BIOLOGICAL OVERVIEW

The transformation from larval to adult is one of the miracles of insect biology. This transformation is orchestrated by hormone secretion and directed by members of the hormone receptor superfamily. Usp acts as the principle dimerization partner for these hormone activated transcription factors. The steroid hormone 20-hydroxyecdysone (20E) initiates metamorphosis in insects by signaling through the ecdysone receptor complex, a heterodimer of the Ecdysone receptor (EcR) and Ultraspiracle.

Usp belongs to the steroid hormone receptor superfamily. Such receptors are also transcription factors, binding hormone ligands and acting together with others in the superfamily to activate transcripts of hormone responsive genes. Usp is a promiscuous factor, meaning it can dimerize with multiple partners. It is homologous to the vertebrate retinoid-X receptor, which demonstrates the ancient lineage of the nuclear receptor superfamily. Like Usp, the RXR functions in numerous hormone driven developmental pathways as a partner with other hormone receptors.

Usp function is best understood by looking at its partners and its targets. Partnering the Ecdysone receptor for the molting hormone in insects, Usp activates genes involved in molting. Another target is the larval serum protein-2 gene whose product accumulates in the fat body tissue. This implicates Usp in the functioning of the fat body in larval flies (Antoniewski, 1994). The fat body serves as an energy store for high levels of activity and for reproduction.

Seven-up acts to inhibit Ultraspiracle signaling involved in photoreceptor determination. Ultraspiracle is required for expression of an R7 photoreceptor-specific opsin, for example. seven-up, a target of the sevenless/ras/raf pathway, prevents the R7 photoreceptor fate in photoreceptor precursor cells R1, R3, R4 and R6., acting to inhibit ultraspiracle-based signaling pathways. seven-up can function either by DNA binding competition or protein-protein interactions (Zelhof, 1995).

Usp's third partner is hormone receptor 38 (DHR38), a homolog of rat nerve growth factor induced protein B. In the fly, DHR38 expression causes a dramatic drop in the ecdysone-induced stimulation of ecdysone responsive genes. The DHR38-Usp interaction fine tunes the response to ecdysone (Sutherland, 1995).

Usp finctions in abdominal cuticle synthesis during midembryogenesis and larval cuticle molting. Adult thoracic and abdominal metamorphosis can occur in the absence of Usp, indicating that these responses are either not regulated by ecdysone or are dependent on the activity of another nuclear receptor (Thummel, 1995).

Juvenile hormones (JH) bind to the nuclear receptor Ultraspiracle. JH are a sesquiterpenoid group of ligands that regulate developmental transitions ininsects. In fluorescence-based binding assays, Usp protein binds JH III and JH III acid with specificity, adopting for each ligand a different final conformational state. JH III treatment of Saccharomyces cerevisiae expressing a LexA-Usp fusion protein stabilizes an oligomeric association containing this protein, as detected by formation of a protein-DNA complex, and induces Usp-dependent transcription in a reporter assay. Juvenile hormone acid induces a different Ultraspiracle conformation than does binding of Juvenile hormone ester. The results strongly support the inference that JH III promotes at least homodimerization of Usp. It is proposed that regulation of morphogenetic transitions in invertebrates involves binding of JH or JH-like structures to Usp. The demonstration that JH ester and JH acid each induce different conformational states to Usp raises the possibility that these two different conformational states may confer different transcriptional activities to Usp (Jones, 1997).

Analysis of usp mutant clones in the wing disc of Drosophila shows that in the absence of Usp, early hormone responsive genes such as EcR, DHR3 and E75B fail to up-regulate in response to 20E, but other genes that are normally expressed later, such as beta-Ftz-F1 and the Z1 isoform of the Broad-Complex (BRC-Z1), are expressed precociously. Sensory neuron formation and axonal outgrowth, two early metamorphic events, also occur prematurely. In vitro experiments with cultured wing discs show that BRC-Z1 expression and early metamorphic development are rendered steroid-independent in the usp mutant clones. These results are consistent with a model in which these latter processes are induced by a signal arising during the middle of the last larval stage but suppressed by the unliganded EcR/Usp complex. These observations suggest that silencing by the unliganded EcR/Usp receptor and the subsequent release of silencing by moderate steroid levels may play an important role in coordinating early phases of steroid driven development (Schubiger, 2000).

In clones lacking Usp, BRC-Z1 is still developmentally regulated but it first appears about 103 hours AED rather than at its expected time of expression at mid-wandering. This precocious appearance of BRC-Z1 corresponds to the so called 'mid-third instar transition', which is a time when changes occur in preparation for metamorphosis. It has been proposed that the gene DHR78 may play a role in this transition and that these changes may occur independently of the 20E titers. It appears that events at this time serve to activate BRC-Z1 expression but a Usp mediated mechanism suppresses expression until rising ecdysteroid titers remove the inhibition late in wandering (Schubiger, 2000).

