9; 140 mM NaCI, 5 mM KCI, 2 mM CaCh), crossing scheme used for generation of insertions of the Becker's Ringer's solution or phosphate-buffered saline P[Z] transposon is described elsewhere (KARPENand SPRA- (PBS) (ASHBURNER1989). Seminal vesicles were torn open DLING 1992). In brief, a P[Z] element on the X chromosome with forceps under a dissection microscope and examined was mobilized by introducing the A2-3 P transposase source for thepresence and motility ofsperm. Testes from mutants (ROBERTSONet al. 1988). Approximately 8000 lines were that failed to produce motile sperm were analyzed by phase generated containing new insertions of P[Z] in an isogenized contrast microscopy and by X-gal and Hoechst staining, as cn second chromosome or a ryYo6 third chromosome. Muta- described below. For those male-sterile mutants that pro- tions were balanced with Cy0 for second chromosome inser- duced motile sperm, transfer of sperm to females was as- tions and with one of four chromosomes (MKRS;TM3, sayed by dissection of the sperm storage organs of females ry;TM6B, ry; or CxD) for third chromosome insertions. with whom they had been mated. Mutants that failed to An additional set of single P element insertion lines car- transfer their motile sperm to females were further analyzed rying P[lac,ry+]A (O'KANE andGEHRINC 1987) was simi- for behavioral or anatomical defects. larly generated by using the A2-3 transposase source and Examination of testis contents by phase-contrast mi- was generously provided by the UCLA fly group. Insertions croscopy: For the examination of testis contents by phase of this transposable element were on unmarked second contrast microscopy, testes from at least four newly eclosed chromosomes or third chromosomes carrying ryTo6. males were dissected out with forceps and placed on a glass Identification of males-sterile mutants: Male sterile mu- slide with a drop of Ringer's solution or PBS. The testes tants were identified among the single P element lines by were torn open and gently squashed under the weight of a performing fertility assays: Eight to ten homozygous males cover slip (KEMPHUESet al. 1980). Spermatogenic cells of were crossed with an equal or greater number of virgin stages before, during and shortly after meiosis were visual- wild-type females. Homozygotes were designated as sterile ized by this procedure. if there were no progeny from these crosses, and as semis- X-gal and Hoechststaining of testes: To determine the terile if they produced fewer than10% the number of pattern of expression of the enhancer-trapping lacZ gene in progeny produced by heterozygous siblings. the P element, testes from newly eclosed adults were stained Viability for each male sterile line was measured among with X-gal as described (G~~NczY,VISWANATHAN and DI- the progeny of a cross between balanced heterozygotes. NAR~1992). To assess the shape and distribution of germ Mutations were deemed to be semilethal if homozygotes cell nuclei, testes from newly eclosed males were fixed with were present at less than 30% the expected frequency. 4% formaldehyde (EM grade, Polyscience) for minin Female fertility was assayed by crossing at least fivehomo- 20 PBX (0.1% Triton X-100 in PBS), washed three times for zygous females from a given line with an equal number of balancer sibling males. 5 min in PBX, stained with 1 rg/ml Hoechst 33258 (Sigma) for 2 min in PBX, rinsed in PBX, washedfor 30 min in PBS In situ hybridization: Larvae for salivary gland squashes were generated by outcrossing each mutant line to wild and mounted in Fluoromount-G (Fisher). type. After 2-3 days at 25", the vials were transferred to Classification of male-sterilemutations: Male-sterile 18" and supplemented with fresh yeast. In situ hybridiza- mutants were first categorized with regard to whether they tions and Giemsa staining were carried out essentially as produced normal amounts of motile sperm in the seminal described (ASHBURNER1989). The P[Z]-containing plasmid vesicle or were defective for spermatogenesis. The mutants pHZ50PL (a gift of Y.HIROMI), the P[lArB] plasmid (WIL- were then grouped into seven general classes, five for those SON et al. 1989), or genomic DNA flanking the insertion site affected in spermatogenesis and two for those producing was used as probe. wild-type levels ofmotile sperm. Classification was based on Complementation and deficiency crosses: Complemen- the earliest stage at which visible defects were readily and tation tests were performed by crossing each male-sterile reproducibly apparent in homozygotes. mutation from the collection to chromosomal deficiencies ProlveratzonPhase defect: A reduction or absence of potentially spanning the insertion site, as well as mutations growth phase spermatocytes is apparent at eclosion or a few that mapped within two lettered polytene divisions, includ- days post-eclosion.Such testes are typically thin and in more ing: 1)other insertion mutations from the collection; 2) extreme examples resemble those of agametic mutants. other cytologically localized, male-sterile mutations induced Growth phase defect: Germ cells complete the proliferation with P elements (COOLEY,BERG and SPRADLING1988); 3) stage and form cysts of 16 early spermatocytes, but undergo alleles of ,f32t and fra (KEMPHUESet al. 1979; GAILEYand aberrant development duringthe primary spermatocyte HALL 1989); and 4) mutations that are both male- and growth phase. female-sterile from the collection of SCH-BACH and WIES- Meioticentry defect: Spermatogenesis is normal through CHAUS (1991). the late spermatocyte stage, but meiosis does not occur. Drosophila Male-sterile Mutants 49 1 Meiotic division defect: Meiosis is initiated but fails to give TABLE 1 rise to the normal patternof 64 spermatids. Postmeiotic differentiation defect: Mutants undergo normal Screen for male sterile mutations among a setof random single meiosis, but fail to produce mature sperm. The mutants P element insertions on the second and third chromosomes which fall into this categoryare readily distinguishable from wild-type by the absenceof sperm in the seminalvesicle. n 36 Total hits Behavioral defect: Homozygousmales produce motile sperm at wild-type levels, but fail to copulate with virgin A. Total insert lines 100 1919 females during 30-minobservation periods (GAILEYand Homozygous lethal insertions 262 14 HALL1989). Female sterile (male fertile)' 48 2.5 Sperm transfer defect: Motile sperm are found in the sem- Male sterile (female fertile)' 24 1.3 inal vesicle at wild-type levels, but no sperm transferoccurs. Male and female sterile' 0.2 4 In one mutant, matingis observed but few or no spermare found in the female storage organs. In other mutants, the B. Phenotypes of male sterile insertions''b n % Male Steriles I. Defects in spermatogenesis externalgenitalia ofhomozygous males are aberrant or absent. Proliferation defect 2 7 Growth phase defect 1 4 Naming of loci: The results of genetic tests governedthe - naming of loci: 1) in cases where our mutations were found Meiotic entry defect 0 Meiotic division defect to be alleles of previously characterized loci, the preexisting 2 7 name, if any, was adopted; 2) in cases where more than one Postmeiotic differentiation defect -19 68 allele was identified in our screen, or where alleles were 24 identifiedamong available male-sterile mutations of un- named loci, we assigned the locus a name according to our 11. Defects in mating phenotypic characterization; 3) we also gave names to any Anatomical 1 4 of the loci for which we isolated transposase-inducedrever- Behavioral 11 -3 tants or identified deficiencies that failedto complement the 4 4) loci male-sterile phenotype; all other single allele were ' Sterile classes include a few semisterile lines which produce a given designations based solely on cytological location. limited number of progeny, no more than 10% the number seen with wild-type flies (seeTable 2). RESULTS Includes all male-sterile lines, both female-fertile and female- sterile. Our collection of male-sterile mutations was gath- ered in two stages. First,a screen for male-sterile ern analysis using a -0.5-kb probe from the 5' endof mutants was carried out such that the frequencies of the P elementrevealed a single band in genomic different classes of mutants(lethal, female-sterile, digests from each of the 28 lines (data not shown). male-sterile, and male- and female-sterile) could be Thus each male-sterile line from this set contains a tabulated (Table 1). This set was then supplemented single P element. with mutantsdrawn from additional screens. Map- To confirm that the P element was responsible for ping, complementation tests,and phenotypic analyses the mutant phenotype, we used a transposase source were performed on the combined collection. We be- to promote excision of the P[Z] element from seven gin by presenting anoverview of the results from the different insertion lines. For all seven lines, two or first mutant screen. We then consider the complete moreindependent ry derivatives were male fertile collection and describe each of the phenotypic classes when homozygous or in trans to theoriginal insertion. in turn. Representative mutants are highlighted in the Moreover, the parental DNA restriction map at the text and figures. site of insertion was restored for the subset of five Isolation of male-sterileinsertion mutations: lines which were analyzed on Southernblots (data not From approximately 8000 autosomal P[Z] element shown). insertion lines, we screened a subsetof 19 19for male To further test whether the single P element inser- sterility with no preselection otherthan for some tion was responsible for therecessive male-sterile phe- degree of homozygous viability (Table 1A). Twenty- notype, we asked whether sterility was observed when eight (1.5%) caused partial or complete male sterility; an insertion was placed in trans to a cytologically 4 of these 28 also affected female fertility, An addi- defined deficiency spanning the insertion site. Such tional 48 inserts(2.5%) were strictly female-sterile. deficiencies were available for six of the seven lines Viability was good (>70% expected) for 26 of the 28 for which we generated revertants, as well as for six male-sterile lines. Twenty-fourmutations affected others from among the 28 lines. In all 12 cases, the spermatogenesis; the remaining 4 led to anatomical insertions were male-sterile in trans to the deficiency. or behavioral defects. Of the spermatogenesismu- In several instances, the insertion in trans to a defi- tants,the majority affectedpostmeiotic differentia- ciency also showed female sterility or decreased via- tion; only 5 of the 24 disrupted earlier stages (Table bility. These results will be discussed in more detail 1B). below. Of the28 events affecting male fertility,15 mapped Analysis of an expanded collectionof male-sterile to the second chromosomeand 13 to the third. South- mutants: In order tostudy the broadestpossible array 492 D. H. Castrillon et al. of phenotypes and to obtainadditional alleles of loci fromyoung diu males. These results suggest that of interest, we identified more male-sterile mutations diaphanous is required in males for both mitotic and among the remaining P[Z] lines, as well as an addi- meiotic germ cell divisions. In trans to a deficiency, tional set of single element insertions generated with diu males have a similar phenotype but females be- P[lac, ?+]A (see MATERIALS AND METHODS). Since the comesemisterile, producing ovaries with few egg additional lines examined did not include all viable chambers. We infer that diu function is also required mutants generated,we did not measure the frequency for germ cell divisions in females. of male-sterile mutations in these cases. In all, we Proliferation phase defects are also apparent in identified 69 P[Z]-induced mutants as well as 14 males homozygous for mutations in the chickadee (chic) induced with P[lac, ry+]A. In situ hybridization with a locus at 26A. chic testes arereduced in size to a transposon-derived probe revealed a single insertion variable extent; an exampleof a mildly affected testis in each of these lines (Table 2). Six male-sterile lines is shown in Figure 2d. The phenotype becomes more for which in situ hybridization detected no element or severe over adeficiency. The testes of chic”/Df males more than one P element were not included in the are short and thin (Figure 2e) and nearly devoid of collection. germinal content. They thus resemble the agametic Mapping and complementation experiments indi- testes of germlineless progeny of oskar mothers (Fig- cated that the83 male-sterile or semisterile mutations ure 20. Some of our chic alleles affect female fertility. affect 63 loci on the second and third chromosomes. All previously isolated alleles of the chic locus are For about three-fourths of these mutations (61/83), female-sterile; some are also male-sterile (SCHUPBACH we have used excision of the P element or comple- and WIESCHAUS199 1; COOLEY, VERHEYENand AYERS mentation tests with deficiencies or other mapped 1992). insertion alleles to confirm that the insertion is re- Comparison of the testis phenotypes of osk, chic and sponsible for the mutant phenotype (Table 3). diu illustrates a relationship between the size of the The mutants have been placed into seven pheno- testis and the extentof germline division and growth. typic classes, five of which involve defects in sperma- When the germline is absent ab initio, as in the prog- togenesis and two of which affect mating or sperm eny of osklosk mothers, the testis is severely stunted transfer. (Figure 20. When germline cells are present in larval, Proliferation phase defects:Mutations in two loci but not adult, gonads, as in the case of chic/Df (data cause early defects in spermatogenesis that result in not shown), the adulttestes is somewhat larger than a testes with reducedgerminal content. In wild-type germlineless testis (compare Figure 2, e and f). Last, testes, the various stages of spermatogenesis are or- when germ cells are present in the gonads of both dered from thetip to thebase of the testis. Individual larvae and young adults butare notreplenished, as in spermatogonia arise by asymmetric divisions of germ- diu males at day 5, the testis is larger still (Figure 2c). line stem cells at the tip of each testis (Figure la). Growth phase defects: Mutations in five loci result Each spermatogonium becomes surrounded by a pair in defects first detectableduring the spermatocyte of cyst cells derived from a distinct,somatic stem cell growth phase that follows the proliferative period. In population (Figure 1b). The spermatogonium and the the wild-type growth phase, high levels of transcrip- enveloping cyst cells comprise a cyst which progresses tion accompany a 25-fold increase in spermatocyte as a unit through spermatogenesis. Within each cyst, volume (Figure 1, c andd). Mature spermatocytes the spermatogonium undergoes four mitotic divisions, have large nuclei with prominent nucleoli (Figure 3a). generating16 spermatocytes, whereas the two cyst Spermatocytes from growthphase mutants are altered cells do not divide (Figure IC). In mutants affecting in appearance, exhibiting defects in the morphology the proliferation phase, a defect in the early stages of of nuclei or other cellular structures. spermatogenesis resultsin the underrepresentation or The bocce (boc) mutation at 5 1D results in sperma- absence of subsequent stages, and the testes are thin tocyte cysts that contain nuclei of variable size (Figure and short. 3b). The number of nuclei per cyst is also somewhat The phenotype of males mutantfor diaphanous variable, but is on average 16. In addition, thereis an (dia) at 38E is unusual in exhibiting a dramatic de- apparent absence ofcell boundaries within a cyst. This pendence on the age of the fly. At eclosion, testes of can be observed both in spermatocyte cysts (Figure diu flies are somewhat thinner and shorter thanwild- 3b), as well as in rare meiotic cysts, in which mito- type (compare Figure 2, a and b). As a diu male ages, chondria that normally segregate to each of the hap- the early stages of spermatogenesis are not replen- loid cells are foundfused into one or two giant masses ished, resulting in very thin testes, devoid of germinal (Figure 3c, arrow;compare with Figure5a). The contents (Figure 2c). The diu mutation is thus likely absence of cell boundaries raises the possibility that to affect the proliferation stage of spermatogenesis. bocce is required for proper cytokinesis during the In addition, defects in meiosis are observed in testes gonia1 divisions. Drosophila Male-sterile Mutants 493

