Locust retinoid X receptors: 9-Cis-retinoic acid in embryos from a primitive

Shaun M. Nowickyj*, James V. Chithalen†, Don Cameron†, Michael G. Tyshenko*, Martin Petkovich†‡, Gerard R. Wyatt*, Glenville Jones†§, and Virginia K. Walker*¶

Departments of *Biology, †Biochemistry, ‡Pathology, and §Medicine, Queen’s University, Kingston, ON, Canada K7L 3N6

Edited by Walter S. Leal, University of California, Davis, CA, and accepted by the Editorial Board April 10, 2008 (received for review December 21, 2007)

The retinoid X (RXR) is activated by its often elusive cognate high-affinity JH receptor, Methoprene-tolerant (Met; Kd ϭ 5.3 nM) , 9-cis-retinoic acid (9-cis-RA). In flies and moths, molting is has been recently identified in Dm (18). mediated by a heterodimer receptor consisting of the Remarkably, the LBD of the corresponding to USP from ecdysone monomer (EcR) and an RXR homolog, ultraspiracle (USP); the primitive insect, Locusta migratoria (Lm), shows greater identity the latter is believed to have diverged from its RXR origin. In the more to the vertebrate RXR than to USPs of more advanced (19). primitive insect, Locusta migratoria (Lm), RXR is more similar to In silico modeling of the LmRXR-L ligand binding pocket also human RXRs than to USPs. LmRXR was detected in early embryos emphasizes amino acid and tertiary structural similarity to the when EcR transcripts were absent, suggesting another role apart from human RXR␥ (hRXR␥) (19). In Lm and the German cockroach, ecdysone signaling. Recombinant LmRXRs bound 9-cis-RA and all- Blattella germanica, RXR transcripts have been detected during ؍ ؍ trans-RA with high affinity (IC50 61.2–107.7 nM; Kd 3 nM), similar early embryonic development, even before the appearance of EcR to human RXR. To determine whether specific binding had functional transcripts (20, 21). Thus at least in these two insects it is possible significance, the presence of endogenous retinoids was assessed. that this EcR binding partner could have a second function. To Embryos were extracted by using modified Bligh and Dyer and explore this possibility, two RXR cDNAs were isolated from solid-phase protocols to avoid the oily precipitate that makes this Locusta. Long and short isoforms, LmRXR-L (GenBank accession material unsuitable for assay. These extracts contained retinoids (5.4 no. AY348873) and LmRXR-S (GenBank accession no. nM) as assessed by RA-inducible Cyp26A1- luciferase re- AF136372), respectively, differ only by the presence/absence of 22 porter lines. Furthermore, the use of HPLC and MS confirmed the aa in their LBDs. presence of retinoids and identified in any embryo, 9-cis-RA, in addition to all-trans-RA. We estimate that whole embryos contain 3 Results nM RA, including 9-cis-RA at a concentration of 1.6 nM. These findings Recombinant LmRXR-S was expressed by using the phage T5 strongly argue for a functional role for retinoids in primitive insects promoter in pQE-32 after transfer to Escherichia coli M15[pREP4] and favor a model where signaling through the binding of 9-cis-RA to cells. Under similar conditions, LmRXR-L expression was disap- its RXR is established relatively early in evolution and embryonic development. pointing (results not shown). Therefore the hinge and ligand binding domains of LmRXR-L were subcloned into pET-15b all-trans-retinoic acid ͉ Locusta migratoria ͉ ultraspiracle [LmRXR-L(DE)] and expressed in BL21(DE3) cells by using the T7lac promoter. As a control for this shorter (Ϸ28 kDa) sequence, the Ϸ29-kDa human-derived sequence, hRXR␣(DE), was similarly nsect development and metamorphosis are directed by two expressed. Typically, the purification protocol (a total of 10 puri- Iprincipal lipophilic : 20-hydroxyecdysone (20-OH-Ec), fications were done for the long and short isoforms) yielded 0.8 the active molting , and juvenile hormone (JH), whose mg/ml LmRXR-S and 1.7 mg/ml LmRXR-L(DE) protein [see titer determines the nature of the molt (1, 2). As demonstrated in supporting information (SI) Fig. S1 A and B; lanes 7 and 8). the fruitfly, Drosophila melanogaster (Dm), 20-OH-Ec binds to the Expression and purification of hRXR␣(DE) normally yielded 1.0 ecdysone receptor (EcR), which in turn is bound to its obligate mg/ml. These preparations were used for assays and antibody heterodimerization partner ultraspiracle (USP), a homologue of production. the vertebrate (RXR) (3–6). As members of the ␣ superfamily, EcR and USP/RXR share a common Polyclonal antibodies made against purified hRXR (DE) and modular structure (7) comprised of a N-terminal variable domain LmRXR-L(DE) were used for immunological detection by West- (A/B), a DNA binding domain (C), hinge (D), and C-terminal ern blotting (see Fig. S1 C—E). The high similarity between the ligand-binding domain (LBD; or domain E/F). expressed human and locust RXR LBDs is demonstrated by the ␣ The vertebrate RXRs are known heterodimeric partners of observation that the hRXR (DE) antibody cross-reacted with several members of the nuclear receptor superfamily, including the purified LmRXR-L(DE) and the LmRXR-L(DE) antibody cross- ␣ retinoid, thyroid, and vitamin D receptors (8). As demonstrated in reacted with the purified hRXR (DE). Western blots further vivo, these RXRs can also form homodimers and conceivably demonstrated the presence of cross-reacting material to the anti- mediate an independent retinoid signaling pathway (9, 10). Indeed, LmRXR-L(DE) in whole embryo extracts (see Fig. S1E). the vertebrate RXRs are known ligand-activated factors that bind 9-cis-retinoic acid (9-cis-RA), a stereoisomer of the Author contributions: S.M.N., J.V.C., G.R.W., G.J., and V.K.W. designed research; S.M.N., vitamin A derivative, all-trans-RA (11, 12). RA receptors (RARs), J.V.C., D.C., M.G.T., G.R.W., and V.K.W. performed research; J.V.C., M.G.T., M.P., G.R.W., reported only in vertebrates, are distinct in that they bind both G.J., and V.K.W. contributed new reagents/analytic tools; S.M.N., J.V.C., D.C., G.R.W., G.J., all-trans-RA and 9-cis-RA with high affinity (13, 6). In contrast to and V.K.W. analyzed data; and S.M.N. and V.K.W. wrote the paper. the vertebrate RXRs, crystal structures reveal that DmUSP and the The authors declare no conflict of interest. USPs from the moth Heliothis virescens and the flour beetle This article is a PNAS Direct Submission. W.S.L. is a guest editor invited by the Editorial Tribolium castaneum probably adopt an inactive conformation Board. (14–16) and are unlikely to have an activating ligand. Nevertheless, ¶To whom correspondence should be addressed. E-mail: [email protected]. Jones et al. (17) have suggested that USP could be a JH receptor This article contains supporting information online at www.pnas.org/cgi/content/full/ because it binds methyl farnesoate, an unepoxidated derivative of 0712132105/DCSupplemental. JH III (Kd ϭ 44 nM) and JH to a lesser extent (Kd ϭ 6700 nM). A © 2008 by The National Academy of Sciences of the USA

