Noncanonical Sequences Involving NHERF1 Interaction with NPT2A Govern Hormone-Regulated Phosphate Transport: Binding Outside the Box

Na+/H+ exchange factor-1 (NHERF1), a multidomain PDZ scaffolding phosphoprotein, is required for the type II sodium-dependent phosphate cotransporter (NPT2A)-mediated renal phosphate absorption. Both PDZ1 and PDZ2 domains are involved in NPT2A-dependent phosphate uptake. Though harboring identical core-binding motifs, PDZ1 and PDZ2 play entirely different roles in hormone-regulated phosphate transport. PDZ1 is required for the interaction with the C-terminal PDZ-binding sequence of NPT2A (-TRL). Remarkably, phosphocycling at Ser290 distant from PDZ1, the penultimate step for both parathyroid hormone (PTH) and fibroblast growth factor-23 (FGF23) regulation, controls the association between NHERF1 and NPT2A. PDZ2 interacts with the C-terminal PDZ-recognition motif (-TRL) of G Protein-coupled Receptor Kinase 6A (GRK6A), and that promotes phosphorylation of Ser290. The compelling biological puzzle is how PDZ1 and PDZ2 with identical GYGF core-binding motifs specifically recognize distinct binding partners. Binding determinants distinct from the canonical PDZ-ligand interactions and located “outside the box” explain PDZ domain specificity. Phosphorylation of NHERF1 by diverse kinases and associated conformational changes in NHERF1 add more complexity to PDZ-binding diversity.

In addition to canonical interactions, NHERF1 PDZ1 and PDZ2 also form noncanonical interactions distal from the PDZ binding groove and located "outside the box". The presence of such noncanonical interactions is illustrated by various domain-selective binding of different proteins that selectively engage NHERF1 PDZ1 or PDZ2 despite the identical canonical binding sites ( Table 1). The non-canonical binding determinants are unique for individual PDZ domains [30]. There are several typical features making "outside the box" interactions specific. Frequently these interactions have an electrostatic nature [31]; some of them are regulated by phosphorylation [32][33][34] or involved in a long-range allosteric network [35]. Further, the helical-turn-helical (α3-turn-α4) extension is an essential structural element common for many PDZ domains and represents a universal element stabilizing PDZ-ligand binding [36,37]. Whether the α3-loop-α4 extension allosterically affects the binding with target ligands or directly contacts upstream N-terminal residues of the bound target remains to be established.
The main feature making NHERF1 unique is its C-terminal tail corresponding to a PDZ-ligand motif (-FS −2 NL 0 ). It is believed that -FS −2 NL 0 may occupy PDZ2 rather than PDZ1 to form a self-inhibited conformation [38]. The physiological role of the selfinhibited conformer remains to be elucidated. The putative ability of NHERF1 to dimerize through its C-terminal PDZ-motif (-FS −2 NL 0 ) was extensively investigated [26,[39][40][41][42]. Based on dynamic light scattering experiments, NHERF1 in solution is monodispersed [43]. Curiously, the NHERF1-NHERF1 association decreased in the presence of okadaic acid suggesting that phosphorylation may regulate dimerization [40]. A comparable effect was observed in the presence of the C-terminal PDZ-ligand motif of the β2-AR (-DSLL) [41]. The biological importance of NHERF1 dimerization may relate to its ability to assemble large multi-protein complexes and serve as an adapter for different classes of proteins. However, this theory has not been analyzed in detail.
NHERF1 is a phosphoprotein harboring 31 Ser and 9 Thr residues. Notably, there is a high-density Ser cluster in the flexible hinge region of NHERF1 linking PDZ2 and the EBD with 17 Ser/Thr residues in this segment. Ser 290 located in this region was identified as a phosphorylation site that allosterically regulates the interaction between NHERF1 and NPT2A through conformational changes near Glu 43 [44], a site that determines the binding specificity between PDZ1 and NPT2A [45]. Furthermore, dephosphorylationphosphorylation cycling of Ser 290 regulates NHERF1 self-assembly and NPT2A-dependent hormone-sensitive phosphate transport [44]. Remarkably, both PTH and FGF23 pathways contribute equally to the stabilization of the open state of NHERF1 via phosphorylation of Ser 290 by G Protein-coupled Receptor Kinase 6A (GRK6A) (Figure 1) [46]. PTH-induced phosphorylation of Ser 77 in conjunction with Thr 95 has a similar effect on regulating phosphate uptake as does Ser 290 phosphorylation. Ser 77 and Thr 95 , located in PDZ1, are predicted PKC phosphorylation sites ( Figure 1) [47,48]. Ser 162 is a defined PKCα phosphorylation site [43,49]. Notably, PKCα is a major pathway of PTH signaling [28,50,51] and is the only PKC isoform with a PDZ-recognition motif at its C-terminus (-SAV 672 ) [52,53]. Two other PKC phosphorylation sites, Ser 339 and Ser 340 , in the C-terminal region ( Figure  2), promote conformational reorganization in NHERF1 and facilitate phosphorylation of Ser 162 [43]. The role of this remarkable cooperativity on hormone-induced phosphorylation of NHERF1 has not yet been examined in vivo. Whether additional phospho-Ser/Thr sites regulate NHERF1-ligand interactions, conformational diversity, or phosphate transport is not known.

