Interaction Between Hormonal Signaling Pathways in Drosophila melanogaster as Revealed by Genetic Interaction Between Methoprene-tolerant and Broad-Complex

Abstract

Juvenile hormone (JH) regulates insect development by a poorly understood mechanism. Application of JH agonist insecticides to Drosophila melanogaster during the ecdysone-driven onset of metamorphosis results in lethality and specific morphogenetic defects, some of which resemble those in mutants of the ecdysone-regulated Broad-Complex (BR-C). The Methoprene-tolerant (Met) bHLH–PAS gene mediates JH action, and Met mutations protect against the lethality and defects. To explore relationships among these two genes and JH, double mutants were constructed between Met alleles and alleles of each of the BR-C complementation groups: broad (br), reduced bristles on palpus (rbp), and 2Bc. Defects in viability and oogenesis were consistently more severe in rbp Met or br Met double mutants than would be expected if these genes act independently. Additionally, complementation between BR-C mutant alleles often failed when MET was absent. Patterns of BRC protein accumulation during metamorphosis revealed essentially no difference between wild-type and Met-null individuals. JH agonist treatment did not block accumulation of BRC proteins. We propose that MET and BRC interact to control transcription of one or more downstream effector genes, which can be disrupted either by mutations in Met or BR-C or by application of JH/JH agonist, which alters MET interaction with BRC.

HORMONAL regulation of insect development involves the relatively well-understood steroid 20-hydroxyecdysone (20E) (Henrich and Brown 1995; Riddiford et al. 2000; Spindler et al. 2001; Thummel and Chory 2002) and the more enigmatic sesquiterpenoid juvenile hormone (JH). 20E orchestrates molting and metamorphosis and regulates reproduction (Wyatt and Davey 1996; Riddiford et al. 2000). Its dimeric receptor ECR/USP is a 20E-dependent transcription factor that regulates the expression of target genes, including both primary response (early) genes, such as Broad-Complex (BR-C), that are directly and rapidly induced by 20E and secondary response (late) genes that are regulated by the primary response gene products (Ashburner et al. 1974; Richards 1997). JH action during larval development, at least in lepidopteran and hemimetabolous insects, is to maintain the “status quo,” probably to allow proper larval molting and prevent premature metamorphosis (Williams 1961; Zhou and Riddiford 2002). At the end of larval development, circulating JH is degraded, enabling 20E to trigger metamorphosis (Riddiford 1996). JH reappears in many adult insects to control both oogenesis and male accessory gland function (Wyatt and Davey 1996). Neither the JH receptor nor its molecular mechanism of action is well understood (Gilbert et al. 2000), although JH, as well as JH agonists (JHA), can regulate gene expression (Jones 1995; Dubrovsky et al. 2000; Sempere et al. 2002). The 20E and JH signaling pathways interact during development. In hemipteran and lepidopteran larvae, withdrawal of JH resulted in 20E induction of precocious metamorphosis, whereas application of JH during the last larval instar resulted in a supernumerary instar or larval-pupal intermediate (Riddiford 1994; Gilbert et al. 2000). These two hormones also interact in adults to control oogenesis (Soller et al. 1999).

In Drosophila melanogaster, metamorphosis is controlled by several waves of 20E secretion: the first at the end of the third larval instar driving pupararium formation, the second 10–12 hr after pupararium formation (APF) triggering pupation, and finally a large wave beginning 25–30 hr APF (Handler 1982; Riddiford 1993). JH has been detected in Drosophila larvae but not pupae (Bownes and Rembold 1987; Sliter et al. 1987) when its absence is thought to permit 20E control of metamorphosis. Application of exogenous JH or JH analog insecticides, such as methoprene and pyriproxyfen which act as JHA (Staal 1975; Wilson 2004), does not block pupararium formation or pupation. Rather, it results in disruptions of the adult abdominal epidermis, male genitalia orientation, and many internal tissues, resulting in death during the late pupal (pharate adult) stage (Ashburner 1970; Madhavan 1973; Postlethwait 1974; Wilson and Fabian 1986; Riddiford and Ashburner 1991; Restifo and Wilson 1998).

The primary response gene BR-C is composed of three lethal complementation groups: broad (br), reduced bristles on the palpus (rbp), and 2Bc, mutations of which cause death during prepupal or pupal stages (Belyaeva et al. 1980; Kiss et al. 1988). Null alleles of the entire gene, called nonpupariating1 (npr1), cause death in late third-instar larvae (Kiss et al. 1988; Gonzy et al. 2002), demonstrating that BR-C function is essential for metamorphic onset. BR-C uses alternative splicing to encode a small family of transcription factors with amino-terminal BTB–POZ domains linked to one of four possible C₂H₂ zinc-finger domains: BRC–Z1, BRC–Z2, BRC–Z3, and BRC–Z4 (Dibello et al. 1991; Bayer et al. 1996). During metamorphosis BRC isoforms regulate numerous downstream effector genes (Guay and Guild 1991; Karim et al. 1993; von Kalm et al. 1994; Hodgetts et al. 1995; Bayer et al. 1996, 1997; Crossgrove et al. 1996; Liu and Restifo 1998; Mugat et al. 2000; Dubrovsky et al. 2001; Dunne et al. 2002; Sempere et al. 2003). Although there is some functional redundancy among the isoforms, in general, BRC–Z1 provides the function for rbp, BRC–Z2 for br, and BRC–Z3 for 2Bc (Crossgrove et al. 1996; Bayer et al. 1997; Sandstrom et al. 1997; Liu and Restifo 1998; Consoulas et al. 2005).

BR-C metamorphosis phenotypes include failures of larval tissue histolysis (Lee and Baehrecke 2001; Kucharova-Mahmood et al. 2002), epidermal morphogenesis (Kiss et al. 1988), and internal tissue remodeling (Restifo and White 1991, 1992; Sandstrom et al. 1997; Consoulas et al. 2005). Several phenotypes, such as a split-brain abnormality (Restifo and White 1991) were common to all BR-C mutants, while others, such as failure of thoracic muscle attachment or persistence of larval salivary glands (Restifo and White 1992) were restricted to one or two complementation groups. This suggested that BRC proteins operate in two pathways, “common,” requiring all three functions and “restricted,” requiring a subset of them (Restifo and White 1991, 1992; Restifo and Wilson 1998). Following metamorphosis BR-C is expressed in egg chamber follicle cells to function during oogenesis (Huang and Orr 1992; Deng and Bownes 1997; Tzolovsky et al. 1999).

