Phosphate Transporter Regulator That Prevents Abnormal Hyperphosphatemia.

25 Renal type II sodium-dependent inorganic phosphate (Pi) transporters NaPi2a and NaPi2c 26 cooperate with other organs to strictly regulate the plasma Pi concentration. A high Pi load induces 27 the phosphaturic hormones parathyroid hormone (PTH) and fibroblast growth factor 23 (FGF23), 28 enhances urinary Pi excretion and prevents the onset of hyperphosphatemia. How FGF23 is 29 induced from the bones by a high Pi load and the setpoint of the plasma Pi concentration, however, 30 are unclear. Here, we investigated the role of transporter-associated protein (TRAP), found in 31 gene co-expression networks in NaPi2a and NaPi2c function. TRAP is localized in the renal 32 proximal tubules and interacts with NaPi2a. In TRAP-knockout (KO) mice, the serum FGF23 33 concentration was markedly increased but increased Pi excretion and hypophosphatemia were not 34 observed. In addition, TRAP-KO mice exhibit reduced NaPi2a responsiveness to FGF23 and PTH 35 administration. Furthermore, a dietary Pi load causes marked hyperphosphatemia and abnormal 36 NaPi2a regulation in TRAP-KO mice. Thus, TRAP is thought to be associated with FGF23 37 induction in bones and the regulation of NaPi2a to prevent an increase in the plasma Pi 38 concentration due to a high Pi load and kidney injury. 39 40 41 42 43 47 associated in the proximal The present study showed that TRAP-knockout (KO) mice 75 abnormal fluctuations in the plasma Pi levels in response to dietary Pi. TRAP binds to NaPi2a on the cell membrane and is considered to be involved in the regulation of NaPi2a by PTH and FGF23. The roles of TRAP in the control of plasma Pi are discussed. by analysis 2 were established by analysis by was considered significant.