TABLE 2 Loci on second and third chromosomes mutating to male sterility

Locus/cytolo~descriptionD Phenotypic Allele numbers

Proliferation defect (2 genes) chickadee/26A chic12, chic", Testes have reduced germinal content. Alleles chic" and chic" are male chic", chic", chic" semisterile. Allele chic" is female semisterile. Allele chic" is female sterile. In trans to a deficiency, inserts are male- and female-sterile and phenotypes are more severe. Locus was discovered by ScHijP- BACH and WIESCHAUS(1991). Encodes profilin (COOLEY,VERHEYEN and AYERS 1992). diaphanous/38E diu' Late-stage cysts are present at eclosion. Testes are empty in 5-day-old males. Insert is male sterile and female semisterile in trans to deficiency. Ovaries contain few egg chambers. Null alleles are lethal (D. CASTRILLONand S. WASSERMAN,unpublished results). Growth phase defect (5 genes) bocce15 1D boc' Nuclear size and number is variable in spermatocyte cysts. Infrequent spermatid cysts are highly aberrant. Testes are small. Females are semisterile; few eggs are laid. Phenotype is more severe in trans to deficiency: no spermatids are found in males; females lay no eggs. cueball162A cueZ Spermatocytes contain cytoplasmic abnormalities. Rare post-meiotic cysts are defective. Testis is short with defective sheath. Females are semisterile; ovaries are small and misshapen. Phenotypes are more severe in trans to a deficiency. Locus was discovered by P. G. WILSON and M. T. FULLER(personal communication). Testes fromfbl/'l males contain spermatids with large nebenkerne and micronuclei at onion stage. Testes fromfbl1Df males contain degenerating spermatocytes. Classification is tentative. Phenotype is more severe in trans to deficiency and earlier spermatogenic stages are affected; testes are short andnearly devoid of germinal content. scratchl63AB stc' Needle-shaped crystals accumulate throughout developing germline. Spermatids contain nuclei and nebenkerne of variable size. Pheno- type resembles that of X0 males. 1(2)26Ab/26A 1(2)26Ab4'b', Testes are short andfilled with cysts of 16-cell spermatocytes, which 1(2)26Ab5 degenerate prior tocompletion of the growth phase. Mutation is allelic to 1(2)gdh-2 (KOTARSKI,PICKERT and MACINTYRE(1983). Meiotic entry defect (2 genes) boule166F boll Some 16-cell cysts resemble those seen in pelota (see below); others contain multiple nuclei in addition to the abnormally large nebenkerne. pelo' Meiosis does not occur, but abnormally large nebenkerne form in late spermatocytes. Mutation is cold-sensitive, female-sterile, and semilethal. Meiotic division defect (3 genes) bobble/82D bob' A few spermatids per cyst have nuclei or nebenkern of abnormal size. Mutation is semilethal. doublefault/32A dV Size varies for both nuclei and nebenkern at onion stage. Spermatid nuclei fail to change shape. shank/82C shk' Spermatids contain two or four nuclei associated with a single large nebenkern. Females are semisterile. Phenotype is less severe in trans to deficiencies. Postmeiotic differentiation defect (46 genes) arrestl33CD aref' Mutation is semilethal in trans to deficiency, but male-sterile phenotype is no more severe. Locus was previously identified as male- and female-sterile (SCHUPBACHand WIESCHAUS,199 1). bellwether/58F blw' Mutation is allelic to ms(2)Pry58F (COOLEY,BERG and SPRADLING 1988). bete1/78D bet' blanbl36AB bln' capon/85A cap' Mutation is allelic to ms(3)Pneo85A (COOLEY,KELLEY and SPRADLING 1988). cashews188B cusl(b' dispersed/46C disd"" Elongated spermatid nuclei are dispersed. Males are semisterile. Mutations are allelic to ms(3)88D (BERGand SPRADLINC199 1)'. 494 D. H. Castrillon et al. TABLE 2-Continued Loci on second and third chromosomes mutatingto male sterility

