Molecular Evolution of Juvenile Hormone Signaling

Insect development proceeds through a series of discrete developmental stages called instars. During hexapod evolution, the development of complete metamorphosis introduced a novel mechanism for separating feeding and reproductive stages (Truman & Riddiford, 2002), facilitating the tremendous evolutionary success of holometabolous insects. In contrast to hemimetabolous insects, which progress through a series of instars that appear as smaller iterations of the adult form, holometabolous insects proceed from egg to adult through a progression of isomorphic larval instars and a pupal transitory stage. In each case, the physical boundary for growth during an instar is established by a chitinous exoskeleton, which must be periodically shed. This molting process is under the control of two counteracting hormones. Toward the end of an instar, a pulse of the insect molting hormone, 20-hydroxyecdysone (20E) initiates a transcriptional cascade that carries the molt to a subsequent instar. However, it is the interaction of 20E and the sesquiterpenoid juvenile hormone (JH) that governs the developmental outcome of each molt. During larval development, an elevated JH titer and 20E directs the sequential progression through larval development until the final larval instar, when the JH titer substantially declines. The removal of circulating JH facilitates a 20E-directed developmental switch that initiates the metamorphic molt. Thus, it was proposed that JH can modulate 20E activity, maintaining the status quo during preadult development.


Introduction
Insect development proceeds through a series of discrete developmental stages called instars.During hexapod evolution, the development of complete metamorphosis introduced a novel mechanism for separating feeding and reproductive stages (Truman & Riddiford, 2002), facilitating the tremendous evolutionary success of holometabolous insects.In contrast to hemimetabolous insects, which progress through a series of instars that appear as smaller iterations of the adult form, holometabolous insects proceed from egg to adult through a progression of isomorphic larval instars and a pupal transitory stage.In each case, the physical boundary for growth during an instar is established by a chitinous exoskeleton, which must be periodically shed.This molting process is under the control of two counteracting hormones.Toward the end of an instar, a pulse of the insect molting hormone, 20-hydroxyecdysone (20E) initiates a transcriptional cascade that carries the molt to a subsequent instar.However, it is the interaction of 20E and the sesquiterpenoid juvenile hormone (JH) that governs the developmental outcome of each molt.During larval development, an elevated JH titer and 20E directs the sequential progression through larval development until the final larval instar, when the JH titer substantially declines.The removal of circulating JH facilitates a 20E-directed developmental switch that initiates the metamorphic molt.Thus, it was proposed that JH can modulate 20E activity, maintaining the status quo during preadult development.

JH and JHAs: Insecticidal use of hormone agonists
Since the first chemical analysis resolved the sesquiterpenoid structure of endogenous JH (Röller et al., 1967), several homologs have been identified, each bearing opposing, terminal epoxide and methyl ester functions.Variation in the degree and identity of alkyl group substitution at C3, C7, and C11 along the carbon skeleton defines the homologs.The evolutionary importance of multiple JH homologs is unclear.JH 0, I, II, and III have all been isolated from lepidopteran insects, whereas JH III, the presumed evolutionary precursor to the higher homologs, is found in all insects.JH bisepoxide (JHB3) has been identified as a product of the corpus allatum (CA) in higher Diptera including Drosophila melanogaster and Sarcophaga bullata (Richard et al., 1989;Bylemans et al., 1998).Nearly identical in structure to JH III, JHB3 is distinguished by an additional epoxide group spanning C6-C7.The physiology and chemistry of JH prompted intense research into the synthesis and commercial-scale production of JH analogs, or juvenoids, for agricultural use.The allure of these compounds was at least twofold.First, juvenoids exhibit extremely low nontarget (in particular, mammalian) toxicity.Second, it was originally thought that insect resistance to JHAs would be unlikely, since an insect was not likely to become refractory to an endogenous hormone (Williams, 1967).Methoprene, a juvenoid structurally similar to endogenous JH, has enjoyed success in the management of larval mosquito populations.However, JHAs need not mimic the chemical structure of endogenous JH, as exemplified by the pyridine-based pyriproxyfen, whose activity exceeds JH by two orders of magnitude in dipteran white puparial and larval assays (Riddiford and Ashburner, 1991).Exogenous JH exposure can elicit classic antimetamorphic activity in both Lepidoptera and Coleoptera (Srivastava & Srivastava, 1983;Konopova & Jindra, 2007), extending larval development through one or more supernumerary instars.Also in these insects, exposure to exogenous JH or to its chemical analogs (JHA) can result in the deposition of a second pupal cuticle (Zhou & Riddiford, 2002).Thus, in Lepidoptera and Coleoptera, JH exposure at an inappropriate time inhibits 20E-directed developmental progression.In Diptera, treatment with exogenous JH produces dose-dependent lethality at the pharate adult stage.All adult structures arise from imaginal discs in flies, and these discs are insensitive to JH during development, unlike Lepidoptera and Coleoptera, in which the reduced oogenesis (~20% compared to Met + ), consistent with a role for JH in this physiology.However, since absence of a JH receptor is expected to preclude normal development, the viability of Met 27 flies challenged the notion of Met as a bona fide JH receptor.Some evidence supports alternative mechanism(s) of JH signaling (see Flatt et al., 2008;Riddiford et.al., 2010).In this chapter, we review data that support the notion of germ cell expressed (gce), the paralog of Met in higher Diptera, as conferring at least partial functional redundancy.