The development of sensory precursor cells in the wing shows a similar response to the loss of Usp. As sensory neurons differentiate and project their axons toward the CNS, the timing of these events is crucial. In order for axons to navigate centripetally, guidance cues must be present at the correct place and time. The synchrony of these events, as has been shown for the developing neurons along the presumptive wing margin, is disrupted in usp mutant clones. Differentiation and axon outgrowth occur precociously, and, as a consequence, the axons originating from these neurons are misrouted. Premature differentiation and faulty axon projection for the presumptive campaniform sensilla on the radius of the wing has been seen, and precocious differentiation of photoreceptors in the eye disc. Thus, Usp appears to play a general role of suppressing early sensory neuron differentiation in the imaginal tissue (Schubiger, 2000).

Both the expression of BRC-Z1 and the differentiation of sensory neurons are dependent on exposure to ecdysteroids. Surprisingly, though, the presumed disruption of the ecdysone receptor complex by removal of Usp allows these processes to proceed in a hormone-independent fashion. Observations on the developing eye disc show that 20E is required for the correct progression of the morphogenetic furrow, but that loss of Usp leads to advancement of the morphogenetic furrow and precocious differentiation of the photoreceptors. It is assumed that in the case of the eye, the events in the morphogenetic furrow are also rendered hormone-independent by the removal of Usp. The data suggest that Usp is involved in two distinct types of ecdysteroid controlled responses and that these responses have distinct developmental roles. In some instances Usp serves as a hormone-inhibited silencer whereas in others it is a hormone-dependent activator. Early metamorphic events in the wing, including neurogenesis and axonal outgrowth, clearly require ecdysteroids in order to occur, but this requirement is carried out through an ecdysteroid-dependent release of Usp-mediated suppression. Thus, in the absence of Usp these events occur in a steroid-independent fashion. Importantly, the rate of development in usp mutant clones is at least as fast or faster than in wild-type tissue exposed to 20E, suggesting that for these developmental processes the effects of 20E are at best permissive (Schubiger, 2000).

A molecular parallel to what is seen for early sensory neuron development is illustrated for the expression of BRC-Z1. Even though BRC-Z1 is expressed in the neurogenic regions of the disc, studies with BRC-Z1 mutants show that altered expression of BRC-Z1 does not interfere with the differentiation of the wing or its sensory neurons. Nevertheless the precocious expression of BRC-Z1 in cells lacking Usp function provides insight into what would be expected for the genes directly involved in neuronal birth and differentiation. BRC-Z1 expression appears to be activated by events during the mid-third instar, but it is suppressed via Usp until the titer of 20E is high enough to remove this Usp-mediated silencing. In this context the presence of the hormone is permissive in that it allows other factors (induced by the mid-third instar transition?) to take control of gene expression. This de-repression contrasts with the other class of ecdysteroid-dependent events, such as the up-regulation of the early response genes (EcR, DHR3, E75B). EcRA, for example, is not up-regulated in the absence of Usp, with or without hormone (Schubiger, 2000).

GENE STRUCTURE

Bases in 5' UTR - 198

Exons - one

Bases in 3' UTR - 164

PROTEIN STRUCTURE

Amino Acids 507

Structural Domains

There are two zinc finger domains in Usp. Toward the N-terminal end there is a transactivation domain. The hormone binding domain is located toward the C-terminal end (Henrich, 1990).

Ultraspiracle (USP) is the invertebrate homolog of the mammalian retinoid X receptor (RXR). RXR plays a uniquely important role in differentiation, development, and homeostasis through its ability to serve as a heterodimeric partner to many other nuclear receptors. RXR is able to influence the activity of its partner receptors through the action of the ligand 9-cis retinoic acid. In contrast to RXR, USP has no known high-affinity ligand and is thought to be a silent component in the heterodimeric complex with partner receptors such as the ecdysone receptor. The 2.4-Å crystal structure of the USP ligand-binding domain is reported in this study. The structure shows that a conserved sequence motif found in dipteran and lepidopteran USPs, but not in mammalian RXRs, serves to lock USP in an inactive conformation. It also shows that USP has a large hydrophobic cavity, implying that there is almost certainly a natural ligand for USP. This cavity is larger than that seen previously for most other nuclear receptors. Intriguingly, this cavity has partial occupancy by a bound lipid, which is likely to resemble the natural ligand for USP (Clayton, 2001).

Two regions of the USP structure differ significantly from both the hRXRalpha and holo hERalpha structures: the (1) loop joining helices 1 and 3 and (2) helix 12. This H1-H3 loop adopts an unusual position in USP not seen in any previous nuclear receptor structure. It wraps over the top of helix 3 and lies between helices 3 and 11. In most of the nuclear receptor structures (e.g., apo-hRXRalpha, holo-H. sapiens peroxisome proliferator-activated receptor gamma, and holo-H. sapiens thyroid receptor beta), this loop passes outside the beta-strand. An exception is the hERalpha structure in which the loop passes between the beta-strand and helix 3. The position of this loop in USP is significant because it prevents helix 12 from adopting the position seen in either the unliganded or 9-cis retinoic acid (9cRA)-bound structures of hRXRalpha. As a consequence, helix 12 adopts a position that closely resembles that seen in the structure of estrogen receptor bound to an antagonist. However, in contrast to estrogen receptor, helix 12 is locked firmly in this inactive position by making extensive contacts to the H1-H3 loop (Clayton, 2001).