Locus/Cytology Allele numbers Phenotypic descriptionu

emmenthalj56E emm' Nebenkerne are vacuolated. geldedl28D gel' gorpl83S Males are semisterile. goulashl97EF goulJ'b' gruyircj98D gru"" Nebenkerne are vacuolated. Mutation is female-sterile and semilethal. halleyj55A Spermatid nuclei fail to elongate; mutation is lethal over deficiency. hephaestus 11OOE heph', heph' Tip of testis is dilated to approximately twice wild-type circumference. jaguarj95F jar', jar2 Males are semisterile. muletl46F mWb) Mutation is allelic to ms(2)ry-3 (BERGand SPRADLING1991). oxenj49C ox ' peanutsf50D pea'(') pistachioj68C pto'(') Males are semisterile. purity of essence12'8E foe', foe' Rb97D Rb97D' Mutation is insertion in RNA-binding protein gene (KARSCH-MIZRACHI and HAYNES1993) seedless/84F sdl', sd12 scatteredj30B scat' Elongated spermatid nuclei are dispersed. thousand points of lightj86E tho' Elongated spermatid nuclei are dispersed. trail mixj57E tmx' Males are semisterile. ms( 2)21 D ms( 2)27C ms( 2)27CD Mutation is semilethal. ms( 2)29F ms(2)3OC Males are semisterile; elongated spermatid nuclei are dispersed. ms(2)42A Spermatid nuclei are not clustered. ms(2)42D Males are semisterile. ms(2)43C Spermatids occasionally contain variably sized nuclei and nebenkern. ms(2)463C Mutation is semilethal and female-semisterile. ms(2)46C Elongated spermatid nuclei are dispersed. ms( 3)6I CD ms(3)65E ms(3)72D Postmeiotic nuclei have sickled appearance. ms(3)73D ms(3)8O ms(3)85D ms(3)89B ms(3)90E Mutation is semilethal. ms( 3)988 Males are semisterile. ms( 3)lOOEF Males are semisterile. Behavioral defect (2 genes) cuckold/28A cud, CUC' Males are semisterile due tofailure to court andmate females. Longev- ity of both males and females is decreased. fruitless19 1 B fru', fru' Males court females and males, but fail to mate. Male-specific abdominal muscle is reduced. Locus was previously described by GILL(1963) and by GAILEYand HALL(1 989). Sperm transfer defect (3 genes) ken and barbiej60A ken ' External genital structures are absent in some males and females. Aris- tae are sparse and unpigmented. Mutation is semilethal and both male- and female-semisterile. pointless16 1 E ptl'(') Males are semisterile, with wild-type levels of motile sperm in seminal vesicle. Little or no sperm is transferred to females. twigj89E Iwig' Anal-genital plate is twisted in both sexes. Females are sterile. Mutations are male-sterile, female-fertile, and have good viability, unless otherwise noted. Phenotype of mutations over deficiency, where available, is similar unless noted. Mutation was induced with the P[lac, ry+]A element. All other mutations were induced with the P[Z]element. Alleles efl' and eff ', as well as two preexisting P insertions in eflete, affect viability and female fertility, but heteroallelic combinations are viable and female-fertile. Drosophila Male-sterile Mutants 495

TABLE 3 Male sterile loci by cytology

Deficiency breakpoints Polytene location Locus name Locus location locus for Deficiency Left Right R"

21D ms(2)2ZD 26A chickadee DA2L)GpdhA 25D7-El 26A8-9 J 26A 1(2)26Ab Dfl2L)GpdhA 25D7-El 26A8-9 J 27C ms(2)27C 27CD ms( 2)27CD 28A cuckold 28D gelded J 28E purity of essence 29F ms(2)29F 30B scattered DA2L)3OA; C 30A 30C 30C pelota DA2L)30A; C 30A 30C J 30C ms(2)3OC 32A doublefault D'2LU2 31B 32A 33CD arrest Dj(2L)escP3-0 33A1-2 33E 36AB blanks DA2L)H20 36A8-9 36E3-4 38E diaphanous Dj(2L)TW84 37F5-38A1 39D3-EI J 42A ms( 2)42A J 42D ms(2)42D Df(2Rlpk78s 4261-7 43F5-8 43c ms(2)43C 46AB ms( 2)46AB 46C ms(2)46C 46C dispersed J 46F mulet 49c oxen DA 2 )ug-CR 49B2-3 49E7-Fl 50D peanuts J 51D bocce ~'2RUPI (1) 51C3 52F5-9 J 55A halley DJ(2R)Pc4 55A 55F 56E emmenthal Df(2R)C100-L141 (2) 56D 56F 57E trail mix DA2R)PuDZ7 57B4 58B 58F bellwether 60A hen and barbie Df12R)bwS46 59D8-11 60A7 61B pointless J 61CD ms(3)6ZCD 62A cuebatl Df(3L)Ac14-8 (3) 61C3,4 62A8 J 63AB scratch J 65E ms( 3)65E 66F boule 66F3 67B1 J 68C pistachio 67F2-3 68D6 72D ms(3)72D 73D ms(3)73D 77B fumble Dfi 3L)rdgC (4) 77A1 77D1 78D betel Dj(3L)Pc-MK 78A3 79E1-2 80 ms(3)80 82C shank D'3R)Z-I (5) 82A5-6 82E4 82D bobble D'3R)6-7 (6) 82D3-8 82F3-6 83B gorp J 84F seedless 85A capon D!3R)P13 (7) 84F2 85B 85D ms(3)ry85D 86E thousand pointsof light Dj(3R)TE32 (8) 86E2-4 87C6-7 J 88B cashews Df(3R)redZ 88B1 88D3-4 88D effete J 89B ms(3)89B 89E twig 89B5-6 89E2-3 90E ms(3)90E 91B fruitless 91B3 91D1 J 95F jaguar 95F7 96A17-18 J 97D Rb97D 97B 97E 97EF goulash 97A 98A 1.2 98B ms(3)98B 98D gruyire J 1OOE hephaestus 1 OOEF ms( 3)ZOOEF

All deficiencies are referenced in LINDSLEYand ZIMM(1992) except as noted: (1) SAXTON etal. (1991); (2) LINDSLEYet al. (1972); (3)J. POSAKONY,personal communication; (4) STEELE and O'TOUSA(1990); (5) LETSOUet a1 (1991); (6) S. ALEXANDERand S. WASSERMAN, un ublished results; (7) K. KEMPHUES,personal communication; (8) GAUSZet al. (1991). B A checkmark under R (revertability) indicates a locus for which transposase-induced excision of the P element was carried out and male- fertile revertants were obtained; absence of a checkmark merely indicates that reversion was not attempted. 496 D. H. Castrillon et al.

germ line stem cell and two cyst e) progenitor cells

asymmerric spermatogonium division cell surrounded by @ two cyst cells 4 roundr of mitotic divisions

early primary C) spermatocyte cyst

/meiosis

spermatid elongarion individualirarion H coiling. and transfer 10 - 10 prn seminal vesicle