9540–9545 ͉ PNAS ͉ July 18, 2008 ͉ vol. 105 ͉ no. 28 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0712132105 Downloaded by guest on September 28, 2021 A 100 B 100 75 0.6 75 50 0.4 50

25 Bound/Free -RA Bound (%)-RA Bound -RA Bound (%)-RA Bound 0.2 cis cis 25 9- 9- 0 0.0 0.0 0.5 1.0 1.5 Bound (fmol/µg) -25 0 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 log [9-cis-RA] (M) log [all- trans-RA] (M)

C 100 D 100

75 75

50 50 -RA Bound (%)-RA Bound -RA Bound (%)-RA Bound cis cis 25 25 9- 9-

0 0 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 log [DHA] (M) log [Methoprene Acid] (M)

Fig. 1. Retinoid binding analysis. Competitive binding of 9-cis-RA (A), all-trans-RA (B), DHA (C), and methoprene acid (D) with [3H]-9-cis-RA to the purified receptors LmRXR-Short (Œ), LmRXR-L(DE) (F), and hRXR␣(DE) (■). (A Inset) A Scatchard analysis of [3H]-9-cis-RA binding to LmRXR-L(DE). Competition binding studies were done three to seven times on a total of two to seven independently prepared recombinant protein purifications.

Ligand Binding Assays. Although control insect protein did not bind respectively, than 9-cis-RA for LmRXR-S (Fig. 1 C and D and to [3H]-9-cis-RA (data not shown), when recombinant LmRXR-S, Table 1). The two nonretinoids were also not as effective compet- 3 LmRXR-L(DE), and hRXR␣(DE) were incubated with [ H]-9- itors for 9-cis-RA binding to LmRXR-L(DE). Overall, the IC50 for cis-RA, there was evidence of binding (Fig. 1). Competition with DHA was similar for both RXR isoforms but the IC50 of metho- unlabeled 9-cis-RA showed that for all three protein preparations, prene acid for LmRXR-S was 30% that of LmRXR-L(DE). binding was specific and very similar. The IC50 for 9-cis-RA by hRXR␣(DE) was 74.2 nM, in reasonable agreement with a previ- Retinoids in Locust Embryos. The analysis of early locust embryos ously published value (5.6 nM) (22). LmRXR-S showed 1.8 times was difficult because of the presence of yolk and a tough chorion or the affinity (IC50 ϭ 61.2 nM) for 9-cis-RA than did LmRXR-L(DE) egg shell. A procedure was developed in several initial trials by (Table 1). When the unlabeled stereoisomer all-trans-RA was used ‘‘spiking’’ extracts with 500 ng of 9-cis-RA and radio-labeled 3 to displace [3H]-9-cis-RA, LmRXR-S again showed slightly more [ H]-9-cis-RA (500:0.15 ng). Many extraction methods using ethyl acetate and hexane and numerous solid-phase extractions were also affinity (IC50 ϭ 75.0 nM) than LmRXR-L(DE) for the retinoid, but both retinoids had similar overall competition curves (Fig. 1 A and tried. Finally, a modified Bligh and Dyer (23) procedure combined B and Table 1). Scatchard analysis resulted in a calculated K for with tandem solid-phase extraction was selected because it resulted d Ϯ ϭ 3 LmRXR-L(DE) of 3.0 nM (Fig. 1). in 42.6 9.9% (n 3) recovery of the added [ H]-9-cis-RA label Competition binding assays with other potential ligands showed (data not shown). Embryos were extracted twice to optimize lower affinity for the . The affinities of the polyunsaturated retinoid recovery and avoid the problem of precipitation of an fatty acid, docosahexaenoic acid (DHA), and the JH analogue, embryo-derived brown viscous oil during evaporation. Retinoid reporter cells were used to detect the presence of methoprene acid, were two and three orders of magnitude lower, endogenous retinoids in Locusta embryo homogenates and extract preparations (24). These cells proved sensitive to very low concen- Table 1. Competitive binding data trations (0.1 nM) of all-trans-RA and all-trans-RA/9-cis-RA sus- pensions, resulting in luciferase reporter activity that was 2.5- and ␣ Competitor LmRXR-L(DE), nM LmRXR-S, nM hRXR (DE) 2.2-fold greater than DMSO control treatments (Fig. 2). Dilutions 9-cis-RA* 107.7 61.2 74.2 nM of locust embryo homogenates (1:10 dilution) did not induce a all-trans-RA† 104.9 75.0 ND significant reporter response. However, preparations of embryo

Methoprene acid‡ 8,743 26,610 ND extracts (see Experimental Procedures) resulted in significantly BIOCHEMISTRY DHA§ 4,048 2,915 ND greater reporter activity (2.5-fold) than DMSO controls (P Ͻ 0.05). Concentration determination using standard curves generated by a ϭ Mean IC50 values (N 3–7) of purified Locusta RXR isoforms [LmRXR-L(DE) range of 9-cis-RA and all-trans-RA diluted in 100% DMSO indi- ␣ and LmRXR–S] and hRXR (DE) were determined by using retinoid or related cated the presence of retinoids in total embryonic tissue at a molecules to competitively displace [3H]-9-cis-RA, used at 5 nM. Data are ϭ Ϸ means from multiple experiments as shown in Fig. 1. ND, not determined. concentration of 1.6 ng/g (n 4) or 5.4 nM. *Tested at concentrations ranging from 0.046 to 50,000 nM. Because in vitro experiments indicated the presence of retinoids †Tested at concentrations ranging from 0.042 to 42,000 nM. in Locusta embryos, HPLC-MS was used to determine the presence ‡Tested at concentrations ranging from 3.1 to 310,000 nM. of specific RAs in embryo extract. Monitoring the eluant from §Tested at concentrations ranging from 1.3 to 1,300,000 nM. HPLC at 350 nm did not resolve a significant retinoid peak relative