Our previous work showed that lethal pharate adults developing from wild-type methoprene-treated larvae showed a selective BR-C phenocopy, including disruptions of the central nervous system and salivary glands (Restifo and Wilson 1998). The methoprene syndrome was striking in that it included none of the restricted-pathway defects, which is inconsistent with methoprene simply blocking the 20E induction of BR-C expression during the larval-to-pupal transition. We interpreted the pathological effects of methoprene to reflect dysfunction of BR-C and probably additional primary response genes during metamorphosis, resulting in aberrant expression of secondary response genes. In contrast, other investigators found that the JHA pyriproxyfen caused abnormal pupal cuticle gene expression, apparently due to abnormal BR-C expression (Zhou and Riddiford 2002).

Methoprene-tolerant (Met) is essential for the manifestation of the toxic and morphogenetic effects of JH/JHA in D. melanogaster (Wilson and Fabian 1986; Riddiford and Ashburner 1991; Wilson 1996; Restifo and Wilson 1998). Met mutants are resistant to these effects of methoprene (Wilson and Fabian 1986). MET can bind JH III with specificity and nanomolar affinity (Shemshedini and Wilson 1990; Miura et al. 2005), suggesting that it is a component of a JH receptor. Met encodes a bHLH–PAS transcriptional regulator family member (Ashok et al. 1998) and MET can activate a reporter gene in transfected Drosophila S-2 cells (Miura et al. 2005).

We hypothesize that BR-C and Met function together in one or more aspects of development. If methoprene disrupts 20E-mediated metamorphic development or oogenesis by acting on BR-C or its downstream genes, one would expect double mutants to show synergistic genetic interaction, such as synthetic lethality of viable alleles, shifts in lethal phase, or enhanced oogenesis defects. We found that animals carrying both Met and BR-C mutations showed just such synergistic interactions. However, we did not detect any major disruption in BRC protein accumulation following methoprene treatment, suggesting that the MET and BRC interact to regulate expression of downstream effector gene(s).

MATERIALS AND METHODS

Stocks:

BR-C mutant alleles used in this study were previously used to investigate internal tissue metamorphosis (Restifo and White 1991, 1992; Restifo and Merrill 1994; Sandstrom et al. 1997; Consoulas et al. 2005). BR-C mutant progeny were identified by visible markers yellow, white, and/or singed. Wild-type BR-C function was provided by second- and/or third-chromosome BRC–Z1, BRC–Z2, or BRC–Z3 cDNA transgenes with hsp70 promoters (Bayer et al. 1997). The third-chromosome hsBRC-Z3 transgene, provided by C. Bayer (University of Central Florida, Orlando, FL) is leaky in a temperature-sensitive manner (R. Spokony, H. J. Clark, and L. L. Restifo, unpublished data). Survival to adult eclosion of 2Bc¹/Y; hsBRC-Z3/+ varied with rearing temperature: 9% at 22–23°, 12% at 25°, and 29% at 29°. Most of the Met alleles used in this study were isolated from a methoprene-susceptible isogenic vermilion (v) strain as previously described (Wilson and Fabian 1987). After isolation, each Met allele was backcrossed to v for 5–7 generations to minimize background genome differences that might impact comparison of phenotypes. A functional copy of Met⁺ was provided by a second-chromosome genomic transgene (Ashok et al. 1998; Wilson and Ashok 1998). The Y-borne duplications y²Y67g 19.1 (hereafter termed y²Y67g), bearing BR-C⁺ (Belyaeva et al. 1980; Lindsley and Zimm 1992) and y⁺Yv⁺B^S− (hereafter termed Y-Met⁺), bearing Met⁺ (Lindsley and Zimm 1992) were used to cover BR-C and Met mutations, respectively.

Double mutants were constructed by genetic recombination, and each was maintained heterozygous with an FM7 balancer chromosome. Each double-mutant chromosome was tested to verify the presence of each mutation and the absence of inadvertently introduced lethal mutations.

D. melanogaster culture and methoprene treatment:

Stocks and crosses were cultured on one of three standard media with mold inhibitors, at 25° with a 12:12 L:D photoperiod unless otherwise specified. For experiments comparing phenotypes, the same culture medium and other conditions were used for all genotypes. For progeny phenotype analysis, cultures were performed in glass vials (Capital Vial) without larval crowding, and all progeny were censused and examined, either as adults or as uneclosed pupae.

Cultures were tested for methoprene resistance on diagnostic doses given in Table 5 of ZR-2008, the biologically active isomer of methoprene [isopropyl-(2E,4E)-11-methoxy-3,7,11-trimethyl-2,4-dodecadienonate], as previously described (Wilson 1996; Wilson and Ashok 1998). At lethal concentrations, mortality occurs in Met⁺ typically during the pharate adult stage. Eclosing survivors were examined for methoprene-induced morphogenetic defects of malrotated male genitalia and defective sternal bristles, particularly on the posterior sternites. For analysis of BRC proteins, 100–120 OreRC or v Met²⁷ eggs were transferred to glass bottles containing Drosophila Instant Food (Carolina Biologicals) with either a high-lethal dose of methoprene or vehicle (acetone) alone, as described in Restifo and Wilson (1998). Control and methoprene-treated cultures were reared in parallel.

TABLE 5.

