Isolation and Characterization of Three Sodium-Phosphate Cotransporter Genes and Their Transcriptional Regulation in the Grass Carp Ctenopharyngodon idella

It is important to explore the regulatory mechanism of phosphorus homeostasis in fish, which help avoid the risk of P toxicity and prevent P pollution in aquatic environment. The present study obtained the full-length cDNA sequences and the promoters of three SLC20 members (slc20a1a, slc20a1b and slc20a2) from grass carp Ctenopharyngodon idella, and explored their responses to inorganic phosphorus (Pi). Grass carp SLC20s proteins possessed conservative domains and amino acid sites relevant with phosphorus transport. The mRNAs of three slc20s appeared in the nine tissues, but their expression levels were tissue-dependent. The binding sites of three transcription factors (SREBP1, NRF2 and VDR) were predicted on the slc20s promoters. The mutation and EMSA analysis indicated that: (1) SREBP1 binding site (−783/−771 bp) negatively but VDR (−260/−253 bp) binding site positively regulated the activities of slc20a1a promoter; (2) SREBP1 (−1187/−1178 bp), NRF2 (−572/−561 bp) and VDR(615/−609 bp) binding sites positively regulated the activities of slc20a1b promoter; (3) SREBP1 (−987/−977 bp), NRF2 (−1469/−1459 bp) and VDR (−1124/−1117 bp) binding sites positively regulated the activities of the slc20a2 promoter. Moreover, Pi incubation significantly reduced the activities of three slc20s promoters, and Pi-induced transcriptional inactivation of slc20s promoters abolished after the mutation of the VDR element but not SREBP1 and NRF2 elements. Pi incubation down-regulated the mRNA levels of three slc20s. For the first time, our study elucidated the transcriptional regulatory mechanisms of SLC20s and their responses to Pi, which offered new insights into the Pi homeostatic regulation and provided the basis for reducing phosphorus discharge into the waters.


Introduction
Aquaculture was considered to be one of the most important food-producing industries and provided enormous excellent protein sources for people over the world. However, the ever-increasing aquaculture is producing some adverse influences on the aquatic environment. One of the main concerns is the discharge of wastes, especially phosphorus. Inorganic phosphate (Pi) is essential for skeletal mineralization, oxidative response and cellular metabolism. Aquafeeds need dietary Pi addition in the amount that meets the requirement of fish. However, phosphorus is also the rate-limiting nutrient for the eutrophication of waters, and excess phosphorus will be excreted by the fish to the aquatic environment and result in eutrophication [1]. Thus, it is very important to investigate Int. J. Mol. Sci. 2020, 21, 8228 2 of 17 the mechanism of Pi absorption and transport in fish, which helps avoid the risk of phosphorus toxicity and prevent phosphorus loading in the aquatic environment.
Pi absorption and transport are mediated by sodium-phosphate cotransporters, containing three families of SLC17/type I NaPi, SLC34/type II NaPi and SLC20/type III NaPi. At present, the structure and functional role of the SLC34 members were established for the vertebrates; in contrast, relatively few studies have been performed for SLC20 proteins [2]. In mammals, the SLC20 members comprise SLC20A1 and SLC20A2, whose cDNA sequences and putative protein sequences have successfully been obtained [3,4]. They are expressed in various tissues and presumably function in Pi uptake and transport across the plasma membrane [2,5]. Moreover, low extracellular Pi content increased slc20a1 and slc20a2 expression in many cells [6,7], suggesting that they mediated the Pi homeostatic regulation [6]. However, despite their important roles in cellular Pi uptake, nothing is known about their precise regulatory mechanism at the transcription level. Elucidation of these mechanisms may help develop new pathways to increase dietary phosphorus utilization of fish and reduce phosphorus excretion into the environment.
The regions of gene promoters contain many important transcriptional regulatory elements which modulate their expression [8]. At present, the promoters of slc20s have been cloned in humans and mice [9,10] and some key transcriptional factors have been reported to regulate their transcriptional expression [9,11]. For example, vitamin D receptor (VDR), one member of the nuclear receptor (NR) superfamily, is important to regulate Pi homeostasis [12]. Nuclear factor-erythroid 2 p45-related factor 2 (NRF2), a central transcription factor regulating oxidative stress, functions in the uptake of extracellular Pi [13]. Sterol regulatory element-binding protein 1 (SREBP1) is an important transcriptional factor which regulates lipid metabolism [14] and vascular calcification [15,16]. However, direct evidence of VDR, NRF2 and SREBP1 mediating Pi homeostasis was absent. In the present study, we hypothesize that these transcriptional factors mediated the control of Pi homeostasis probably by directly targeting the slc20s promoters.
Teleost fish have approximately 30,000 species and they possess the most diverse group among the vertebrates [17,18], which is attributable to a teleost-specific whole genome duplication (TSGD) event during their evolution [19]. These duplicated genes offer a unique framework to explore the differentiation and diversification of their functions. Thus, teleost fish are very attractive models to study the gene evolution, structures and functions [20,21]. This study isolated the full-length cDNA and promoter sequences of three slc20s genes from grass carp (Ctenopharyngodon idella). Their mRNA tissue expression, transcriptional regulation and their function in Pi homeostasis were determined. Grass carp, a herbivorous freshwater fish, is one of the most important aquaculture species in the world and is selected as the experimental animals. The present study will increase our knowledge into the Pi homeostatic regulation and provide new knowledge for improving phosphorus utilization and reducing phosphorus pollution in the waters.