v motile spermatocyte cyst spenn

FIGURE1 .-Overview and schematic of key stages of spermatogenesis in Drosophila melanogaster. Structures in this figure are portrayed similarly but not exactly as they appear in phase contrast microscopy of unfixed testis contents. The two cyst cells surrounding each group of germ cells are shown for the sake of clarity but are not easily visualized by phase contrast. Nuclei are shown as open circles against a background of gray (germline cells) or white (somatic cyst cells) cytoplasm. Black spots in nuclei in a-d are nucleoli; black dots in nuclei in e are protein bodies. Structures are drawn approximately to scale. Spermatogenesis entails the differentiation of cells of both germinal and somatic origin. The precursors for these cells, the germline stem cells and the somatic cyst-progenitor cells, are attached to a specialized somatically-derived structure, the hub. at the tip of the testis (HARDYef al. 1979). At the hub, each stem cell is closely associated with two cyst progenitor cells (a). These three stem cells undergo unequal cell divisions: the parental stem cells remain attached to the hub, while the three daughter cells (one spermatogonium and two cyst cells) detach and initiate a program of differentiation (b). The transformation of a single spermatogonium to 64 sperm occurs within a thin envelope formed by these two cyst cells; this group of germline and somatically derived cells is defined as a cyst. Four rounds of mitotic division result in a cyst of 16 early primary spermatocytes interconnected by cytoplasmic bridges or ring canals (c). These spermatocytes enter a growth phase in which they undergo a 25-fold volume increase (d). During this growth phase a great deal of transcription takes place (OLIVIERIand OLIVIERI1965); the bulk of these mRNAs are stored and not translated until after meiosis (BRINK1968). The late primary spermatocytes enter meiosis, giving rise to 64 haploid spermatids. The nucleus in each spermatid reforms (open circle) and the mitochondria fuse to form the mitochondrial derivative or nebenkern (solid circle). At the onion stage, so-called because of the multilamellate appearance of the nebenkern in electron micrographs, both of these structures are closely associated and highly uniform in size and shape (e). Each young spermatid will be transformed into a mature sperm through a complex process of cytodifferentiation (TOKAYASUet al. 1972a.b). This process involves nuclear condensation aswell as the dramatic elongation of the axoneme and the mitochondrial derivative to form the sperm tail. The final steps of spermatogenesis are individualization, the process by which each elongated spermatid becomes tightly invested in its own membrane, and coiling, which results in the cyst of 64 sperm being drawn to the base of the testis. During this coiling process, grossly abnormal sperm are segregated away and are subsequently degraded. The remaining sperm are transferred to the seminal vesicle for storage. The bocce mutation is pleiotropic. Females homo- matocytes are filled with small, refractile cytoplasmic zygous for boc are semisterile. Furthermore, in trans inclusions(Figure 3d, arrows); late spermatocytes to a deficiency, the boc mutation becomes semilethal often contain up to three hook-shaped structures in and the defect in spermatogenesis becomes more se- the cytoplasm (Figure 3e, arrows). vere. Defective spermatocytecysts are similar to those The morphology of cue testes is also abnormal. seen in homozygotes; however, no postmeiotic cysts Testes are short,with incomplete coiling. The sheath are found. of pigmentcells that covers the muscle layer surround- Mutationsat the cueball (cue) locus in 62A, first ing the testis frequently peels away, particularly from identified by P. G. WILSONand M. T. FULLER(per- the region near the tip. Fertility is also affected in sonal communication), cause an accumulation of sper- females. The severity of phenotypes in both sexes is matocytes with cytoplasmic abnormalities. Early sper- moreextreme in trans to adeficiency: in females, Drosophila Male-sterile Mutants 497

I / i

/ d .#‘ f

FIGURE2.-Testis morphology of proliferation defect mutants. Testes are from newly eclosed males unless otherwise noted. Arrows at the base of the testis indicate junction between testis and seminal vesicle (shown only for a and f). Bar represents 100 pm. (a) diu/+ testis exhibiting wild type morphology. Proliferating cysts are present at thetestis tip, while growing spermatocytes are found along theinside wall of the testis through the first coil. Elongated sperm tails can be seen extending from the base almost to the tip of the testis. (b) diuldiu testis, somewhat smaller than wild type. yet with some elongated cysts, as evidenced by the presence of sperm tails. (c) diuldiu testis, 5-day old male. Testis is very thin and appears devoid of germinal contents. (d)chic”/chic” testis, slightly smaller than wild type. exhibiting some elongated sperm tails (e) chic”/DJ2L)CpdhA testis, much shorter than wild type. with severely reduced germinal contents. (f) testis from progeny of osk3’”/osk3’” mother. When raised at 18” such mothers give rise to viable progeny lacking pole cells (LEHMANNand NWSLEIN-VOLHARD 1986). The resulting germlineless adult testis is stunted. gonads that appearto lack a germlineare occasionally spermatogenesis. As for 1(2)26Ab homozygotes,fbl/Df found. males have short testes containing degenerating sper- Testes of males mutant for scratch(stc) at 63AB matocytes. In addition,with low penetrance, cysts with contain needle-shaped crystals in spermatocytes, as morethan 16 mitotically dividing cells have been well as spermatids (Figure 30. Thecrystals are similar observed in 5-bromodeoxyuridine labelling experi- in size, shape, and distribution to the proteinaceous ments (P. GONCZY,unpublished results). When homo- crystals found in the testes of X0males (MEYER1968; zygous, fumble insertionsresult in meiotic division COX, WHITEand KIEFER 1976). In addition, meiosis defects (see below). Given this range of phenotypes, in scratch males is defective. Nuclear size is variable classification of this mutant is conditional and awaits (Figure3f, arrows), as is the size of the spherical the generation of a null allele. mitochondrial mass called the nebenkern (Figure 3f, Meiotic entry defects: Mutations in two loci prevent arrowheads). These defects are also observed in X0 entry into meiosis. In wild-type testes, the two meiotic males (LIFSCHYTZ1987). Thus, the scratch insertion divisions result in cysts of 64 roundspermatids, which is a recessive-third chromosome mutation which re- are foundapproximately one-third of the distance sults in a male-sterile phenotype similar to that seen along the lengthof the testis (Figure 4, a andc). These in males lacking a Y chromosome. spermatids then move toward the base of the testis, The two alleles of 1(2)26Ab cause degeneration of while sperm tails form and elongate until they span spermatocytes priorto completion of the growth almost the entire testis length (Figure 4a). In mutants phase (Figure 3g, arrowheads) and no meiotic prod- defective for meiotic entry, haploid cells are not ucts are observed. In addition, 1(2)26Ab testes are formedand the absence of bothunelongated and shorter thanwild type. The 1(2)26Ab insertions, which elongated spermatids is readily apparent through the map to 26A, are allelic to the lethal mutation l(2)gdh- testis sheath. The entire length of testes from these 2 (KOTARSKI,PICKERT and MACINTYRE1983). In trans mutants is filled with spermatocyte cysts (Figure 4, b to l(2)gdh-2, the 1(2)26Ab insertions are male-sterile and d). but viable, indicating that these male-sterile 1(2)26Ab Spermatogenesis in males mutant for pelota (pelo), alleles do not represent a completeloss of function. which maps to 30C, appears wild-type up to the 16- Whereas males homozygous for fumble cfbl) at 77B cell spermatocyte stage. Thereafter, pelota spermato- show no defects prior to meiosis, thefbl insertion in cytes fail to undergo meiotic nuclear reduction, but trans to a deficiency causes defects at earlier stages of nonetheless differentiate further. In mutant sperma- 498 D. H. Castrillon et al.

FIGURE3.-Phenotypes of growth phase defect mutants, visualized by phase contrast microscopy of unfixed testis contents. Bars represent 10 bm; panels a, b, c, f and g are at the same magnification. (a) Wild-type spermatocytes have prominent nuclei, whose diameter increases during the growth phase (compare younger spermatocytes in top half of panel with more mature spermatocytes at bottom). (b) boclboc spermatocyte cyst. Nuclei are of variable size, and the cytoplasmic organization apparent in the wild-type is lost, giving the appearance that cell boundaries are absent. The number of nuclei is somewhat variable, but is on average 16; at least 17 nuclei are visible in this cyst. (c) boc/ boc spermatid cyst. Existence of a single giant mitochondrial mass (arrow) is indicative of a lack of cell boundaries. (d) cue2/cuez early spermatocytes. Cytoplasm of most cellscontains several refractile inclusion bodies (arrows). (e) cue2/cue2 late spermatocyte. Cytoplasm typically contains 1-3 hook-shaped structures (arrows). (f) stclstc late spermatocytes and early spermatids. Needle-shaped crystals are prominent in late-stage spermatocytes as well as in spermatids. Meiosis is severely defective, resulting in nuclei (arrows) and nebenkern (arrowheads) of variable size. (g) f(2)26Abs/f(2)26Absspermatocytes. Young spermatocytes (arrows) appear normal, but degenerate prior to completion of the growth phase. Degenerating spermatocytes lyse, as evidenced by the loss of cytoplasm surrounding thenuclei (arrowheads). tid cysts, the nuclei become pale and homogeneous, The phenotype of some cysts from males mutant but remain as large as spermatocyte nuclei. The mi- for boule (bol) at 66F is very similar to that seen for tochondria fuse to form an abnormal, large nebenk- pelota. Such cysts are comprised of 16 cells, each ern (Figure 4e). Little differentiation takes place be- containing a large nucleus and nebenkern resulting yond this stage, as very few elongated cysts are seen. from postmeiotic differentiation in the absence of The male-sterile pelota phenotype is reminiscent of meiosis. In other cysts, 16 large nebenkerne areasso- that of a mutation in twine, a Drosophila cdc25 hom- ciated with a greater number of variably sized nuclei. olog (COURTOTet al. 1992; ALPHEYet al. 1992), Thus, in this subset of cysts, aberrant nuclear division certain heteroallelic combinations of haywire muta- appears to takeplace in the absence of cytokinesis. tions (REGAN andFULLER 1990) and three X-linked Meiotic division defects: Mutations in three loci male-sterile mutations (LIFSCHYTZand HAREVEN result in defects first detectable during the meiotic 1977), in that aspects of postmeiotic differentiation divisions. Following wild-type meiosis, each spermatid occur in the absence of nuclear reduction. The pelota nucleus appears as a pale sphere (Figures le and 5a). mutation also results in female sterility; pelota females Next to each nucleus, and equal in size, lies a dark, have small ovaries and lay very few eggs. At 18", spherical nebenkern. Since nuclear diameter in early pelota becomes semilethal and both the spermatogen- spermatids correlateswith chromosome content(GON- esis and oogenesis phenotypes are moresevere. ZALEZ,CASAL and RIPOLL1989), a defectin chromo- Drosophila Male-sterile Mutants 499

FIGURE4.-Phenotypes of a meiotic entry defect mutant. Panels ad viewed with Nomarski optics. Bar in b represents 100 pm (panel a is at same magnification); bars in c and d represent 50 pm; bar in e represents 10 pm. (a) pelo/+ testis with elongating spermatids; sperm tails can be seen stretching throughout the entiretestis. (b) pelolpelo testis filled with premeiotic cysts. No elongated cysts are present. (c) Higher magnification of framed region in panel a. Nuclei of wild-type unelongated spermatids, the immediate products of meiosis, appear as small spheres and arefound approximately one-third of the distance along the length of the testis. Note the neighboring sperm tails, which belong to older cysts. (d) Higher magnification of framed region in panel b. Meiosis does not occur in the pelo testis and the lumen of the testis fills with large. spherical premeiotic germ cells. (e) contents of pelolpelo testis, viewed in phase contrast. pelo spermatocytes mature fully, then form a nebenkern (arrow)adjacent to the nucleus.

