Retinoic acid synthesis in the prevertebrate amphioxus involves retinol ...

Dev Genes Evol (2002) 212:388–393 DOI 10.1007/s00427-002-0254-z

S H O R T C O M M U N I C AT I O N

Diana Dalfó · Ricard Albalat · Andrei Molotkov Gregg Duester · Roser Gonzàlez-Duarte

Retinoic acid synthesis in the prevertebrate amphioxus involves retinol oxidation Received: 9 April 2002 / Accepted: 23 May 2002 / Published online: 13 July 2002 © Springer-Verlag 2002

Abstract All-trans-retinoic acid (RA) contributes to the establishment of the anterior-posterior (AP) axis in chordates. In vertebrates, all-trans-retinol is oxidized to RA by two oxidative steps. However, the controversy about the enzymes responsible for retinol oxidation (ADH vs RDH) and the fact that some candidates are absent in cephalochordates questioned retinol oxidation in this lineage. Retinoid quantitation has revealed that Branchiostoma floridae adults contain both retinol and retinoic acid as well as retinal, the intermediate in the metabolic pathway. Furthermore, our data show that the developmental effects of retinol treatment are comparable to those reported for RA. SEM analysis revealed mouth and gill slit aberrations due to a posteriorization effect, also visualized by changes in the β-galactosidase pattern. Overall, these findings support the idea that amphioxus metabolizes endogenous retinol to retinoic acid and suggest a common oxidative pathway for RA in the chordate phylum. Keywords Amphioxus · Cephalochordate · Retinoic acid · Retinol oxidation · SEM

Introduction All-trans-retinoic acid (RA) is involved in vertebrate embryonic development. RA binding to RAR and RXR nuclear receptors regulates transcription of developmental genes involved in anterior-posterior (AP) axis forma-

Edited by J. Campos-Ortega D. Dalfó · R. Albalat · R. Gonzàlez-Duarte (✉) Departament de Genètica, Facultat de Biologia, Universitat de Barcelona, Av. Diagonal 645, 08028 Barcelona, Spain e-mail: [email protected] Fax: +34-93-4110969 A. Molotkov · G. Duester Gene Regulation Program, Burnham Institute, 10901 North Torrey Pines Road, La Jolla, CA 92037, USA

tion (reviewed in Maden 1998). Profound anatomical and functional effects are observed when embryos develop in either the absence or excess of RA. While RA-deprivation is associated with absence or defective development of posterior branchial arches and posterior hindbrain (White et al. 2000), exogenous RA “posteriorizes” structures causing truncation of head and loss of eyes and brain (Durston et al. 1989). Moreover, the severity and extent of these traits are concentration-, stage- and, to some extent, species-dependent (Shimeld 1996). Overall, it has been assumed that RA acts as a graded signal, with long-range effects on gene expression and cell fates in vertebrates (Maden 1999). Interestingly, this AP pattern regulation by RA is not restricted to vertebrates, but constitutes a chordate innovation, probably linked to the development of a central nervous system (Shimeld 1996). Amphioxus (subphylum Cephalochordata) and ascidians (subphylum Urochordata) contain retinoic acid receptors (Escrivà et al. 1997; Hisata et al. 1998) and show a vertebrate-like response to RA treatments (Holland and Holland 1996; Katsuyama et al. 1995). RA synthesis proceeds through two oxidative steps, from all-trans-retinol (retinol) to all-trans-retinal and then further to RA. Retinol (vitamin A) is acquired from the diet as retinyl esters (RE) or β-carotenes. In the intestine, RE are hydrolyzed to retinol and β-carotenes are cleaved to retinal, which is subsequently reduced to retinol. Retinol is then bound to cellular retinol-binding protein type II (CRBP-II) and esterified to RE, which are incorporated into chylomicrons and secreted into the lymph. In the liver, RE are hydrolyzed to retinol, which is then bound to CRBP-I and secreted into the circulation to target tissues. The bulk of retinol is, however, stored in the liver as RE (reviewed in Gottesman et al. 2001; Napoli 1996). A clear picture of the enzymes involved in RA synthesis has not yet emerged. Whereas it is widely accepted that oxidation of retinal to RA is catalyzed by members of the aldehyde dehydrogenase family including RALDH1, RALDH2, and RALDH3 (Mic et al. 2002;

