The Drosophila melanogaster Sex Determination Gene sisA Is Required in Yolk Nuclei for Midgut Formation

Corresponding author: James W. Erickson, Department of Biological Sciences, Columbia University, 600 Fairchild, 1212 Amsterdam Ave., Mail Code 2402, New York, NY 10027., erickson{at}cubsps.bio.columbia.edu (E-mail)

Communicating editor: T. SCHÜPBACH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

During sex determination, the sisterlessA (sisA) gene functions as one of four X:A numerator elements that set the alternative male or female regulatory states of the switch gene Sex-lethal. In somatic cells, sisA functions specifically in sex determination, but its expression pattern also hints at a role in the yolk cell, a syncytial structure believed to provide energy and nutrients to the developing embryo. Previous studies of sisA have been limited by the lack of a null allele, leaving open the possibility that sisA has additional functions. Here we report the isolation and molecular characterization of four new sisA alleles including two null mutations. Our findings highlight key aspects of sisA structure-function and reveal important qualitative differences between the effects of sisA and the other strong X:A numerator element, sisterlessB, on Sex-lethal expression. We use genetic, expression, clonal, and phenotypic analyses to demonstrate that sisA has an essential function in the yolk nuclei of both sexes. In the absence of sisA, endoderm migration and midgut formation are blocked, suggesting that the yolk cell may have a direct role in larval gut development. To our knowledge, this is the first report of a requirement for the yolk nuclei in Drosophila development.

IN Drosophila melanogaster the genetic elements that signal chromosomal sex encode a diverse set of proteins that regulate the early transcription of Sex-lethal (Sxl), the regulatory switch gene that controls all aspects of sex determination and dosage compensation. Central among the signal elements are the X-linked sisterlessA (sisA), sisterlessB (sisB), sisterlessC, and runt genes, which serve as the direct dose-sensitive determinants of X chromosome number (CLINE 1988 , CLINE 1993 ; DUFFY and GERGEN 1991 ; TORRES and SANCHEZ 1992 ; reviewed in CLINE and MEYER 1996 ). In female embryos, the diplo-X dose of these four "X:A numerator" gene products, in conjunction with several maternally supplied "X:A signal transduction" gene products including Daughterless (Da), activate the Sxl establishment promoter Sxl_Pe, initiating the female developmental program (CLINE 1988 ; KEYES et al. 1992 ). In male embryos, the haplo-X dose of X:A numerator genes is insufficient to overcome the effects of maternal and autosomal inhibitors of Sxl_Pe. Consequently, Sxl is left in its off, or male, state.

Although four genes function as X:A numerator elements, two, sisA and sisB (also called scute with reference to its proneural function), account for most of the dose sensitivity of the X counting process. They also are required to regulate Sxl expression throughout the embryo, unlike the weaker runt element (BOPP et al. 1991 ; DUFFY and GERGEN 1991 ; ERICKSON and CLINE 1993 ; CLINE and MEYER 1996 ). The sisB gene product is a basic helix-loop-helix (bHLH) transcription factor that dimerizes with maternally supplied Da to directly regulate Sxl_Pe. The sisA gene encodes a divergent member of the basic leucine-zipper (bZIP) class of transcription factors, which is thought to bind Sxl_Pe in conjunction with an as-yet-unidentified protein partner (ERICKSON and CLINE 1993 , ERICKSON and CLINE 1998 ).

While sisB has been intensively studied genetically because of its role in neurogenesis, sisA is less completely characterized, because only a single sisA mutation exists (CLINE 1986 ). This allele, sisA¹, alters the DNA-binding region of the protein, rendering it defective in Sxl_Pe activation and female-lethal when homozygous (ERICKSON and CLINE 1993 ). While strongly female-lethal, sisA¹ is without effect in males, raising the possibility that sisA could be unique among the sex signal elements in being a specific regulator of sex determination and dosage compensation.

An unusual feature of sisA is its high-level expression in the embryonic yolk nuclei or vitellophage (ERICKSON and CLINE 1993 , ERICKSON and CLINE 1998 ). This expression pattern, which has been conserved for the 40 million years separating D. melanogaster and D. virilis, indicates that if sisA has a second function, it likely involves the yolk nuclei. Although both yolk and yolk nuclei are found in the eggs of most insect species, little is known of the role the yolk nuclei play in development. As their alternative name implies, the vitellophage may participate in yolk utilization and BOWNES et al. 1988 have proposed that they may be involved in the timed release of ecdysteroids stored in the yolk; but no direct data support either contention. Ultimately the yolk and yolk nuclei are enclosed within the developing midgut, where yolk protein degradation occurs (BOWNES 1982B ). Prior to the engulfment of the yolk in the midgut, however, yolk nuclei are found on the surface of the yolk sac in proximity to regions of morphogenetic movements, hinting that they could be involved in these processes (BOWNES 1982A ).

To further address the role of sisA in sex determination, to study its structure-function relationships, and to answer the question of alternative developmental functions, we undertook a genetic screen to isolate new sisA alleles. We obtained four sisA mutations, including two nulls and a temperature-sensitive allele. Analysis of these mutations reveals that sisA may have a lesser role in the expression of Sxl than does the other strong X:A numerator sisB, as considerable Sxl expression occurs in the absence of sisA protein. Our data also demonstrate that sisA has a second essential function, one that for the first time reveals a role for the yolk nuclei in the development of Drosophila.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila:
Flies were grown on a standard cornmeal, yeast, and molasses medium in uncrowded conditions. The criterion for viability was eclosion. Mutations and chromosomes are described at http://flybase.bio.indiana.edu:7083.

Isolation of sisterless mutations:
Males of genotype y cm ct v/Y were fed 0.006 M or 0.012 M EMS in 1% sucrose (GRIGLIATTI 1986 ) and crossed to attached-X females bearing a Y chromosome duplication of the sisA⁺ region (v⁺Yy⁺). Single F₁ y cm ct v /v⁺Yy⁺ male progeny were mated with v sisA¹ shi^ts/FM3, y B females. After 5–7 days at 25°, adults were removed and progeny shifted to 29° to kill shi^ts males. As FM3 is male-lethal, only y cm ct v/v sisA¹ shi^ts and y cm ct v/FM3 virgin female F₂ progeny were recovered. Complementation was assessed by comparing the ratio of vermilion (y cm ct v*/v sisA¹ shi^ts) to yellow Bar (y cm ct v*/FM3, y B) progeny. Ratios of <1:2 were taken as evidence of noncomplementation and retested. Rather than count all progeny, phenotypic ratios were assessed visually through the sides of the vials. Vials with apparent distortions in the progeny ratio were examined in detail. We could reliably detect twofold deviations from unity scoring unanesthetized flies through the vial; however, some weak candidates may have been discarded. Candidate y cm ct v/FM3 virgin females were retested and stocks created by crossing to y cm ct sisA¹, sc^sisB3-1 w, cm Sxl^f1ct, and FM7c males. This also served as an initial mapping strategy as homozygous sisA, sisB, or Sxl mutants are generally less viable than the transheterozygous combinations. The sisA⁺ and sisA^- P element transgenes used in complementation experiments were P[A10m1, sisA⁺, l(1)10Bb^-] and P[A10m2 sisA^-, l(1)10Bb⁺] (ERICKSON and CLINE 1993 ).