Met homologs across holometabola
Reports of methoprene resistance in mosquito populations (Dame et al., 1998;Cornel et al., 2000;Cornel et al., 2002) led us to investigate the Met orthologs of three mosquito species: Aedes aegypti, Culex pipiens, and Anopheles gambiae.Using a combination of degenerate RT-PCR and genomic database mining, we isolated a single Met homolog from each of these mosquitoes.Sequence analysis of these genes showed that they share high identity with both Met and the closely related gce from Drosophila, as expected.However, a comparison of the genomic structures among DmMet, Dmgce, and the three putative mosquito Met genes revealed higher structural conservation between each mosquito Met and Dmgce.Importantly, the intron number of these genes is more consistent with that of Dmgce than DmMet (from six to nine, versus one in DmMet).Furthermore, several introns in each mosquito gene are positionally conserved with those in Dmgce.This led to our proposal that the Met gene of higher Diptera originated via retrotransposition of a basal, gce-like gene of lower Diptera (Wang et al., 2007).Retrotransposition, or retroposition is a mechanism of gene duplication that proceeds through an mRNA intermediate.Following post-transcriptional splicing, the parental message is reintegrated into the genome.Ultimately, for the duplicate copy to escape the fate of becoming a pseudogene, it must reintegrate with associated regulatory elements intact or incorporate into a suitable transcriptional environment elsewhere in the genome.Following duplication, the increase in copy number of the parental gene affords a relaxation of selective constraint, facilitating functional divergence.This may manifest as subfunctionalization, in which a modification of the parental function evolves, or neofunctionalization, which refers to attainment of a novel function (MacCarthy & Bergman, 2007).DmMet retains a strong diagnostic feature of retroposition: a paucity of introns relative to gce, which is consistent with splicing and genomic reintegration of an ancestral gce-like transcript.A conserved gce-like gene appears to be conserved across holometabolan genomes, including the red flour beetle, Tribolium castaneum, and the honeybee, Apis mellifera.An independent gene duplication within the Lepidoptera has given rise to two Met-like proteins, presently called Methoprene tolerant proteins I and II, whose functions are currently under investigation (i.e.Li et al., 2010).Despite a demonstrated sequence conservation favoring the Met-like genes of more primitive Holometabola as ancestral to gce, we will continue to refer to these genes as Met-like in this text.

Met and gce within the genus Drosophila
When the genomes of 12 representative Drosophila species became available (Ashburner, 2007), we chose to examine the molecular evolution of Met and gce within this genus of flies.Both paralogs are conserved in each species, indicating that the origin of Met predates that of the genus Drosophila, some 63 million years ago (Tamura et al., 2004).The architecture of these genes is generally conserved in each species, with a few notable exceptions.A single conserved intron is present in Met in the PAS B domain of 11 species.In addition to this conserved intron, independent intron gains have occurred in the lineages leading to D. simulans and D. willistoni.A single Met ortholog exists in each Drosophila genome examined, but D. persimilis harbors two separate, consecutive loci on the X chromosome, currently called GL13106 and GL13107, that align to distinct regions of DmMet.The 5' putative gene GL13106 contains a complete PAS A domain followed by a severely truncated PAS B domain.We performed RT-PCR across these two genes and failed to obtain a single PCR product, suggesting that GL13106 and GL13107 indeed code for two distinct open reading frames.Eleven of the 12 representative gce orthologs contain at least six conserved introns, with independent intron gains evident in the lineages leading to D. melanogaster, D. pseudoobscura, and D. mojavensis, whereas a substantial deletion in D. persimilis gce has eliminated the central portion of this gene, including the PAS repeats.In addition to the bHLH, PAS, and PAC domains, putative transactivation domains (TAD) are evident in Met and gce orthologs.TADs are glutamine and/or aspartic acid-rich motifs whose amino acid sequences are broadly defined and generally reside in the C-terminal region of PAS proteins (Ramadoss & Perdew, 2005).Met homologs show Q-and D-rich motifs between the PAS B and PAC domains, while alignments of gce homologs indicate a D-rich region C-terminal to the PAC domain.Miura et al. (2005) suggest the presence of a Cterminal TAD in recombinant MET protein, but this region has yet to be functionally defined.Using DmMet and Dmgce as query sequences, we conducted homology searches under tBLASTx criteria (translated nucleotide query to search a translated nucleotide database) against the publicly available EST library of Glossina morsitans, the tsetse fly.Our search recovered several clones, which were imported into the Sequencher program to produce two independent contigs.These composite nucleotide sequences were used to infer a gene tree with other holometabolan Met and gce orthologs, including those of two representative Drosophila species (Figure 2).This preliminary analysis reveals the presence of distinct Met and gce orthologs in the G. morsitans genome, indicating that the origin of Met predates the divergence of the Aschiza and Schizophora.These two taxonomic groups, which are estimated to have diverged more than 85 million years ago (Bertone & Wiegmann, 2009), reside within the brachyceran infraorder Muscomorpha.