389

Niederreither et al. 1999), the rate-limiting oxidation of retinol to retinal can be catalyzed by distinct enzyme families: cytosolic medium-chain dehydrogenase/reductase (MDR) referred to as alcohol dehydrogenase (ADH) or microsomal short-chain dehydrogenase/reductase (SDR) also known as RDH (reviewed in Duester 2000). In cephalochordates, there is no evidence yet on the functional contribution of RDH to retinol oxidation, and previous analyses had shown that the two classical ADH active with retinol – ADH1 and ADH4 – were absent in these animals (Cañestro et al. 2000, 2002). The ancestral ADH3 form, which has recently been shown to have retinol activity in mouse (Molotkov et al. 2002), remains as the only ADH candidate in amphioxus to conduct retinol oxidation. However, this would not be easily reconcilable with the gut-specific pattern we have shown for amphioxus ADH3 (Cañestro et al. 2000). Therefore, in the absence of a clear candidate in cephalochordates for retinol metabolism, the oxidative step itself deserved further analysis. To address this issue, we have assessed retinoid concentrations in Branchiostoma floridae adults and studied the morphological and molecular effects on retinol-treated embryos.

Materials and methods Quantitation of retinoids For retinoid quantification, amphioxus adults (B. floridae) were collected from Old Tampa Bay (Florida, United States) and kept frozen until used. For sample preparation, 2.63 g amphioxus tissue (whole animal) was homogenized in 5 ml of a buffer containing 0.05 M HEPES pH 6.0/methanol 4:1 (v/v) and extracted twice with 6 ml chloroform/methanol 2:1 (v/v). Chloroform phases were collected, combined, and evaporated under vacuum. Pellets were dissolved in 110 µl dimethylsulfoxide/methanol 1:1 (v/v) and injected into the HPLC system. Reversed-phase HPLC analysis was performed using a MICROSORB-MVTM 100 C18 column (4.5×250 mm; Varian) at a flow rate of 1 ml/min. Mobile phase consisted of 0.5 M ammonium acetate-methanol-acetonitrile (25:65:10 v/v/v; solvent A) and acetonitrile (solvent B). The A:B (v/v) gradient composition was: 100:0 at the time of injection; 70:30 at 1 min; 65:35 at 14 min; 0:100 at 16 min. UV detection was carried out at 340 nm. For quantitation of retinoids, standards included all-trans-retinoic acid, 9-cis-retinoic acid, 13-cis-retinoic acid, all-trans-retinol, and all-trans-retinal (Sigma, St. Louis, Miss.). Retinol treatment and scanning electron microscopy Spawning of ripe animals of the Florida lancelet was induced by electrical shock and subsequent in vitro fertilization was performed. Embryos were raised at 21°C as previously described (Stokes and Holland 1995). Stock solutions of 10–3 M RA and 10–1 M, 10–2 M and 10–3 M retinol (Sigma) in DMSO were prepared. Embryos were treated with 1:1,000 dilution of each stock as described (Holland and Holland 1996). DMSO and untreated controls were allowed to develop. Treatments with 10–6 M RA, 10–4 M, 10–5 M and 10–6 M retinol and DMSO controls were started at early gastrula and terminated by dilution at the hatched neurula stage. Larvae of 36 h and 48 h were fixed in 4% paraformaldehyde, 0.1 M MOPS (3-[Nmorpholino]propanesulfonic acid) pH 7.5, 0.5 M NaCl, 2 mM EGTA (ethylene glycol-bis[β-aminoethyl ether]-N,N,N',N'-tetra-

acetic acid) and 1 mM MgSO4 overnight at 4°C, then washed two times with 70% ethanol and stored at –20°C. The fixed samples were processed for SEM as previously described (Stokes and Holland 1995) and viewed in a Hitachi S 2500 scanning electron microscope. β-Galactosidase activity Amphioxus embryos were treated at early gastrula stage with a single application of RA and retinol to a final concentration of 10–6 M and 10–4 M, respectively, and raised until the 3-day-larvae stage. DMSO and untreated controls were also allowed to develop. Afterwards, animals were fixed and stained for over 48 h as described (Cañestro et al. 2001).