Survival and morphological defects in various BR-C and Met flies following treatment with methoprene

	Methoprene dose
Genotype	0.05	0.01	0.005	0.001
Oregon-RC	0	0	11 (100)	62 (69)
rbp2/rbp2/Y	0	0	7.6 (100)	70 (63)
br1/br1/Y	0	0	13 (89)	75 (74)
2Bc1/FM7	0	0	3.1 (100)	58 (48)
npr13/FM7	ND	0	14 (100)	67 (72)
Met3/Met3/Y	45 (0)	67 (0)	92 (0)	ND
br1 Met3/br1 Met3/Y	40 (0)	74 (0)	81 (0)	ND

Mean N = 63, range 40–79, individuals were evaluated at each methoprene dose applied to at least triplicate cultures of 30 individual larvae. Survival is expressed as percentage of individuals of the indicated genotype surviving to adulthood. Numbers in parentheses represent the percentage of survivors having abnormal sternite bristle/male genitalia. Met³ was used as the Met allele because of high viability with br¹. ND, not determined.

BR-C phenotype examination:

Well-described BR-C phenotypes, including lethality and epidermal defects of wing and maxillary palpus morphology (Kiss et al. 1988) were examined. Lethality was assigned to the prepupal, pupal, or pharate adult stage on the basis of external appearance (Bainbridge and Bownes 1981). Homozygous or hemizygous Met²⁷ pupae can be identified by a slightly elongated pupal case (T. G. Wilson, unpublished data). For quantitative lethal-phase analysis, white or very young brown prepupae were transferred to moistened ashless filter paper (Whatman no. 42) in small glass petri dishes and allowed to continue developing in a humid chamber. Developmental stage was monitored daily until eclosion or death was evident. We found some variability (<5%) in the survival rates from pupariation to eclosion of Met br¹ and Met rbp² in different genetic backgrounds. To accurately compare the phenotypes of the various allele combinations, crosses were standardized using FM7-balanced mothers whenever possible.

Ovipositional rate and ovary examination:

Females isolated within 4–6 hr after eclosion were provided with wild-type (OreRC) males in food vials sprinkled with baker's yeast. Egg counts were made at 2-day intervals when the medium was changed, and fertility of the eggs was noted. Oogenesis was assessed by dissecting ovaries from females at several times after eclosion and examining for the presence of stages 8–14 vitellogenic oocytes as previously described (Wilson and Ashok 1998).

Analysis of BRC protein accumulation by immunoblotting:

White prepupae (WPP) were collected and either homogenized immediately or placed in humid chambers for further development. In experiments with late-pupal stages, animals were resynchronized at head eversion. Protein extraction was based on the method of Emery et al. (1994). For each time point, 5–10 animals were homogenized in 50–100 μl sample buffer with a Teflon pestle in a microcentrifuge tube. Sample buffer consisted of 75 mm Tris-HCl, pH 6.8, 6% SDS, 15% glycerol, 10% β-mercaptoethanol and protease inhibitors (0.1 μg/μl pepstatin A, 0.5 μg/μl leupeptin, and 10 mm PMSF, Sigma). Following centrifugation for 10 min at 14,000 rpm (Eppendorf 5415C), the supernatant was used immediately for electrophoresis or stored at −80° for up to 3 weeks, which did not compromise BRC protein stability (data not shown).

Extracts representing 0.25- or 0.5-animal equivalents were heated for 5 min at 90°, quick chilled on ice for 10 min, and separated by SDS–PAGE (Towbin et al. 1979). For optimal band separation and size assessment, we used large (16 × 18 cm) 10% acrylamide gels on a Hoefer SE 600 Ruby electrophoresis apparatus at constant current (30 mA) for 5 hr. Otherwise, 12% acrylamide gels (7 × 8 cm) were run on a Bio-Rad mini-PROTEAN II apparatus at constant voltage (195 V) for 50 min. Proteins were transferred to nitrocellulose membranes (Protran, 0.45 μm, Schleicher & Schuell) by electroblotting at 4° overnight. Overall protein pattern was detected by staining the membrane with 0.5% Ponceau-S (Sigma, St. Louis).

Nonspecific binding sites were blocked with phosphate-buffered saline (PBS) plus 0.1% Tween 20 (TPBS) and 5% w/v nonfat dry milk powder (Bio-Rad) for 90 min at room temperature. Blots were probed with either anti-BRcore (mAb25E9) or anti-Z1 (mAb3C11) mouse monoclonal antibodies (Emery et al. 1994), diluted 1:2,500 or 1:100, respectively, in PBS + 5% w/v milk powder, for 2 hr at room temperature or overnight at 4°. After three 10-min washes in TPBS, the blots were incubated with HRP-conjugated goat anti-mouse IgG (Cappel, ICN) at 1:10,000 (with anti-BRcore) or 1:2,500 (with anti-Z1) for 1 hr at room temperature. After another three 10-min washes, the signal was revealed by chemilumiscent detection of HRP (ECL detection kit; Amersham-Pharmacia) and exposure to X-ray film (Kodak X-OMAT AR). To evaluate lane loading, the blot was stripped (by serial 10-min washes in water, 0.2N NaOH, water, and TPBS), reblocked, and reprobed with anti-actin (mAb1501; Chemicon) at 1:20,000 (Mackler and Reist 2001).

RESULTS

Interaction between Met and rbp alleles:

We first investigated the impact of reduced Met function on rbp mutants with marginal viability. Heteroallelic mutants carrying the viable allele rbp² with the BRC-null allele npr1³ showed ∼60% eclosion (Figure 1). Careful examination of rbp²/npr1³ mutants, starting at puparium formation and continuing throughout metamorphosis, showed that 100% of them were able to pupate, after which ∼15% died as early pupae (Figure 1). Eighty percent of the original cohort survived to late pharate adult stage, with ∼20% failing to eclose. In contrast, when the hypomorphic-viable allele Met³ was crossed onto each of the BR-C mutant chromosomes, the survival curve of the resulting heteroallelic double mutants was shifted (Figure 1), showing a dramatic reduction in viability.

Figure 1.