Nowickyj et al. PNAS ͉ July 18, 2008 ͉ vol. 105 ͉ no. 28 ͉ 9541 Downloaded by guest on September 28, 2021 2000 * comigrated with standards for 9-cis-RA and all-trans-RA, respec- * tively (Fig. 3 Lower). These experiments were repeated several * times with isomer standards. The locust extract (3 g) preparation 1500 yielded 1.17 Ϯ 0.27 ng (n ϭ 3) of total retinoid. Diluted locust extracts (1.5 g BSA) yielded similar peaks and retention times (32.8 1000 and 33.3 min, respectively) but approximately half the mass (0.64 Ϯ 0.17 ng; n ϭ 3; data not shown). Retinoid peaks were not seen when 500 no embryo extracts or3gofBSAwere chromatographed. Fur- Luciferase Activity thermore, blank runs of samples involving a single isomer did not result in the isomerization of one isomer into the other, suggesting * 0 that 9-cis-RA was not a methodological artifact. Thus both M) M) 9-cis-RA and all-trans-RA were present in embryos. Integration of -10 -10 DMSO DMSO DMSO the area under the curves suggested that the majority (54.6%) of the retinoid was 9-cis-RA. Using this method and based on total

-RA (0.1nM) -RA retinoid mass and embryo volume, a calculated 1.3-nM retinoid was Embryo Extract Embryo Extract Embryo cis detected. Analysis of radiolabeled 9-cis-RA in extracted samples -RA (0.1nM) -RA trans -cis-RA (10 -cis-RA rans-RA (10 rans-RA omogenate (1:100) Homogenate (1:10) t 9

H indicated a recovery of 43% of 9-cis-RA; accordingly, there would all- -RA/9- be Ϸ3 nM of identified total retinoid, including 1.6 nM of 9-cis-RA in whole Locusta embryos. trans Embryo Homogenate (1:10) Embryo Homogenate (1:100) all- Discussion Fig. 2. Response of retinoid reporter cell lines to locust embryo extracts. Locust RXR Characterization. Recently, we described the cloning and Mouse P19 cells were incubated with locust embryo homogenates or extracts. Retinoids were suspended in 100% DMSO, and this solvent was used as a characterization of two locust RXR isoforms (20) and noted that control, along with all-trans-RA, and equimolar amounts of all-trans and 9-cis the LBDs showed greater identity to the vertebrate RXR LBDs RA. Embryos were either used as whole homogenates or subjected to retinoid than to the USP receptors from moths and . Furthermore, extraction as described. Luciferase activity was determined in a luminometer. RXR transcripts appeared to be present throughout embryogene- Vertical error bars represent standard deviations (n ϭ 4). * indicate significant sis, despite the absence of message for EcR until midembryogenesis differences from the controls. (20). This finding appeared curious because EcR is RXR’s sup- posed obligate heterodimerization partner in insects. Similarly, in the cockroach, RXR-L transcripts are detected at the beginning of to control protein samples (Fig. 3 Upper). However, single reaction embryogenesis, before the initial ecdysteroid pulse and RXR-S monitoring and multiple reaction monitoring (MRM) of the tran- message appears later in midembryogenesis (21). Together, these sition from 299 to 255 in chromatograms of the Locusta extract observations suggested that RXR in locusts and possibly other preparations showed a significant peak with retention time of 32.8 primitive insects and might have an additional role min with a shoulder at 33.4 min. The peak components exactly during early development, similar to the function of vertebrate

0.02 X8

350 0.01 AU

0.00 0 5 10 15 20 25 30 35 40

5000 9-cis-RA all-trans-RA Intensity

2500

0 5 10 15 20 25 30 35 40 45 Retention Time (minutes)

Fig. 3. HPLC-MS analysis of locust embryo extracts. A typical chromatograph of control protein samples (BSA; broken lines) and day-5 Locusta embryo extract preparations (solid lines). (Upper) Elution profile of the HPLC, monitored at 350 nm, with the scale shown. An expanded view (ϫ8) of the fractions of interest indicates that the retinoids were outside of the detection range. (Lower) A typical chromatogram for the negative ion LC-MS/MS in the MRM mode using the 299-255 transition showing the retention time positions of all-trans-RA and 9-cis-RA as previously determined by numerous reference experiments (results not shown) and a nonspecific peak at Ϸ21 min. Exact retention times of standards were as follows: 13-cis-RA, 32.52 min; 9-cis-RA, 32.97 min; all-trans-RA, 34.41 min.

