Calcitriol results from hydroxylation of 25(OH)D3 by 25-hydroxycholecalciferol-1-hydroxylase (1-OHase). Both freshwater and seawater fish display renal 1-OHase activity (Henry & Norman 1975; Table 2). The only exception is a study of Hayes et al. (1986) who could not detect activity of renal 1-OHase in rainbow trout. Fish also have hepatic 1-OHase activity and some researchers state that liver is the main organ for 1,25(OH)2D3 synthesis in fish (Sundell et al. 1992; Takeuchi 1994; Takeuchiet al. 1991a). The presence of 1,25(OH)2D3 in fish plasma was first documented by Avioli et al. (1981). They observed that, in contrast to 25(OH)D3 levels, 1,25(OH)2D3 levels in plasma are up to 10 times higher in fish than in mammals. This observation was later confirmed by others (Table 1), and now even 1,25(OH)2D3 plasma levels up to 55 times higher than humans have been documented in seawater adapted Atlantic salmon (Fjelldalet al. 2009). Plasma levels of 1,25(OH)2D3 vary widely between species. The variation within species seems to be less pronounced, although the number of observations in general is limited. Additional information on the time of sampling, feeding, and the physiological state of the fish is lacking in studies on most fishes, although these conditions and parameters could have profound effects on the plasma 1,25(OH)2D3 values as demonstrated in, e.g. humans and horses (Halloran et al. 1985; Piccione et al. 2004).
Table 2 25-Hydroxycholecalciferol-1-hydroxylase activity in fish
24,25(OH)2D3 results from hydroxylation of 25(OH)D3 by 25-hydroxycholecalciferol-24-hydroxylase (24-OHase). This enzyme has, next to biosynthetic properties, also catabolic properties and is likely to be involved in the breakdown of 1,25(OH)2D3 to the inactive form 1,24,25(OH)2D3 (Henry 2001). For 24,25(OH)2D3 two chiral forms occur: the enantiomers 24R,25(OH)2D3 and 24S,25(OH)2D3. The naturally occurring form is 24R,25(OH)2D3 and, if not stated otherwise, '24,25(OH)2D3' in this review refers to this enantiomer. Initially 24,25(OH)2D3 was not detected in fish (Takeuchi et al. 1991a). However, in the past two decades, the Fish Endocrinology Group in Göteborg has advanced firm evidence for this metabolite in fish and shed light on its physiological actions. Sundh et al. (2007) measured increased hepatic and renal production of 24,25(OH)2D3 in rainbow trout after SW transfer. Larsson et al. (2003) observed increased binding of 24,25(OH)2D3 to enterocyte basal lateral membranes (BLMs) in rainbow trout when adapted to seawater (SW). Simultaneously, specific binding of 1,25(OH)2D3 to enterocyte BLMs preparations decreased and correlated with inhibited intracellular calcium uptake. It is possible that the change in sensitivity of enterocyte BLMs to vitamin D3 metabolites for an anadromous fish is part of a strategy to adapt to the calcium-rich environment of seawater (>10 mM Ca). It should be noted that seawater fish drink copious amounts of seawater (and thus imbibe significant amounts of calcium) to compensate the osmotic water loss in this hypertonic medium (Flik et al. 2002).
Other vitamin D3 metabolites
In mammalian research other vitamin D3 metabolites, such as 1,25(OH)2-3-epi-D3 (Brown et al. 1999), and C23 epimers (Lee et al. 2000), and P450 cytochromes involved in their synthesis have been identified (e.g. Bouillon et al. 1995; Tuckey et al. 2008). Understanding the role of these metabolites can, e.g. assist medical research to develop new vitamin D3 analogues. In research on fishes, the development of specific vitamin D3 analogues has received no priority. However, for the understanding of the complexity of the vitamin D endocrine system, it is important to realize that many more metabolites of vitamin D3 than described above may exist in fish and could be of physiological significance. A picture that emerges in other steroids and in thyroid hormones (THs) (Moreno et al. 2008), where multiple TH derivatives have metabolic effects, seems to apply for vitamin D3 as well.