BRC and Met mutations interact to cause synergistic enhancement of lethality during metamorphosis. Survival curves were obtained by selecting individuals as very young prepupae, culturing them at 25° in a humid chamber, and evaluating their developmental progress, with staging based on Bainbridge and Bownes (1981). The x-axis is not strictly proportional to time because the stages are of unequal duration and mutants of different genotypes develop at different rates. Single-mutant BRC progeny, y rbp² w/y npr1³ w sn³ (n = 156) were generated by crossing y npr1³ w sn³/Binsn females to y rbp² w/y²Y67g males. Control siblings from the same cross, y rbp² w/Binsn (n = 157) showed 100% eclosion. Double mutants, y rbp² w v Met³/y npr1³ w v Met³ (n = 115) were generated by crossing y npr1³ w v Met³/FM7, y^31d B v females to y rbp² w v Met³/Y males. Sibling controls from that cross, y rbp² w v Met³/FM7, y^31d B v (n = 114) had 100% eclosion. Data for v Met³ (n = 284) and v Met²⁷ (n = 125) include similar numbers of hemizygous male and homozygous female progeny from the respective inter-sib matings.

We next examined survival in double mutants carrying Met²⁷, a bono fide null allele (Wilson and Ashok 1998), with each of two alleles of rbp: rbp², a weak allele, and rbp¹, a severe allele (Belyaeva et al. 1980; Kiss et al. 1988). Met²⁷ flies show good survival, with ≤15% mortality during pupal development (Figure 1). Flies homozygous for rbp² and carrying one copy of Met²⁷ showed good survival to adults (Table 1). In contrast, survival of rbp² Met²⁷ homozygotes to the adult stage was poor, dying typically during the pharate adult stage (Table 1).

TABLE 1.

Survival to adulthood of females carrying various combinations of BR-C and Met alleles

Mutant progeny of interest	N	Estimated survival of mutant (%)
rbp alleles
rbp2 Met27/rbp2 Met27	189	<2
rbp2 Met27/rbp2	498	86
rbp2 Met27/rbp2 Met27; p[Met+]/+	298	84
rbp1 Met27/rbp1 Met27	204	0
rbp1 Met27/rbp1	350	0
rbp1 Met27/rbp2 Met27	278	0
rbp2 Met27/rbp1	265	100
rbp2 Met27/rbp1 Met27; p[Met+]/+	384	80
br alleles
br1 Met27 br1 Met27	238	<2
br1/br1 Met27	195	84
br1 Met27/br1Met27; p[Met+]	277	64
br5 Met27/br1 Met27	135	0
br5 Met27/br1	247	54
br5 Met27/rbp1 Met27	239	0
br5 Met27/rbp1	271	38
br5 Met27/rbp2 Met27	294	10
br5 Met27/rbp2	125	82
br1 Met27/rbp1 Met27	182	8
br1 Met27/rbp1	242	88
2Bc alleles
2Bc1 Met27/2Bc1 Met27	119	0
2Bc1 Met27/2Bc1	247	0
2Bc1 Met27/rbp2 Met27	360	78
2Bc1 Met27/rbp1 Met27	133	0
2Bc1/rbp1 Met27	238	92
2Bc1 Met27/br1 Met27	350	82
2Bc1 Met27/br5 Met27	208	24
2Bc1/br5 Met27	178	104

Flies were generated from FM7-bearing females carrying the first chromosome listed in the above genotypes. N is the total female progeny examined from a particular cross, and estimated survival is the percentage of non-FM7 females of interest divided by 0.5. Homozygous Met²⁷ female survival is presented first, followed by the corresponding heterozygous Met²⁷ female survival, and finally representative p[Met⁺] transgenic survival.

Double mutants homozygous for Met²⁷ and rbp¹ did not survive to adulthood, as expected (Table 1). The lethal phase was shifted from the readily discernable pharate adult stage seen in rbp¹ pupae (Belyaeva et al. 1980; Kiss et al. 1988) to an earlier, less well-defined prepupal/early pupal stage, judging from visual observation of the pupae. Heteroallelic rbp² Met²⁷/rbp¹ Met²⁷ individuals, in contrast to surviving rbp² Met²⁷/rbp¹ individuals, were also prepupal/early pupal lethals (Table 1). Therefore, loss of Met⁺ gene product resulted in pharate adult lethality in rbp² and shifted the lethal phase of rbp¹ to an earlier stage.

Lethality in rbp² Met²⁷ homozygotes or rbp² Met²⁷/rbp¹ Met²⁷ heteroallelic pupae could be rescued by one copy of Met⁺ as the p[EN71] transgene (Table 1).

Interaction between Met and br alleles:

We next tested for interaction between Met and br. A viable br allele, br¹, exists (Morgan et al. 1925; Kiss et al. 1988). Individuals homozygous for br¹ Met²⁷ were usually lethal, but escaper adults eclosed in small numbers (Table 1), and hemizygotes showed slightly higher viability (Table 3) and were fertile. Examination of br¹ Met²⁷ pupae showed lethality in both pupal and especially pharate adult stages. Transgenic br¹ Met²⁷; p[EN71]/+ flies readily survived (Table 1). Individuals homozygous or hemizygous for the severe br⁵ allele were lethal in early pupal development (Kiss et al. 1988) and double mutants of Met²⁷ with br⁵ were likewise lethal in prepupal/early pupal development. Heteroallelic br⁵ Met²⁷/br¹ pupae survived well, but br⁵ Met²⁷/br¹ Met²⁷ heteroallelic individuals were lethal in prepupal/early pupal development (Table 1), an effect of Met²⁷ similar to that seen in rbp² Met²⁷/rbp¹ Met²⁷.

TABLE 3.

Oviposition by females of various BR-C and Met genotypes

		Eggs laid/female/2-day period: day after eclosion
Genotype	N	2	4	6	8
br1 v Met27/br1 v Met27	63	0	1.1	<1	<1
br1 v Met27/br1	20	6.6	30.9	42.6	38.0
p[Met+]/+	20	1.0	13.0	14.0	14.6
v/v	30	9.3	76.6	129	136
	43	0	0	<1	<1
	20	4.8	17.8	17.6	16.6
p[Met+]/+	30	4.1	29.8	34.0	32.3

Each value is the mean of egg counts from females of the indicated genotypes isolated from at least two cultures. Due to low preadult survival, 5–7 cultures were required to produce the indicated numbers of the exceptional double mutant homozygotes.