12/13/21 12/13/21 h after FA treatment compared with TRAP +/+ mice, and the markedly high levels of serum FGF23 241 were maintained only in TRAP -/mice until 7 days after FA treatment (Fig. 10f, 10g). Renal 242 NaPi2a protein levels were significantly reduced in TRAP +/+ mice by FA administration, but only 243 slightly reduced in TRAP -/mice (Fig. 10h). In the present study, we investigated the roles of a strongly correlated molecule (TRAP) in 247 the GCNs and Pi metabolism. The TRAP protein was extremely highly expressed in the kidney 248 and localized at the apical membrane of the renal proximal tubules. Dietary Pi content regulates 249 the renal TRAP protein levels the same as NaPi2a. Immunoprecipitation experiments suggest that 250 TRAP interacts with the NaPi2a/NHERF1 complex (directly with NaPi2a, but not with NHERF1). 251 TRAP -/mice showed markedly increased serum FGF23 levels and significantly increased plasma 252 PTH levels. Interestingly, renal NaPi2a protein levels were not decreased despite the marked 253 increase in serum FGF23 levels. Based on studies of dietary Pi responses, the marked increase in 254 plasma Pi levels in TRAP -/mice are due to high Pi loading and may result from a resistance to 255 NaPi2a regulation by PTH and FGF23. Thus, an abnormal dietary Pi response in TRAP -/mice 256 leads to dysregulation of plasma Pi levels. 257 TRAP -/mice are characterized by 1) abnormally high serum FGF23 levels and 2) dietary 258 Pi response abnormalities in the regulation of the plasma Pi concentration. The present findings 259 suggest that TRAP binds to NaPi2a on the cell membrane and is involved in the internalization of 260 NaPi2a. In TRAP -/mice, no decrease in NaPi2a was observed despite high FGF23 and PTH 261 concentrations. On the other hand, NaPi2c was significantly decreased. In addition, NHERF1 was 262 significantly elevated in TRAP -/mice. For these reasons, TRAP is considered to be a component 263 of the NaPi2a/NHERF1 complex that receives signals from PTH and FGF23. 12/13/21 Although TRAP deficiency affects the NaPi2a/NHERF1 system, vitamin D-metabolizing 265 enzymes and NaPi2c regulation are considered normal. PTH and FGF23 downregulate the 266 NaPi2a/NHERF1 binary complex by activating 2 distinct signaling pathways that converge at 267 NHERF1 1 . The internalization and degradation of NaPi2a increase Pi excretion and depend on 268 activation of the ERK1/2 and serum/glucocorticoid-regulated kinase-1 pathways, resulting in 269 phosphorylation of NHERF1 29 . PTH signals activate protein kinases A and C 13, 30, 31 . Triggered by 270 phosphorylation of NHERF1, NaPi2a dissociates from NHERF1 and is then internalized 13 . Not 271 all signals from the receptor have been examined in detail, but some signals were activated. 272 Considering the changes in NaPi2c and vitamin D-metabolizing enzymes (cyp24a1, cyp27b1) in 273 TRAP -/mice and the effects of PTH/FGF23 administration, it is expected that they are 274 functioning normally except for the signal to NaPi2a. The interaction of NaPi2c with NHERF3 275 (PDZK1) is more important than that with NHERF1 32, 33 . In fact, NaPi2c expression is suppressed 276 in NHERF3-KO mice 33 . We previously reported differences in signals between the phosphaturic 277 action of FGF23 and the inhibitory effect on vitamin D synthesis 34 . Therefore, it is considered 278 that the effect of TRAP deficiency is limited to the control function of NaPi2a. More detailed 279 studies on the role of TRAP in NaPi2a regulation, such as the effect of NHERF1 on 280 phosphorylation, are needed. 281 Another feature of TRAP -/mice is enhanced FGF23 induction from the bone. High serum 282 FGF23 levels cause the pathology observed in a mouse model of X-linked hypophosphatemia 283 rickets (Hyp mice) 35, 36 . On the other hand, TRAP -/mice did not exhibit the abnormal bone 284 morphology seen in Hyp mice and we speculate that this is because TRAP -/mice do not develop 285 hypophosphatemia. The bone analysis data suggest that a high PTH concentration affects 286 fluctuations in the numbers of osteoblasts and osteoclasts. More detailed studies will help to 287 clarify the role of TRAP in bone. 288 12/13/21 High FGF23 induction in TRAP -/mice is improved by a low Pi diet. Therefore, renal TRAP 289 is expected to signal dietary Pi levels to bone FGF23. On the other hand, in an FA-induced renal 290 disorder model, a further increase in serum FGF23 concentration was observed in TRAP -/mice. 291 FGF23 induction is known to be independent of dietary Pi signals in an FA-acute kidney injury 292 model 37 . Therefore, the increase in FGF23 in TRAP -/mice may be independent of the signal of 293 renal damage. 294 The relationship between a-Klotho and TRAP as a mediator from the kidney remains 295 unclear. a-Klotho plays an important role in phosphate regulation by FGF23 as a co-receptor 296 for FGFR1 in the kidney 38, 39 . In TRAP -/mice, renal a-Klotho levels are reduced by 297 approximately 50%. Previous studies reported that a decrease in a-Klotho in the kidney triggers 298 the induction of FGF23 from the bone 38, 39 . Pi retention, progressive hyperphosphatemia, rising 299 FGF23 levels, and low a-Klotho expression are all observed in patients with progressive chronic 300 kidney disease [40][41][42] . The cause of the decrease in a-Klotho in the TRAP -/mice kidney may be the 301 high concentration of serum FGF23. We reported that renal a-Klotho levels are significantly 302 reduced in NaPi2a/NaPi2c double-KO mice, but no increase in plasma FGF23 is observed 27 . The 303 reason for the increase in FGF23 in TRAP -/mice cannot be explained by the decrease in a-Klotho 304 alone. A low-Pi diet suppresses the induction of FGF23 from the bone in a-Klotho-KO mice as 305 well, however, suggesting that it is mediated by the same signal 43,44 . The relationship between 306 TRAP and a-Klotho in the kidney was not investigated in the present study. 307 Finally, phosphaturic hormone is secreted in response to an excessive Pi load and acts on 308 the kidneys to promote Pi excretion. The NaPi2a/NHERF1 complex has an important role. TRAP 309 is expected to regulate the amount of NaPi2a in response to a Pi deficiency or excess and regulates 310 the responsiveness of phosphaturic hormone. For example, vitamin D treatment in Hyp mice 311 restores serum Pi levels by causing FGF23 resistance to NaPi2a/NHERF1 34, 45 . Thus, TRAP, a 12/13/21 strongly correlated molecule with NaPi2a in the GCNs, is thought to be involved in the regulation 313 of NaPi2a by phosphaturic hormones in the kidney and the prevention of hyperphosphatemia in 314 response to a high dietary Pi load ( Figure 11). 315 316

Experimental animals 318
All experimental procedures involving animals were conducted in accordance with the 319 Tokushima University School of Medicine and Osaka University Graduate School of Medicine 320 guidelines. This study was also carried out in compliance with the ARRIVE guidelines. All 321 procedures involving the use of animals were subjected to approval from Tokushima University Male and female C57B6/J mice were purchased from Charles River Laboratories Japan 325 (Yokohama, Japan). NaPi2a-KO and NaPi2c-KO mice were maintained as described previously 7, 326 46 . Mice were weaned at 4 weeks of age and provided free access to water and standard mouse 327 chow containing 0.8% phosphorus (Oriental MF; Oriental Yeas t , Osaka, Japan). 328 329

Generation of TRAP-KO mice 330
TRAP-deficient mice were generated by gene targeting. A targeting vector was constructed by 331 replacing the 129 genomic TRAP loci (exon1 through part of exon2) with neo r (Supplemental Fig.  332 S1a). The targeting vector was introduced into 129 days 14 embryonic stem (ES) cells by 333 electroporation, and clones that underwent homologous recombination were confirmed by 334 Southern blot analysis (Supplemental Fig. S1b). Genomic DNA was extracted from tail clippings 335 and amplified by PCR using specific primers (Supplemental Fig. S1c TransIT in vivo gene delivery system (Takara, Osaka, Japan) as described previously 26,27 . For 344 PTH injection, mice were administered bovine PTH (amino acid 1-34; Sigma) at a dose of 7.5 345 µg/100 g body weight, as described previously 50, 51 . 346 347