FIGURE5.-Phenotypes of meiotic division defect mutants visualized by phase contrast microscopy of unfixed testis contents. Bar represents 10 pm. (a) wild-type spermatids. Each spermatid contains a single nucleus (appear as clear circles) and a single nebenkern (appear as dark circles) of uniform size. (b) shR/shR spermatids. Typically two or four nuclei of uniform size are associated with abnormally large nebenkern (arrows); some abnormally small nuclei are also present (arrowhead). (c)fbl'/fbl' spermatids. Large, pale nebenkerne are associated with two to four very small nuclei of uniform size. (d) dbfldbf spermatids. Both nuclei and nebenkern are variable in size. (e) boblbob spermatids. Meiosis is mildly defective. Most spermatids appear normal but some contain abnormally small nuclei (arrow). The nebenkerne in this cyst appear variable in shape because the cyst as slightly older and thespermatids have begun to elongate. 500 D. H. Castrillon et al. some segregation manifests itself by variable nuclear Postmeiotic differentiation defects: The majority size. Similarly, since inhibition of cytokinesis leads to of the male-sterile mutations result in defects in post- the aggregation of the four nebenkerne intoa single meiotic differentiation. Following completion of mass (LIEBRICH1982), a defect in this process is ap- meiosis, wild-type haploid germ cells differentiate and parent from a disruption in the normal one-to-one elongate. The clustered spermatid nuclei are reduced ratio between nebenkerne and nuclei. For mutations in volume and transformed in shape from spheres to with defects in eitherchromosome segregation or long, thin cylinders as they reachthe base of the testis cytokinesis, spermatid differentiation proceeds none- (Figure 6a). At the same time, an axoneme develops, theless and elongated cysts are produced. along with the associated mitochondrial derivative. At The shank (shk)mutation at 82Cresults in abnormal the end of the differentiation program, the mature spermatids. A single large nebenkern (Figure 5b, ar- spermatozoaexit the testis andenter the seminal rows) associated with two or four nuclei is indicative vesicle for storage. Mutants in this class successfully of a defect in cytokinesis. The nuclei in most sper- carry out meiosis, but produce no motile sperm for matids have asize and shape similar to wild-type; other passage from the testis into the seminal vesicle. spermatidscontain a large number of abnormally For several mutants we have been able to detect small nuclei (Figure5b, arrowhead). A phenotype defects in nuclear shaping or clustering by staining resembling that of shank has been observed for the fixed testes with the DNA-binding dye Hoechst male-sterilefour-wheel drive locus (FULLER 1993).The 33258. The halley (hal) mutation at 55A results in a shank mutation has less severephenotypic conse- failure of most spermatid nuclei to change shape; the quences in trans to a deficiency. clustered nuclei remain round even as they reach the Whereas fbl/Df males exhibitdefects priorto base of the testis (Figure 6b, arrow). In addition,some meiosis (see above), testes fromfbl/’l males contain of the nuclei are scattered (arrowhead). Thesedefects early spermatids in which groups of two or four nuclei are also detectable in testis squashes. Round nuclei are associated with a single, abnormally large nebenk- are found in halley spermatid cysts (Figure 6e) at a ern (Figure 5c). The fbl locus hence appears to be stage when wild-type nuclei are no longervisible, and required for cytokinesis. The nuclei in many of these some nuclei are scattered throughout the length of fumble spermatids are smaller than wild type, but are the cyst (Figure 6e, arrows). uniform in size.The uniform size of these micronuclei The ms(2)46C mutation results in nuclei that be- may indicate that they result from abnormal conden- come cylindrical, as in wild-type, butare not fully sation rather thandefective chromosome segregation. condensed and fail to remain clustered (Figure 6c). In Nuclear defects persist duringfbl spermatid differen- ms(3)72D homozygotes, the nuclei appear contorted tiation; nuclei in later spermatids fail to change shape into various shapes rather than extended in the nor- but appearinstead as clear circularbodies throughout mal configuration (Figure 6d). Other mutations with the length of elongating cysts. Such spherical nuclei nuclear defectsapparent in Hoechst stainingare listed are also seen among late stage spermatids in several in Table 2. other mutants from our collection (see Table 2 and Mutations in emmenthal (emm) at 56E and gruyire Figure 6), as well as sterile males with Xautosome (gru) at 98Daffect nebenkern morphogenesis, as vis- translocations (SHOUP1967) and tra2 males (BELOTE ualized by phase contrast microscopy. Nebenkerne and BAKER 1983). from emm males are highly vacuolated, but are oth- An insertion in doublefault (dbf) at 32A results in erwise normal in size and shape, as well as in their abnormally sized nebenkerne and nuclei in ahigh association with nuclei (Figure 6f, arrows). gru gives proportion of cysts (Figure 5d). Both cytokinesis and rise to a similar phenotype, but fewer nebenkerne are chromosome segregation may therefore be defective. vacuolated. A number of D. hydei mutants with similar Abnormal morphology is also apparent in elongating phenotypes have beendescribed (HACKSTEINet al. spermatids: the nuclei fail to change shape and are 1990). found scattered throughout the cysts. Behavioral defects: Mutations in two loci result in The bobble (bob) mutation at 82D exhibits mildly behavioral sterility. Males mutant for the previously defective spermatid cysts.Most spermatids are nor- described fruitless V;u) locus at 91B exhibit aberrant mal, but a few areaberrant, containing nuclei of courtship behavior, courting females and males indis- variable size (Figure 5e, arrow). The lack of uniform criminately (GILL1963; GAILEYand HALL 1989).The nuclear size may reflect abnormal chromosome seg- P[Z] alleles are sterile as homozygotes or in trans to a regation. If so, however,a variable nuclear DNA deficiency for the locus, but do not appear tobe null: content cannot of itself be responsible for the lack of unlike such a deficiency, they show some fertility in motile sperm in bob males, since even complete ab- trans to an inversion infru. In addition, homozygotes sence of chromosomal content doesnot block produc- for both alleles lack most of the male-specific abdom- tion of functional sperm (LINDSLEYand GRELL 1969). inal muscle, whereas males carrying some other allelic Drosophila Male-sterileMutants 50 1