Slc20s mRNA Expression among the Tissues of Grass Carp
The slc20a1a mRNA amounts were the highest in the kidney and liver, followed by the spleen and brain, and the lowest in other tissues. The slc20a1b mRNA expression was predominant in the kidney, and no significant differences were found among other tissues. The slc20a2 mRNA amounts were predominant in the brain, followed by the skin and heart, and were lowest in other tissues ( Figure 1).   1813  180  1935  377  644  2264  slc20a1b  1804  76  1995  410  664  2453  slc20a2  1775  140  1989  79  662  2125 In grass carp, the amino acid (AA) sequences of SLC20A1A and SLC20A1B showed 61.9% and 60.0% identities with SLC20A2, respectively (Table 2). Moreover, the SLC20A1A, SLC20A1B and SLC20A2 sequences of grass carp shared the identities of 66.9-94.7%, 69.8-91.7% and 76.6-95.6% to those of other species, respectively ( Table 2). The protein sequences of grass carp SLC20A1A, SLC20A1B and SLC20A2 possessed 11, 11 and 10 transmembrane domains (TMDs), respectively ( Figure S2). Some conserved AA residues, such as D28, E55, D506, E575 and N81, were also found in grass carp SLC20s ( Figure S2), and they were important for SLC20s function.

Slc20s mRNA Expression among the Tissues of Grass Carp
The slc20a1a mRNA amounts were the highest in the kidney and liver, followed by the spleen and brain, and the lowest in other tissues. The slc20a1b mRNA expression was predominant in the kidney, and no significant differences were found among other tissues. The slc20a2 mRNA amounts were predominant in the brain, followed by the skin and heart, and were lowest in other tissues ( Figure 1). (C), across heart (H), brain (B), intestine (I), kidney (K), liver (L), muscle (M), gill (G), spleen (Sp) and skin (Sk) of grass carp. Data (mean ± SEM, n = 3 replicate tanks. For each tank, three fish were sampled for analysis) were expressed relative to expression of housekeeping gene (gapdh and 18srRNA). Expression of slc20s in heart was regarded as the relative expression 1. Data were subjected to one-way ANOVA and Tukey's multiple range test. Bars that do not share a common letter mean significant difference among different tissues (p < 0.05). and slc20a2 (C), across heart (H), brain (B), intestine (I), kidney (K), liver (L), muscle (M), gill (G), spleen (Sp) and skin (Sk) of grass carp. Data (mean ± SEM, n = 3 replicate tanks. For each tank, three fish were sampled for analysis) were expressed relative to expression of housekeeping gene (gapdh and 18srRNA). Expression of slc20s in heart was regarded as the relative expression 1. Data were subjected to one-way ANOVA and Tukey's multiple range test. Bars that do not share a common letter mean significant difference among different tissues (p < 0.05).

5 -Deletion Sequence Analysis of the Slc20s Promoters
To investigate the functional regulatory regions and the Pi-induced response of slc20 promoters, we performed the 5 deletion assays. For grass carp slc20a1a promoter, the −1813/−1355 bp sequence deletion did not markedly affect its relative luciferase activity, but the following deletion to −795 bp and the −795/−263 bp increased the luciferase activity significantly (Figure 2A), indicating that the −1355/−263 bp region negatively controlled its promoter activity.
For the slc20a2 promoter, the −1775/−1296 bp sequence deletion significantly increased its relative luciferase activity. The subsequent deletion to −835 bp did not influence its relative luciferase activity but the −835/−191 bp sequence deletion significantly increased the luciferase activity, indicating that the −1775/−1296 bp and −835/−191 bp region positively regulated its promoter activity. ( Figure 2C).
For the slc20a1b promoter, the−1804/−1374 bp sequence deletion did not influence the relative luciferase activity, but the subsequent deletion to −899 bp significantly increased the luciferase activity, indicating that the −1374/−899 bp region negatively controlled its promoter activity. However, the −899/−383 bp sequence deletion significantly reduced the luciferase activity, indicating that the −899/−383 bp region positively regulated its promoter activity ( Figure 2B).
For the slc20a2 promoter, the −1775/−1296 bp sequence deletion significantly increased its relative luciferase activity. The subsequent deletion to −835 bp did not influence its relative luciferase activity but the −835/−191 bp sequence deletion significantly increased the luciferase activity, indicating that the −1775/−1296 bp and −835/−191 bp region positively regulated its promoter activity. ( Figure 2C).