Sequence comparison of the various USP proteins reveals a sequence motif within the H1-H3 loop that is conserved in both diptera (flies) and lepidoptera (moths) USP proteins but not in hymenoptera (honey bee, GenBank accession no. AF263459), orthoptera (migratory locust), or other arthropods such as crab and tick. The structure shows that every residue within this conserved motif plays an architectural role in mediating interactions with helices 3, 4/5, 11, and 12 as well as the loop between helices 11 and 12. This sequence conservation and extensive pattern of interactions clearly indicate that the position of the H1-H3 loop and of helix 12 are bona fide features of the structure that are not the result of crystal-packing interactions. This finding is supported further by the evidence that all six copies of the USP in the crystal have identical conformations in this respect (Clayton, 2001).

At first sight, it would seem unlikely that the lipid bound to the bacterially expressed USP is the authentic ligand for USP, especially because it seems that the USP is bound to a mixture of different lipids. However, the hydrophobic cavity revealed in the structure implies that the USP is very likely to have a natural ligand in Drosophila. This implication raises a question: what might this ligand be? Importantly, there is only one polar side chain, Q288, that is in a position to interact with the ligand and suggests that the ligand would need to be largely nonpolar in character. It has been suggested that JH might serve as a ligand for USP. Although this may make biological sense, such a small ligand would leave most of the cavity empty and therefore does not seem to be the most likely possibility. Recently it has been reported that RXR is activated in vivo by docosahexaenoic acid (a long chain fatty acid). Furthermore, a mutant mouse RXR also has been observed to bind oleic acid and is presumed to be activated by this ligand. Could the natural USP ligand be a fatty acid? USP does not have the basic arginine needed to interact with the carboxyl group, and therefore a fatty acid ligand for USP is perhaps not so likely. On balance, it seems that a diacylglycerol-based ligand is well suited to binding in the cavity seen in the USP structure. It is striking that the protein has retained the ligand through multiple chromatographic purifications, suggesting that the dissociation constant, or at least the off rate, must be very low (Clayton, 2001).

The observation that the crystal lattice contains three LBDs with significantly lower ligand occupancy, which does not correlate with any changes in the conformation of helix 12 or the H1-H3 loop, strongly suggests that lipid binding has little effect on the overall structure of the USP LBD. Of course it is not possible to exclude the possibility that a cognate ligand might cause significant structural perturbation. However, the large number of backbone-backbone and side-chain-side-chain contacts between H12, H1-H3, and H3 would not be consistent with a substantial rearrangement after binding any ligand (Clayton, 2001).

Because it seems that the USP is locked into an antagonized conformation irrespective of bound ligand, a role for a USP ligand is not immediately obvious. The consensus view is that nuclear receptor ligands cause dissociation of any bound corepressor and facilitate binding of LxxLL-containing coactivators. This effect is achieved by modulating the position and dynamics of helix 12. Clearly this mechanism cannot operate in USP, because not only would the position of helix 12 prevent binding of LxxLL-containing coactivators but also because ligand binding does not influence the position of helix 12. These conclusions are consistent with the fact that no direct-activation activity of USP has been reported so far and that indeed coactivators of the p160 family have not been identified in Drosophila (Clayton, 2001).

There are however several possible mechanisms through which a USP ligand might regulate transcription independently of helix 12 and p160 coactivators. Such mechanisms require that a cofactor protein recognize a region on the surface of USP that is different from the canonical LxxLL-binding pocket. If this surface was to involve the exposed portions of the ligand or a region whose stability is influenced by ligand, then the surface could serve readily to distinguish liganded from unliganded receptor. In principle, such a cofactor could be one of a range of proteins including different coactivators or corepressors, components of the general transcriptional machinery, or even a partner receptor. It is even possible that ligand binding influences the half-life of the receptor. Indeed, there is precedent for receptors interacting with the basal transcriptional machinery; for ligand binding influencing receptor half-life and for communication between heterodimeric partners. In particular it has been shown that the ecdysone receptor absolutely requires its USP partner for activity. Understanding whether and how USP ligands regulate transcription remains to be established (Clayton, 2001).

Evidently this structure does not provide all of the answers concerning the mechanism of USP action. However, it seems clear that USP almost certainly must have a ligand, but the mode of action of that ligand may well differ from the canonical model. The structure also gives a remarkable view of the mechanisms through which protein structure and function can be changed through evolution (Clayton, 2001).

The crystal structure of the USP LBD from the lepidoptera Heliothis virecens has also been reported (Billas, 2001). This structure shows many of the same features as the structure described in this study, including the antagonist conformation of helix 12 and the bound lipid. However, in the H. virecens crystals, there is only one LBD in the asymmetric unit and it is fully occupied by lipid. Accordingly, it is suggested that the antagonist conformation may be a consequence of the bound lipid (Clayton, 2001).

date revised: 25 October 2023  

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.