Results and discussion Retinoids content in amphioxus In order to investigate the retinoids normally present in untreated amphioxus, we determined the concentrations of several retinoids commonly found in higher organisms (Table 1). Among the five retinoids tested, we found that all-trans-retinol was the most abundant, with all-trans-retinal being present at about 40% the level of retinol. All forms of retinoic acid were at least tenfold lower in concentration than retinol. All-trans-retinoic acid was the most abundant acid form with 9-cis and 13-cis forms being present at 45% and 30%, respectively, of the level of all-trans. Our studies have determined that untreated amphioxus does indeed contain both retinol and retinoic acid as well as retinal, the intermediate in the metabolic pathway. These findings suggest that amphioxus normally metabolizes endogenous retinol to retinoic acid. The different isomers of retinoic acid observed (all-trans, 9-cis, and 13-cis) have all been previously observed in tissues of mammals, although it is unclear if they have unique functions for retinoid signaling or simply perform the same function. Effect of retinol on amphioxus development The most striking morphogenetic effect of excess RA reported on amphioxus larvae was the failure of mouth and gill slits to form properly. Moreover, the expression domain of Hox-1 extended anteriorly in the nerve cord and that of Pax-1 compressed and shifted forward in regions of pharyngeal endoderm (Holland and Holland 1996). RA also caused an expansion of β-galactosidase activity from the gut to the most anterior regions of the digestive Table 1 Retinoid content in amphioxus whole animals

Retinoid

ng/g

All-trans-retinol All-trans-retinal All-trans-retinoic acid 9-cis-retinoic acid 13-cis-retinoic acid

5.60 2.30 0.60 0.27 0.17

390

391

Fig. 2 Concentration-dependence of excess retinol on development of B. floridae. Embryos were treated with 10–6 M, 10–5 M and 10–4 M retinol, 10–6 M RA and DMSO (controls) from early gastrula to the hatched neurula stage, and allowed to develop up to the 36-h and 48-h-larva stage. Mouth and gill slit morphology was analysed under a stereoscopic microscope, and the percent of normal (black bar) and affected (grey bar) larvae was determined. Values for the 36-h larvae were: 185 treated with DMSO (185 normal, 0 affected); 237 with 10–6 M retinol (123 normal and 114 affected); 415 with 10–5 M retinol (83 normal and 332 affected); 169 with 10–4 M retinol (29 normal and 140 affected) and 163 with 10–6 M RA (23 normal and 140 affected). Values for the 48-h larvae: 153 treated with DMSO (153 normal, 0 affected); 122 with 10–6 M retinol (70 normal and 52 affected); 264 with 10–5 M retinol (66 normal and 198 affected); 124 with 10–4 M retinol (11 normal and 113 affected) and 76 with 10–6 M RA (13 normal and 63 affected)

▲

tract (Cañestro et al. 2001). Remarkably, retinol caused similar developmental effects and generated aberrant morphologies (Fig. 1). However, retinol was needed at higher concentrations than RA to produce comparable morphological phenotypes (Fig. 2). SEM analysis of 36 h and 48 h retinol-treated larvae revealed mouth and gill slit malformations. The severity of the traits and the number of affected larvae (Fig. 2) were concentrationFig. 1A–R SEM of 48-h Branchiostoma floridae larvae treated with DMSO, retinoic acid (RA: 10–6 M) and retinol (10–4 M). A–I Right side showing gill slits of animals treated with DMSO (A), RA (B) and retinol (C); bar 200 µm. Enlargements of anterior end of larvae showing (arrowed) regular (DMSO; D, G), reduced (RA; E, H) and absence of gill slits (retinol; F, I). Bar D–F 100 µm; G–I 20 µm. J–R Left side showing mouth and ciliated pit of animals treated with DMSO (J), RA (K) and retinol (L); bar 200 µm. Enlargements of anterior end of larvae showing mouth (arrowhead) and ciliated pit (arrow; DMSO; M, P) and ciliated pit but no mouth (RA; N, Q; retinol; O, R). Bar M–O 100 µm; P–R 20 µm

and stage-dependent. At 48 h, amphioxus larvae are clearly asymmetric (Stokes and Holland 1995). The mouth and the ciliated pit open on the left of the ventral midline (Fig. 1J, M, P). Two or three gill slits open on the right, with the third remaining in a more ventral position (Fig. 1A, D, G). Our data showed that retinol-treated larvae suffered a dramatic decrease in number and/or size of gill slits, which were even absent in some animals (Fig. 1C, F, I). On the left, the mouth either did not appear (Fig. 1L, O, R) or was smaller than controls. The ciliated pit frequently opened, although its size decreased in some animals (Fig. 1R). Finally, the anterior end of the larvae was sometimes slightly reduced (Fig. 1L, O). Endogenous β-galactosidase activity is a histochemical marker for the amphioxus digestive system. Untreated and DMSO-control larvae showed β-gal activity from the midgut to the anus, whereas the pattern expanded anteriorly in RA-treated animals (Cañestro et al. 2001). Exogenous retinol produced similar effects (Fig. 3B, C). The signal was not uniformly distributed after either treatment; it was more intense in the gut, low or absent in the pharyngeal region, and intense again towards the anterior tip (Fig. 3D–G). Interestingly, an AP discontinuous expression pattern has also been described for the RA-regulated AmphiHox-1 in RA-treated larvae (Holland and Holland 1996). In summary, our data showed that exogenous RA and retinol induced similar posteriorization effects in amphioxus development, which were clearly detectable at the morphological level, observed as mouth and gill slit abnormalities, and at the functional level, followed by the β-galactosidase expansion. Also, to produce a similar degree of alterations, higher concentrations of retinol than RA were needed. Therefore, it could be assumed that it