Interaction between br¹ and other Met alleles:

It is possible that the effects of Met²⁷ in the double mutants are allele specific. To determine if other alleles of Met also show an interaction with br¹, double mutants were constructed and the resultant hemizygotes examined. Table 2 shows hemizygote survival values for each allele combination. Examination of the pupae showed the lethal phase generally to be pharate adult, as found for br¹ Met²⁷. The only Met-weak allele recovered to date is Met^E1, and the double mutant showed good survival. The remaining alleles resulted from mutagenesis screens employing a variety of mutagens (Wilson and Fabian 1987; Ashok et al. 1998) and generally show poorer survival with br¹. Therefore, the interaction between Met and br¹ is not allele specific for Met²⁷.

TABLE 2.

Survival of br¹ hemizygotes and oviposition by homozygotes carrying various Met alleles

Genotype of X chromosome	Hemizygotes (% of F1 adults)	Oviposition (eggs/female/10 days)
br1	27	880 ± 50
Met3	25	773 ± 81
br1 Met	19	179 ± 25
br1 Met3	25	134 ± 9
br1 MetA3	5.4	16 ± 7
br1 MetE1	26	987 ± 79
br1 Met128	14	26 ± 8
br1 Met253	15	12 ± 3
br1 MetW3	7.8	27 ± 7
br1 Met27	2.8	3.4 ± 0.8
br1 MetD29	21	32 ± 5

For each chromosome, F₁ progeny were generated from at least three separate cultures of FM7-balanced females × FM7/Y or br¹ Met/y²Y67g males, the latter to generate homozygous females for oviposition determination. Survival is expressed as the percentage of adult hemizygotes for the indicated X chromosome present among the F₁ progeny. Oviposition (eggs laid ± SEM) was determined over a 10-day period beginning at 2–4 day post eclosion for females that were homozygous for the indicated genotype. When >100 eggs were laid during a 2-day period on the food surface, the total oviposition was estimated by extrapolating from one counted quadrant of the food surface. The oviposition rate of Met³ was similar to that of the other Met alleles, except for the lower rate of Met²⁷ Wilson and Ashok (1998).

Interaction between Met and 2Bc alleles:

Finally, we tested for interaction between Met and 2Bc. Both alleles of 2Bc, 2Bc¹ and 2Bc², have a similar phenotype of prepupal lethality (Belyaeva et al. 1980; Kiss et al. 1988). Individuals homozygous for 2Bc¹ Met²⁷ (Table 1) or 2Bc² Met²⁷ (data not shown) were also lethal as expected, and visual examination of the pupae showed the stage of prepupal lethality to be indistinguishable from that of 2Bc individuals. Therefore, either Met does not interact with 2Bc or the prepupal/early pupal-lethal phase cannot be shifted to an earlier stage in individuals carrying Met²⁷.

Altered BR-C complementation patterns in the absence of MET:

BR-C complementation group mutants complement one another to varying extents, with rbp partially complementing br, and 2BC fully complementing both rbp and br (Belyaeva et al. 1980; Kiss et al. 1988). To determine if Met²⁷ influences complementation for viability, double mutants for both the weak and severe alleles of br and rbp and for 2Bc were crossed and survival of the F₁ examined. The absence of Met⁺ resulted in failure of br⁵ and rbp¹ to complement one another as well as reduced complementation between br⁵ and rbp² (Table 1). Complementation was reduced between 2Bc¹ and either br⁵ or rbp¹. However, complementation in the Met²⁷ double mutants continued to be robust between either of the weak alleles br¹ or rbp² with 2Bc¹ (Table 1). Therefore, an absence of MET reduced complementation to varying degrees, depending on the alleles examined, being less evident with the weak alleles and with the 2Bc¹ mutation.

Oogenesis in BR-C Met double mutants:

Both of these genes are involved in oogenesis in Drosophila (Huang and Orr 1992; Wilson and Ashok 1998; Tzolovsky et al. 1999). Normally, vitellogenic oocyte development begins within 12 hr following eclosion, and mature eggs begin to be oviposited within 48 hr (King 1970; Wilson and Ashok 1998). BR-C homozygotes for the weak alleles br¹ and rbp² showed good oviposition, although below that of v/v females (Table 3). Met²⁷ females show oogenesis reduced to ∼20% of wild type (Wilson and Ashok 1998). To determine if interaction between BR-C and Met is apparent during this process in adults, oviposition was examined in BR-C Met homozygotes at 2-day intervals following eclosion. Homozygous br¹ Met²⁷ and rbp² Met²⁷ females occasionally survived to adulthood, and survivors showed strong reductions in oviposition: both br¹ Met²⁷ and rbp² Met²⁷ homozygotes laid only a few eggs during an 8-day examination period, and casual examination of these females for another week showed no change in the ovipositional pattern.

To determine if the ovipositional failure was due to a defect in oogenesis or in oviposition, both br¹ Met²⁷ and rbp² Met²⁷ females were dissected at 5–7 days following eclosion and their vitellogenic oocytes staged and censused. Despite having access to ample yeast and courting wild-type males, there were few (usually 0–3) vitellogenic oocytes in the ovaries of these females, showing that the defect resides in oogenesis, not oviposition. However, no degenerating oocytes were present, a condition suggestive of hormonal disruption (Wilson 1982; Soller et al. 1999).

Oviposition was also measured in the br¹ Met double-mutant combinations of the various Met alleles. Some of the allele combinations resulted in severe disruption of oviposition (Table 2), presumably due to defects in oogenesis, as seen for Met²⁷. Other combinations resulted in ovipositional rates higher than those of br¹ Met²⁷. In summary, the severe depression in oviposition seen for some Met allele combinations represents a strong allele interaction with br¹ and is not specific for the Met²⁷ allele.

The link between BRC and methoprene:

We addressed the enigmatic relationship between BR-C and methoprene in three contexts: (i) the overlap between methoprene-induced defects and BR-C common phenotypes, (ii) the influence of BR-C function on sensitivity to methoprene, and (iii) the effect of methoprene on BR-C expression.