Site-Mutation Analysis of SREBP1, NRF2 and VDR on the Promoters of Slc20s
In order to identify the roles of SREBP1, NRF2 and VDR binding sites on the promoter regions of slc20s in grass carp, we performed the site-directed mutagenesis. Compared to the wild-type  Results are expressed as the mean ± SEM arbitrary units of three independent experiments. Student's t-test was undertaken to detect the differences between two groups. Symbol (*) indicates significant differences between the two groups.

EMSA Analysis of SREBP1, NRF2 and VDR Binding with the Slc20s Promoters
Next, we examined whether SREBP1, NRF2 and VDR functionally bind with their corresponding regions of slc20s promoters using EMSA analysis.
For the SREBP1 binding assay, the 100-fold unlabeled SREBP1 binding sequence competed for the binding when we used biotin-labeled SREBP1 binding sequences (−783/−771 bp of slc20a1a promoter) as the probe, while the 100-fold unlabeled mutated SREBP1 binding sequences markedly reduced this competition, indicating that the SREBP1 binding sequence (−783/−771 bp of slc20a1a promoter) was functionally bound by SREBP1 ( Figure 4A). The SREBP1 binding elements at −1187/−1178 bp of the slc20a1b promoter ( Figure 4B) and at −987/−977 bp of the slc20a2 promoter ( Figure 4C) were functionally bound by SREBP1. Student's t-test was undertaken to detect the differences between two groups. Symbol (*) indicates significant differences between the two groups.

EMSA Analysis of SREBP1, NRF2 and VDR Binding with the Slc20s Promoters
Next, we examined whether SREBP1, NRF2 and VDR functionally bind with their corresponding regions of slc20s promoters using EMSA analysis.
For the SREBP1 binding assay, the 100-fold unlabeled SREBP1 binding sequence competed for the binding when we used biotin-labeled SREBP1 binding sequences (−783/−771 bp of slc20a1a promoter) as the probe, while the 100-fold unlabeled mutated SREBP1 binding sequences markedly reduced this competition, indicating that the SREBP1 binding sequence (−783/−771 bp of slc20a1a promoter) was functionally bound by SREBP1 ( Figure 4A). The SREBP1 binding elements at −1187/−1178 bp of the slc20a1b promoter ( Figure 4B) and at −987/−977 bp of the slc20a2 promoter ( Figure 4C For the VDR binding assay, the 100-fold unlabeled VDR binding sequence competed for the binding when we used biotin-labelled VDR binding sequence (−260/−253 bp of slc20a1a promoter) as the probe, while the 100-fold unlabeled mutated VDR binding sequence markedly reduced this competition, indicating that VDR binding sequence (−260/−253 bp of slc20a1a promoter) was bound by VDR ( Figure 4F). The VDR binding elements located at −615/−609 bp of slc20a1b promoter ( Figure 4G) and at −1124/−1117 bp of the slc20a2 promoter ( Figure 4H) were functionally bound by VDR. Figure 4. EMSA of putative transcription factors binding sequence. The 5′-biotin labeled double-stranded oligomers were incubated with HEK293T nuclear extract. A 100-fold excess of the competitor and mutative competitor oligomers was added to the competition and mutant competition assay, respectively. NP represents nuclear protein; BLP represents biotin-labeled-probe; C represents competitor oligomers; Mut-C represents mutative competitor oligomers. The oligonucleotide sequences are given in Table S4