392

Fig. 3 β-Galactosidase activity of 4-day DMSO- (A), 3-day RA(B) and retinol- (C) treated B. floridae larvae. Magnified view of the anterior end of RA- (D) and retinol- (E), and posterior end of RA- (F) and retinol- (G) treated larvae (a anterior end, g gut, mg midgut, n notochord, ph pharynx, p posterior end, pps primary pigment spot)

is the in vivo oxidation of retinol to RA that is responsible for the phenotypic changes.

Alternatively, a putative microsomal retinol dehydrogenase recently described in amphioxus (Dalfó et al. 2001) or a member of the aldo-keto reductase family (Crosas et al. 2001), not yet characterized in cephalochordates, stand as promising candidates to take over this function. Irrespective of the enzymes involved, our data support the idea that the oxidative process leading to RA predated the cephalochordate-vertebrate split, and suggest coevolution of metabolic pathways and signal transduction mechanisms for the acquisition of the new chordate features.

Retinol metabolism in the Cephalochordata Our data provide support for the metabolism of retinol to RA in amphioxus in vivo. On one hand, the functional and morphological effects of retinol treatment were similar to those for RA treatment suggesting that amphioxus might convert retinol to RA. On the other hand, the HPLC analysis indicated the presence of endogenous retinol and RA in amphioxus as well as retinal, the intermediate in the metabolic pathway. Also, the all-trans-RA concentration in amphioxus adults was similar to that reported for mammalian plasma (Kurlandsky et al. 1995) suggesting that the RA we detected may be enough to function in RA signaling. All of the enzymes responsible for oxidation of retinol to retinal in vivo may not yet be identified, but recent genetic studies on ADH have shown that ADH1, ADH3 and ADH4 provide this function in mice (Molotkov et al. 2002). As ADH3 has ubiquitous expression in mammals (Ang et al. 1996; Haselbeck and Duester 1997), this could provide a source of retinal for any of the RALDHs. As ADH3 is the only ADH conserved between mammals and cephalochordates (Cañestro et al. 2000, 2002), this could mean that ADH3 provides retinol oxidation for amphioxus. However, it is not easy to reconcile this assumption with the gut-specific expression we have shown for amphioxus ADH3 (Cañestro et al. 2000), clearly different from the ubiquitous vertebrate pattern compatible with RA localization in mesodermal cells.

Acknowledgements We are indebted to J. Garcia-Fernàndez and E. Benito for their help in the β-gal experiments. We thank C. Cañestro for helpful discussions and valuable comments, L.Z. Holland and N.D. Holland for useful suggestions, J.M. Lawrence for laboratory facilities and Ray Martinez for technical support in Tampa, Florida. We are indebted to the Serveis Científico-Tècnics (UB) for the assistant in SEM analysis. This work was supported by a grant from DGICYT (Ministerio de Educación y Cultura, Spain, BMC2000–0536), Generalitat de Catalunya grant 2001SGR00103 and a FPI fellowship to D.D. from the MEC (Ministerio de Educación y Cultura) and National Institutes of Health grant GM62848 to G.D.