Evidence for differential selective constraint imposed on Met and gce
Based on an a priori hypothesis that Met and gce were subject to differential post-duplication selective constraint, we performed analyses of nonsynonymous-to-synonymous (dN/dS) substitution ratios on codon alignments of these Drosophila paralogs.Datasets were analyzed using the DataMonkey tool (Kosakovsky- Pond & Frost, 2005), a web-based implementation of the HyPhy package (Kosakovsky Pond et al., 2005).dN/dS analyses can be used to infer the relative selective pressure along entire coding sequences or in a site-specific manner.A substantially depressed dN/dS ratio (i.e.zero or close to zero) implies purifying (negative) selection.That is, nonsynonymous changes are stringently selected against.In contrast, when dN/dS is nearly one, neutral evolution is inferred.A dN/dS value far in excess of one implies positive selection, or adaptive evolution.In this case, nonsynonymous substitutions confer a selective advantage.
The results of our dN/dS analyses showed dramatic dissimilarity in the relative selective pressures that have shaped the coding sequences of Met and gce.In the case of Met, dN/dS was generally suppressed along the entirety of the coding sequence, indicating strong selection against nonsynonymous codon substitution.This is perhaps surprising, since MET deficiency has no effect on viability (Wilson & Ashok, 1998).Possibly, mutations that alter amino acid identities are selected against in Met due its involvement in reproduction.In the absence of methoprene selection, Met mutants are quickly out-competed by wild type flies despite the seemingly slight fitness cost of Met loss (Minkhoff III & Wilson, 1992).In contrast, dN/dS values close to one dominate the C-terminal half of gce, indicating a substantial relaxation of selective constraint in this region.The N-terminal region of this gene, containing the canonical bHLH and PAS functional domains, shows a strongly depressed dN/dS.Based on functional data from other PAS proteins, this region is assumed to harbor DNA and ligand binding activity, whereas the C-terminal region contains putative TADs.C-terminal degeneracy was shown to confer differential target gene specificity between the Ahr homologs of mice and humans (Ramadoss & Perdew, 2005;Flaveny et al., 2010).Similarly, the disparate selective constraints evident in the C-terminal regions of Met and gce may partially define these genes' functions.

Toward a functional definition of Dmgce
A functional characterization of gce, named for its expression in a subset of embryonic germ cells (Moore et al., 2000), is in its infancy.Column pulldown assays showed MET, in addition to forming homodimers, forms heterodimers with GCE, and addition of JH or either of two JHAs significantly impaired these interactions (Godlewski et al., 2006).It is unknown whether GCE forms homodimers, like MET, or whether GCE can bind JH or its analogs.GAL4/UAS-driven (Brand & Perrimon, 1993) overexpression of Met + from actin or tubulin promoters results in larval lethality in the absence of methoprene (Barry et al., 2008), perhaps by upsetting the stoichiometry of MET and GCE dimers, favoring MET homodimerization at inappropriate times or in inappropriate tissues.Recently, JH was shown to inhibit MET and GCE in D. melanogaster by preventing caspase-driven programmed cell death (PCD) and histolysis of the larval fat body.DRONC and DRICE, evolutionarily conserved caspase genes involved in this physiology at the onset of metamorphosis, were shown to be downregulated in Met and gce deficient flies (Liu et al., 2009).Similarly, methoprene interferes with caspase-driven midgut remodeling in A. aegypti (Nishiura et al., 2003;Wu et al., 2006) and T. castaneum (Parthasarathy et al., 2008;Parthsarathy et al., 2009), showing that this mechanism of JH action is evolutionarily conserved.It is noteworthy that recombinant MET can repress reporter gene expression in the absence of JH (presumably, MET forms homodimers in this system; Miura et al., 2005); transcriptional repression has previously been reported in other PAS proteins (Dolwick et al., 1993).Therefore, the JH-dependent, stage-specific formation of alternative MET/GCE dimers may have unique regulatory consequences on distinct suites of target genes.