The BR-C common phenotypes include malrotation of male genitalia:

In our previous study, we showed that methoprene treatment of wild type causes a specific partial phenocopy of BRC-associated internal defects (Restifo and Wilson 1998). Here, we tested whether BR-C mutants of each complementation group show the well-known methoprene-induced malrotation of the male genitalia (Postlethwait 1974; Wilson and Fabian 1986). In some cases, this required combining strong mutations with moderate wild-type transgene activity, an established method for revealing late developmental functions (Hall and Thummel 1998). The malrotation phenotype is of particular interest because a genetic interaction between Met and spin, a Fas2 mutation with a malrotation phenotype, has recently been described (Adam et al. 2003).

For rbp, we found that 100% of rbp¹/Y hemizygotes have malrotated genitalia (Table 4). This fully penetrant phenotype was rescued by a BR-C⁺ Y-borne duplication y²Y67g and uncovered by the Y-borne duplication with an interstitial deletion y²YSz280 that lacks all BR-C sequences, confirming that it maps to the BR-C region. To generate br mutant males that die as pharate adults (when the genitalia are pigmented), we partially rescued br-null mutants br⁵/Y using two transgenic copies of heat-shock-inducible BRC–Z2. Optimal heat-shock protocols rescue lethality (Bayer et al. 1997), gene expression (Liu and Restifo 1998), and CNS morphogenesis (R. F. Spokony and L. L. Restifo, unpublished data). To obtain partial rescue, we heat-shocked unsynchronized third-instar larvae once (37°, 1 hr), and then twice more 18 and 23 hr later. This resulted in very small numbers of pharate adults, 93% of which (14/15) had malrotated genitalia (Table 4). For 2Bc, we first observed malrotation in two very rare, late-dying 2Bc²/Y mutants (Restifo and White 1991, 1992; Consoulas et al. 2005). To examine larger numbers, we used a “leaky” BRC–Z3 transgene whose expression is dependent on temperature (see materials and methods). We found malrotated genitalia, inversely related to rearing temperature: 14% at 25° and 40% at 22–23° (Table 4).

TABLE 4.

Malrotation of the male genitalia is a developmental phenotype of all BR-C complementation groups

Genotype	Culture conditions	Phenotype: % with malrotation
reduced bristles on palps
rbp1/Y	25° standard	100 (n = 31)
rbp1/y2Y67g	25° standard	0 (n = 128)
rbp1/y2YSz280	25° standard	100 (n = 23)
broad
br5/Y; hsZ2/+; hsZ2/+	37° heat shocks; partial rescue	93 (n = 15)
y w sn3/Y; hsZ2/+; hsZ2/+	37° heat shocks	0 (n = 16)
br5/y2Y67g	37° heat shocks	4 (n = 48)
lethal(1)2Bc
2Bc2/Y	25° standard	100 (n = 2)
2Bc1/Y; hsZ3/+	25°; modest partial rescue	14 (n = 28)
2Bc1/Y; hsZ3/+	22–23°; modest partial rescue	40 (n =53)
2Bc1/y2Y67g	22–23°	0.8 (n = 261)
yw/Y; hsZ3/+	22–23°	0 (n =582)

Neither ubiquitous expression of BRC–Z2 nor BRC–Z3 in wild type caused malrotation. Genetic controls, br⁵/y²Y67g or 2Bc¹/y²Y67g, exposed to the corresponding temperature protocol showed only very rare malrotation (4 or <1%, respectively), confirming the mapping of the phenotype to BR-C and suggesting the possibility of a very small heat-shock effect. In summary, BR-C mutants of all three complementation groups have malrotated male genitalia, which adds this methoprene-induced defect to the list of BR-C common phenotypes.

Sensitivity of BR-C mutants to methoprene:

Met mutations confer semidominant resistance to both the toxic and morphogenetic effects of methoprene (Wilson and Fabian 1986; Restifo and Wilson 1998). To determine if BR-C mutations, either singly or in double-mutant combination with Met, affect the response to methoprene treatment, larvae were raised in the presence of one of four diagnostic concentrations of methoprene and evaluated for survival and the external morphology of surviving adults. Viable BR-C mutations do not change the susceptibility to methoprene from that of wild type (Table 5). Similarly, reducing the dose of 2Bc⁺ or BR-C⁺ by 50% (2Bc¹/FM7 or npr1³/FM7, respectively) did not shift the sensitivity to methoprene. Moreover, in the double mutant, br¹ did not affect the resistance conferred by Met³. Hence, BR-C⁺ function does not appear to impact the MET-dependent methoprene-sensitivity mechanism.

Effect of Met and methoprene on BRC protein accumulation:

We investigated BRC protein expression for two reasons. First, a plausible mechanistic explanation for the genetic interaction between Met and BR-C would be that Met⁺ upregulates BR-C expression, and that reduced BRC levels in Met mutants would enhance the lethality of partial-loss-of-function BR-C genotypes. Second, reported effects on BR-C transcript levels caused by JH/JHA treatment (Zhou et al. 1998; Zhou and Riddiford 2002) have failed to provide an explanation for the methoprene phenocopy of BR-C common defects, especially those involving internal structures of the head and thorax (Restifo and Wilson 1998).

The BRC family of proteins, which migrate as three size groups (Emery et al.,1994; Mugat et al. 2000), is readily detected by immunoblotting of proteins extracted from whole white prepupae (Figure 2A). The largest group, Emery's p118, is thought to represent BRC–Z4; the middle group, Emery's p91 and p81 contains BRC–Z1 and –Z3; the smallest group, Emery's p64 and p57 contains BRC–Z2. Over the subsequent 24 hr, especially after head eversion (∼12 hr APF), BRC protein levels declined (Figure 2B). In Met²⁷ mutants, the pattern and relative abundance of BRC isoforms detected over this first day of metamorphosis was indistinguishable from those of wild type (Figure 2B). Likewise, methoprene treatment of wild-type animals did not change the overall quantities and isoform patterns of BRC proteins (Figure 2C).

Figure 2.