FIGURE6.-Phenotypes of post-meiotic differentiation defect mutants. (a-d) Testes stained with Hoechst 33258. Photographs are of the base of the testis with the base toward the right. Phase contrast views of unfixed testis contents (e andf). For both sets of panels, bar represent 10 pm. (a) Elongated spermatid nuclei from hall+ testis exhibiting wild-type morphology. Several clusters of threadlike spermatid nuclei are visible (some clusters are out of focus). (b) Nuclei from hallha1 testis fail to elongate. Although most nuclei remain clustered (arrow), others are scattered (arrowhead). (c) Nuclei from ms(2)46C/ms(2)46C testis are completely scattered; these nuclei have elongated but are not fully condensed. (d) Nuclei from ms(3)72D/ms(3)72D testis adopt defective shapes but remain clustered. (e) hallha1 elongating spermatid cyst with round nuclei (appearing as clear spheres). Most nuclei are located at the head end of the cyst (left), but some are scattered throughout the cyst (arrows). [Note that cyst cell nuclei are easily seen in this intact cyst (arrowheads).] (f) Spermatids from emmlemm testis. Nebenkerne contain one or more vacuoles (arrows). combinations lack the muscle entirely (GAILEY,TAY- asa rotated anal-genital plate, in both males and LOR and HALL199 1). females. Males homozygous for either of the cuckold (cuc) The seminal vesicles of males homozygous for the alleles at 28A are semisterile. When individual pairs pointless (ptl)mutation at 6 1B containwild-type levels were observed for 30 min, cuc males failed to court of motile sperm. However,dissection of virgin females or mate with virgin females, whereasheterozygous immediately after mating with ptl males reveals that control males courted, and in most cases, mated dur- few or no sperm are transferred to the female sperm ing the observation period. Males and females homo- storage organs. zygous for either allele have decreased longevity; al- Expression of enhancer trap reporter genes:Each most all flies die within 8 days after eclosion. These of the P elements used to generate our male-sterile results might suggest that mutations in any gene re- collection contains a lacZ enhancer trap, a bacterial sulting in general metabolic debilitation would be reporter gene that can be brought under the control male-sterile due toa failure to carry out courtship(see of regulatory elements nearthe site of insertion in the also CARTHEW andRUBIN 1990). We note, however, genome (O'KANE and GEHRINC 1987). The pattern that theonly two male-sterile mutants in our collection of expression of the lacZ reporter gene often reflects which do not court areallelic to one another. that of the disruptedlocus and may therefore provide Sperm transfer defects:Mutations at two loci result information about thedomain of expression of a gene in structural barriers to copulation. These mutations of interestbefore it is cloned(FASANO, CORL and cause partial or complete sterility, despitethe presence KERRIDCE1988; BIERet al. 1989; WILSONet al. 1989). of mature gametes. The ken and barbie (ken)mutation The pattern of expression of the lacZ enhancer trap at 60A causes semilethality as well as an incompletely was examined for the 77 male-sterile lines in which penetrant sterile phenotype. Sterilemales and females spermatogenesis was defective. In 7 1 of these lines, /3- lack external genitalia, but dissection frequently re- galactosidase expression was observed in specific sub- veals the presence of internalized genitalia. In addi- sets of somatic and germline cells of the testis (data tion, aristae are sparse and unpigmented,giving them not shown). We considered the possibility that some a blond appearance. The twig mutation at 89E is also mutations might alter cell fate and that such altera- pleiotropic, leading to unpigmented bristles, as well tions in cell fate would be revealed by a change in the 502 D. H. Castrillon et al. expression of the lacZ reporter gene in the mutants. Consideration of allele frequencies indicates that We therefore looked for differences in the X-gal our collection does not represent the majority of loci staining pattern of heterozygotes and homozygotes, mutating to a male-sterile phenotype. Only two loci, but did not find any. Nor did we find that theexpres- chickadee and effe, are represented by more than two sion of the reporter gene was restricted to the stages alleles; both represent previously identified hotspots of spermatogenesis affected by the insertion. for P element insertion (BERGand SPRADLING1991 ; We did find a striking enrichment for one expres- COOLEY,VERHEYEN and AYERS 1992).Of theremain- sion pattern among the lines defective for spermato- ing loci, 53 out of 61 are represented by only a single genesis: staining in all germline cells of the testis. allele. It is not surprising, therefore, that we did not Among the 77 lines, 21 expressed fl-galactosidase in obtain mutationsin a number ofautosomal loci known the germline beginning in stem cells and persisting tomutate to male sterility, such as bag-ofmarbles, through the spermatid stage. By comparison, in an benign gonia1 cell neoplasm, ms(3)sa,twine or BTtubulin, examination of 700 viable fertileinserts of the particularly since some loci may not be mutable with P[lacW] element, only two lines had expression P elements. Nevertheless, we did isolate mutations in throughout the adult male germline (GONCZY,VIS- previously characterized male-sterile loci, including WANATHAN and DINARDO1992; P. GONCZYand s. arrest, chickadee and fruitless. We also identifieda DINARDO,unpublished results). mutation in a locus foran RNAbinding protein (Rb97D), for which there were no known mutations DISCUSSION (KARSCH-MIZRACHIand HAYNES 1993),as well as a male-sterile allele of apreviously identified lethal com- Spermatogenesis in Drosophila offers a variety of plementation group (1(2)gdh-2). challenges and opportunities to developmental biolo- Classification of male-sterilemutations: The gists. Arange of fundamental processes, including range of male-sterile phenotypes we observed is simi- cell-cell interaction, mitosis, meiosis, and morphoge- lar to that seen in screens with chemical mutagens nesis, all occur in a stereotyped spatial and temporal (LIFSCHYTZ1987; HACKSTEINet al. 1990; HACKSTEIN pattern. Moreover, cells from all stages of spermato- 199 1). As in these previous screens, mutationsaffect- genesis can be simultaneously observed in a single ing germline proliferation, spermatocyte growth and testis. meiosis were relatively rare, while mutations affecting From a genetic standpoint, spermatogenesisis readily tractable, since sterile mutations can be easily gen- postmeiotic differentiation were the most common. erated, identified, and investigated. The presence or We classified male-sterile phenotypes with regard absence of a germline, the number and quality of to thedistinct stages that characterize spermatogenesis mitotic and meiotic divisions, and the production of and grouped themutations based on the earliest phe- motile spermcan all be scored in the light microscope. notypic deviation from wild-type differentiation ob- Mutations provide the key to initiating molecular bi- servable in the light microscope. These classes are ological studies, a transition greatly facilitated by the broadand surely encompass a number of cellular use of mutations induced with single P elements. processes. Nevertheless, we believe that the current Mutation frequencies: Male-sterile mutations were classification of mutants represents a good starting identified in our screen about one-tenth as frequently point. By focusing on loci in which mutations affect a as lethal mutations. Given that the number of loci in given developmentalstage, it should be feasible to Drosophila melanogaster that can mutate to lethality is dissect particular processes in spermatogenesis,de- about half the number of polytene bands (PERRIMON, spite the large number of genes required overall. ENGSTROMand MAHOWALD1989), or approximately Two lines of evidenceindicate thatthe earliest 2500, it can be calculated that at least 250 genes spermatogenesis phenotype detected in our mutants should mutate to male-sterility. In fact, as discussed is a property of the locus rather than of particular by others (LINDSLEYand TOKAYASU1980; PERRIMON mutant alleles. First, in the eight cases for which et al. 1986; SCHUPBACHand WIESCHAUS1991), the multiple alleles of a locus were isolated, all alleles number of genes required for fertility in either sex is belong to the same phenotypic category. Second,with likely to be much greater. Germline clones for a large one exception, when a mutation was placed in trans fraction of zygotic lethal mutations on theX chromo- to a deficiency, the resulting defects in spermatogen- some are sterile in females (PERRIMON,ENGSTROM esis fell into thesame phenotypic class as observed for and MAHOWALD1989). Moreover, flies carrying tem- the homozygous P element. This was true even in perature-sensitive lethal mutations often become ster- those cases where the insertionwas homozygous viable ile when shifted to non-permissive conditions as adults and female fertile, but became semilethal or female (SHELLENBARCERand CROSS1979). Thus many loci sterile in trans to the deficiency. required for fertility may mutate only rarely, or not Some mutationsblock the transition from one stage at all, to give viable, sterile phenotypes. to the next. In such cases, the mutation may define a Drosophila Male-sterile Mutants Male-sterile Drosophila 503 step in a dependent series along the spermatogenic ing sequences in the 5' end of a gene, they may be pathway. For example,mutations in 1(2)26Ab block particularly effective at inducing tissue or sex-specific development before the late spermatocyte stage. In alleles of essential loci. For example, many P element the wild-type, therefore, subsequentprocesses depend insertions in chickadee that are viable but sterile lie on the execution of the step requiring 1(2)26Ab func- within analternative 5' exon (COOLEY,VERHEYEN tion. and AYERS 1992). Spermatogenesis, is not, however, simply a linear Proliferation phase defects:Mutations that disrupt series of dependent steps (LIFSCHYTZ1987). For ex- male germline proliferation provide an opportunity ample, males lacking a Y chromosome exhibit defects to begin to dissect processes of inherent interest in in the spermatocytegrowth phase, but still initiate spermatogenesis: the control of stem cell identity, the meiosis and spermatid differentiation (MEYER1968). synchronization of germlineand somatic stem cell A number of mutations in our collection also affect a divisions, and theregulation of the numberof mitoses. given stagewithout blocking furtherdevelopment. Germline proliferation is affected by mutations in For example, the scratch mutation results in the for- two of our male-sterile loci, chickadee and diaphanous. mation of crystals apparent at the growth phase as The phenotypes of these mutations suggest that these well as additional defectsat meiosis. These phenotypes loci act prior to the growth phase. For chic, a prolif- may reflect a complete loss of function in one of two eration defect has been demonstrated directly by an or more parallel processes in the spermatogenicpath- examination of 5-bromodeoxyuridine labeling in mu- way. It is also possible, however, thatthe allele is tant and wild-type testes (P. GONCZYand S. DINARDO, hypomorphic and thata null mutation would result in unpublished results). For dia, null alleles are lethal at a complete developmental arrest. the pupal stage and produce aberrantmitotic figures Sterile mutations in essential genes: A number of in larval brains (D. CASTRILLONand S. WASSERMAN, our male-sterile mutationsaffecting early stages of unpublished results). Since P elementmutations in spermatogenesis disrupt gamete development in both either genelead to a loss of germline cells in the testis, sexes (Table 2). This is not unexpected, since there chic and dia are complementary in phenotype tobenign are many similarities in the early steps of male and gonia1 cell neoplasm and bag-ofmarbles, mutations that femalegametogenesis and thus many opportunities cause an overproliferation of germline cells (GATEFF for shared gene function(SCHUPBACH and WIESCHAUS 1982; MCKEARINand SPRADLING 1990). It remainsto 199 1). Similarly, the fact that many of these mutations be proven whether mutations in these loci affect cell also affect viability presumably reflects the existence fate determinationor the subsequent execution of the of functionscommon to cell growth or division in mitotic cell cycle. both the germline and the soma. Indeed, for females, Meiotic entry defects:Male-sterile mutations defec- at least, it appears that most genes which mutate to sterility are not exclusively required for gametogene- tive for entry into meiosis are also of experimental sis (PERRIMONet al. 1986). interest. Whereas the regulation of the meiotic cell Male-sterile or female-sterile alleles of essential loci cycle has been studied extensively in yeast (MALONE offer the opportunity to study phenotypes that are 1990), much less is known about this process in mul- frequently more readily interpretable than the lethal- ticellular organisms. Although gametogenesis in flies ity that characterizes null mutations. For example, in has been studied more extensively in females than in the case of the Drosophila homolog of the epidermal males, spermatogenesis offers a significant advantage growth factor (EGF) receptor, a requirement for the for the study of meiosis. The two meiotic divisions in locus in dorsoventral pattern formation was apparent spermatocytes occur in rapid succession in a discrete from characterization of female-sterile torpedo alleles portion of theadult testis. Incontrast, meiosisin (Egfr'),but not zygotic lethalfaint little ball mutations females begins in the nuclei of pro-oocytes but is not (EgfrI) in the same locus (SCHUPBACH1987; PRICE, completed until after fertilization, at least three days CLIFFORDand SCHUPBACH1989; SCHEJTERand SHILO later. 1989). Similarly, female-sterile alleles of chickadee, an The premeiotic cysts found in testes from pelota essential locus (E. VERHEYEN andL. COOLEY, personal males undergo incomplete chromosome condensation communication), revealed the participation of profi- and thus have arrested at the earliest steps of meiosis lin, an actin-binding protein,in the anchoringof nurse (C. EBERHARTand S. WASSERMAN,unpublished re- cell nuclei within the developing eggchamber sults). Nevertheless, in pelotu males, cells from these (COOLEY,VERHEYEN and AYERS1992). Male-sterile arrested cysts forma nebenkern, a mitochondrial chickadee alleles, which result in testes nearly devoid form normally found only in postmeiotic spermatids. of germinal content,are likely to offeradditional This uncoupling of meiosis and postmeiotic differen- insight into profilin function. tiation has also beenobserved for twine, as well as Since P elements frequently insert upstreamof cod- several other male-sterile mutations (ALPHEYet al. 504 D. H. Castrillon et al.