Dmgce substitution for DmMet
To evaluate the notion that gce might confer viability to Met null flies, we manipulated gce expression using a binary UAS/GAL4 system to drive either a gce cDNA or an RNAi construct designed to target gce transcript.We carried these experiments out in a variety of genotypic contexts in order to examine the effect of gce transcript abundance on several methoprene conditional and non-conditional phenotypes (Baumann et al., 2010b).First, we explored the effect of gce over-and under-expression on a Met-specific non-conditional phenotype that manifests as a variable number of grossly malformed posterior facets of the compound eye (Figure 3).This phenotype is visible in Met 27 and Met w3 flies, and is enhanced in the latter genotype.In our experiments, we found that gce overexpression in a Met w3 genetic background can rescue the Met-specific eye phenotype, suggesting functional overlap of gce and Met.Notably, when gce was overexpressed in a Met 27 background from the GawB}dan[AC116] promoter, targeting transgene expression to the compound eye, the eye phenotype was completely rescued (Baumann et al., 2010b).The Met 27 phenotype mimics a set of defects resulting from genetic ablation of the JHproducing corpus allatum (CAX), including a heterochronic shift in EcR-B1 expression in the optic lobe (Riddiford et.al., 2010).Exogenous JH application rescues the entire suite of defects in CAX prepupae, while JH provision to Met 27 flies rescues only a subset of these defects, suggesting an alternate mechanism of JH signal transduction (Riddiford et al., 2010).Based on our findings that gce can substitute for Met in the compound eye, further study of GCE involvement in eye development may provide a link between these phenomena.For instance, GCE may partially substitute for MET as a ligand binder to mediate JH signaling when this hormone is supplied in excess.We also explored the effect of gce overexpression on several methoprene-conditional phenotypes.Overexpressed gce rescued both the diagnostic malrotation of male genitalia and sensitivity to the toxic effects of methoprene exposure.Sublethal doses of methoprene can induce malrotation of the male genital disc in D. melanogaster, resulting in terminalia that are improperly oriented for copulation (Bouchard & Wilson, 1987).Met 27 males are resistant to this phenotype.We found that global gce overexpression in a Met 27 background rescues blockage of the malrotation phenotype in Met 27 ; UAS-gce/ tubulin-GAL4 flies.When these flies were exposed to methoprene, we observed malrotation close to levels seen in Met + flies (Baumann et al., 2010b).Met and gce are generally co-expressed in JH target tissues, but we detected insignificant amounts of gce transcript in late third instar larval fat body.When gce was expressed from a construct targeting expression to this tissue, partial rescue of JH-induced pupal lethality was achived, perhaps as a result of supplying gce to a tissue in which its expression is normally depressed at this time in development.gce expression in the larval fat body was unable to rescue either the eye phenotype or to prevent methoprene-induced malrotation of the male genitalia, indicating that gce substitution for Met is tissue specific (Baumann et al., 2010b).

Functional partitioning of DmMet and Dmgce in D. melanogaster reproduction
Following metamorphosis, the interaction of 20E and JH is crucial in insect reproduction.JH was first isolated in large quantities from the MAG of Hyalophora cecropia (Williams, 1956), suggesting a role in male reproductive biology.In D. melanogaster, JH controls MAG protein accumulation (Yamamoto et al., 1988) and male apterous (ap) mutants court females less vigorously than wild-type flies (Tompkins, 1990).In females, the activity of these counteracting hormones is critical for ovarian development and oocyte maturation.Development of the D. melanogaster oocyte is under the control of JH through previtellogenic stages 8-9.Female D. melanogaster apterous 4 mutants are sterile owing to reduced levels of JH synthesis (Bownes, 1989); provision of exogenous JH rescues vitellogenic oocyte development in ap females (Postlethwait & Weiser, 1973).In A. aegypti, JH also controls previtellogenic ovarian development (Clements, 1992).In this case, JH signaling is necessary to promote 20E competence in the fat body, the site of post-blood meal vitellogenin synthesis.In contrast, vitellogenesis is retarded by JH treatment in the gypsy moth, Lymantria dyspar (Davis et al., 1990).Thus, there is variation in hormonal control in insects.GCE clearly compensates for MET deficiency in preadult development (Baumann et al., 2010b).In our experiments, over-expressed gce failed to rescue both the documented behavior of reduced courtship in Met 27 ; UAS-gce/tubulin-GAL4 males and the reduction in oocyte development and oviposition in these females.Therefore, it appears that excess gce cannot compensate for Met-induced reduction of reproductive capacity.This result suggests that the functional roles for MET and GCE are incompletely partitioned between preadult development and reproduction in adults.In A. aegypti, AaMet regulates the transcription of several JH target genes in newly eclosed, previtellogenic adult females (Zhu et al., 2010).Presumably, the MET-like gene product in lower Diptera serves an analogous function both MET and GCE in JH signaling, but through the action of a single gene.This is perhaps accomplished by virtue of its modular architecture of DmMet-and Dmgce-specific domains.Higher sequence identity exists between the bHLH and PAS B of Dmgce and more primitive holometabolous Met-like genes, while the PAS A and PAC domains share higher sequence identity with DmMet.These domains may confer a discriminating Met-like function that may partially underlie the functional divergence of Met and gce in higher Diptera.