9542 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0712132105 Nowickyj et al. Downloaded by guest on September 28, 2021 RXR, and in contrast to the orthologous USP of more advanced derived from the yolk, which readily precipitated as reported insect species. previously for eggs of another locust, Schistocerca gregaria (42). Recombinant LmRXR-L(DE) and hRXR␣(DE) were ex- Notably, those authors reported that the presence of the thick oil pressed and used to generate antibodies. Immunochemical assays prevented analysis by HPLC. We were able to circumvent this showed that the sequence similarity of the locust and human problem by performing a second, solid-phase extraction followed by proteins was also reflected in epitope determinants because there tandem MS. Even with a limited amount of insect tissue for was strong antigen–antibody cross-reactivity with antisera derived extraction, the addition of the preparation to the mouse cells from either insect or vertebrate proteins. These antibodies were bearing two synergistic endogenous RA response elements used to detect cross-reacting material in locust embryos during early (RAREs) (24) was positive; the Cyp26A1 promoter-induced lucif- embryogenesis (see Fig. S1E), and indicated that RXR transcripts erase activity indicated the presence of retinoids. Because this seen early in embryonic development (20) were translated and particular cell line is responsive to both 9-cis-RA and all-trans-RA therefore of likely functional significance. In vertebrates, activated (Fig. 2), these experiments indicated that locust embryos contain RXR (as opposed to heterodimerized RXR), is important for measurable levels of retinoids (1.6 ng/g) similar to those reported neurological promotion in embryos (25, 26). Of the various forms in mouse serum (0.7–1.1 ng/ml) (43). The subsequent assignment of of vertebrate RXRs, RXR␣ is thought to be of the greatest 9-cis-RA and all-trans-RA as the embryonic retinoids was achieved functional significance during mouse embryogenesis (27). by their retention time on RP- HPLC and by the specific 299–255 fragment because of a loss of the carboxyl group as determined by Ligand Binding and Retinoids. To evaluate whether Locusta RXR has tandem MS, derived from RA isomers. Thus, this unambiguous a role similar to that of the vertebrate receptor during development, determination of 9-cis-RA in any embryo could only be accom- it was crucial to determine whether it could bind ligand. Purified, plished by using highly sensitive HPLC-MS/MS in the MRM mode. recombinantly expressed locust RXRs competitively bound Previously, all-trans-RA, and its precursor, all-trans-retinol, has 9-cis-RA (IC50 ϭ 61.2–107.7 nM) similar to the high affinity shown been identified in several vertebrate tissues and embryos (e.g., refs. by hRXR␣(DE) in our assays (Fig. 1A and data not shown) and as 39 and 44–47). There have also been several reports that there are previously published (22). It is noteworthy that although RXRs insignificant levels of 9-cis-RA in these same species (48, 49). from vertebrates have been reported to bind 9-cis-RA (ref. 28 but Indeed, some have suggested that 9-cis-RA is not the endogenous see also refs. 29 and 30), USP from Dm does not (6). USP may ligand for RXR (28, 29). However, Luria and Furlow (26) argued function solely as a dimerization partner for EcR (6, 16, 31). Unlike that 9-cis-RA was a ‘‘logical candidate’’ for the RXR ligand but that LmRXR-L, a potential capping loop between helices 1 and 3 of DHA could also be the important morphogen in Xenopus. Here, we LBD appears to be missing in LmRXR-S, which initially suggested show that DHA does not bind to locust RXR with such high affinity that the small locust isoform might not bind ligand. However, both compared to 9-cis-RA and that 9-cis-RA is present, representing isoforms showed high affinity for 9-cis-RA. Among invertebrates, Ϸ55% of the retinoids in the locust embryo. As in where only the RXRs from the jellyfish, Tripedalia cystophora, and the sea compartmentalization of RAs can occur (50), we expect that snail, Biomphalaria glabrata, have been previously shown to bind 9-cis-RA and all-trans-RA would not be evenly distributed through- this ligand (32, 33). We have shown the binding of LmRXR to a out the egg. Certainly by day 5 approximately half the egg is yolk, ligand, 9-cis-RA, in terrestrial invertebrates. with the embryonic insect occupying the posterior portion. Never- In chordates, the active vitamin A metabolite, all-trans-RA binds theless, the presence of RA (Ϸ3 nM) coinciding with the presence the chordate-exclusive RARs. The Dm genome does not contain an of RXR (Kd ϭ 3.0 nM) at a period of rapid embryonic growth just RAR gene (34), and no RAR has been identified in any inverte- before dorsal closure and the production of 20-OH-Ec and JH by brate. Indeed, phylogenetic analysis compellingly argues that RARs developing prothoracic glands and corpora allata, respectively (51, evolved from duplication events that took place at the origins of 52), argues for its importance during development. vertebrates (35). Nevertheless, both LmRXR-L(DE) and Lm- Although the role of 9-cis-RA in vertebrates remains unclear, the RXR–S bind all-trans-RA with high affinity, indicating a role for significance of this isomer as an important signaling molecule in the receptors in transducing the signal of both retinoid isomers and invertebrates is better established. 9-Cis-RA has been detected in suggesting that all-trans-RA and 9-cis-RA could act as morphogens adults of the cephalochordate, amphioxus, Branchiostoma floridae in these primitive insects. (53), and the limb blastemas of an , the crab, Uca pugilator Another potential ligand, the polyunsaturated fatty acid, DHA, (54). Furthermore, 9-cis-RA acts as a potent morphogen in the has been detected in the mouse brain and activates mammalian gastropod, Thais clavigera, where exogenous 9-cis-RA induced male RXR␣ with micromolar affinity (36, 37). Similarly, DHA bound genital tracts in females (55). In insects, exogenous all-trans-RA competitively to both LmRXR-L(DE) and LmRXR–S isoforms inhibits metamorphosis and arrests embryogenesis (56). Therefore, but at a lower affinity compared with the retinoids. As well, the JH our identification of 9-cis-RA in Locusta embryos argues strongly analogue, methoprene acid, binds to the locust RXRs with approx- that RXR and this natural ligand are also active morphogens in imately the same low affinity as it does to hRXR␣(DE) (Fig. 1) (38). primitive insect embryogenesis. Thus locusts may be a good system Thus, it appears that locust RXRs are retinoid-specific receptors to investigate the mode of synthesis of this retinoid, which is still not and unliganded partners for EcR. clear (Table 2) (57). We further speculate that similar to the To determine whether RAs were present in locust embryos at a function of all-trans-RA in vertebrate anterior-posterior patterning time when RXR transcripts were detected, retinoid-sensitive re- mechanisms, 9-cis-RA may also have a role in early mesoderm porter cells (24) were used. When whole locust embryo homoge- patterning in invertebrates along with a primitive signaling pathway. nates were incubated with these cells no significant reporter activity Therefore, locust RXR may not merely serve as a silent partner

was detected. As well, the addition of homogenates to known with EcR as does the USP of flies and moths, but it reasonably has BIOCHEMISTRY concentrations of retinoids lowered the apparent reporter activity a major role in early embryonic development. Indeed, this role is (results not shown), suggesting that locust homogenates interfered likely ancestral and conserved (57, 58) in invertebrates, but may with the cell line assay. Therefore, efforts were made to extract have been partially usurped by the loss of retinoid binding by the retinoids from the early embryos. USPs of the more advanced insects and the evolution of RARs in Although procedures for the extraction of retinoid isomers have vertebrates (Table 2). been previously described for many vertebrate embryo types and tissues including mouse (39), Xenopus (40), and chicken (41), Experimental Procedures Locusta embryo extracts presented additional challenges. The main Animals. African migratory locusts, L. migratoria migratorioides, were reared difficulty was the presence of a viscous brown oil, presumably under gregarious colony conditions, which are standard for our experiments (59).