Sequencing of sisA alleles:
Genomic DNA was prepared from sisA⁴ males and from sisA², sisA³, and sisA⁵ males carrying transgenes with either the D. virilis sisA⁺ coding sequence (ERICKSON and CLINE 1998 ) or a D. melanogaster sisA gene in which the first 15 sisA codons had been replaced by the HA epitope tag (unpublished). DNA from male-lethal mutants was PCR amplified using a 5' primer specific for the endogenous sisA locus CTTGAATTCACCATGGAACGGAGTC and nonspecific 3' primer CTAATGCTGGACTAGTCAGGGATCCG. The entire sisA⁴ coding sequence was amplified using 5' primer CCCAGTGACCGCCATTCATGAAGGTAGAAACCGCGAA. The sisA⁴ hobo insertion and flanking DNA was PCR amplified using primers CGCCCTTATCAGTTGCATTGAATTCG and CTGAATTCACCATGGAACGGAGTC. For each mutation products of three independent amplifications were sequenced using Sequenase 2.0 (Amersham Pharmacia Biotech, Inc., Piscataway, NJ) with similar results.

Determination of the sisA⁴ TSP:
Crosses were y cm ct v sisA⁴/FM7c females x FM7c/Y males. For downshifts, egg collection periods (and number of control females) were 0 (408), 0.5–1.5 (66), 1.5–2.5 (198), 2–3 (172), 2.5–3.5 (260), 3–4 (454), 3.5–4.5 (138), 4–5 (191), 4–6 (434), 5–7 (361), 6–8 (317), and 7–9 hr after oviposition (447). For upshifts, embryos were aged (number of control females) 0–4 (97), 3–5 (34), 4–6 (73), 5–7 (76), 6–8 (145), 7–9 (68), 8–10 (98), 9–11 (134), 10–12 (81), 12–16 (151), and 14–19 hr (186). Temperature shifts were made by moving vials from an air incubator to a water bath. After 48 hr, progeny were shifted to 25° to complete development.

Clonal analysis:
Gynandromorphs were generated using the mitotically unstable ring-X chromosome In(1)w^vC. In(1)w^vC/y w lz females were crossed to y w sn v sisA²/v⁺Yy⁺ or y w sisA¹/Ymales. Scoring of cuticular structures was done on anesthetized adults. Due to the low viability and fecundity of our In(1)w^vC/y w lz stock (now extinct), experimental and control gynandromorphs were generated over several months, during which time the ring-X stabilized. Consequently, not all data were obtained under identical conditions. This prevented a summation of overall viability; nonetheless, data from parallel crosses suggest that there was little difference between the survival of experimental and control gynandromorphs. The most comparable crosses produced 238 In(1)w^vC/sisA² females and 6 sisA²-gynandromorphs and 150 In(1)w^vC/sisA¹ females and 3 sisA¹-gynandromorphs.

Immunocytochemistry and in situ hybridization:
Embryos were processed, stained, and mounted according to PATEL 1994 . Mouse anti-SXL polyclonal (BERNSTEIN et al. 1995 ) and anti-ß-galactosidase monoclonal (Promega, Madison, WI) antibodies were used at 1:1000 dilution. Embryos were fixed in heptane/PEM-4% formaldehyde for 20–40 min or in heptane/PBS-10% formaldehyde for 50 min (in situ fixation). In situ hybridizations were as described (ERICKSON and CLINE 1993 , ERICKSON and CLINE 1998 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of new sisterless mutations:
We searched for new sisA alleles by screening for EMS-induced X-linked mutations that failed to complement the female-specific lethal allele sisA¹. Although our goal was to isolate new sisA mutations, we also expected to recover sisB(sc) and Sxl alleles because these are male-viable and exhibit a dominant-lethal synergy in heterozygous sisA¹ females (CLINE 1988 ). An important feature of our screen was that it allowed the recovery of lethal sisA alleles, since the F₁ males carried a duplication of the sisA⁺ region on their Y chromosomes (v⁺Yy⁺). From 17,034 fertile F₁ males, we identified 11 mutations that failed to complement sisA¹ upon retesting. Eight were male-viable. Five male-viable lines exhibited bristle defects characteristic of sc mutations and were found to map very close to yellow, indicating that they were sisB alleles. Two male-viable mutations partially complemented sisA¹ and sisB^3-1, but failed to complement the null Sxl^f1 allele. They were confirmed as Sxl alleles by genetic mapping that placed both near the flanking cm and ct genes.

Of the four remaining mutations one was partially male-viable while the others proved male-lethal. The male-lethality was completely rescued by P-element transgenes containing a functional sisA⁺ gene indicating that these were new sisA alleles and revealing the existence of a novel lethal sisA phenotype. The male-lethal alleles were designated as sisA², sisA³, and sisA⁵, and the partially male-viable line as sisA⁴.

Molecular characterization of sisA alleles:
The putative sisA², sisA³, and sisA⁵ mutations were cloned by PCR amplification. Genomic DNA was prepared from sisA mutant males that had been rescued by P-element transgenes carrying a modified sisA⁺ gene. Using a 5' primer specific for the endogenous sisA locus, we amplified sequences extending from the sixth codon into the 3' untranslated region of sisA and sequenced the products.

The sisA² mutation has a C-to-T nucleotide substitution that changes codon 161 (Gln) into a TAG stop signal. Although the sisA² change occurs late in the 189-amino-acid coding sequence, the mutant protein is likely to be nonfunctional because it is truncated within the leucine zipper dimerization domain (Fig 1A). The sisA⁵ allele contained a single-base deletion (TACGCA TACCA) that creates a frameshift after the 20th codon, suggesting that it also produces a nonfunctional polypeptide (Fig 1A).

View larger version (61K): In this window In a new window Download PPT slide   Figure 1. (A) Mutations affecting sisA. The sisA5 allele is a frameshift mutation that changes the reading frame after the 20th residue. sisA3 deletes amino acids 108–110 and replaces 111–118 with out-of-frame residues. sisA1 changes lys-128 to glu (ERICKSON and CLINE 1993 ). sisA2 is an amber mutation that truncates the protein after ala-160. The sisA sequences from D. melanogaster, pseudoobscura, and virilis (m p v) are shown (ERICKSON and CLINE 1998 ). The basic leucine zipper is labeled. (B) The sisA3 mutation. Nucleotide and amino acid sequences for the wild-type (top) and sisA3 (bottom) alleles. Identities are blocked. Deleted sequences are denoted by dots. (C) The sisA4 mutation is a 1.4-kb hobo insertion flanked by a 12-bp duplication that includes the sisA transcription start site.