BRC protein accumulation during the first day of metamorphosis is not affected by lack of MET or by treatment with methoprene. Immunoblotting of whole-body protein extracts with anti-BRcore; 0.5-animal equivalents per lane. (A) Wild-type (OreRC) white prepupae. The migration positions of molecular weight markers are shown on the far left. The individual boxes on the right show optimized images of each group of BRC proteins, obtained by changing exposure times or amounts of protein loaded. The indicated molecular weight estimates are averages based on three or more independent experiments. (B and C) Time course of BRC accumulation in animals collected at puparium formation and sampled every 8 hr. The rightmost lanes contain protein from BR-C-null mutant (npr1³/Y) wandering third-instar larvae as a negative control to verify the specificity of the antibody. Each blot was reprobed for actin as an indicator of protein loading. hAPF, hours after puparium formation. (B) Wild-type (OreRC) and Met-null mutant (v Met²⁷). The BRC protein profiles are qualitatively and quantitatively indistinguishable. (C) Control and methoprene-treated OreRC. The BRC protein profiles are very similar.

In independent experiments, BRC proteins were evaluated over a 3-day interval, representing ∼75% of metamorphosis, during which animals were resynchronized at head eversion. At 24 hr APF (12 hr after head evesion), BRC proteins were detectable in the methoprene-treated animals, but not in the controls (Figure 3A). Similarly, methoprene-treated animals showed mild persistence of BRC–Z1 during midpupal stages, following a peak in accumulation at the normal time, 8 hr APF. There was no reappearance of BRC–Z1 or any other BRC isoforms during mid-to-late pupal stages (Figure 3A), even on very long exposures of the immunoblots (data not shown). In summary, chronic larval treatment with lethal doses of the JH agonist and mimic methoprene does not block BRC protein accumulation during the first day of metamorphosis, a developmental window in which BR-C function is essential.

Figure 3.

BRC protein accumulation is mildly prolonged by methoprene treatment. Animals were collected at puparium formation, sampled at 0 and 8 hr, resynchronized at head eversion (∼12 hr APF in control wild-type animals), and then sampled 4 hr after head eversion (= 16 hr APF) and at 8-hr intervals thereafter. Immunoblotting with anti-BRcore and anti-Z1; 0.5-animal equivalents per lane. In each experiment, control and methoprene-treated animals were reared and processed in parallel, the gels were run simultaneously in the same apparatus, and the antibody labeling and detection were performed in parallel. Each of the blots was reprobed for actin as an indicator of protein loading. (A) Wild type (OreRC). Two independent immunoblotting experiments are shown. Treatment with a lethal dose of methoprene causes a mild prolongation of BRC protein accumulation, especially of the abundant 75–85-kDa bands, which contain primarily BRC–Z1. Methoprene does not cause reappearance of BRC in the midpupal period (40– 64 hr APF). (B) Met-null mutant (Met²⁷). Sequential immunoblotting for BRC–Z1 and BRcore. As in wild type, BRC accumulation is mildly enhanced by treatment with a lethal dose of methoprene.

Over the 3-day interval, BRC protein profiles in Met²⁷ mutants were normal, in both the presence and absence of methoprene (Figure 3B). These data demonstrate that the accumulation of BRC proteins at the onset of metamorphosis is MET independent. Hence, altered BRC expression, at least at the level of the whole-body protein accumulation, cannot explain the genetic interactions between BR-C and Met mutations.

DISCUSSION

In this work, we have examined two genes required for signaling by 20E (BR-C) and JH (Met) to probe for interaction between these pathways. We have found evidence for interaction between Met and BR-C as reflected by synergistically reduced viability and oogenesis seen in double mutants. Consistent results were seen with different combinations of Met and br or Met and rbp alleles (Figure 1; Tables 1 and 2), indicating that the interactions are not allele specific in either direction.

Met interacted with both the weak-viable alleles br¹ and rbp² and with the severe alleles br⁵ and rbp¹ (Table 1) during pupal development. Each of the weak alleles possesses sufficiently functional gene product to permit completion of pupal development; but this amount is insufficient when MET is absent or defective. The more severe rbp¹ homogygotes are pupal lethal, but only at late metamorphosis, in the pharate adult stage (Belyaeva et al. 1980). Lethality was shifted in rbp¹ Met²⁷ pupae to prepupal/early pupal development, suggesting that MET absence causes the residual rbp¹ function to be insufficient during these earlier stages in pupal development. Homozygotes of br⁵ and 2Bc die in the early and late prepupal stage, respectively (Kiss et al. 1988), and the double mutants with Met²⁷ showed a similar phenotype, demonstrating that the interaction cannot shift lethality to an earlier stage, late third-instar larvae. Our observations are consistent with the interaction between BR-C and Met beginning in prepupal or early pupal development. While we interpret the Met–BR-C interaction as enhancing the lethality of br and rbp mutations, it is also possible that Met becomes an essential gene when BR-C function is reduced, or that the interaction is mutual, such that both mutations become more severe in phenotype when they are present together.

Genetic interaction became strikingly evident when complementation failures between mutant alleles from different BR-C complementation groups occurred in the presence of Met²⁷ (Table 1). Without MET, developing animals may be less able to make use of the partial functional redundancy among BRC isoforms that has been documented previously (Bayer et al. 1997).

The interaction between mutant alleles of BR-C and Met was also evident in the adult stage when oogenesis was examined. Both the rate of oviposition and the paucity of vitellogenic oocytes in ovaries of br¹ Met²⁷ and rbp² Met²⁷ females reflects almost complete failure of oogenesis, with only a few eggs oviposited during the lifetime of the female.

Previous studies have also detected BR-C interaction with other genes. Double mutants of BR-C with another primary response gene E74 show interaction for some but not all of the phenotypic characters (Fletcher and Thummel 1995). In addition to interactions among transcription regulators of the ecdysone cascade, br alleles interact with genes involved in imaginal disc morphogenesis, including those encoding an atypical serine protease, Stubble-stubboid (Beaton et al. 1988; Appel et al. 1993), nonmuscle myosin II heavy chain (zipper), the Drosophila serum response factor transcription factor [blistered: (Gotwals and Fristrom 1991)], the small GTPase Rho1, cytoplasmic tropomyosin, and 22 others (Ward et al. 2003).