Nowickyj et al. PNAS ͉ July 18, 2008 ͉ vol. 105 ͉ no. 28 ͉ 9543 Downloaded by guest on September 28, 2021 Table 2. A summary of retinoid synthesis and signaling pathways in mammals and insects Evidence for RA synthesis and signaling Human/mouse Moth Locust

USP Ϫϩ(6) ϩ (15) Ϫ RXR ϩ (13) ϪϪ ϩ(18, 19) † USP/RXR possible ligand 9-cis and all-trans-RA (Kd ϭ 5.6 nM) Methyl farnesoate ?9-cis and all-trans-RA (Kd ϭ 3.0 nM)* (22) (IC50 ϭ 74 nM)* (Kd ϭ 44 nM) (17) (IC50 ϭ 61–130 nM)* Exogenous RA is a morphogen ϩ (59) ? Ϫϩ?* Total embryonic RA ϳ2 ng/g‡ (44) ϪϪ Ϸ3 ng/g* Endogenous all-trans-RA ϩ (64) Ϫ (66) Ϫϩ* Endogenous 9-cis-RA ? (65) Ϫ (66) Ϫϩ* RA signaling via nuclear receptor ϩ (13) ϪϪ ?

The fly (D. melanogaster), moth (H. virescens), and locust (L. migratoria) represent the insects and human and mouse represent the mammals (portion of the table modified from ref. 57). ϩ, Substantial data exist and are generally accepted. Ϫ, Data are absence. ?, Not determined or not enough evidence exists. Numbers in parentheses are references. *Data from the present work. †LmRXR-L(DE) value from the present work. ‡Determined as total retinoids from porcine embryo forelimbs early in gestation.

Reproductively mature female locusts oviposited into plastic cups lightly packed luciferase reporter gene (24). The cells were seeded in 24-well plates at a density

with wet sand. Cups were replaced daily and embryos were kept at 31°C where of 30,000 cells per well and incubated for 24 h at 37°C in 5% CO2. The cells were first-instar larvae emerged on day 14. Embryos were recovered from egg pods in then washed with 1 ml of 1ϫ PBS and treated with 0.5 ml MEM (pH 7.3, Ϫ the sand cups at day 5 and stored at 80°C. supplemented with 0.37% NaHCO3, 10% FCS, 0.5% penicillin-streptomycin, 0.1% gentamicin, and 0.1% fungizone; Invitrogen Life Technologies). The MEM was Plasmids, Protein Purification, and Western Blotting. The full-length isoform supplemented with DMSO, all-trans-RA, a mixture of all-trans-RA and 9-cis-RA, LmRXR-L (GenBank accession no. AY348873) was transformed into M15[pREP4] day-5 Locusta embryo homogenates, or extracts in the amounts listed in Results. cells (Qiagen), but also subcloned to generate the LBD (domain DE) for recom- The cells were incubated for an additional 24 h, washed as before, and lysed with binant expression and functional studies. LmRXR-L (corresponding to Thr-176 to 200 ␮l of passive 1ϫ lysis buffer (Promega) for 20 min. Lysates were scraped from Ser-411) was PCR-amplified, subcloned, and expressed (see SI Text). The sequence the bottom of the wells, 20-␮l samples were treated with luciferase reagents corresponding to the DE domain was then obtained by restriction enzyme (Promega), and enzyme levels were determined immediately in a Berthold lumi- digestion and used to transform BL21(DE3) cells (Novagen) by using pET-15b as a nometer. Data were analyzed by using Student’s t test. vehicle. The resulting LmRXR-L(DE) construct was sequenced twice in both direc- tions to ensure veracity. Embryo Retinoid Extraction. All dissections, extractions, and downstream anal- LmRXR-S (GenBank accession no. AF136372) in M15[pREP4] has been de- yses were performed under dim amber lighting, and glassware was used in scribed (19). Both LmRXR-S and LmRXR–L constructs were designed to produce sample preparation and handling. Day-5 locust embryos were ground in homog- recombinant locust proteins bearing six His residues on the amino-terminal end. enization buffer (PBS containing 100 ␮M 1,2-dianilinoethane) for 2 min by using ␣ The hRXR (DE) (GenBank accession no. NM002957) also with His residues has a Rotor Stator homogenizer (KikaWerk). Embryo samples included3gofem- been previously used for crystallographic studies (60). E. coli cells [LmRXR-L(DE) bryos, 1.5 g of embryos with 1.5 g of BSA, and3gofBSA, all in triplicate; all and hRXR␣(DE)] and M15[pREP4] (LmRXR-L and LmRXR-S) were used for the samples were extracted according to a modified Bligh and Dyer (23) protocol. production of recombinant RXRs by using standard procedures (see SI Text). The Homogenates were adjusted to pH 9.5 to partition the deprotonated RA to the purified proteins were used to generate antibodies for Western blots (see SI Text) aqueous phase and nonpolar lipids to the organic phase. The homogenates were and ligand binding assays. extracted with 2 ϫ 5 ml 100% (vol/vol) methanol, 2 ϫ 5 ml dichloromethane, and 5 ml of saturated KCl and centrifuged at 1,000 ϫ g at 4°C for 15 min. The aqueous Ligand Binding Assay. Displacement binding experiments were performed ac- phase was removed with a Pasteur pipette and subsequently dried in a sample cording to Allegretto (61). Briefly, LmRXR-L(DE) and LmRXR-S purified as de- evaporator (37°C) under a gentle stream of N to reduce the content of organic scribed (see SI Text) were incubated in borosilicate glass tubes containing binding 2 solvent. Samples were resuspended in 1 ml of 5% (vol/vol) methanol, vortexed, buffer [0.15 M KCl, 10 mM Tris⅐HCl (pH 7.4), 8% (vol/vol) glycerol, and 0.5% and subjected to a second solid-phase extraction. Waters Oasis hydrophilic- (wt/vol) CHAPS detergent] with Ϸ5nM[3H]-9-cis-RA (1.44 TBq/mmol or 39.0 lipophilic balanced cartridges (6 ml) were washed with 6 ml of 100% methanol Ci/mmol; Amersham) to a final volume of 300 ␮l. Purified, recombinant Dm and then 6 ml of ddH O. Fractions were collected by vacuum filtration, washed dihydrofolate reductase was used as a control protein. Competitor ligands were 2 once with 6 ml of 5% (vol/vol) methanol, and eluted with 6 ml of 100% (vol/vol) dissolved in 100% (vol/vol) ethanol and added in increasing concentrations and methanol. Eluants were dried in a sample evaporator (37°C) under a gentle incubated for4hat4°C. Competitors included 9-cis-RA (0.046–50,000 nM), all-trans-RA (0.042–42,000 nM), methoprene acid (3.1–310,000 nM), and cis- stream of N2, resuspended in 1 ml of 100% (vol/vol) ethanol, dried again, and ␮ 4,7,10,13,16,19-DHA (1.3–1,300,000 nM). Hydroxylapatite (HAP; Bio-Rad) was resuspended in 30 l of 60% (vol/vol) acetonitrile. Sample standards (500 ng of used to separate bound from free ligand (62). Three-hundred microliters of HAP all-trans-RA and 500 ng of 9-cis-RA) were dried under N2 and resuspended in 30 ␮ slurry (0.1 mg/ml) was added to the binding reactions and incubated for 45 min l of 60% (vol/vol) acetonitrile. Standards and samples were centrifuged for 15 ϫ on ice (with mixing every 5 min). The samples were then diluted with 0.9 ml of min at 1,500 g at 4°C and transferred to glass vials with a polyspring insert for binding buffer and centrifuged at 857 ϫ g for 5 min at 4°C. The HAP pellet was HPLC analysis. Concentrations were calculated by using the formula weight of washed three more times and transferred to a scintillation vial with 0.5 ml of all-trans-RA and 9-cis-RA (300.44 g/mol) and the embryonic volume of Locusta binding buffer and 4.5 ml of liquid scintillant (Amersham Canada). Saturation (0.01 ml). binding experiments using LmRXR-L(DE) were performed with increasing con- centrations of [3H]-9-cis-RA in the presence and absence of 300-fold unlabeled HPLC and MS. HPLC/MS analysis of extracted locust embryos was performed as 9-cis-RA. Specific binding, determined by subtracting nonspecific binding from described by Chithalen et al. (63). Briefly, RP-HPLC separation of extracted sam- total binding, was subjected to Scatchard analysis. Each sample was counted ples was performed on a Waters 2695 separation module with a 996-photodiode (Beckman LS6500 liquid scintillation counter), and the binding constant and IC50s array (monitoring at 350 nm) and a Zorbax-SB C18 column (Agilent; 3.5 ␮m, 150 ϫ were calculated by using GraphPad Prism 3.0 software. 2.1 mm) by using a gradient solvent system of ddH2O-acetonitrile-glacial acetic acid [74.99:25:0.01 (vol/vol/vol) to 0.99:99:0.1 (vol/vol/vol) at a flow rate of 200 Luciferase Assays. A P19 mouse embryomal carcinoma cell line was transfected ␮l/min for 35 min with final conditions held for 5 min]. Samples were further with pGL3-Basic (Promega) containing a 2.6-Kb Cyp26A1 promoter sequence analyzed by the most sensitive LC-MS/MS MRM system available to us. A Micro- (Cyp26A1 is a member of the cytochrome P450 gene family) fused to a firefly mass Quattro Ultima (Waters) mass spectrometer (equipped with a Z-Spray