The sisA³ allele is complex, consisting of a base substitution, an 8-bp frameshifting deletion, and a single-base deletion that restores the normal reading frame (Fig 1B). The alteration occurs just after a run of six gly-ser repeats and replaces 11 amino acids with 8 novel residues prior to the restoration of the normal reading frame. The altered amino acids occur in an evolutionarily conserved N-terminal extension of the basic DNA-binding domain (Fig 1A).

Males carrying the sisA⁴ allele were partially viable and the entire sisA⁴ coding region was PCR amplified from genomic DNA. The sisA⁴ coding sequence was wild type. Initial PCR and Southern analysis suggested an insertion mutation upstream of sisA and this was confirmed using PCR primers located upstream and downstream of the predicted insertion site. The DNA sequence revealed a 1.4-kb hobo transposable element flanked by a 12-bp target site duplication that includes the sisA transcription start site (Fig 1C).

sisA has a vital function in embryos of both sexes:
The most striking phenotype of the four new sisA alleles is male-lethality, indicating that sisA has a second vital function in Drosophila. The putative null alleles, sisA² and sisA⁵, are invariably male-lethal. No sisA² or sisA⁵ males have been observed in controlled experiments (Table 1) or in balanced stocks at any temperature. The complex sisA³ allele is also strongly male-lethal; however, escapers are occasionally observed, suggesting that the sisA³ protein retains residual function (Table 1). A small number of sisA⁴ males survive to adulthood, indicating that the hobo insertion does not eliminate all transcription of sisA (Table 1). The surviving sisA³ and sisA⁵ males were fertile and of normal morphology. The male-lethal effects of all four sisA alleles were completely rescued by sisA⁺, but not sisA^-, transgenes (ERICKSON and CLINE 1993 ), confirming that the male-lethality was due solely to the sisA gene (data not shown).

Previously, the only known effect of loss of sisA function was female-specific lethality, due to the failure to activate Sxl_Pe during the initial stage of sex determination (CLINE 1986 , CLINE 1988 ; ERICKSON and CLINE 1993 ). The four sisA alleles reported here are also female-lethal; however, the lethality cannot simply be ascribed to defects in Sxl activation if females also require sisA's second vital function. To determine if sisA has a second function in females, we asked if the strong constitutive Sxl^M4 allele could rescue the viability of sisA² females. Sxl^M4 bypasses the requirement for the X-counting elements in sex determination by activating the female mode of Sxl splicing independent of transcription from Sxl_Pe (BERNSTEIN et al. 1995 ). We found that Sxl^M4 was incapable of rescuing sisA² females even though it efficiently suppressed the lethal sex determination defect in sisA¹/sisA² females (Table 2), suggesting that sisA's second function is needed in both sexes.

To determine the developmental stage at which sisA acts we examined the lethal period for sisA² males. Embryos were collected from crosses between y w v sisA²/y w ++ females and wild-type (Oregon-R) males and monitored for viability at different developmental stages. Hemizygous sisA² males accounted for one-fourth of the progeny in these crosses with the remaining animals expected to be fully viable. We examined 954 fertilized eggs and found that 245 (25.7%) failed to hatch, suggesting that sisA² mutant males die as embryos. Very little lethality was observed at later stages (93% of collected first-instar larvae emerged as adults with a 2:1 ratio of females to males), showing that the lethal phase is limited to the embryonic period.

Effects of sisA on sex determination:
The sisA¹ mutation is fully male-viable but strongly female-lethal. The viability of homozygous sisA¹ females varies from <0.1 to ~3%, depending on culture conditions and undefined aspects of genetic background, but is always greater than that of sisA¹ hemizygous females, suggesting that sisA¹ retains some sex determination function (CLINE 1986 , CLINE 1988 ). This inference, however, is only valid if a single nondosage-compensated copy of sisA¹ provides females with a sufficient amount of sisA's second function. Otherwise, the greater lethality of sisA¹ hemizygotes could be explained as the additive effect of two developmental defects. Data in Table 2 (lines 2–4) show that the majority of females hemizygous for sisA¹ are viable if they also carry the Sxl^M4 allele (also shown by BERNSTEIN et al. 1995 ). This confirms that one copy of sisA¹ is sufficient to provide the non-sex-specific function to females and thus confirms the previous inference that sisA¹ retains partial sex determination function.

Because sisA¹ is partially functional, the relative viability of sisA¹ transheterozygotes can be used as a sensitive measure of any residual sex determination function in the new sisA alleles. By this criterion sisA², sisA³, and sisA⁵ are sex determination null mutations because the viability of these alleles, heterozygous with sisA¹, is indistinguishable from a deletion of the sisA region under the most permissive conditions (Table 1). In contrast, the hobo insertion mutation sisA⁴ partially complements sisA¹, demonstrating that sisA⁴ possesses residual sex determination activity. Although sisA⁴ has nearly the same sex determination function as the sisA¹ allele, we have not observed any homozygous sisA⁴ females, indicating that the combination of the two partial functional defects is invariably lethal to females.

To determine whether sisA is absolutely required for Sxl expression, we examined sisA² and sisA⁵ mutant embryos using anti-SXL antibodies and in situ hybridization. The results indicate that while sisA is necessary for proper Sxl regulation, Sxl_Pe can be expressed in some nuclei even in the complete absence of sisA (Fig 2). Most homozygous sisA⁵ embryos exhibited residual SXL staining in the anterior thoracic and mid-ventral regions with the phenotypes ranging from no expression to patchy high-level SXL staining. In contrast, severe reductions in the other strong numerator element sisB have been reported to eliminate expression of Sxl protein in all parts of the embryo (cited in BOPP et al. 1991 ), a result we confirmed using the Achaete-scute complex deletion Df(1)sc¹⁹ (data not shown).

View larger version (81K): In this window In a new window Download PPT slide   Figure 2. SXL expression in sisA null mutant embryos. Sxl protein was detected with anti-SXL antibody. Cross was: y cm ct v sisA5/FM7c x y cm ct v sisA5/v+Yy+. (A) Female with uniform SXL staining. Yolk is labeled. (B–D) Homozygous sisA5 female embryos with low-level (B), intermediate (C), and high-level (D) residual SXL expression. Female sisA5 embryos were identified by their partial SXL staining and/or abnormal morphology. Embryos were fixed for 30 min. Embryos are positioned anterior/left and dorsal/up in all figures unless otherwise indicated.