References Ang HL, Deltour L, Hayamizu TF, Zgombic-Knight M, Duester G (1996) Retinoic acid synthesis in mouse embryos during gastrulation and craniofacial development linked to class IV alcohol dehydrogenase gene expression. J Biol Chem 271:9526– 9534 Cañestro C, Hjelmqvist L, Albalat R, Garcia-Fernàndez J, Gonzàlez-Duarte R, Jörnvall H (2000) Amphioxus alcohol dehydrogenase is a class 3 form of single type and of structural conservation but with unique developmental expression. Eur J Biochem 267:6511–6518 Cañestro C, Albalat R, Escrivà H, Gonzàlez-Duarte R (2001) Endogenous β-galactosidase activity in amphioxus: a useful histochemical marker for the digestive system. Dev Genes Evol 211:154–156 Cañestro C, Albalat R, Hjelmqvist L, Godoy L, Jörnvall H, Gonzàlez-Duarte R (2002) Ascidian and amphioxus Adh genes correlate functional and molecular features of the ADH family expansion during vertebrate evolution. J Mol Evol 54:81–89

393 Crosas B, Cederlund E, Torres D, Jörnvall H, Farrés J, Parés X (2001) A vertebrate aldo-keto reductase active with retinoids and ethanol. J Biol Chem 276:19132–19140 Dalfó D, Cañestro C, Albalat R, Gonzàlez-Duarte R (2001) Characterization of a microsomal retinol dehydrogenase gene from amphioxus: retinoid metabolism before vertebrates. Chem Biol Interact 130–132:359–370 Duester G (2000) Families of retinoid dehydrogenases regulating vitamin A function: production of visual pigment and retinoic acid. Eur J Biochem 267:4315–4324 Durston AJ, Timmermans JP, Hage WJ, Hendriks HF, de Vries NJ, Heideveld M, Nieuwkoop PD (1989) Retinoic acid causes an anteroposterior transformation in the developing central nervous system. Nature 340:140–144 Escrivà H, Safi R, Hanni C, Langlois MC, Saumitou-Laprade P, Stehelin D, Capron A, Pierce R, Laudet V (1997) Ligand binding was acquired during evolution of nuclear receptors. Proc Natl Acad Sci USA 94:6803–6808 Gottesman ME, Quadro L, Blaner WS (2001) Studies of vitamin A metabolism in mouse model systems. BioEssays 23:409– 419 Haselbeck RJ, Duester G (1997) Regional restriction of alcohol/retinol dehydrogenases along the mouse gastrointestinal epithelium. Alcohol Clin Exp Res 21:1484–1490 Hisata K, Fujiwara S, Tsuchida Y, Ohashi M, Kawamura K (1998) Expression and function of a retinoic acid receptor in budding ascidians. Dev Genes Evol 208:537–546 Holland LZ, Holland ND (1996) Expression of AmphiHox-1 and AmphiPax-1 in amphioxus embryos treated with retinoic acid: insights into evolution and patterning of the chordate nerve cord and pharynx. Development 122:1829–1838 Katsuyama Y, Wada S, Yasugi S, Saiga H (1995) Expression of the labial group Hox gene HrHox-1 and its alteration induced

by retinoic acid in development of the ascidian Halocynthia roretzi. Development 121:3197–3205 Kurlandsky SB, Gamble MV, Ramakrishnan R, Blaner WS (1995) Plasma delivery of retinoic acid to tissues in the rat. J Biol Chem 270:17850–17857 Maden M (1998) Vertebrate development: a nervous vitamin. Curr Biol 8:R846–R849 Maden M (1999) Heads or tails? Retinoic acid will decide. BioEssays 21:809–812 Mic FA, Haselbeck RJ, Cuenca AE, Duester G (2002) Novel retinoic acid generating activities in the neural tube and heart identified by conditional rescue of Raldh2 null mutant mice. Development 129:2271–2282 Molotkov A, Fan X, Deltour L, Foglio MH, Martras S, Farrés J, Parés X, Duester G (2002) Stimulation of retinoic acid production and growth by ubiquitously-expressed alcohol dehydrogenase Adh3. Proc Natl Acad Sci USA 99:5337–5342 Napoli JL (1996) Biochemical pathways of retinoid transport, metabolism, and signal transduction. Clin Immunol Immunopathol 80:S52–S62 Niederreither K, Subbarayan V, Dolle P, Chambon P (1999) Embryonic retinoic acid synthesis is essential for early mouse post-implantation development. Nat Genet 21:444–448 Shimeld SM (1996) Retinoic acid, HOX genes and the anteriorposterior axis in chordates. BioEssays 18:613–616 Stokes MD, Holland ND (1995) Embryos and larvae of lancelet, Branchiostoma floridae, from hatching through metamorphosis: growth in the laboratory and external morphology. Acta Zool 76:105–120 White JC, Highland M, Kaiser M, Clagett-Dame M (2000) Vitamin A deficiency results in the dose-dependent acquisition of anterior character and shortening of the caudal hindbrain of the rat embryo. Dev Biol 220:263–284