Effect of Pi Incubation on the Luciferase Activity of the Site-Mutagenesis Plasmids
To further determine whether Pi-induced decrease of slc20s promoter activities could be mediated by SREBP1, NRF2 and VDR elements, we conducted the site-directed mutation at their corresponding regions of slc20s promoters and used Pi to incubate the cells. Pi incubation significantly decreased three slc20s promoter activities. However, compared to the wild type pGl3-1813/+226 vector of slc20a1a, Pi-induced transcriptional inactivation of slc20a1a was abolished after the mutation of VDR element but not SREBP1 element ( Figure 5A). Compared with the wild-type pGl3-1801/+143 vector of slc20a1b, Pi-induced transcriptional inactivation of slc20a1b was abolished after the mutation of VDR element but not SREBP1 and NRF2 elements ( Figure 5B). Similarly, compared with the wild-type pGl3-1775/+211 vector of slc20a2, the Pi-induced transcriptional inactivation of slc20a2 was abolished after the mutation of VDR element but not the mutation of SREBP1 and NRF2 elements ( Figure 5C). For the NRF2 binding assay, the 100-fold unlabeled NRF2 binding sequence competed for the binding when we used biotin-labeled NRF2 binding sequence (−572/−561 bp of slc20a1b promoter) as the probe, while the 100-fold unlabeled mutated NRF2 binding sequence markedly reduced this competition, indicating that NRF2 binding sequence (−572/−561 bp of slc20a1b promoter) was bound by NRF2 ( Figure 4D). However, no band was found when we used NRF2 binding sequence (−1469/−1459 bp of slc20a2 promoter) as the probe, suggesting that this sequence was not bound by any factors (Figure 4E).
For the VDR binding assay, the 100-fold unlabeled VDR binding sequence competed for the binding when we used biotin-labelled VDR binding sequence (−260/−253 bp of slc20a1a promoter) as the probe, while the 100-fold unlabeled mutated VDR binding sequence markedly reduced this competition, indicating that VDR binding sequence (−260/−253 bp of slc20a1a promoter) was bound by VDR ( Figure 4F). The VDR binding elements located at −615/−609 bp of slc20a1b promoter ( Figure 4G) and at −1124/−1117 bp of the slc20a2 promoter ( Figure 4H) were functionally bound by VDR.

Effect of Pi Incubation on the Luciferase Activity of the Site-Mutagenesis Plasmids
To further determine whether Pi-induced decrease of slc20s promoter activities could be mediated by SREBP1, NRF2 and VDR elements, we conducted the site-directed mutation at their corresponding regions of slc20s promoters and used Pi to incubate the cells. Pi incubation significantly decreased three slc20s promoter activities. However, compared to the wild type pGl3-1813/+226 vector of slc20a1a, Pi-induced transcriptional inactivation of slc20a1a was abolished after the mutation of VDR element but not SREBP1 element ( Figure 5A). Compared with the wild-type pGl3-1801/+143 vector of slc20a1b, Pi-induced transcriptional inactivation of slc20a1b was abolished after the mutation of VDR element but not SREBP1 and NRF2 elements ( Figure 5B). Similarly, compared with the wild-type pGl3-1775/+211 vector of slc20a2, the Pi-induced transcriptional inactivation of slc20a2 was abolished after the mutation of VDR element but not the mutation of SREBP1 and NRF2 elements ( Figure 5C). and VDR site mutagenesis. (C) Effect of Pi incubation on the promoter activities of slc20a2 after SREBP1, NRF2 and VDR site mutagenesis. Student's t-test was undertaken to detect the differences between two treatments. Asterisk (*) indicates significant differences between control and Pi treatment group in the same plasmid (p < 0.05). Hash symbol (#) indicates significant differences between different 5′ unidirectional deletion plasmids under the same treatment (p < 0.05).

Effect of Pi Incubation on the mRNA Level of Three slc20s Gene in CIK Cells of Grass Carp
To identify whether the expressions of slc20s were regulated by Pi in grass carp, we cultured CIK cells and used Pi to incubate the cells. Compared to the control, 3 mM Pi incubation markedly reduced the mRNA abundances of three slc20s genes ( Figure 6). and VDR site mutagenesis. Student's t-test was undertaken to detect the differences between two treatments. Asterisk (*) indicates significant differences between control and Pi treatment group in the same plasmid (p < 0.05). Hash symbol (#) indicates significant differences between different 5 unidirectional deletion plasmids under the same treatment (p < 0.05).

Effect of Pi Incubation on the mRNA Level of Three slc20s Gene in CIK Cells of Grass Carp
To identify whether the expressions of slc20s were regulated by Pi in grass carp, we cultured CIK cells and used Pi to incubate the cells. Compared to the control, 3 mM Pi incubation markedly reduced the mRNA abundances of three slc20s genes ( Figure 6). Figure 6. Effect of Pi incubation on the mRNA level of slc20s in CIK cells. Student's t-test was undertaken to detect the differences between two treatments. Asterisk (*) indicates significant differences between control and Pi treatment group (p < 0.05).