To obtain a more direct measure of the effect of the loss of sisA on Sxl_Pe, we examined transcription from Sxl during nuclear cycles 12–14 using in situ hybridization. Although we could not distinguish nonstaining sisA⁵ females from males in precellular embryos, we identified a number of diplo-X embryos that expressed low levels of Sxl_Pe transcript in a patchy mosaic pattern reminiscent of the later Sxl protein pattern (data not shown), confirming that the SXL staining accurately reflects the residual activity of Sxl_Pe in sisA null embryos.

After cellularization sisA is expressed in the yolk nuclei:
The earliest sisA expression has been described in detail for both D. melanogaster and D. virilis (ERICKSON and CLINE 1993 , ERICKSON and CLINE 1998 ). There are no maternal sisA messages. Zygotic expression begins during the cleavage stage and continues until the beginning of cellularization in nuclear cycle 14. At that time expression in periphery stops, and there is an abrupt increase in expression from the yolk nuclei or vitellophages. Thereafter, sisA appears to be expressed exclusively in the yolk, being maintained at high levels through late germ-band retraction, until ceasing during stages 13 or 14 (Fig 3). This temporal progression suggests either that the non-sex-specific vital function of sisA acts prior to cellularization, perhaps coincident with sisA's sex-determining function, or that it involves the yolk nuclei.

View larger version (97K): In this window In a new window Download PPT slide   Figure 3. sisA is expressed exclusively in the yolk nuclei after cellularization. Transcripts were detected by in situ hybridization. Wild-type embryos at stages 6 (gastrulation), 11 (germ-band extended), 12 (early germ-band retraction), and 13 (retracted germ band) (CAMPOS-ORTEGA and HARTENSTEIN 1997 ).

The sisA temperature-sensitive period occurs early in embryogenesis:
The hobo insertion mutation sisA⁴ is temperature sensitive for male viability. This allowed us to ask when the non-sex-specific function of sisA is required by monitoring the viability of sisA⁴ males after timed temperature shifts. We found that the non-sex-specific sisA⁴ temperature-sensitive period (TSP) begins at about 2.75 hr after fertilization, a time corresponding to the end of the syncytial blastoderm phase, and ends <2 hr later (normalized to 25°) during extension of the germ band (Fig 4). Since sisA⁴ retains residual function at the restrictive temperature, it is formally possible that this TSP represents only the most sensitive period and that sisA is required at other times as well. Nevertheless, the data indicate that males require sisA from a time just after sex is determined (1–2.5 hr after fertilization; TORRES and SANCHEZ 1991 ; ERICKSON and CLINE 1993 ) through gastrulation and the early stages of germ-band extension (CAMPOS-ORTEGA and HARTENSTEIN 1997 ). Thus, the TSP, in combination with the in situ hybridization data, implicate the yolk nuclei as the probable site of the novel sisA function.

View larger version (17K): In this window In a new window Download PPT slide   Figure 4. The TSP for sisA function in males includes gastrulation and germ-band extension. The viability of sisA4/Y males was compared to their fully viable sisA4/+ sisters as a function of temperature and time. Mean embryo age at shift is indicated in hours after oviposition. The abscissa scale has been adjusted for developmental time. The approximate developmental stages are indicated between the graphs. The estimated times of the TSPs for sex determination (TORRES and SANCHEZ 1991 ; ERICKSON and CLINE 1993 ) and for sisA's hypothesized yolk function are indicated by the slashed boxes.

Clonal analysis supports the sisA yolk function model:
The yolk nuclei originate from ~100 cleavage-stage energids that drop from the periphery into the central yolk mass between the 8th and 10th nuclear cycles (FOE et al. 1993 ). These nuclei originate from all regions of the embryo and none appear to possess a predetermined yolk fate. When the membranes surrounding the peripheral nuclei fuse to form the 6000 cells of the cellular blastoderm, they enclose 175–225 polyploid yolk nuclei within a large syncytial yolk cell. Each yolk nucleus has an associated cytoplasm that fuses with that of other yolk nuclei and with a thin layer of cytoplasm on the surface of the yolk mass. As gastrulation and germ-band extension proceed, the yolk nuclei migrate to form an interconnected network on the inner surface of the yolk cell (POULSON 1950 ). The yolk nuclei persist until after the yolk sac is engulfed by the developing midgut, where they eventually degenerate without producing any larval or adult structures (CAMPOS-ORTEGA and HARTENSTEIN 1997 ).

The development and structure of the yolk cell suggests that sisA should function nonautonomously within this syncytial cell. In other words, a sisA defect should be tolerated in mosaic animals, provided a sufficient number of functional yolk nuclei are present within the yolk cell. To test this idea, we generated sisA² mosaics using the mitotically unstable ring-X chromosome In(1)w^vc. The In(1)w^vc chromosome is lost at high frequency in the earliest nuclear divisions to generate gynandromorphs containing large clones of male (haplo-X) and female (diplo-X) tissue (ZALOKAR et al. 1980 ). In the absence of lethal effects male and female clones survive equally well, with the percentage of male adult tissue determined by the divisions in which the ring-X was lost and by the position of the clones with respect to the blastoderm fate map. If the second function of sisA is required exclusively in the yolk, it should be possible to isolate large-patch gynandromorphs in which any surface structure is mutant, because nearly all should be yolk nuclear mosaics (ZALOKAR et al. 1980 ). In contrast, if low-level expression elsewhere were responsible for sisA's vital function, one would expect the recovery of sisA²-gynandromorphs to be reduced, and the distribution of surface clones to be biased against those structures originating closest to the site of sisA action (BRYANT and ZORNETZER 1973 ; see also ZUSMAN and WIESCHAUS 1985 ).

We analyzed the tissue and size distribution of sisA mutant clones in 57 sisA²-gynandromorphs and 21 control sisA¹-gynandromorphs. sisA² null mutant clones were recovered in every adult structure with similar frequency (Table 3), indicating that there is no single region of the blastoderm surface that must be sisA⁺ for embryonic viability. The average "maleness" index for the cuticular structures was 43% for the sisA² mosaics and 34% for the sisA¹ controls, both within the range published for nonlethal controls, and much greater than that seen for most lethal alleles (BRYANT and ZORNETZER 1973 ; ZUSMAN and WIESCHAUS 1985 ; CLINE 1988 ). The average size of the sisA null mutant patches was quite large. Many (24/57) sisA² mosaics had more than half male tissue, and four were almost entirely male (97%), suggesting that animals are viable even if they develop from the first few divisions with most of their nuclei mutant for sisA. The recovery of these large-patch gynandromorphs also demonstrates that sisA function is not needed later during imaginal disc development, in accordance with the sisA⁴ TSP determination (Fig 3) and larval/pupal expression analysis (ERICKSON and CLINE 1993 ).