1992; COURTOTet al. 1992; LIFSCHYTZand HAREVEN L. ACKERMAN,R. CARRETTO, T. UEMURA,E. GRELL,L. Y. JAN 1977; RECANand FULLER1990). and Y. N. JAN, 1989 Searching for pattern and mutation in the Drosophila genome with a P-lac2 vector. Genes Dev. 3 Some mutations thatwe have placed into thegrowth 1273-1287. phase class may also define genes required for entry BRINK,N. G., 1968 Protein synthesis during spermatogenesis in into meiosis. For example, in 1(2)26Ab mutant testes, Drosophila melanogaster. Mutat. Res. 5: 192-194. maturing spermatocytes degenerate and no meiotic CARTHEW,R. W., and G. M. RUBIN, 1990seven in absentia, a gene products are observed. The possibility exists that germ required for specification of R7 cell fate in the Drosophila eye. Cell 63: 561-577. cell degeneration is a secondary consequence of an COOLEY,L., C. BERG and A. C. SPRADLING,1988 Controlling P alteration in cell fate due to the absence of a signal element insertional mutagenesis. Trends Genet. 4: 254-258. required for initiation of meiosis. COOLEY,L., R. KELLEY and A.C. SPRADLING,1988 Insertional Prospects for further genetic analysis of sperma- mutagenesis of the Drosophila genome with single P elements. togenesis: It is noteworthy that mutations in which Science 239 1121-1 128. COOLEY,L., E. VERHEYEN andK. AYERS,1992 chickadee encodes there is an underproliferation of the germline, or a a profilin required for intercellular cytoplasm transport during degeneration of thegermline following develop- Drosophila oogenesis. Cell 69: 173-184. mental arrest, alter testis morphology. Testes from COURTOT,C., C. FRANKHAUSER,V. SIMANIS and C.F. LEHNER, males homozygous for mutations in diaphanous, chick- 1992 The Drosophila cdc25 homolog twine is requiredfor adee, bocce, cueball, boule and 1(2)26Ab are substantially meiosis. Development 116: 405-416. Cox, G. N., J. D. WHITEand B. I. KIEFER, 1976 Isolation and reduced in size. Mutations in bag-ofmarbles and benign partial characterization of X0 spermatocyte crystals in Dro- gonia1 cell neoplasm also result in small testes. That the sophila melanogaster. Genetics 83: SI 7. size of the testis is highly correlated with the fate of EDGAR,B. A., and G. SCHUBIGER,1986 Parameters controlling germline cells is also evident in the tiny testes from transcriptional activation during early Drosophila develop- the germlineless progeny of osRar females. ment. Cell 44871-877. FASANO,L., N. CORE and S. KERRIDGE,1988 Expression of a The short or shriveled testis phenotypecould reporter gene resembles that of its neighbour: an insertion in greatly facilitate the isolation of mutations in loci the hairy gene of Drosophila. Roux’s Arch. Dev. Biol. 197: 507- required for germline proliferation and growth. This 512. phenotype is immediately obvious upon dissection of FULLER,M., 1993 Spermatogenesis, in DevelopmentofDrosophila, homozygous males, as is the absence of elongating edited by A. MARTINEZ-ARIASand M. BATE. Cold Spring Harbor Laboratory Press, New York. cysts inthe testis. Thus onecould screenfor mutations GAILEY,D. A,, and J. C. HALL, 1989 Behavior and cytogenetics in genes required for early stages of spermatogenesis offruitless in Drosophila melanogaster: Different courtship de- without the need for labor-intensive fertility tests. fects caused by separate, closely linked lesions. Genetics 121: 773-785. We would like to thank MONICA BOYLE,LYNN COOLEY, SCOTT GAILEY,D. A., B. J. TAYLOR,and J. C. HALL, 1991 Elements of DOUGAN,MINX FULLER, CATHYLAURIE, CAI LI, KATHYMATTHEWS thefruitless locus regulate development of the muscle of Law- and the Bloomington Drosophila Stock Center, DENNISMCKEARIN, rence, a male-specific structure in the abdomen of Drosophila JOHN MERRIAMand the UCLA fly laboratories, MICHELLESANI- melanogaster adults. Development 113: 879-890. COLA, ALLANSPRADLING, TRUDISCHUPBACH, ESTHERVERHEYEN GATEFF,E., 1982 Gonia1 cell neoplasm of genetic origin affecting and especially TERESAGALLARDO, SUSAN GILLESPIE and CHRIS both sexes ofDrosophila melanogaster,pp. 621-632 in Embryonic SHELTONfor assistance with this project. S.D. is, and S.A.W. was, a Development, Cellular Aspects, edited by M. M. BURGERand R. Lucille P. Markey Scholar, and this work was supported in part by WEBER.Alan R. Liss, New York. the Lucille Markey Charitable Trust. We also gratefully acknowl- GAUSZ,J., H. GYURKOVICS,G. BENCZE,A. A. M. AWAD,J. J. edge predoctoral support for C.G.E. from the March of Dimes and HOLDENand D. ISH-HOROWICZ,1991 Genetic characteriza- for D.H.C. through Medical Scientist Training grant GM-08014 tion of the region between 86F1,2 and 87B15 on chromosome from the National Institutes of Health. 3 of Drosophila melanogaster: Genetics 98: 775-789. Note: A representative allele for each of the male-sterile loci GILL, K. S., 1963 Amutation causing abnormal courtship and described in this paper is available from theBloomington Drosophila mating behavior in males of Drosophila melanogaster (abstract). Stock Center. American Zoology 3: 507. GONCZY,P., S. VISWANATHANand S. DINARW, 1992 Probing LITERATURECITED spermatogenesis in Drosophila with P-element enhancer detec- tors. Development 114: 89-98. ALPHEY,L., J. JIMENEZ,H. WHITE-COOPER,I. DAWSON,P. NURSE GONZALEZ, C., J. CASALand P. RIPOLL,1989 Relationship be- and D.M. GLOVER,1992 twine, a cdc25 homolog that func- tween chromosome content and nuclear diameter in early tions in the male and female germline of Drosophila. Cell 69 spermatids of Drosophilamelanogaster. Genet. Res. 54: 205- 977-988. 221. ASHBURNER,M., 1989 Drosophila: A LaboratoryHandbook. Cold HACKSTEIN,J. H. P., 1991 Spermatogenesis in Drosophila. A ge- Spring Harbor Laboratory, Cold Spring Harbor, N.Y. netic approach to cellular and subcellular differentiation. Eur. BELOTE,J. M., and B. BAKER,1983 The dual functions of a sex J. Cell Biol. 56 151-169. determination gene in Drosophila melanogaster. Dev. Biol. 95: HACKSTEIN,J. H. P., H. BECK,R. HOCHSTENBACH,H. KREMER and 512-515. H. ZACHARIAS 1990 Spermatogenesis in Drosophila hydei: a BERG,C. A,, and A. C. SPRADLING,1991 Studies on the rate and genetic survey. Roux’s Arch. Dev. Biol. 199 251-280. site-specificity of P element transposition. Genetics 127: 5 15- HARDY,R. W., K. T. TOKUYASU,D. L. LINDSLEY and M. GARAVITO, 524. 1979 The germinal proliferation center in the testis of Dro- BIER,E., H. VAESSIN,S. SHEPHERD,K. LEE,K. MCCALL,S. BARBEL, sophila melanogaster. J. Ultrastruct Res. 69 180-190. M ale-sterile MutantsDrosophila Male-sterile 505