Although BR-C expression and function overlap the JH/JHA-sensitive period, data presented above demonstrate that methoprene treatment does not block BRC expression in either wild-type or Met-null mutants (Figures 2 and 3). Furthermore, the methoprene phenocopy, which excludes complementation group-specific defects (e.g., larval salivary gland persistence, which is rbp-restricted), is not consistent with methoprene simply reducing BRC expression (Restifo and Wilson 1998). We propose that JH application results in abnormal function of BRC proteins, thus phenocopying certain characteristics common to all BR-C mutants. Therefore, the link between BR-C mutant phenotypes and JH-induced defects could be abnormal regulation of target genes, resulting in the phenotypic characteristics observed (Figure 4). Previously, we suggested several possibilities to explain methoprene pathology and BR-C phenocopy, including BRC interaction with an unidentified partner, perhaps MET (Restifo and Wilson 1998). We believe that the Met–BR-C genetic interaction reported in this work can be explained best by this hypothesized protein–protein interaction between MET and BRC to regulate one or more target genes. Supporting our hypothesis are the following findings: first, both proteins are located in the nucleus (Emery et al. 1994; Restifo and Hauglum 1998; Pursley et al. 2000; Renault et al. 2001), so there is no compartmental barrier to interaction. Second, both proteins appear to be transcription factors: BRC isoforms bind specific DNA sequences (von Kalm et al. 1994; Dubrovsky et al. 2001) and regulate transcription (Dubrovsky et al. 1994; Hodgetts et al. 1995; Crossgrove et al. 1996; Mugat et al. 2000). BR-C mutants have misexpressed secondary-response and other target genes (Guay and Guild 1991; Karim et al. 1993; Mugat et al. 2000; Renault et al. 2001; Dunne et al. 2002; Sempere et al. 2003). MET is a member of the bHLH–PAS family of transcription factors (Ashok et al. 1998) and was recently shown to act as one (Miura et al. 2005). Third, both are found at common times during development, such as prepupae (Karim et al. 1993; Ashok et al. 1998) and during vitellogenic oocyte development (Tzolovsky et al. 1999; Pursley et al. 2000). Finally, PAS domains in bHLH–PAS proteins are thought to promote protein–protein interaction (Heery et al. 1997), either with other PAS proteins or as coactivators with nuclear receptor proteins (Xu et al. 1999), and the BTB/POZ domain of BRC has been implicated in protein–protein interaction (Zollman et al. 1994; Melnick et al. 2002).

Figure 4.

Proposed scheme for regulation by MET and BRC–Z1 of a target gene necessary for pupal viability. Illustrated are three hypothesized transcriptional situations for rbp: top, wild type; center, in the presence of wild-type MET, strong hypomorphic rbp alleles cause lethality in the pharate adult stage; bottom, in rbp Met double mutants, lethality is shifted to the prepupal stage. No interaction between MET and Z1 is shown, but formation of heterodimers is possible. Likewise, each is shown binding DNA when the protein is wild type but not when mutant, although DNA binding by lesioned protein is possible. The presumed level of target-gene transcription is reflected by arrow thickness.

In Met²⁷ mutants, BRC protein accumulation profiles are normal (Figures 2 and 3). Since metamorphosis is not derailed in Met²⁷ pupae, BRC⁺ function in these pupae does not seem to be adversely affected. The fly may be protected from absence of MET by functional redundancy (Wilson and Ashok 1998). A candidate for the redundant substitute is the PAS gene germ cell expressed (gce), a gene with high (∼70% amino acid identity) homology to Met (Moore et al. 2000) that could substitute for MET to rescue larval and/or pupal development. However, this substitute does not appear to be satisfactory if BR-C is mutant. When a gce mutant becomes available, its phenotype could help evaluate this hypothesis.

How does the application of exogenous JH act to phenocopy BR-C? It is clear that the action of these compounds occurs through MET, probably acting as a JH receptor component (Wilson and Fabian 1986; Shemshedini et al. 1990; Shemshedini and Wilson 1990; Ashok et al. 1998; Miura et al. 2005). JH is present during larval development when it presumably acts to prevent premature metamorphosis resulting from each wave of 20E secretion that triggers a molt. This failsafe mechanism may occur by JH binding by and conformational change of MET, resulting in regulation of genes necessary for molting or perhaps simply blocking expression of metamorphic genes. Studies with Drosophila S-2 cells have implicated the transcription factor E75A in promoting JH regulation of larval development (Dubrovsky et al. 2004). At metamorphosis, when little or no JH is present (Bownes and Rembold 1987; Sliter et al. 1987), BR-C is expressed, and we propose that BRC dimerizes with the nonliganded MET protein to regulate a different set of target genes, promoting the initiation of metamorphosis. If exogenous JH is present during this time, it binds to MET and results in a more larval conformation, resulting in inappropriate binding to BRC and leading to a change in target-gene expression patterns consequently seen as defects characteristic of BR-C mutants.

Other work has implicated BR-C in the action of the JH agonist pyriproxyfen during metamorphic disruption. Zhou and Riddiford (2002) showed that application of this compound to white prepupae resulted in re-expression of BRC–Z1 in the abdomen during late pupal development, which in turn caused abnormal development of abdominal epidermis, including bristle disturbances. Those findings differ from ours with methoprene in two significant ways. First, a lethal dose of methoprene caused a mild enhancement and prolongation of BRC protein accumulation in young pupae, but no re-expression at later times (Figure 3A). Second, the modest effect of methoprene on BRC protein profiles cannot mediate the developmental effects of this JHA because the same mild persistence of BRC was seen in Met²⁷ mutants (Figure 3B), which are protected against methoprene-induced defects. It is not clear what underlies the difference in response of BR-C to methoprene and pyriproxyfen. We note that pyriproxyfen is a more powerful JH agonist than methoprene (Riddiford and Ashburner 1991), but qualitative differences in the actions of the two compounds may exist as well.

In summary, our results provide genetic evidence that supports other studies implicating BR-C as a focal point for interaction of JH and 20E signaling pathways, and they suggest that BRC and MET interact to regulate expression of one or more effector genes involved in metamorphic development.