The embryonic phenotypes of sisA null mutants:
To characterize the phenotypes of the sisA alleles without complications due to altered dosage compensation, we examined male sisA null mutant embryos for developmental defects. Mutant males were generated from attached-X females (X^X/Y) that had been mated with sisA males carrying a Y chromosome duplication of sisA⁺. The sisA male embryos could be distinguished from their 2X and 3X sisters because they failed to express the female-specific Sxl protein and from nullo-X embryos by their postblastoderm development.

The earliest obvious defect in fixed sisA null embryos was the presence of a space, or gap, between the yolk cell and the adjacent soma. The gap was first evident in stage 8 embryos and the separation between the yolk and surrounding layers persisted in later stages (Fig 5). The gap was most clearly seen in embryos that had been fixed for <30 min, but could also be observed in more heavily fixed preparations. The aberrant junction between the yolk and the surrounding tissue appears to be an area of structural weakness as many sisA null mutant embryos break there when placed under a coverslip (Fig 5C and Fig E).

View larger version (73K): In this window In a new window Download PPT slide   Figure 5. Phenotypes of sisA null mutant embryos are suggestive of defects in the yolk cell. (A) Normal female expressing SXL at stage 8 or 9. (B–I) sisA mutant male embryos. Yellow arrows point to gaps between the yolk mass and surrounding cells. Black arrows to unenclosed or ectopic yolk. (B) Stage 8, when structural defects are first evident. Note pole cells in proctodeum. (C) Stage 9 or 10 embryo broken at gap between yolk and somatic tissue. (D) Stage 10: pole cells have migrated out of the proctodeum, which has twisted laterally. (E) Stage 12. (F) Stages 13–15. Note normal fore- and hindgut structures and unenclosed yolk. (G–I) Terminally arrested sisA embryos. (G) Exemplifies the earliest arrest and (I) the most developed. Fixation was for 20 min (A–D, I) or 40 min (E–H).

Examination of lethally arrested sisA null embryos revealed a variety of terminal phenotypes. Some arrested late in germ-band retraction, while others developed a wide range of cuticular and internal structures occasionally nearly completing development (Fig 5, G–I). The only identifiable common feature was a variably positioned yolk mass that often exhibited large globular yolk droplets. Similar ectopic accumulations of unenclosed yolk have been described in mutants that fail to form endoderm or midgut (BRONNER et al. 1994 ), suggesting that sisA mutants may be defective in the ability to form the endoderm-derived midgut.

sisA null mutations prevent endoderm migration and midgut formation:
The Drosophila larval midgut originates from two spatially distinct groups of cells, the anterior midgut (AMG) and posterior midgut (PMG) primordia (CAMPOS-ORTEGA and HARTENSTEIN 1997 ). The physical separation of the primordia requires that the AMG and PMG migrate over the visceral mesoderm and around the yolk before they fuse and form the midgut. Late in germ-band extension, the primordia form lobed structures that extend toward each other as they move along the visceral mesoderm and over the yolk cell (REUTER et al. 1993 ; TEPASS and HARTENSTEIN 1994 ). By the end of germ-band retraction the AMG and PMG have joined laterally on both sides of the yolk to form an epithelium that will then wrap around the yolk, enclosing it within the newly formed midgut.

To determine if sisA null mutants are defective in endoderm specification, migration, or other aspects of midgut formation, we examined the development of the endoderm and midgut in sisA⁵ male embryos using the lacZ-expressing enhancer trap A490.2M3 (REUTER et al. 1993 ; TEPASS and HARTENSTEIN 1994 ) to mark the cells that contribute to the midgut. We found that the midgut primordia formed normally and that the AMG began to adopt its characteristic two-lobed structure during stage 10 in sisA⁵ mutants, but that subsequent development of the AMG and PMG was abnormal (Fig 6). In particular, there was little extension of the AMG lobes and no migration of the AMG or PMG precursors in sisA⁵ mutants. Consequently, the yolk was left as an unenclosed mass that moved to various ectopic locations as other aspects of development continued (Fig 5).

View larger version (159K): In this window In a new window Download PPT slide   Figure 6. Endoderm precursors form but their migration is blocked in sisA null mutants. Anterior (AMG) and posterior midgut (PMG) primordia were identified using the lacZ-expressing enhancer trap A490.2M3. The wild-type ß-gal expression pattern is shown in germ-band extension, stage 10 (A), early germ-band retraction, stage 12 (B), and in two embryos at the completion of germ-band shortening, stage 13 (C and D). In sisA5 males, the AMG and PMG primordia form and express the lacZ marker at stage 10 (E), but migration is prevented during and after germ-band retraction when development becomes abnormal (F–H). Embryos were fixed as for in situ hybridization.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A screen for sisterless mutations:
Previous attempts to identify new sisterless mutations have failed to generate new sisA alleles, although sisB, Sxl, and sisC mutations were recovered (CLINE 1993 ; TIMMER 1995 ; CLINE and MEYER 1996 ). Here we used a noncomplementation screen to identify mutations that interacted with sisA¹ to produce a female-lethal phenotype. The key feature of our screen was that it allowed the recovery of lethal sisA alleles, the importance of this being demonstrated by the male-lethal phenotypes of all four new sisA mutations. That, combined with the fact that no sisA alleles were recovered from large-scale screens where lethals would not have survived (TIMMER 1995 ), suggests that it may be difficult to mutate the 600-bp sisA gene in a way that creates sex determination-specific alleles such as sisA¹.

A notable sidelight of the screen was the efficient recovery of sisB or sc alleles. Although the Achaete-scute complex has been extensively studied genetically, few point mutations exist. A recent screen for bristle defects, for example, produced one EMS-induced sc point mutation from >28,000 F₁ chromosomes screened (GOMEZ-SKARMETA et al. 1995 ). Here we recovered five EMS-induced sc alleles, each possessing a different bristle phenotype, suggesting that the most effective means of isolating novel sc point mutations may be to search for new sisB alleles.

Sequence and function of the new sisA alleles:
Two of our mutations, sisA² and sisA⁵, are nulls by molecular and genetic criteria. While the sisA⁵ frameshift occurs early in the coding sequence and reveals no structural information about the protein, the null character of the late nonsense mutation sisA² further supports the importance of the proposed leucine zipper domain for SISA function. A third strong mutation, sisA³, affects eight residues of a conserved N-terminal extension of the proposed DNA-binding region, demonstrating the functional importance of this segment of the sisA protein.

The sisA³ mutation is complex, consisting of a base substitution and two frameshifting deletions. The changes are located downstream of a highly diverged region that separates the conserved N- and C-terminal halves of the protein. This structural characteristic of SISA, conserved sequence blocks separated by randomized segments of varying length, is commonly observed among homologous proteins in Drosophila species, but the mechanisms that generate such randomized segments are obscure. In this regard, the sisA³ allele, although itself nonfunctional, may be illustrative of how compensating frameshifts can produce randomized segments and yet maintain conserved sequence blocks on either side.