Dmgce is a vital gene
Overepxression studies demonstrated that gce can substitute for Met in a tissue specific manner to rescue several preadult Met mutant phenotypes.Hence, our results empirically support the notion of functional redundancy between Met and its paralog gce.To further explore the relationship between Met and gce i n J H s i g n a l i n g , w e c a r r i e d o u t underexpression studies in Met + and Met mutant backgrounds by driving the expression of a gce RNAi construct.First, we examined the consequence of gce deficiency in a Met mutant background under the justification that, if gce is responsible for Met 27 viability, then concomitant reduction of Met and gce could result in lethality.Interestingly, Met 27 ; UAS-gce-dsRNA/tubulin-GAL4 flies died as early pupae (0-2 days), whereas expression of the dsRNA construct from an actin-GAL4 promoter caused lethality in the pharate adult stage.Next, we assessed the effects of gce reduction in Met + flies.Surprisingly, Met + ; UAS-gce-dsRNA/tubulin-GAL4 flies died as pharate adults, indicating that even in the presence of functional MET, gce is a vital gene.Driving the transgene from an actin-GAL4 promoter allowed some degree of adult survival, but these adults were clearly affected by insufficient gce, dying within two to three days.Differential intensity of transgene expression from actin and tubulin promoters was previously reported in our lab (Barry et al., 2008).gce underexpression had no observable effect on embryonic development, a stage during which no role for JH has been demonstrated.We have shown that gce transcription begins after about eight hours in early embryos, in contrast to Met, which is supplied as a maternal message (Baumann et al., 2010a).The importance of such divergence in temporal expression profiles is unclear.

Evolutionary conservation of JH signaling mechanisms
Numerous JH target genes have been identified throughout Holometabola.Importantly, many of these genes are known components of the early 20E response.The majority of the work done in our lab has been carried out on D. melanogaster, in which DmMet clearly plays a role in JH signaling: its absence both interferes with methoprene toxicity (Wilson & Fabian, 1986) and hinders JH-driven reproductive physiology (Wilson, 1992;Wilson et al., 2003).However, Met involvement in metamorphosis has been difficult to demonstrate in Drosophila (Riddiford, 2008).As previously stated, JH exposure has no effect on dipteran entry into metamorphosis, unlike other insects (Williams, 1961;Zhou & Riddiford, 2002).In recent years, researchers have turned to the model coleopteran, T.
castaneum.These beetles are both amenable to genetic manipulation and gene knockdown owing to the dramatic effects of systemic RNAi, and the larvae of this species are very sensitive to JH, unlike D. melanogaster larvae.Exposure to JH or a number of its chemical analogs precipitates supernumerary larval instars, similar to the effects of JH on the model lepidopteran, Manduca sexta (Parthasarathy & Palli, 2009).T. castaneum, like mosquitoes, has a single Met-like gene.In their seminal paper, Konopova and Jindra (2007) demonstrated that RNAi-mediated knockdown of TcMet results not only in a methoprene resistance phenotype, but also in the precocious metamorphosis of early instar larvae.A long soughtafter result, the genetic reduction of TcMet provided the phenotype frustratingly absent in D. melanogaster: metamorphic disruption.Reproductive roles for TcMet have also been shown; TcMet knockdown results in a substantial decrease in vitellogenin transcription, (Parthasarathy, et al., 2010) consistent with Met deficiency in D. melanogaster females (Wilson & Ashok, 1998).These results demonstrate that the single Met-like genes in primitive Holometabola function in both development (metamorphosis) and reproduction.Further functional characterization of TcMet (and the single Met-like gene of lower Diptera) could lead to a better understanding of how DmMet has apparently co-opted reproductive functional roles from a gce-like ancestor in higher Diptera