Vitamin D binding protein
Vitamin D binding protein (DBP) is important for binding, solubilization and transport of vitamin D and its metabolites. The metabolic clearance rate of vitamin D metabolites depends partly on their binding to plasma proteins. In humans 85–88% of the vitamin D metabolites in plasma are bound to DBP; 12–15% are associated with albumin and only a small fraction (<1%) is unbound (Speeckaert et al. 2006). In chicken vitamin D2 and 25(OH)D2 have a lower binding affinity to plasma proteins than vitamin D3 and 25(OH)D3 (Belsey et al. 1974), which results in a faster clearance from plasma. A similar preference of DBP for vitamin D3 over vitamin D2 is present in fish as well (Hay & Watson 1977b). The first report on DBP in fish (Hay & Watson 1976a) demonstrated that cartilaginous and bony fish transport 25(OH)D3 via a lipoprotein. Only bony fish, both FW and SW species, have specific 25(OH)D3?-globulin transport proteins. Considerable variation in binding proteins of vitamin D3 and 25(OH)D3 in vertebrates exists, including ?-globulins, ?-globulins, lipoproteins and albumin (Edelstein et al. 1973; Hay & Watson 1976b). Hay & Watson (1977a) observed equal binding of 25(OH)D3, 24R,25(OH)2D3 and 24S,25(OH)2D3 to plasma proteins in lungfish (Protopterus sp.) and goldfish (Carassius auratus). These authors also demonstrated that in reptilian, avian and mammalian species a divergent binding of these metabolites exists, where 24R,25(OH)2D3 and 25(OH)D3 were bound with a higher affinity than 24S,25(OH)2D3. Hay & Watson (1977a) concluded that the properties of these proteins have changed during vertebrate evolution. Sundell et al. (1992) reached a similar conclusion for Atlantic cod (Gadus morhua) DBP. Both 25(OH)D3 and 1,25(OH)2D3 were bound to one specific DBP isolated from cod plasma. Compared with mammals and birds, cod DBP had a lower affinity for 25(OH)D3 and a sedimentation coefficient (3.4S), lower than the one found in humans (4.1S). In comparison with mammals and birds, cod DBP was unable to form a protein complex with globular actin (G-actin). Studies on Nile tilapia (Oreochromis niloticus) DBP (3.6S) resulted in similar observations, a low affinity protein for 25(OH)D3 was found that did not interact with G-actin (Allewaertet al. 1988a). As in mammals 25(OH)D3 has to be transported from the liver to the kidney, the necessity for a plasma DBP with a high affinity for 25(OH)D3 seems obvious. As both hydroxylations in most fish can take place in the liver there is little need for a plasma binding component with high affinity for 25(OH)D3. A few fish species, however, do have a DBP that shows great resemblance with mammalian DBP. For example, Marcocci et al. (1982) described a specific plasma binding protein for both 1,25(OH)2D3 and 25(OH)D3 in European eel and common carp (Cyprinus carpio). Allewaert et al. (1988b) showed two 25(OH)D3 binding proteins in serum of common carp, one with a sedimentation coefficient similar to Nile tilapia and cod (3.7S), with a low affinity for 25(OH)D3 and no binding to G-actin, and a second binding protein (4.2S) with a high affinity but low capacity for 25(OH)D3 and ability to form a complex with G-actin (5.2S). The latter protein exhibits similar physical and biochemical characteristics as the DBP from terrestrial vertebrates (Allewaert et al. 1988a). Clearly, the DBP in vertebrates is not a single protein but a group of proteins which differs markedly in composition between species.
The nuclear 1,25(OH)2D3 receptor
The nuclear 1,25(OH)2D3 receptor (VDR) is a ligand-activated transcription factor. Upon activation by 1,25(OH)2D3 it will form a heterodimer with the retinoid-X-receptor (RXR) and regulates gene expression by interacting with specific DNA sequences upstream of vitamin D responsive genes (Fig. 2). Marcocci et al. (1982) were the first to demonstrate specific binding of 1,25(OH)2D3 in various eel tissues and the first mRNA sequence of a nuclear VDR in a teleostean fish appeared in Suzuki et al. (2000). They sequenced two subtypes of VDR in the Japanese flounder (Paralichthys olivaceus) that share 86% amino acid identity. The authors suggested that an independent genome duplication that has occurred in ray-finned fishes phylogeny is the underlying cause for these gene subtypes. Recently, a study by Howarth et al. (2008) on a VDR in medaka (Oryzias latipes) demonstrated the existence of two functional paralogs with different transactivation activity and sensitivity to 1,25(OH)2D3, which suggests a functional divergence of the receptors. Alpha- and ?-isoforms of the VDR were identified in fugu (Takifugu rubripes), tetraodon (Tetraodon nigriviridis) and stickleback (Gasterosteus aculeatus; Maglich et al. 2003). In European seabass (Dicentrarchus labrax) two VDR isoforms are identified. One isoform was sequenced from the liver (AM040727), the other from intestine (AM040728). It is unclear if these isoforms exist due to gene duplication. Isoforms of the VDR may also result from alternative splicing. The Atlantic salmon (Salmo salar) VDR (NM001123557) is spliced because of intron retention at two locations (AM238619, AM238620) (Lock et al. 2007). Also the common carp (C. carpio) VDR (AJ784084) can be spliced at the same introns as occurs in Atlantic salmon (our unpublished results). Whether a similar inhibitory regulatory activity applies to these splice variants as described for mice (Ebihara et al. 1996) needs further experimentation. Other species with known VDR sequences include: zebrafish (Danio rerio, NM130919) and rainbow trout (O. mykiss, AY526906).