Discussion
In mammals, the SLC20 family includes two members (SLC20A1 and SLC20A2), which are encode by slc20a1 and slc20a2 genes [13]. In contrast, the present study obtained three members of the SLC20 family (slc20a1a, slc20a1b and slc20a2) from grass carp. Thus, grass carp show more slc20a1 isoforms than mammals, which is attributable to the TSGD events [22]. Moreover, our study indicated that grass carp SLC20A1A, SLC20A1B and SLC20A2 possessed 11, 11 and 10 TMDs, respectively ( Figure S2). Similarly, 10-12 TMDs of SLC20s were also predicted in mammals [23,24]. Many amino acid residues important for Pi transport function in human, such as D28, E55, D506 and E575, were also conserved in the SLC20A2 from grass carp [25][26][27]. Additionally, Virkki et al. [25] suggested a conserved N-glycosylated site Asp−81, which was also found in the Asp−81 in grass carp of SLC20A2.
It is very important to explore the mRNA tissue expression of genes, which provides the knowledge into their potential functions. The present study indicated that the slc20a1a, slc20a1b and slc20a2 mRNA amounts were ubiquitously existent in tissues but varied (Figure 1), in consistence with other reports [9,28,29]. As a matter of fact, due to their broad distribution, they are considered to act as "housekeeping" roles for Pi transport into the cells [2,25]. Moreover, we found that slc20a1a mRNA amounts were the highest in the kidney and liver and the slc20a1b mRNA expression was predominant in the kidney, and that the slc20a2 mRNA amounts were predominant in the brain, suggesting their importance in these tissues. Similarly, Beck et al. [30] reported that slc20a1 was a critical gene for the normal growth and development of the liver. The predominant slc20a1a and slc20a1b mRNA expression in the kidney suggested that they shared the similar function in Pi reabsorption in the tissue. For slc20a2, Inden et al. [5] pointed out that the mRNA levels of slc20a2 were higher in the brain than those in the liver and kidney in mammals. Similarly, Jensen et al. [31] indicated that slc20a2 played important roles in keeping Pi homeostasis in the calcification of the brain.
In eukaryotes, in order to explore the mechanism of transcriptional initiation of genes, the first step is to identify their core promoter regions [32]. In the present study, three slc20s genes had different structures of the core promoters. For example, the core promoters of slc20a1a had both CCAAT-box (−39 bp) and TATA-box (−12 bp), the core promoters of slc20a1b only had TATA-box (−14 bp), but the core promoters of slc20a2 had CCAAT-box (−93 bp) and Sp1 sites (−72 bp, −10 bp), indicating that three slc20s had different regulatory mechanism for their transcriptional initiation ( Figure S1). Usually, TATA-less promoters possessed various Sp1 binding sites in their promoter regions [33]. In order to decipher the transcriptionally regulatory mechanisms, it is very important to identify the TFBSs of gene promoters [34]. SREBP1 is an important transcriptional factor that Figure 6. Effect of Pi incubation on the mRNA level of slc20s in CIK cells. Student's t-test was undertaken to detect the differences between two treatments. Asterisk (*) indicates significant differences between control and Pi treatment group (p < 0.05).