Metabolic cages to collect urine and fecal samples 348
Mice were individually placed in the metabolic cages at 10:00 for quantitative urine and fecal 349 collection for 24 h with free access to food and water. Fecal samples were ashed according to a 350 modified protocol, as described previously 34,47,52,53 . 351

Protein sample purification and immunoblotting 374
BBMVs prepared using the Ca 2+ precipitation method, cortical membrane, and whole homogenate 375 were obtained from mouse kidneys and used for immunoblotting and immunoprecipitation 376 analyses as described previously 34,49,52,53 . Protein samples were heated at 95°C for 5 min in 377 sample buffer in the presence of 2-mercaptoethanol and subjected to sodium dodecyl sulfate-378 polyacrylamide gel electrophoresis (SDS-PAGE). The separated proteins were transferred by 379 electrophoresis to Immobilon-P polyvinylidene difluoride (Millipore, Billerica, MA, USA) and 380 treated with diluted antibodies. Signals were detected using Immobilon Western (Millipore) 34 O.C.T. Compound (Sakura Finetek Japan Co. Ltd., Tokyo, Japan), and frozen in hexane at -80°C. 387 Frozen sections (5 µm thick) were collected onto MAS-coated slides (Matsunami Glass IND,Ltd.,388 Osaka, Japan) and air-dried 48 . For immunofluorescence microscopy, serial frozen sections were 389 incubated with primary antibodies overnight at 4°C. Alexa Fluor 488 anti-rabbit (Molecular 390 Probes, Eugene, OR, USA) and Alexa Fluor 568 anti-mouse (Molecular Probes) were used as 391 secondary antibodies for 60 min at room temperature 48, 49, 51-53 . Thereafter, the sections were 392 mounted with DAPI Fluoromount G TM (Southern Biotech, Birmingham, UK). Images were taken 393 with an A1R confocal laser scanning microscope system (Nikon, Tokyo, Japan). 394 395 Immunoprecipitation 396 Renal BBMVs of mice were lysed for 30 min at 4°C in TNE lysis buffer (20 mM Tris-HCl, 150 397 mM NaCl, 1 mM EDTA, 1mM EGTA, 1% TritonX-100, pH7.5), centrifuged for at 12,000 x g, 398 10 min at 4°C, and then supernatants were collected for immunoprecipitation. 399 Immunoprecipitation samples were adjusted to 200 µg proteins/ ml in tubes, and anti-NHERF1 400 (LS-C46891, LIFESPAN BIOSCIENCES, Seattle, WA, USA) or anti-TRAP antibodies were 401 added to tubes and rotated at 4°C overnight. Next, protein A agarose beads (Santa Cruz 402 Biotechnology, Dallas, TX, USA) were added to the tubes and rotated at 4°C for 1 h. Protein A 403 agarose beads were centrifuged at 3000 x g for 1 min at 4°C and washed with TNE lysis buffer 4 404 times before removing the supernatant and eluting in SDS sample buffer. Loading samples were 405 heated at 95°C for 5 min and then analyzed by SDS-PAGE using antibodies against NHERF1, 406 TRAP, and NaPi2a. 407 408 12/13/21

Antibodies 409
Rabbit anti-NaPi2a and NaPi2c polyclonal antibodies were generated as described previously and 410 used for immunoblotting and immunohistochemistry 7, 48, 51, 52 . Rabbit anti-TRAP polyclonal 411 antibodies were generated against mouse TRAP C-terminal fragments in rabbits as described 412 previously 54 . Briefly, purified GST-TRAP C-terminal (residues 90-243) fusion proteins were used 413 as antigens in rabbits. The purified IgG fractions were obtained from serums of the immunized 414 animals. Mouse anti-actin monoclonal antibodies (Millipore) were used as an internal control. 415 Horseradish peroxidase-conjugated anti-rabbit or anti-mouse IgG was utilized as the secondary 416 antibody (Jackson Immuno Research Laboratories, Inc, West Grove, PA, USA). The diluted 417 antibodies for immunoblotting were as follows: anti-NaPi2a (1:15000), anti-NaPi2c (1:1500),

Statistical Analysis 452
Data are expressed as means ± SE. Differences among multiple groups were analyzed by analysis 453 of variance followed by the Scheffe test. Differences between 2 experimental groups were 454 established by analysis of variance followed by Student's t test. A P value of less than 0.05 was 455 considered significant.

Competing interests 473
The authors declare no competing interests 474

TRAP interacts with NaPi2a and is involved in internalization by PTH and FGF23. Under a high 813
Pi load, urinary Pi excretion is enhanced by internalization of NaPi2a induced by PTH and FGF23,814 and hyperphosphatemia is prevented. In TRAP-KO mice, internalization of NaPi2a is resistant to 815 PTH and FGF23. As a result, Pi reabsorption is maintained, and plasma Pi concentration increases. 816 TRAP is expected to be a molecule that is associated with both the induction of phosphaturic 817 hormone (FGF23) and the regulation of NaPi2a in the proximal tubules. G; glomerulus, PT; 818 proximal tubule.