9544 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0712132105 Nowickyj et al. Downloaded by guest on September 28, 2021 electrospray interface in the negative mode set with a capillary voltage of Ϫ3.70 was run by using single reaction monitoring so that the transition from 299 to 255 kV and a cone voltage of Ϫ55 V) was used. The source and desolvation temper- (loss of a carboxyl group) was detected, ensuring a specificity for RA. atures were maintained at 80°C and 350°C, respectively, and the cone and desolvation gases were kept at 50 and 562 liters/h, respectively. Argon, used as the ACKNOWLEDGMENTS. We thank Drs. S. Zhou, D. Hayward and E. Ball for collision gas, was maintained in the chamber at 1.6 ϫ 10Ϫ3 mBar with a collision their encouragement, Dr. H. Gronemeyer (Institut de Ge´ne´ tique et de energy of 14 V. MRM was used to observe the transition of m/z 299–255. The area Biologie Mole´culaire et Cellulaire/CNRS/INSERM/Universite´Louis Pasteur, Strasbourg, France) for the generous gift of the hRXR␣(DE), the reviewers under the curve was calculated for each chromatogram and averaged. Isomer for their suggestions. The majority of this work was supported by a Natural standards (including radiolabeled 9-cis-RA) were used, and all-trans-RA, 9-cis-RA, Sciences and Engineering Research Council (Canada) grant (to V.K.W.), and and 13-cis-RA were run both before and after each of the samples to ensure there Canadian Institutes of Health Research Operating and Equipment Mainte- was no drift in retention times and no isomerization of the retinoids. The LC/MS nance Grants MT-9475 and SMFSA-260938 (to G.J.) are also acknowledged.