Discussion
In mammals, the SLC20 family includes two members (SLC20A1 and SLC20A2), which are encode by slc20a1 and slc20a2 genes [13]. In contrast, the present study obtained three members of the SLC20 family (slc20a1a, slc20a1b and slc20a2) from grass carp. Thus, grass carp show more slc20a1 isoforms than mammals, which is attributable to the TSGD events [22]. Moreover, our study indicated that grass carp SLC20A1A, SLC20A1B and SLC20A2 possessed 11, 11 and 10 TMDs, respectively ( Figure S2). Similarly, 10-12 TMDs of SLC20s were also predicted in mammals [23,24]. Many amino acid residues important for Pi transport function in human, such as D28, E55, D506 and E575, were also conserved in the SLC20A2 from grass carp [25][26][27]. Additionally, Virkki et al. [25] suggested a conserved N-glycosylated site Asp−81, which was also found in the Asp−81 in grass carp of SLC20A2.
It is very important to explore the mRNA tissue expression of genes, which provides the knowledge into their potential functions. The present study indicated that the slc20a1a, slc20a1b and slc20a2 mRNA amounts were ubiquitously existent in tissues but varied (Figure 1), in consistence with other reports [9,28,29]. As a matter of fact, due to their broad distribution, they are considered to act as "housekeeping" roles for Pi transport into the cells [2,25]. Moreover, we found that slc20a1a mRNA amounts were the highest in the kidney and liver and the slc20a1b mRNA expression was predominant in the kidney, and that the slc20a2 mRNA amounts were predominant in the brain, suggesting their importance in these tissues. Similarly, Beck et al. [30] reported that slc20a1 was a critical gene for the normal growth and development of the liver. The predominant slc20a1a and slc20a1b mRNA expression in the kidney suggested that they shared the similar function in Pi reabsorption in the tissue. For slc20a2, Inden et al. [5] pointed out that the mRNA levels of slc20a2 were higher in the brain than those in the liver and kidney in mammals. Similarly, Jensen et al. [31] indicated that slc20a2 played important roles in keeping Pi homeostasis in the calcification of the brain.
In eukaryotes, in order to explore the mechanism of transcriptional initiation of genes, the first step is to identify their core promoter regions [32]. In the present study, three slc20s genes had different structures of the core promoters. For example, the core promoters of slc20a1a had both CCAAT-box (−39 bp) and TATA-box (−12 bp), the core promoters of slc20a1b only had TATA-box (−14 bp), but the core promoters of slc20a2 had CCAAT-box (−93 bp) and Sp1 sites (−72 bp, −10 bp), indicating that three slc20s had different regulatory mechanism for their transcriptional initiation ( Figure S1). Usually, TATA-less promoters possessed various Sp1 binding sites in their promoter regions [33]. In order to decipher the transcriptionally regulatory mechanisms, it is very important to identify the TFBSs of gene promoters [34]. SREBP1 is an important transcriptional factor that regulates the genes involved in lipogenesis [35]. Several reports suggested that Pi influenced lipid deposition [36,37] and that SREBP1 increased lipogenesis and accelerated vascular calcification in mammals [15,16]. Interestingly, SLC20s also play key roles in cell calcification in mammals [5,38,39]. Our study identified the functional SREBP1 binding sites on the slc20s promoters ( Figure 4A-C). We speculated that SREBP1 mediated the transcriptional regulation of SLC20s expression. Moreover, we found that SREBP1 positively regulated the activities of slc20a1b and slc20a2 promoters, and negatively regulated the promoter activity of slc20a1a ( Figure 3A), suggesting that SREBP1 differentially regulated Pi transport by targeting different slc20s. Our study also indicated the differentially transcriptional regulatory mechanism for duplicated genes slc20a1a and slc20a1b. Studies point out that duplicated genes will result in neofunctionalization and subfunctionalization [40,41]. NRF2 is a crucial transcription factor that influences intrinsic resistance to oxidative stress [42]. Pi regulates the production of reactive oxygen species (ROS) [43]. Phosphate load increased oxidative stress and led to tissue damages and cytoxicity in various models [44,45]. In this study, we found that NRF2 positively regulated the promoter activity of slc20a1b, but not slc20a1a and slc20a2 (Figures 3B and 4D,E), indicating that slc20a1b might play important role than slc20a1a and slc20a2 under oxidative stress. Michigami et al. [13] suggested that NRF2 mediated the direct influences on extracellular Pi transport. Wei et al. [46] found that the activation of NRF2 signaling alleviated high phosphate-induced calcification of vascular smooth muscle cells by suppressing the ROS production. Moreover, several studies also suggested that NRF2 was involved in osteogenic differentiation and mineralization [47,48]. Considering that the important functions of SLC20 in osteogenic differentiation and cell calcification, we speculated that NRF2 probably regulated osteogenic differentiation and cell calcification by targeting slc20a1. The VDR is a zinc-finger transcription factor regulating the expression of several genes and important for the regulation of Pi homeostasis [12]. Our site-mutation and EMSA analysis identified the functional VDR binding sequences on slc20a1a, slc20a1b and slc20a2 promoters ( Figure 4F-H), and found that the mutation of VDR decreased the luciferase activities of three slc20 promoters (Figure 3). Taken together, it seems that three slc20 genes possessed not only the similar but also some different transcriptional regulatory mechanism.
Although slc20a1 and slc20a2 are expressed in various tissues, evidence was scarce about whether these slc20 transporters controlled the phosphate homeostasis at the cellular levels. Using cell lines from mammals, studies found that slc20a1 and slc20a2 mediated Pi transport and their effects are regulated by Pi levels [49]. In the present study, we found that 3 mM Pi incubation markedly reduced the mRNA abundances of three slc20s ( Figure 6) and decreased the activities of three slc20s promoters ( Figure 2). Similarly, Chien et al. [50,51] reported that phosphate starvation increased slc20a1 and slc20a2 gene expression and induced the posttranslational modifications of slc20a2 that accordingly activated phosphate transport. Other studies indicated that slc20a1 and slc20a2 had redundant roles in the regulation of Pi transport. For example, in mouse VSMCs, Crouthamel et al. [52] found that the slc20a2 knockdown in the slc20a1-deficient mouse VSMCs decreased Pi uptake, but the deletion of the slc20a1 gene alone had no effects on Pi uptake, probably due to the compensatory upregulation of slc20a2. In contrast, other studies indicated different functions in the two SLC20 transporters. For example, Villa-Bellosta and Sorribas [53] found that SLC20A1, not SLC20A2, was the only Pi transporter in the kidney that was not up-regulated in response to dietary Pi deprivation. Thus, it would be important to clarify the redundancy of SLC20A1 and SLC20A2 in Pi transport. Moreover, we found that VDR, but not SREBP1 and NRF2, mediated Pi-induced transcriptional inactivation of slc20a1a, slc20a1b and slc20a2 ( Figure 5), suggesting that VDR binding sites directly mediated the slc20s transcriptional response to Pi. Similarly, other studies indicated that slc20a1 and slc20a2 expression can be positively regulated by VDR [9,12]. Although SREBP1 and NRF2 did not directly mediate Pi-induced transcriptional inactivation of slc20s, they may be mediated by some other factors. For example, Kawamoto et al. [54] reported that high-fat diets increased the expression of slc20a1 and provoke phosphorus absorption in rats, which may directly mediated by SREBP1.