Figure 2 Transcriptional control of gene expression by 1,25(OH)2D3 (CTR). In plasma CTR is bound to vitamin D binding protein (DBP) with only a small unbound fraction. This unbound CTR passes the cell membrane and enters the nucleus where it interacts with the vitamin D receptor (VDR). After dimerization with the retinoid X receptor (RXR), the complex binds to vitamin D responsive element (VDRE) and induces or represses gene transcription.
[Normal View ]
VDR mRNA expression has been demonstrated in a (wide) variety of tissues and organs of several fish species (Suzuki et al. 2000; Maglich et al. 2003; Lock et al. 2007). The most extensive study on the distribution of the VDR mRNA was carried out by Craig et al. (2008) in zebrafish. Specific VDR staining was found in tissues and organs including renal tubular cells, enterocytes and chloride cells of the gills. The VDR was further demonstrated in the bile duct epithelial cells, in line with an involvement of 1,25(OH)2D3 in transport of solutes and water in this organ (Makishima et al. 2002). The positive VDR mRNA staining found in Sertoli cells points to a role for the vitamin D endocrine system in male reproduction. Bone specific staining of the VDR was found in osteoblasts. Also chondrocytes in gill filaments, the nervous system (brain, spinal cord and retina) stained positively for VDR mRNA. Thus, the VDR appears to be widely expressed in epithelial endocrine and neural tissues. It is expressed already during early development (Fleming et al. 2005). The ubiquitous expression of the VDR in a wide range of tissues reflects the pleiotropic nature of the vitamin D endocrine system, far beyond its actions on mineral status.
A membrane 1,25(OH)2D3 receptor
Membrane-initiated signalling by steroid hormones (via G protein-coupled receptors) is now widely accepted (Hammes & Levin 2007). Activation of nongenomic signal transduction pathways in target cells involves a membrane vitamin D receptor (mVDR). Nemere et al. (2000) found specific binding of 1,25(OH)2D3 to enterocyte BLMs of common carp. In vitro a rapid (within minutes) increase of protein kinase activity was seen after 1,25(OH)2D3 exposure. Both 24,25(OH)2D3 and 25(OH)D3 competitively bound to the mVDR, which distinguishes this protein from the nuclear VDR. In Atlantic cod (G. morhua) no specific binding of 1,25(OH)2D3 to BLM was found, which is in line with the findings of Larsson et al. (2003) in this teleost. Rapid responses to 1,25(OH)2D3 can be explained by various mechanisms. Norman (2006) proposed a conformational ensemble model for VDR activity in mammals. He suggested that chirality of 1,25(OH)2D3 may translate into either slow classical genomic responses or more rapid responses (Norman et al. 2002; Huhtakangas et al. 2007), the receptor involved in both actions is the nuclear VDR. A second theory involves the recently identified membrane binding protein 1,25D3-MARRS (membrane-associated, rapid-response steroid-binding) also known as Erp57 (Nemere et al. 2004a,b; Rohe et al. 2007). 1,25D3-MARRS appears to play a role in intestinal phosphate uptake in young poultry chicks. As both theories are based on mammalian models, it is outside the scope of this study to discuss these pathways in more detail; for more information please see Khanal & Nemere (2007).
A membrane 24,25(OH)2D3 receptor
The existence of a membrane 24,25(OH)2D3 remains putative, however the physiological significance of this metabolite is presently established. Specific binding of 24,25(OH)2D3 to enterocyte BLMs of Atlantic cod (G. morhua), common carp (C. carpio) and SW-adapted rainbow trout (O. mykiss) was demonstrated by Larsson et al. (2001, 2003). Within seconds after administration of a physiological dose of 24,25(OH)2D3, Ca2+ influx across the mucosa of in vitro-perfused intestinal preparations in Atlantic cod was reduced (Larsson et al. 1995). These rapid responses indicate a functional nongenomic pathway for 24,25(OH)2D3. As mentioned earlier, 24,25(OH)2D3 may fulfill a greater role in SW than in FW fish. Indeed, in FW adapted rainbow trout Larsson et al. (2003) found no specific binding of 24,25(OH)2D3 to enterocyte BLMs. Also, 24,25(OH)2D3 had no effect on enterocyte calcium homeostasis.