JH regulation of the E-20 transcriptional cascade
The molecular networks that link JH and 20E signaling pathways form the foundation of multiple aspects of insect physiology, as evidenced by the criticality of both hormones in development, reproduction, and diapause (Zhou & Riddiford, 2002;Soller et al., 1999;Denlinger, 1985).Broad Complex (Broad or BR-C) is an early gene in the 20E cascade that encodes a family of alternatively spliced zinc finger transcription factors (four in D. melanogaster, Z1-Z4) fused to a common core protein.Certain Broad alleles phenocopy the morphogenetic defects incurred by methoprene exposure in D. melanogaster.Wilson et al (2006) showed phenotypic synergism in Met and broad double mutants, demonstrating JHsensitive MET and BROAD interaction (BROAD protein accumulation is comparable to that of wild type flies, suggesting physical interaction with, rather than transcriptional regulation by Met), and providing a link between JH and 20E signaling (Wilson et al., 2006).
In a hemimetabolous insect, Oncopeltus fasciatus, continuous Broad expression directs progressive development through nymphal instars (Erezyilmaz et al., 2006).In Holometabola, Broad expression is confined to the prepupal stage, acting as a pupal specifier (Zhou & Riddiford, 2002).Loss of Broad expression, characteristic of the npr1 mutant (nonpupariating; a deletion of the entire complementation group), results in the namesake phenotype of failure to enter the pupal program.Consequently, a restriction of Broad expression during this developmental stage may have contributed to the evolution of complete metamorphosis.During larval development in D. melanogaster, JH represses broad.At pupariation, exogenous JH induces a second wave of broad expression in the abdominal epidermis, resulting in the deposition of a second pupal cuticle (Zhou & Riddiford, 2002), demonstrating that the networks underlying these signaling mechanisms are complex.In T. castaneum, methoprene exposure induces Broad expression and this upregulation is ablated upon TcMet knockdown.Therefore, TcMet is upstream of Broad in JH signaling in these beetles (Konopova & Jindra, 2008).Krüppel homolog 1 (Kr-h1) is upstream of Broad in D. melanogaster JH signaling, where its expression in abdominal epidermis produces sternal bristle disruption similar to that seen following low dose JHA exposure (Minakuchi et al., 2008).Genetic suppression of TcKr-h1 induces precocious metamorphosis, similar to TcMet deficiency; TcMet knockdown in combination with JHA treatment demonstrated that TcKr-h1 exists downstream of TcMet and upstream of TcBroad (Minakuchi et al., 2009).Similarly, Kr-h1 upregulation in newly eclosed A. aegypti females depends on AaMet expression (Zhu et al., 2010).Therefore, the relationships among these genes are generally conserved within holometabolan evolution.
While Kr-h1 also has demonstrated roles in JH-influenced social behavior of honeybees, it has not been reported whether AmKr-h1 is under the transcriptional control of an A. mellifera Met-like protein.However, there is evidence for conservation in the sets of genes regulated by JH between flies and bees.This is perhaps unsurprising given the deep evolutionary conservation of these genetic mechanisms; Kr-h1 and Broad expression profiles in two species of hemimetabolous thrips, whose life histories involve pupa-like, quiescent or nonfeeding stages, are compatible with the expression profiles of Broad and Kr-h1 in holometabolous insects (Minakuchi et al., 2011).Microarray data from D. melanogaster and A. mellifera identified a subset of conserved, JHinducible genes (Li et al., 2007).In the promoter region of 16 of the D. melanogaster orthologs, a conserved JH response element (JHRE) was identified.RNAi-driven reduction of the expression of two proteins identified as JHRE binders, FKBP39 and Chd64, inhibits JHIIIinduced expression of a reporter construct, suggesting their involvement in JH-dependent transcriptional machinery.Bitra and Palli (2009) demonstrated physical interaction of MET with both ECR and USP.Furthermore, column pulldown assays showed FKBP39 and CHD64 as binding partners of D. melanogaster E C R , U S P , a n d MET, providing a more robust framework for a protein complex involving constituents of both JH and 20E signaling pathways (Li et al., 2007).FKBP39, which is present at the onset of metamorphosis (Riddiford, 2008), is an inhibitor of autophagy in D. melanogaster; FKBP39 overexpression precludes the developmental autolysis of larval fat body cells in wandering third instar larvae (Juhász et al., 2007), a physiology shown to be partially dependent on MET/GCE regulation of caspase gene expression (Liu et al., 2009).A role for GCE in any of these protein complexes has yet to be reported.Chd64 is expressed during larval molts, but not in the third instar or during metamorphosis (Riddiford, 2008).Accordingly, putative regulatory complexes consisting of different combinations of these elements may assemble in a stage-or tissue-specific manner.Assembly of differential protein complexes in response to JH, 20E, or both could be a strategy for the tight regulation of the activities of these counteracting hormones.
The Met-like genes of Tribolium and Drosophila appear to act in similar genetic environments to regulate the expression of members of the 20E induced transcriptional cascade, including EcR (Riddiford et al., 2010) and USP (Xu et al., 2010), the heterodimeric components of the ecdysone receptor, various orphan nuclear receptors involved in 20E activity, and 20E-induced caspase genes involved in PCD (Liu et al., 2009).Knockdown of seven nuclear receptors (E75, HR3, EcR, USP, SVP, FTZ-F1, and HR4) results in a significant reduction of vitellogenin prouction in T. castaneum (Xu et al., 2010), a phenotype similar to that obtained via TcMet knockdown (Parthasarathy & Palli, 2009).The data presented in this section therefore strongly support for the action of Met-like genes as crucial to 20E/JH crosstalk.