Experimental Grass Carp and Reagents
The grass carp juveniles were bought from a farm in Wuhan (China) and transported to our wet laboratory in Huazhong Agricultural University (HZAU) for the culture. CIK and HEK293T cell lines came from the Cell Resource Center in Fishery College of HZAU. All these media and reagents for cell cultures were purchased from Gibco company (ThermoFisher Scientific, Waltham, MA, USA). All these animal experiments followed the guideline of the Animal Experimentation Ethics Committee of HZAU (Wuhan, Hubei, China) and were approved by Ethics Committee of HZAU (Fish-2019-0418, 18 April 2019).

Sequence Clone and Plasmids Construction
Total RNA and genomic DNA were extracted from nine tissues of grass carp (heart, brain, intestine, kidney, liver, muscle, gill, spleen and skin) by using TRIzol RNAiso reagent (Invitrogen, Carlsbad, CA, USA) and commercial DNA extracted kit (Omega, Norcross, GA, USA). Total RNA was then reverse-transcribed to cDNA by using the oligo-dT primers and commercial cDNA Synthesis Kits (TaKaRa, Otsu, Japan). Core cDNA and promoter sequences of slc20a1a, slc20a1b and slc20a2 were obtained by RT-PCR, with the reference of the genome of grass carp [55]. The 3 -and 5 -end sequences were cloned via the nested 3 -and 5 -RACE PCR. Table S1 showed all these primers.

Sequence Analysis
We used BLAST network service at the NCBI (http://blast.ncbi.nlm.nih.gov/) to compare the nucleotide sequences with DNA sequences from the GenBank database. TMpred (http://www. ch.embnet.org/software/TMPRED_form.html) was utilized to predict the putative transmembrane regions, and the Clustal-W multiple alignment algorithm for analyzing the sequence alignments and percentage of amino acid conservation. Several online software programmes, such as MatInspector (http://www.genomatix.de/), the JASPAR database (http://jaspar.genereg.net/) and the TFSEARCH database (http://www.cbrc.jp/research/db/TFSEARCH.html), were utilized to analyze the potential transcription factor binding sites (TFBS).

Plasmid Transfections and Analysis of Luciferase Activities
HEK293T cells were cultured in DMEM medium with 10% fetal bovine serum (FBS) (Gibco, Carlsbad, CA, USA) in an incubator (5% CO 2 and 37 • C). Prior to the transfection, HEK293T cells were seeded at a density of 1.2 × 10 5 in 24-well plate. They were cultured for 24 h to reach 70-80% confluence. Lipofectamine™2000 (Invitrogen, Carlsbad, CA, USA) was utilized to transfect all these plasmids into HEK293T cells, based on the manufacture's protocol. The 400-ng reporter plasmids were co-transfected with 20 ng pRL-TK, the internal control with a Renilla luciferase reporter vector.
HEK293T cells were incubated for 24 h and collected to analyze the luciferase activity (Dual-Luciferase Reporter Assay System, Promega, San Luis Obispo, CA, USA). For an internal control, we added 20 ng of Renilla luciferase vector (pRL-TK) per well to normalize the efficiency of the transfection. After 4h, we replaced the transfection medium by 10% FBS-DMEM or 10% FBS-DMEM + 3 mM Pi (Pi was added as a mixture of NaH 2 PO 4 and Na 2 HPO 4 , pH7.4). The form and the concentration of Pi was selected according to our pilot trial and to previous in vitro studies [56][57][58][59]. Then, after 24-h incubation, cells were collected to determine the promoter activity, based on the manufacturer's instruction of the Dual-luciferase Reporter Assay System (Promega, San Luis Obispo, CA, USA). The relative luciferase activities were obtained by calculating the ratio of Firefly luciferase activity to Renilla luciferase activity. We conducted all these experiments in triplicates.