The sisA⁴ mutation was caused by the insertion of a hobo element in the sisA promoter region. This partial loss-of-function allele is temperature sensitive (ts) both for yolk function and sex determination, suggesting that it may have a ts effect on sisA expression. While ts mutations generally affect protein structure, there are precedents for cis-acting ts alleles in Caenorhabditis elegans (PERRY et al. 1994 ) and for the Drosophila sex determination elements da and sisB (CRONMILLER et al. 1988 ; ERICKSON and CLINE 1991 , ERICKSON and CLINE 1993 ; PARKHURST et al. 1993 ). Alternatively, the ts nature of sisA⁴ might reflect an underlying temperature sensitivity in the two developmental processes. The sex-determining X-counting mechanism is inherently sensitive to temperature regardless of the particular alleles employed (CLINE 1988 ; TORRES and SANCHEZ 1991 ). Whether a similar, natural heat sensitivity applies to the sisA yolk function remains to be determined.

sisA is needed for proper Sxl expression but is not essential for Sxl_Pe promoter activity in all cells:
By genetic criteria sisA and sisB are the most potent zygotic determinants of X chromosome dose (CLINE and MEYER 1996 ) and both are needed for the proper activation of Sxl_Pe in all regions of the embryo (BOPP et al. 1991 ; TORRES and SANCHEZ 1991 ; ERICKSON and CLINE 1993 ). However, previous studies left open the question of what happens to Sxl expression in the complete absence of sisA activity. Here we show that a surprisingly high level of Sxl expression remains. While the effects of the loss of sisA are qualitatively different from the effect of the loss of sisB, sisA is nonetheless critical for proper Sxl activation throughout the embryo. A number of sisA null females express trace levels of SXL, while only a small fraction express wild-type levels of SXL in more than a few cells.

Curiously, the residual Sxl expression in sisA null embryos is not obviously different from that seen with the partial loss-of-function sisA¹ allele (ERICKSON and CLINE 1993 ). While careful quantitation might reveal the expected difference, the implication is that, while not null, sisA¹ is severely defective for sex determination.

Given the strong effect of sisA¹ on sex determination, and the expectation that the lys-to-glu change in the basic region should seriously hinder SISA's DNA binding, why is the sisA¹ allele male-viable? One possibility is that the sisA yolk function might not depend on DNA binding. If SISA were to function negatively by sequestering other proteins in an inactive complex, the mutant sisA¹ protein might be able to perform this function without binding DNA, in a manner analogous to the Id proteins in muscle development (MURRE and BALTIMORE 1992 ). An alternative is that the polyploid yolk nuclei produce an excess of sisA protein and that the viability of sisA¹ males is explained by high-level expression of a partially functional protein.

Pleiotropy and the evolution of the sex-signal elements:
Primary sex-determining mechanisms are among the most rapidly evolving developmental processes. Sxl, for example, appears to have been recruited to sex determination relatively late in the Dipteran lineage (see SACCONE et al. 1998 ). If sisA, -B, -C, and runt coevolved with Sxl as signal elements, it was likely due to their X-linkage and because they possessed preexisting expression patterns and functions that could be easily modified or directly coopted for use in sex determination. This would account for their pleiotropy, why their primordial functions were unrelated to sex, and why each is expressed in a fairly broad pattern coincident with, or slightly after, the early burst of Sxl expression. While nearly all the maternal and zygotic sex signal elements have alternative functions in the nervous system, we suspect that their acquisition as signal elements may be more directly related to other, earlier-acting functions. runt, for example, functions as a broadly expressed gap gene just after sex is determined (DUFFY and GERGEN 1991 ), while sisB and maternal da participate in dorsal-ventral axis specification during the period when Sxl_Pe is most active (GONZALEZ-CRESPO and LEVINE 1993 ). Our discovery that sisA has a vital developmental function in the yolk nuclei suggests that sisA was likely chosen as a signal element by virtue of its early uniform yolk expression. Whether that required the acquisition of new regulatory information or simply exploited an underlying prepattern necessary for yolk-specific expression remains to be determined.

sisA has a vital function in the yolk nuclei:
Taken together, the expression and TSP analyses strongly implicate the yolk nuclei as the site of sisA function. While we cannot exclude the possibility that undetected low-level sisA expression occurs outside the yolk, our in situ hybridization data were particularly clear during gastrulation and germ-band extension, minimizing the chances that low-level expression was missed or obscured by the strong yolk signal during these critical periods (Fig 3).

On the basis of the obvious diffusion of sisA mRNA (Fig 3) and the extensive cytoplasmic connections (POULSON 1950 ) between the yolk nuclei, we suspected that sisA would act nonautonomously within the yolk cell. If sisA's second vital function is restricted to the yolk cell, one would predict that gynandromorphs carrying large patches of sisA null mutant tissue could be easily isolated, as these animals should be yolk nuclear mosaics, and thus viable, regardless of their cuticle phenotype. The 57 sisA² gynandromorphs perfectly matched these predictions. There was no apparent difference in the viability of control sisA¹ and null sisA² gynandromorphs (MATERIALS AND METHODS) and little or no regional bias as to where sisA function could be lost (Table 3). Although the gynandromorph data are most simply explained by a nonautonomous function of sisA within the yolk cell, we cannot exclude certain alternative explanations for these results. The most important limitation was that because we scored only the adult cuticle phenotype in the mosaic flies, we were able to directly infer the genotype of only a fraction of the blastoderm surface. This leaves open the possibility that sisA is required in other regions of the embryo that are not represented in the adult cuticle (ZUSMAN and WIESCHAUS 1985 ).

Perhaps the strongest support for the hypothesis that sisA functions in the yolk nuclei is the phenotype of sisA null embryos. The alteration in yolk cell structure, or in its attachment to the surrounding tissues, which is visible shortly after the beginning of the sisA TSP (Fig 2 and Fig 5), links the mutant phenotype directly to the cell in which sisA is expressed and to the stage when it is active. The subsequent failure of the AMG and PMG primordia to migrate over the yolk cell surface to create the midgut (Fig 6) further strengthens the connection between gene expression and phenotype.

Yolk nuclei have a developmental role in embryogenesis:
The role of the yolk has been one of the more neglected areas of study in Drosophila development. Early transplantation experiments suggested that the yolk played no inductive role in development (HADORN and MULLER 1974 ), a conclusion supported by the absence of any previous reports of developmental mutations acting in the yolk. While sisA has an essential function in the yolk cell, its role does not appear to be inductive, as both the endodermal and mesodermal primordia of the midgut, the organ most directly affected by the loss of sisA, form in sisA null mutants.