Discovery of an evolutionarily conserved Met binding partner
Recent biochemical data from A. aegypti indicate that AaMet binds another bHLH PAS gene, AaFISC, and that this interaction requires a high JH titer.FISC is a coactivator of EcR/USP (Chen, J.D., 2000;Zhu, et al., 2006), providing yet another link between 20E and JH signaling.
The authors also report that coexpression of DmMet or Dmgce with DmTaiman (the D. melanogaster AaFISC ortholog) in the presence of JH III induced reporter gene expression in L57 cells (Li et al., 2011).Furthermore, this interaction has also been demonstrated in T. castaneum between TcMet and the FISC/TAI homolog, TcSRC (Steroid receptor coactivator; Zhang et al., 2011).This gene has previously been implicated in metamorphic activity.T. castaneum larvae treated with SRC RNAi fail to achieve critical weight and consequently die before the larval-pupal transition (Bitra et al., 2009).Therefore, MET interaction with FISC/SRC/TAIMAN underpins key transcriptional events of JH signaling holometabolous insects.Structure-function analyses performed using site-directed mutagenesis identified regions of MET that are necessary for homodimerization and GCE binding.Point mutations in the bHLH and PAS A domains (Met 1 and Met w3 alleles, respectively) had no effect on partner binding, whereas N-and C-terminal truncations, deletions in the HLH or PAS A domains, and a point mutation in the PAS B domain (Met 128 allele) all inhibited dimerization (Godlewski, et al., 2006).Structure-function data for AaMET and AaFISC binding illustrates that the criticality of PAS domains for protein-protein interaction.Interestingly, two-hybrid assays showed that MET/FISC interaction increased when AaMET lacked a bHLH domain (Li et al., 2011).Therefore, this domain is unnecessary for MET-FISC interaction, suggesting that the sole function of the AaMet bHLH may be in DNA binding.In contrast, deletion of the bHLH domain in FISC hindered the JH-induced interaction with AaMET.Mixespression of DmTaiman in a variety of Met/gce genetic backgrounds will be valuable from both physiological and evolutionary perspectives.How do these proteins interact in the context of hormonal control of D. melanogaster development?Presumably, during larval development JH secreted from the CA inhibits MET and GCE interaction while promoting MET and GCE binding with TAIMAN.Are MET:TAI and GCE:TAI dimers functionally congruent in D. melanogaster or do these complexes preferentially regulate disparate target genes?Is the Met-like gene in A. aegypti and T. castaneum functionally analogous to Met/gce or are other, unidentified proteins involved?How has the interaction of these proteins changed during dipteran evolution following the origin of Met?

Conclusions
RT-PCR analysis with degenerate primers identified a single Met-like homolog in the genome of each of the three mosquito species, Aedes aegypti, Anopheles gambiae, and Culex pipiens.Likewise, a single Met-like ortholog exists in the beetle, T. castaneum, (Konopva & Jindra, 2007).Phylogenetic analysis and comparison of intron number and position in each of the identified mosquito genes indicates that the mosquito Met orthologs share higher sequence identity with Dmgce than DmMet, suggesting that DmMet arose from the duplication of an ancestral, gce-like gene in lower Diptera.To examine the evolutionary history of Met and gce within the Diptera, we mined the public G. morsitans EST library, recovering unique putative Met and gce orthologs in this fly, showing conservation of a Met homolog within the Schizophora.We also recently isolated a putative gce homolog from a Bombyliid, Bombylius major (A.A.B., unpublished).Taxonomically, this group of flies exists in the Asilomorpha, a paraphyletic sister taxon to the Muscomorpha within the dipteran infraorder, Brachycera.While this study is in its preliminary stages, only a single Met-like gene has thus far been obtained from this fly using degenerate PCR with cDNA and genomic DNA templates, suggesting the possibility that the Met/gce duplication occurred within the Brachycera.Met function is evolutionarily conserved in Diptera; consistent with independent reports (Zhu et al., 2010) we observed that RNAi-driven reduction of AaMet results in concomitant reduction of JH-inducible genes (Figure 4).RT-PCR analysis of selected D. melanogaster tissues shows that gce is generally co-expressed with Met in known JH target tissues, including ovary and MAG.Overexpression of gce in a Met mutant background results in a dramatic enhancement of methoprene-conditional toxic and morphogenetic defects, similar to those seen in wild type (Met + ) flies after methoprene exposure.Met mutant flies overexpressing gce show rescue of a non-conditional adult phenotype, that of defective development of posterior facets in the compound eye.Our results therefore support the notion of functional redundancy that has been hypothesized to account for Met 27 viability flies.On the other hand, we have also shown that these paralogs have undergone evolutionary subfunctionalization since their origin; gce overexpression fails to rescue the phenotypes of deficient oogenesis or reduced male courtship characteristic of Met adults, showing that Met has co-opted therole as the mediator of JH-regulated reproductive functions in Drosophila.
RNAi-driven reduction of gce expression from either an actin or tubulin promoter demonstrates that unlike Met, gce is a vital gene.Gce underexpression in both Met + and Met genetic backgrounds results in lethality.Met + ; UAS-gce-dsRNA / tubulin-GAL4 homo-/hemizygotes do not survive to adulthood, and die primarily at the pharate adult stage, while the same gce RNAi construct expressed in a Met mutant background shifts lethality to early pupae (Baumann et al., 2010a).