1. Riddiford LM, Cherbas P, Truman JW (2000) Ecdysone receptors and their biological 36. de Urquiza AM, et al. (2000) Docosahexaenoic acid, a ligand for the retinoid X receptor actions. Vitamins and Hormones, ed Litwack G (Academic, San Diego), Vol 60, pp 1–73. in mouse brain. Science 290:2140–2144. 2. Wyatt GR (1997) Juvenile hormone in insect reproduction: A paradox? Eur J Entomol 37. Lengqvist J, et al. (2004) Polyunsaturated fatty acids including docosahexaenoic and 94:323–333. arachidonic acid bind to the retinoid X receptor ␣ ligand-binding domain. Mol Cell 3. Koelle MR, et al. (1991) The Drosophila EcR gene encodes an ecdysone receptor, a new Proteomics 3:692–703. member of the steroid receptor superfamily. Cell 67:59–77. 38. Harmon MA, Boehm MF, Heyman RA, Mangelsdorf DJ (1995) Activation of mammalian 4. Yao T-P, Segraves WA, Oro AE, McKeown M, Evans RM (1992) Drosophila ultraspiracle retinoid X receptors by the insect growth regulator methoprene. Proc Natl Acad Sci modulates ecdysone receptor function via heterodimer formation. Cell 71:63–72. USA 92:6157–6160. 5. Yao T-P, et al. (1993) Functional ecdysone receptor is the product of EcR and Ultraspi- 39. Horton C, Maden M (1995) Endogenous distribution of retinoids during normal racle genes. Nature 366:476–479. development and teratogenesis in the mouse embryo. Dev Dyn 202:312–323. 6. Oro AE, McKeown M, Evans RM (1990) Relationship between the product of the 40. Pijnappel WWM, et al. (1993) The retinoid ligand 4-oxo-retinoic acid is a highly active Drosophila ultraspiracle locus and the vertebrate retinoid X receptor. Nature 347:298– modulator of positional specification. Nature 366:340–344. 301. 41. Thaller C, Eichele G (1990) Isolation of 3,4-didehydroretinoic acid, a novel morphoge- 7. Segraves WA (1994) Steroid receptors and other transcription factors in ecdysone netic signal in the chick wing bud. Nature 345:815–819. response. Recent Prog Horm Res 49:167–195. 42. Dinan LN, Rees HH (1981) The identification and titres of conjugated and free ecdys- 8. Kliewer SA, Umesono K, Mangelsdorf DJ, Evans RM (1992) Retinoid X receptor interacts teroids in developing ovaries and newly laid eggs of Schistocerca gregaria. J Insect with nuclear receptors in retinoic acid, thyroid hormone, and vitamin D3 signaling. Physiol 27:51–58. Nature 355:446–449. 43. Schmidt CK, Brouwer A, Nau H (2003) Chromatographic analysis of endogenous 9. Mangelsdorf DJ, et al. (1991) A direct repeat in the cellular retinol-binding protein type retinoids in tissues and serum. Anal Biochem 315:36–48. II gene confers differential regulation by RXR and RAR. Cell 66:555–561. 44. Scadding SR, Maden M (1994) Retinoic acid gradients during limb regeneration. Dev 10. Mader S, et al. (1993) The patterns of binding of RAR, RXR, and TR homo- and Biol 162:608–617. heterodimers to direct repeats are dictated by the binding specificities of the DNA 45. Costaridis P, Horton C, Zeitlinger J, Holder N, Maden M (1996) Endogenous retinoids in binding domains. EMBO J 12:5029–5041. the zebrafish embryo and adult. Dev Dyn 205:41–51. 11. Levin AA, et al. (1992) 9-Cis retinoic acid stereoisomer binds and activates the nuclear 46. Satre MA, Kochhar DM (1989) Elevations in the endogenous levels of the putative ␣ receptor RXR . Nature 355:359–361. morphogen retinoic acid in the embryonic mouse limb buds associated with limb 12. Heyman RA, et al. (1992) 9-Cis retinoic acid is a high-affinity ligand for the retinoid X dysmorphogenesis. Dev Biol 133:529–536. receptor. Cell 68:397–406. 47. Schweigert FJ, Siegling C, Tzimas G, Seeger J, Nau H (2002) Distribution of endogenous 13. Chambon P (1996) A decade of molecular biology of retinoic acid receptors. FASEB J retinoids, retinoid binding proteins (RBP, CRABPI) and nuclear retinoid X receptor ␤ 10:940–954. (RXR ␤) in the porcine embryo. Reprod Nutr Dev 42:285–294. 14. Clayton GM, Peak-Chew SY, Evans RM, Schwabe JWR (2001) The structure of the 48. Blumberg B, et al. (1996) Novel retinoic acid receptor ligands in Xenopus embryos. Proc ultraspiracle ligand-binding domain reveals a nuclear receptor locked in an inactive Natl Acad Sci USA 93:4873–4878. conformation. Proc Natl Acad Sci USA 98:1549–1554. 49. Ulven SM, Gundersen TE, Sakhi AK, Glover JC, Blomhoff R (2001) Quantitative axial 15. Billas IML, Moulinier L, Rochel N, Moras D (2001) Crystal structure of the ligand-binding profiles of retinoic acid in the embryonic mouse spinal cord: 9-cis retinoic acid only domain of the ultraspiracle protein USP, the ortholog of retinoid X receptors in insects. detected after all-trans-retinoic acid levels are super-elevated experimentally. Dev Dyn J Biol Chem 276:7465–7474. 222:341–353. 16. Iwema T, et al. (2007) Structural and functional characterization of a novel type of 50. Vernet et al. (2006) Retinoic acid metabolism and signaling pathways in the adult and ligand-independent RXR-USP receptor. EMBO J 26:3770–3782. developing mouse testis. Endocrinology 147:96–110. 17. Jones G, Jones D, Teal P, Sapa A, Wozniak M (2006) The retinoid-X receptor ortholog, 51. Hartmann R, Jendrsczok C, Peter MG (1987) The occurrence of a juvenile hormone Ultraspiracle, binds with nanomolar affinity to an endogenous morphogenetic ligand. binding protein and in vitro synthesis of juvenile hormone by the serosa of Locusta FEBS J 273:1–14. migratoria embryos. Dev Genes Evol 196:347–355. 