KARPEN,G. H., and A. C. SPRADLING,1992 Analysis of subtelom- in Drosophila: its application to oeurobiology. Methods Neu- ericheterochromatin in the Drosophila minichromosome rosci. 9: 397-414. Dpll87 by single Pelement insertional mutagenesis. Genetics O'KANE,C. J., and W. J. GEHRING,1987 Detection insitu of 1.92: 737-753. genomic regulatory elements in Drosophila. Proc. Natl. Acad. KARSCH-MIZRACHI,I., and S. R. HAYNES,1993 The Rb97D gene Sci. USA 84: 9123-9127. encodes a potential RNA-binding protein required for sper- OLIVIERI,G., and A. OLIVIERI,1965 Autoradiographic study of matogenesis in Drosophila. Nucleic Acids Res. (in press). nucleic acid synthesis during spermatogenesis in Drosophila KEMPHUES,K. J., R.A. RAFF,T. C. KAUFMANand E. C. RAFF, melangaster. Mutat. Res. 2: 366-380. 1979 Mutation in a structural gene for a @-tubulinspecific to PERRIMON,N., L. ENGSTROM~~~A. P. MAHOWALD, 1989 Zygotic testis in Drosophilamelanogaster. Proc. Natl. Acad. Sci. USA lethals with specific maternal effect phenotypes in Drosophila melanogaster. I. Loci on the X chromosome. Genetics 121: 76 3991-3995. 333-352. KEMPHUES,K. J., E. C. RAFF,R. A. RAFFand T. C. KAUFMAN. PERRIMON,N., D. MOHLER,L. ENGSTROMand A. P. MAHOWALD, 1980 Mutation in a testis-specific @-tubulin in Drosophila: 1986 X-linked female-sterile lociin Drosophilamelanogaster. analysis of its effects on meiosis and map location of the gene. Genetics 113: 695-7 12. Cell 21: 445-451. PRICE,J. V., R. J. CLIFFORDand T. SCHUPBACH,1989 The mater- KOTARSKI,M. A., S. PICKERTand R. J. MACINTYRE,1983 A nal ventralizing locus torpedo is allelic to faint little ball, an cytogenetic analysis of the chromosomal region surrounding embryonic lethal and encodes the Drosophila EGF receptor the a-glycerophosphate dehydrogenase locus of Drosophila mel- homolog. Cell 56: 1085-1092. anogaster. Genetics 105: 371-386. REGAN,C. L., and M. T. FULLER,1990 Interacting genes that LEHMANN,R., and C. NUSSLEIN-VOLHARD,1986 Abdominal seg- affect microtubule function in Drosophilamelanogaster: two mentation, pole cell formation, and embryonic polarity require classes of mutation revert the failure to complement between the localized activity of oskar, a maternal gene in Drosophila. hay"" and mutations in tubulin genes. Genetics 125 77-90. Cell 47: 141-152. ROBERTSON,H. M., C. R. PRESTON,R. W. PHILLIS,D. M. JOHNSON- LETSOU,A. L., S. ALEXANDER,K. ORTH and S. A. WASSERMAN, SCHLITZ,W. K. BENZand W. R. ENGELS,1988 A stable 1991 Genetic and molecular characterization of tube, a Dro- genomic source of P element transposase in Drosophila mela- sophila gene maternally required for embryonic dorsoventral nogaster. Genetics 118: 461-470. polarity. Proc. Natl. Acad. Sci. USA 88: 8 10-8 14. SAXTON,W. M., J. HICKS,L. S. B. GOLDSTEINand E. C. RAFF, LIEBRICH,W., 1982 The effects of cytochalasin B and colchicine 1991 Kinesin heavy chain is essential for viability and neuro- on the morphogenesis of mitochondria in Drosophila hydei muscular functions in Drosophila, but mutants show no defects during meiosis and early spermatogenesis. Cell Tissue Res. in mitosis. Cell 64: 1093-1 102. 224 161-168. SCHEJTER,E. D., and B. 2. SHILO, 1989The Drosophila EGF LIFSCHYTZ,E., 1987 The developmental program of spermioge- receptor homolog (DER) gene is allelic tofaint little ball,a locus nesisin Drosophila: a genetic analysis. Int. Rev. Cytol. 109: essential for embryonic development. Cell 56: 1093-1 104. 211-257. SCHUPBACH,T., 1987 Germline and soma cooperate to establish LIFSCHYTZ,E., and D. HAREVEN, 1977Gene expression and the the dorso-ventral pattern of egg shell and embryo in Drosophila control of spermatid morphogenesis in Drosophila melanogaster. melanogaster. Cell 49: 699-707. Dev. Biol. 58: 276-94. SCHUPBACH,T., and E. WIESCHAUS,1991 Female sterile mutations on the second chromosome of Drosophila melanogaster. LINDSLEY,D. L., and E. H. GRELL,1969 Spermiogenesis without Mutations blocking oogenesis or alteringegg morphology. chromosomes in Drosophila melanogaster. Genetics 61 (Suppl. Genetics 129: 11 19-1 136. 1): ~69-~78. SHELLENBARGER,D. L., and D. P. CROSS, 1979 Genetic dissection LINDSLEY,D. L., and K. T. TOKUYASU,1980 Spermatogenesis, of fertility in Drosophila males: Properties of temperature- pp. 225-294 in The Geneticsand Development of Drosophila, Vol. sensitive lethal-temperature-sensitive male-sterile mutations. 2, edited by M. ASHBURNERand T. R. F. WRIGHT.Academic Dev. Biol. 71: 308-322. Press, London. SHOUP,J., 1967 Spermiogenesis in wildtype and in a male sterility LINDSLEY,D. L., and G. G. ZIMM,1992 The Genome of Drosophila mutant of Drosophila melangaster. J. Cell Biol. 32: 663-675. melanogaster. Academic Press, New York. STEELE,F., and J. E. O'TOUSA,1990 Rhodopsin activation causes LINDSLEY,D. L., L. SANDLER,B. S. BAKER,A. T. C. CARPENTER, retinal degeneration in Drosophila rdgc mutant. Neuron 4: R. E. DENELL,J. C. HALL,P. A. JACOBS, G. L. G. MIKLOS,B. 883-890. K. DAVIS, R. C. GETHMANN,R. W. HARDY,A. HESSLER,S. M. TOKAYASU,K. T., W. J. PEACOCK,and W.R. HARDY, MILLER, H. NOZAWA,D. M. PARRY andM. GOULDSOMERO, 1972a Dynamics of spermiogenesis in Drosophilamelano- 1972 Segmental aneuploidy and the genetic gross structure gaster. I. Individualization process. 2. Zellforsch. 124 479- of the Drosophila genome. Genetics 71: 157-184. 506. MALONE,R. E., 1990 Dual regulation of meiosis in yeast. Cell 61: TOKAYASU,K. T., W. J. PEACOCK,and W.R. HARDY, 375-378. 1972b Dynamics of spermiogenesis in Drosophilamelano- MCKEARIN,D. M., and A. C. SPRADLING,1990 bag-of-marbles: a gaster. 11. Coiling process. Z. Zellforsch. 127: 492-525. Drosophiia gene requiredto initiate both male and female WILSON,C., R.K. PEARSON,H. J. BELLEN,C. J. O'KANE, U. gametogenesis. Genes Dev. 4: 2242-225 1. GROSSNIKLAUS~~~W. J. GEHRING, 1989 P-element-mediated MEYER,G. F., 1968 Spermiogenese in normalen und Y-defizien- enhancer detection: anefficient method for isolating and char- ten mannchen von Drosophilamelanogaster und D. hydei. Z. acterizing developmentally regulated genes in Drosophila. Zellforsch. 84 141-175. Genes Dev. 3: 1301-1313. MLODZIK,M., and Y. HIROMI,1992 The enhancer trap method Communicating editor: M. T. FULLER