The yolk nuclei are commonly known as vitellophages for their presumed function in digesting yolk. Given that yolk protein metabolism does not begin until after the yolk has been engulfed by the midgut and continues long after the yolk nuclei have been degraded (BOWNES 1982B ), we suspect that their primary function is developmental. Considering the ultimate fate of the yolk—to be devoured within the midgut—it is reasonable to think that the yolk nuclei might play a role in gut formation. The idea is not new. In his studies on embryonic development, POULSON 1950 hypothesized that the close contact between the midgut rudiments and the yolk nuclei was essential for gut formation. Although Poulson's claim that some yolk nuclei are incorporated directly into the midgut epithelium has not been borne out by subsequent investigation (CAMPOS-ORTEGA and HARTENSTEIN 1997 ), the implication of the present work is that the yolk nuclei are indeed important for midgut development.

The failure of endoderm migration in sisA mutants suggests that there must be proper contact between the endoderm and the yolk sac for migration to occur. A particularly intriguing idea is that the yolk cell, like the visceral mesoderm, serves as a track, or substratum, for the migrating endoderm (AZPIAZU and FRASCH 1993 ; REUTER et al. 1993 ; TEPASS and HARTENSTEIN 1994 ). sisA could be involved directly by regulating the expression of cell adhesion proteins within the yolk cell. Alternatively, sisA might be required to maintain the structural integrity of the yolk cell, perhaps by regulating the expression of cytoskeletal components. In that case, the structural collapse of the yolk sac might prevent normal recognition of the yolk membrane by the AMG and PMG primordia in sisA mutants.

Although the failure to form a midgut is the most prominent early developmental defect observed, sisA mutants exhibit other abnormalities, such as mislocalized yolk, a twisting of the proctodeum, and abnormal germ-band retraction (Fig 2, Fig 5, and Fig 6), which hint that other morphological movements might depend on proper contact between the yolk cell and the surrounding viscera and soma. The presence of dense plaques and gap junctions between cells in the extended germ-band and the yolk sac (RICKOLL and COUNCE 1980 ) is consistent with the idea that there may be interactions between these tissues.

sisA null mutant embryos exhibit a wide range of terminal phenotypes: some are missing head components and some lack posterior structures, while others form nearly normal larvae (Fig 5, G–I). We suspect that most such late defects are a secondary consequence of the earlier failure of endoderm migration. In the absence of the enveloping midgut, yolk is distributed to a variety of locations, and the ectopic yolk masses likely interfere with many structures by simple displacement. This phenotypic variation may explain why sisA was not identified in saturation screens for X-linked mutations affecting embryonic development (WIESCHAUS et al. 1984 ; EBERL and HILLIKER 1988 ), although low mutability may also have played a part.

Our results clearly implicate sisA and the yolk nuclei in the normal development of the larval midgut. Interestingly, several other genes required for midgut specification, maintenance, or migration are also expressed in the yolk nuclei, including serpent, hunchback, and fork head (REUTER 1994 ; REHORN et al. 1996 ). Little is known about the roles of these genes in the yolk in part because their complex expression patterns and pleiotropy make it difficult to identify a specific yolk function, but primarily because there has been little indication that the yolk nuclei have any important function. Now that the yolk nuclei have been shown to be important for development, further analysis of the regulation and function of these and other genes in the yolk should lead to a greater understanding of the yolk cell's function in midgut formation and may even reveal additional roles of the yolk nuclei in Drosophila. Whether they will uncover unexpected similarities between the developmental roles of the yolk cells of insects and the visceral yolk sacs, yolk cells, or yolk syncytial layers of vertebrates is a question of considerable interest.

	FOOTNOTES

¹ Present address: Department of Molecular, Cellular, and Developmental Biology, University of Colorado, Boulder, CO 80309-0347.

	ACKNOWLEDGMENTS

We are indebted to our Barnard College colleague Jym Mohler for valuable assistance with embryo phenotypes. He, Dan Kalderon, Tulle Hazelrigg, and Teresa Lamb provided thoughtful advice throughout the course of this work and helpful comments on the manuscript. We thank Tom Cline for comments on the manuscript, and him, Mary Bownes, Adelaide Carpenter, and the Bloomington Drosophila Genetic stock center for fly stocks. We thank Tracey Reynolds for help with the screen and Cristina Rumbaitis-del Rio for encouragement. K.K.L. is an I.I. Rabi Scholar of Columbia University. R.N.D. was supported in part by the Summer Undergraduate Research Fellowship. This research was funded by American Cancer Society grant RPG-97-079-01-DB and by a Searle Community Trust Award to J.W.E.

Manuscript received November 13, 1999; Accepted for publication January 24, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AZPIAZU, N. and M. FRASCH, 1993  tinman and bagpipe: two homeo box genes that determine cell fates in the dorsal mesoderm of Drosophila. Genes Dev. 7:1325-1340[Abstract/Free Full Text].

BERNSTEIN, M., R. A. LERSCH, L. SUBRAHMANYAN, and T. W. CLINE, 1995  Transposon insertions causing constitutive Sex-lethal activity in Drosophila melanogaster affect Sxl sex-specific transcript splicing. Genetics 139:631-648[Abstract].

BOPP, D., L. R. BELL, T. W. CLINE, and P. SCHEDL, 1991  Developmental distribution of female-specific Sex-lethal proteins in Drosophila melanogaster.. Genes Dev. 5:403-415[Abstract/Free Full Text].

BOWNES, M., 1982a Embryogenesis, pp. 67–94 in A Handbook of Drosophila, edited by R. RANSOM. Elsevier Biochemical Press, Amsterdam.

BOWNES, M., 1982b  Hormonal and genetic regulation of vitellogenesis in Drosophila. Q. Rev. Biol. 57:247-274[Medline].

BOWNES, M., A. SHIRRAS, M. BLAIR, J. COLLINS, and A. COULSON, 1988  Evidence that insect embryogenesis is regulated by ecdysteroids released from yolk proteins. Proc. Natl. Acad. Sci. USA 85:1554-1557[Abstract/Free Full Text].

BRONNER, G., Q. CHU-LAGRAFF, C. Q. DOE, B. COHEN, and D. WEIGEL et al., 1994  Sp1/egr-like zinc-finger protein required for endoderm specification and germ-layer formation in Drosophila. Nature 369:664-668[Medline].

BRYANT, P. J. and M. ZORNETZER, 1973  Mosaic analysis of lethal mutations in Drosophila. Genetics 75:623-637[Abstract/Free Full Text].

CAMPOS-ORTEGA, J. A., and V. HARTENSTEIN, 1997 The Embryonic Development of Drosophila melanogaster. Springer-Verlag, Berlin.

CLINE, T. W., 1986  A female-specific lethal lesion in an X-linked positive regulator of the Drosophila sex determination gene, Sex-lethal.. Genetics 113:641-663. (corrigendum Genetics 114: 345).[Abstract/Free Full Text].