18. Miura K, Oda M, Mikita S, Chinzei Y (2005) Characterization of the Drosophila Metho- 52. Tawfik AI, Vedrova´A, Sehnal F (1999) Ecdysteroids during ovarian development and prene–tolerant gene product: Juvenile hormone binding and ligand-dependent gene embryogenesis in solitary and gregarious Schistocerca gregaria. Arch Insect Biochem regulation. FEBS J 272:1169–1178. Physiol 41:134–143. 19. Hayward DC, et al. (1999) The sequence of Locusta RXR, homologous to Drosophila 53. Dalfo´ D, Albalat R, Molotkov A, Duester G, Gonza`lez-Duarte R (2002) Retinoic acid Ultraspiracle, and its evolutionary implications. Dev Genes Evol 209:564–571. synthesis in the prevertebrate amphioxus involves retinol oxidation. Dev Genes Evol 20. Hayward DC, et al. (2003) Ligand specificity and developmental expression of RXR and ecdysone receptor in the migratory locust. J Insect Physiol 49:1135–1144. 212:388–393. 21. Maestro O, Cruz J, Pascual N, Martı´nD, Belle´s X (2005) Differential expression of two 54. Hopkins PM (2001) Limb regeneration in the fiddler crab, Uca pugilator: Hormonal and RXR/ultraspiracle isoforms during the life cycle of the hemimetabolous insect Blattella growth factor control. Am Zool 41:389–398. germanica (Dictyoptera, Blattellidae). Mol Cell Endocrinol 238:27–37. 55. Nishikawa J, et al. (2004) Involvement of the retinoid X receptor in the development 22. Cheng L, et al. (1994) Characterization of the ligand binding domain of human retinoid of imposex caused by organotins in gastropods. Environ Sci Technol 38:6271–6276. X receptor ␣ expressed in Escherichia coli. J Biol Chem 269:18662–18667. 56. Nı`mecV, Kodrı´kD, Matolı´nS, Laufer H (1993) Juvenile hormone-like effects of retinoic 23. Bligh EG, Dyer WJ (1959) A rapid method of total lipid extraction and purification. Can acid in insect metamorphosis, embryogenesis, and reproduction. J Insect Physiol J Biochem Physiol 37:911–917. 39:1083–1093. 24. MacLean G, Li H, Metzger D, Chambon P, Petkovich M (2007) Apoptotic extinction of 57. Simo˜es-Costa MS, Azambuja AP, Xavier-Neto J (2008) The search for nonchordate germ cells in testes of Cyp26b1 knockout mice. Endocrinology 148:4560–4567. retinoic acid signaling: Lessons from chordates. J Exp Zool 306:54–72. 25. Solomin L, et al. (1998) Retinoid X receptor signaling in the developing spinal cord. 58. Marletaz F, Holland LZ, Laudet V, Schubert M (2006) Retinoic acid signaling and the Nature 395:398–402. evolution of chordates. Int J Biol Sci 2:38–47. 26. Luria A, Furlow JD (2004) Spatiotemporal retinoid-X receptor activation detected in 59. Chinzei Y, White EN, Wyatt GR (1982) Vitellogenin mRNA in locust fat body: Identifi- live vertebrate embryos. Proc Natl Acad Sci USA 101:8987–8992. cation, isolation, and quantitative changes induced by juvenile hormone. Can J Bio- 27. Krezel W, et al. (1996) RXR␥ null mice are apparently normal and compound RXR␣ϩ/ chem 60:243–251. Ϫ/RXR␤Ϫ/Ϫ/RXR␥Ϫ/Ϫ mutant mice are viable. Proc Natl Acad Sci USA 93:9010–9014. 60. Egea PF, Moras D (2001) Purification and crystallization of the human RXR␣ ligand- 28. Allenby G, et al. (1993) Retinoic acid receptors and retinoid X receptors: Interactions binding domain–9-cis-RA complex. Acta Crystallogr D 57:434–437. with endogenous retinoic acids. Proc Natl Acad Sci USA 90:30–34. 61. Allegretto EA (1998) Detection of RARs and RXRs in cells and tissues using specific 29. Calle´jaC,et al. (2006) Genetic and pharmacological evidence that a retinoic acid cannot ligand-binding assays and ligand-binding immunoprecipitation techniques. Retinoid be the RXR-activating ligand in mouse epidermis keratinocytes. Genes Dev 20:1525– Protocols, ed Redfern CPF (Humana, Totowa, NJ) pp 219–232. 1538. 62. Wecksler WR, Norman AW (1979) An hydroxylapatite batch assay for quantitation of 30. Wolf G (2006) Is 9-cis-retinoic acid the endogenous ligand for the retinoic acid X 1␣,25-dihydroxyvitamin D3-receptor complexes. Anal Biochem 92:314–323. receptor? Nutr Rev 64:532–538. 63. Chithalen JV, Luu L, Petkovich M, Jones G (2002) HPLC-MS/MS analysis of the products BIOCHEMISTRY 31. Tzertzinis G, Malecki A, Kafatos FC (1994) BmCF1, a Bombyx mori RXR-type receptor generated from all-trans-retinoic acid using recombinant human CYP26A. J Lipid Res related to the Drosophila ultraspiracle. J Mol Biol 238:479–486. 43:1133–1142. 32. Kostrouch Z, et al. (1998) Retinoic acid X receptor in the diploblast, Tripedalia cysto- 64. Napoli JL, Pramanik BC, Williams JB, Dawson MI, Hobbs PD (1985) Quantification of phora. Proc Natl Acad Sci USA 95:13442–13447. retinoic acid by gas-liquid chromatography-mass spectrometry: Total versus all-trans- 33. Bouton D, et al. (2005) A conserved retinoid X receptor (RXR) from the mollusk retinoic acid in human plasma. J Lipid Res 26:387–392. Biomphalaria glabrata transactivates transcription in the presence of retinoids. JMol 65. Arnhold T, Tzimas G, Wittfoht W, Plonait S, Nau H (1996) Identification of 9-cis-retinoic Endocrinol 34:567–582. acid, 9,13-di-cis-retinoic acid, and 14-hydroxy-4,14-retro-retinol in human plasma after 34. Adams MD, et al. (2000) The genome sequence of Drosophila melanogaster. Science liver consumption. Life Sci 59:PL169–PL177. 287:2185–2195. 66. Sun D, Chen DM, Harrelson A, Stark WS (1993) Increased expression of chloramphenicol 35. Escriva H, et al. (1997) Ligand binding was acquired during evolution of nuclear acetyltransferase by carotenoid and retinoid replacement in Drosophila opsin promo- receptors. Proc Natl Acad Sci USA 94:6803–6808. tor fusion stocks. Exp Eye Res 57:177–187.

Nowickyj et al. PNAS ͉ July 18, 2008 ͉ vol. 105 ͉ no. 28 ͉ 9545 Downloaded by guest on September 28, 2021