Directions
Previously, USP was proposed as a candidate JH receptor (Jones, et al., 2001).Yet, USP only binds JH with micromolar affinity, requiring a hormone concentration that exceeds endogenous titers by orders of magnitude (Bownes & Rembold, 1987).It is now known that USP binds methyl farnesoate (MF), a precursor in the biological synthesis of JH III (Figure 1), with nanomolar affinity both in D. melanogaster and in A. aegypti (Jones et al., 2006;Jones et al., 2010).Recent studies on natural farnesoid derivatives including MF, JH III, and JHB3 (the main farnesoid secretion product of dipteran ring glands cultured in vitro) have teased out the relative activities of each of these compounds during development in a series of biological assays.Two recent studies have demonstrated that the activity series of these three compounds changes during development.Dietary MF and JH III (MF > JH III) were both more active than JHB3 in delaying larval attainment of the wandering stage.In contrast, JH III applied to prepupae (white puparial assay; Riddiford & Ashburner, 1991) showed much higher activity than MF or JHB3 in blocking adult eclosion (Jones et al., 2010;Harshman et al., 2010).Topical application or dietary provision of these compounds adds to endogenous hormone titers.Therefore, just as USP binds JH III at concentrations exceeding physiological levels, it is possible that MET nonspecifically binds MF or JHB3 under these conditions.It is an intriguing proposition that MET and USP, which interact both with each other and JHRE binding proteins (Bitra & Palli, 2009), may partner in a stage-specific manner throughout development in response to a fluctuating mélange of methyl farnesoids.Does GCE participate in the assembly of the molecular machinery that facilitates the crosstalk between JH and 20E signaling?This protein has been largely ignored in studies regarding the molecular interaction of these hormones.Further, it is unknown whether GCE binds any of the farnesoid products of the CA.There appears to be a correlation between the presence of paralogous Met-like genes and multiple JH isoforms in higher Diptera.If each of the farnesoids JH III, MF, and JHB3 indeed has a unique receptor protein, the possibility arises that GCE fills the role of JHB3 binder.Or perhaps in MET/GCE dimers, MET is the sole ligand binder, while GCE and MET are both necessary for target gene transcription.Clearly, further functional characterization of GCE is necessary to unravel the mechanisms through which JH signaling has evolved from the basal holometabola to the most evolutionarily diverged insects, the higher Diptera.

Acknowledgments
We extend thanks to Drs.John Freudenstein and H. Lisle Gibbs for their guidance in our various evolutionary analyses.Additionally, we would like to recognize Dr. Shaoli Wang for her expertise and contributions to much of the work presented herein, and for her mentorship of AAB.Work presented in this text was supported by NIH grant AI052290 to T.G.W.

Fig. 2 .
Fig. 2. A gene tree of some holometabolous Met-like genes, showing placement of two distinct G. morsitans sequences as putative Met and gce orthologs.D. melanogaster Tango (tgo), the homolog of the vertebrate Aryl hydrocarbon receptor (Ahr), is used as an outgroup sequence.

Fig. 3 .
Fig. 3. Left: malformed facets in the posterior compound eye of Met w3 flies appear dark under light microscopy.Right: EMS-induced production of an unidentified enhancer gene dramatically intensifies the Met w3 phenotype (T.G.W., unpublished).

Fig. 4 .
Fig. 4. Expression of three JH-inducible genes following RNAi-induced knockdown of AaMet (dsMet: orange bars) vs. controls (dsB-gal: red bars).AaMet reduction produced concomitant suppression of the A. aegypti homologs of DmJHE, DmJhI-1, and DmJhI-26.Analysis of the nonsynonymous-to-synonymous substitution ratios (dN/dS) of Met and gce orthologs within the genus Drosophila indicates a substantial relaxation of selective constraint on the C-terminal half of gce, downstream of the functional domains.Conversely, nonsynonymous substitutions in the N-terminal half are stringently selected against.Depressed dN/dS values across the Met coding sequence indicate strong selective constraint over the entire open reading frame(Baumann et al., 2010b).RT-PCR analysis of selected D. melanogaster tissues shows that gce is generally co-expressed with Met in known JH target tissues, including ovary and MAG.Overexpression of gce in a Met mutant background results in a dramatic enhancement of methoprene-conditional toxic and morphogenetic defects, similar to those seen in wild type (Met + ) flies after methoprene exposure.Met mutant flies overexpressing gce show rescue of a non-conditional adult phenotype, that of defective development of posterior facets in the compound eye.Our results therefore support the notion of functional redundancy that has been hypothesized to account for Met 27 viability flies.On the other hand, we have also shown that these paralogs have undergone evolutionary subfunctionalization since their origin; gce overexpression

Table 1 .
Table 1 lists some representative JH-inducible genes.Representative JH-inducible genes.Many of these genes have evolutionarily conserved roles in JH signaling in holometabolous insects.In addition, several are known components of the 20E transcriptional cascade.