CLINE, T. W., 1988  Evidence that sisterless-a and sisterless-b are two of several discrete "numerator elements" of the X/A sex determination signal in Drosophila that switch Sxl between two alternative stable expression states. Genetics 119:829-862[Abstract/Free Full Text].

CLINE, T. W., 1993  The Drosophila sex determination signal: How do flies count to two? Trends Genet. 9:385-390[Medline].

CLINE, T. W. and B. J. MEYER, 1996  Vive la difference: males vs females in flies vs worms. Annu. Rev. Genet. 30:637-702[Medline].

CRONMILLER, C., P. SCHEDL, and T. W. CLINE, 1988  Molecular characterization of daughterless, a Drosophila sex determination gene with multiple roles in development. Genes Dev. 2:1666-1676[Abstract/Free Full Text].

DUFFY, J. B. and J. P. GERGEN, 1991  The Drosophila segmentation gene runt acts as a position-specific numerator element necessary for the uniform expression of the sex-determining gene Sex-lethal.. Genes Dev. 5:2176-2187[Abstract/Free Full Text].

EBERL, D. F. and A. J. HILLIKER, 1988  Characterization of X-linked recessive lethal mutations affecting embryonic morphogenesis in Drosophila melanogaster.. Genetics 118:109-120[Abstract/Free Full Text].

ERICKSON, J. W. and T. W. CLINE, 1991  Molecular nature of the Drosophila sex determination signal and its link to neurogenesis. Science 251:1071-1074[Abstract/Free Full Text].

ERICKSON, J. W. and T. W. CLINE, 1993  A bZIP protein, SISTERLESS-A, collaborates with bHLH transcription factors early in Drosophila development to determine sex. Genes Dev. 7:1688-1702[Abstract/Free Full Text].

ERICKSON, J. W. and T. W. CLINE, 1998  Key aspects of the primary sex determination mechanism are conserved across the genus Drosophila.. Development 125:3259-3268[Abstract].

FOE, V. A., G. M. ODELL and B. A. EDGAR, 1993 Mitosis and morphogenesis in the Drosophila embryo: point and counterpoint, pp. 149–300 in The Development of Drosophila melanogaster, edited by M. BATE and A. MARTINEZ-ARIAS. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

GOMEZ-SKARMETA, J. L., I. RODRIGUEZ, C. MARTINEZ, J. CULI, and D. FERRES-MARCO et al., 1995  Cis-regulation of achaete and scute: shared enhancer-like elements drive their coexpression in proneural clusters of the imaginal discs. Genes Dev. 9:1869-1882[Abstract/Free Full Text].

GONZALEZ-CRESPO, S. and M. LEVINE, 1993  Interactions between dorsal and helix-loop-helix proteins initiate the differentiation of the embryonic mesoderm and neuroectoderm in Drosophila. Genes Dev. 7:1703-1713[Abstract/Free Full Text].

GRIGLIATTI, T., 1986 Mutagenesis, pp. 39–58 in Drosophila: A Practical Approach, edited by D. B. ROBERTS. IRL Press, Oxford.

HADORN, E. and G. MULLER, 1974  Has the yolk an influence on the differentiation of embryonic Drosophila blastemas? Roux Arch. EntwMech. Organ. 174:333-335.

KEYES, L. N., T. W. CLINE, and P. SCHEDL, 1992  The primary sex determination signal of Drosophila acts at the level of transcription. Cell 68:933-943. (Corrigendum, Cell 69: following 572).[Medline].

MURRE, C., and D. BALTIMORE, 1992 The helix-loop-helix motif: structure and function, pp. 861–879 in Transcriptional Regulation, edited by S. L. MCKNIGHT and K. R. YAMAMOTO. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

PARKHURST, S. M., H. D. LIPSHITZ, and D. ISH-HOROWICZ, 1993  achaete-scute feminizing activities and Drosophila sex determination. Development 117:737-749[Abstract].

PATEL, N. H., 1994 Imaging neuronal subsets and other cell types in whole-mount Drosophila embryos and larvae using antibody probes, pp. 446–487 in Methods in Cell Biology, edited by L. S. B. GOLDSTEIN and E. A. FYRBERG. Academic Press, San Diego.

PERRY, M. D., C. TRENT, B. ROBERTSON, C. CHAMBLIN, and W. B. WOOD, 1994  Sequenced alleles of the Caenorhabditis elegans sex-determining gene her-1 include a novel class of conditional promoter mutations. Genetics 138:317-327[Abstract].

POULSON, D. F., 1950 Histogenesis, organogenesis, and differentiation in the embryo of Drosophila melanogaster meigen, pp. 168–274 in Biology of Drosophila, edited by M. DEMEREC. Wiley, New York.

REHORN, K. P., H. THELEN, A. M. MICHELSON, and R. REUTER, 1996  A molecular aspect of hematopoiesis and endoderm development common to vertebrates and Drosophila. Development 122:4023-4031[Abstract].

REUTER, R., 1994  The gene serpent has homeotic properties and specifies endoderm versus ectoderm within the Drosophila gut. Development 120:1123-1135[Abstract].

REUTER, R., B. GRUNEWALD, and M. LEPTIN, 1993  A role for the mesoderm in endodermal migration and morphogenesis in Drosophila. Development 119:1135-1145[Abstract].

RICKOLL, W. L. and S. J. COUNCE, 1980  Morphogenesis in the embryo of Drosophila melanogaster-germ band extension. Wilhelm Roux's Archives Dev. Biol. 188:163-177.

SACCONE, G., I. PELUSO, D. ARTIACO, E. GIORDANO, and D. BOPP et al., 1998  The Ceratitis capitata homologue of the Drosophila sex-determining gene Sex-lethal is structurally conserved, but not sex-specifically regulated. Development 125:1495-1500[Abstract].

TEPASS, U. and V. HARTENSTEIN, 1994  Epithelium formation in the Drosophila midgut depends on the interaction of endoderm and mesoderm. Development 120:579-590[Abstract].

TIMMER, J. R., 1995 A screen for new mutations in numerator elements of the X:A signal in Drosophila, pp. 183. Ph.D. Thesis, University of California, Berkeley.

TORRES, M. and L. SANCHEZ, 1991  The sisterless-b function of the Drosophila gene scute is restricted to the stage when the X:A ratio determines the activity of Sex-lethal.. Development 113:715-722[Abstract].

TORRES, M. and L. SANCHEZ, 1992  The segmentation gene runt is needed to activate Sex-lethal, a gene that controls sex determination and dosage compensation in Drosophila. Genet. Res. 59:189-198[Medline].

WIESCHAUS, E., C. NUSSLEIN-VOLHARD, and G. JURGENS, 1984  Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster. III. Zygotic loci on the X-chromosome and fourth chromosome. Wilhelm Roux Arch. Dev. Biol. 193:296-307.

ZALOKAR, M., I. ERK,