Site-Mutation Analysis of SREBP1, NRF2 and VDR Binding Sites on the Regions of Three Slc20s Promoters
In order to characterize the SREBP1, NRF2 and VDR binding sites on the slc20a1a, slc20a1b and slc20a2 promoters of grass carp, we used QuickChange II Site-Directed Mutagenesis Kit (Vazyme, Piscataway, NJ, US) to undertake the site-directed mutagenesis, and pGl3-1813/+226 of the slc20a1a vector, pGl3-1804/+143 of the slc20a1b vector and pGl3-1775/+221 of the slc20a2 vector as the templates, respectively. Several mutations were performed, as shown below: (1) Table S3. The DNA sequencing was utilized to confirm these mutations. Then, the Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA, USA) was utilized to co-transfect these plasmids and pRL-TK into HEK293T cells. After 4-h transfection, the medium was substituted with 10% FBS-DMEM or 10% FBS-DMEM + 3 mM Pi (Pi was added as a mixture of NaH 2 PO 4 and Na 2 HPO 4 , pH7.4). After 24-h incubation, we harvested the cells to determine the luciferase activities, based on the procedures mentioned above.

Electrophoretic Mobility-Shift Assay (EMSA)
The EMSA was undertaken to confirm the functional SREBP1, NRF2 and VDR binding sites on the three slc20s promoters according to our recent publications [60][61][62]. Nuclear and cytoplasmic extracts were prepared for after the methods of Read et al. [63]. Protein contents were determined by the BCA method [64]. The oligonucleotide probes were synthesized in the Tsingke company (Wuhan, China). Nuclear extracts (10 µg) were incubated for 30 min at the room temperature by using the binding buffer (20 mM HEPES, pH7.9, 1 mM MgCl 2 , 0.5 mM DTT, 4% Ficoll, 110 mM KCl, 0.2 µg Poly(dI-dC)). Then, the biotin-labeled double-stranded oligo nucleotides (Table S4) were added. The reaction continued for 30 min and then the electrophoresis was performed on 6% native polyacrylamide gels. For the competitive binding analysis, a 100-fold excess of unlabeled double-stranded DNA oligo with mutant binding site (Table S4) was added with the corresponding labeled one.

Effect of Pi Incubation on the mRNA Level of Slc20s in CIK Cells of Grass Carp
We cultured grass carp CIK cells in DMEM medium plus 10% FBS (Gibco, Carlsbad, CA, USA) in a SANYO incubator (28 • C and 5% CO 2 ). The CIK cells were seeded at a density of 1.2 × 10 5 in the 6-well plate. They were cultured until the 70-80% confluence was reached. Then we replaced the medium by 10% FBS-DMEM or 10% FBS-DMEM + 3 mM Pi (Pi was added as a mixture of NaH 2 PO 4 and Na 2 HPO 4 , pH7.4). After 24 h, we collected the CIK cells. The total RNAs were isolated from these CIK cells and reverse-transcribed to cDNA as the templates.

mRNA Expression after Quantitative Real-Time PCR (Q-PCR) Analysis
The slc20s mRNA levels in nine tissues and CIK cells were determined by Q-PCR analysis, based on our published protocols [60,61]. All of these primers are given in Table S5. Q-PCR assays were undertaken in a quantitative thermal cycler (MyiQ™ 2 Two Color Quantitative PCR Detection System, BIO-RAD, USA). Five housekeeping genes (β-actin, 18s-rrna, b2m, gapdh and ef-1α) were selected to analyze their transcription stability. Two most stable control genes (gapdh and 18s-rrna) were selected by geNorm software analysis [65]. The mRNA expression levels were calculated via the "delta-delta Ct" method [66].

Statistical Analysis
We showed the results with the means ± standard error of mean (SEM) at three independent experiments. Prior to statistical analysis, all data were tested for normality of distribution using the Kolmogornov-Smirnov test. The homogeneity of variances was tested using Barlett's test. For the slc20 tissue mRNA expression patterns analysis, the data were subjected to one-way ANOVA and Tukey's multiple range test. For assay of promoters' activities, Student's t-test was undertaken to detect the differences between two treatments. The significant level was set at p < 0.05. All of these data were analyzed by using the SPSS 19.0 software (SPSS Inc., Chicago, IL, USA).

Conclusions
For the first time, the full cDNA lengths and the promoters of three slc20 members (slc20a1a, slc20a1b and slc20a2) were isolated and characterized from fish. They shared similar TMDs to mammals and their mRNAs were existent among various tissues. Three transcription factors (SREBP1, NRF2 and VDR) regulated the transcription activities of three slc20s promoters and the transcriptional inactivations of slc20s were mediated by VDR under Pi treatment. Our study partially elucidated the mechanism of Pi homeostatic regulation of vertebrates and provided a good basis for dietary Pi addition and for the reduction of Pi discharge into the waters.

Conflicts of Interest:
The authors declare no conflict of interest.