NHERF1 PDZ1 and PDZ2 Domains Are Not Interchangeable
The association between NHERF1 PDZ1 and its natural ligand NPT2A (SLC34A1) is a perfect example representing NHERF1 PDZ-ligand specificity. The NHERF1 PDZ1-NPT2A complex is required for NPT2A-mediated hormone-sensitive renal phosphate transport [15,54,55]. The interaction between PDZ1 and NPT2A occurs through the PDZligand C-terminal -T −2 R -1 L 0(639) sequence of NPT2A. The PDZ1-NPT2A complex is not assembled in the presence of NPT2A-Leu 639 Ala, a variant that has a defective C-terminal PDZ recognition motif (-T −2 R −1 L 0 /A) [4]. PDZ1 interacts with the C-terminal peptide ligand of NPT2A within the low micromolar range (3-5 µM), whereas the association between PDZ2 and NPT2A is insignificant [4,45]. The biological enigma is why and how PDZ1 and PDZ2 with identical conserved binding sites (GYGF and His) interact uniquely with NPT2A PDZ1. Analysis of possible binding determinants "outside the box" pointed to Glu 43 located in the α1 helix of PDZ1. MD simulations applied in our study predict the formation of an ionic pair between Glu 43 of PDZ1 and Arg −1 of the NPT2A C-terminal -T −2 R −1 L 0 motif [45]. Long-range MD simulations (~100-ns) provide clear evidence that the negatively charged side chain of Glu 43 and the side chain of His 27 are involved in electrostatic interactions with the positively charged side chain of Arg −1 of NPT2A (-T −2-R −1-L 0 ) and, more importantly, these interactions are persistent on the MD simulation time scale. In contrast to PDZ1, PDZ2 possesses Asp 183 and Asn 167 at the homologous positions. However, these residues do not form a stable interaction with Arg −1 [45]. We propose that the side chain of Asp 183 , which is relatively short compared to Glu 43 , is thereby unable to support direct interaction with the side chain of Arg −1 . Our MD simulations [45] and available NMR structure [36] demonstrate that the sidechain of Asp 183 is flexible and not involved in a stable interaction with Arg −1 and provide provisional support of this theory. Notably, the Glu 43 Asp mutation in PDZ1 leads to dramatic loss of affinity with a similar C-terminal -TRL motif of CFTR [36], another biological partner of NHERF1 [56]. CFTR, like NPT2A, largely associates with PDZ1 but not with PDZ2 [57][58][59].
Recently published X-ray structures and MD simulations of the PDZ2-CFTR complex suggest that the interaction between Asp 183 and Arg −1 is underestimated and can be an important element of the complex [60]. However, how the association between Asp 183 and Arg −1 regulates binding affinity between PDZ2 and CFTR remains unresolved.
To explore specificity of Glu 43 and His 27 on the interaction between PDZ1 and NPT2A we generated a PDZ1 variant where Glu 43 and His 27 were replaced by Asp and Asn, respectively. As anticipated, PDZ1 with the Glu 43 Asp/His 27 Asn mutations eliminates binding between PDZ1 and NPT2A. The relevance of Glu 43 and His 27 on NPT2Adependent PTH-sensitive phosphate transport was validated by measuring phosphate uptake in OKH cells expressing the Glu 43 Asp/His 27 Asn-NHERF1 variant. As we presumed, Glu 43 Asp/His 27 Asn-NHERF1 blocks basal phosphate transport and is refractory to PTH [44]. Notably, when Asp 183 and Asn 167 in PDZ2 were replaced respectively by Glu and His, the corresponding residues in PDZ1, the Asp 183 Glu/Asn 167 His rescue variant bound the NPT2A C-terminal -TRL motif ( Figure 3) [45]. However, Asp 183 Glu/Asn 167 His-NHERF1 did not support NPT2A-dependent PTH-sensitive phosphate uptake (unpublished observations). As expected, mutation of Glu 43 to Gly in PDZ1 of NHERF1 blocked phosphate uptake (Figure 4) [44]. Experimental study confirms that PDZ1 is required for formation of the NPT2A complex, and regulation of NPT2A-mediated hormone-sensitive phosphate transport. This finding strongly supports the concept that the specificity of the PDZ1 domain is not determined by the conserved subset of residues (-23 GYGF 26 -and His 72 ) but rather by the "outside the box" determinant Glu 43 . Support for this conclusion also comes from thermodynamic parameters determined by isothermal microcalorimetry [45]. Substitution of Glu 43 by Asp increases enthalpy (∆H o ) and provides a jump in entropy (∆S o ) making the interaction unfavorable. Mutation of His 27 to Asn has a minor effect. A small change in the free energy (∆∆G o ), 0.8 and 0.4 kcal/mol for Glu 43 →Asp and His 27 →Asn, respectively, is attributed to the enthalpy-entropy compensation [45] illustrated in Figure 5.
NMR structure [36] demonstrate that the sidechain of Asp is flexible and not inv in a stable interaction with Arg −1 and provide provisional support of this theory. No the Glu 43 Asp mutation in PDZ1 leads to dramatic loss of affinity with a similar C-ter -TRL motif of CFTR [36], another biological partner of NHERF1 [56]. CFTR, like NP largely associates with PDZ1 but not with PDZ2 [57][58][59]. Recently published X-ray tures and MD simulations of the PDZ2-CFTR complex suggest that the interactio tween Asp 183 and Arg −1 is underestimated and can be an important element of the com [60]. However, how the association between Asp 183 and Arg −1 regulates binding af between PDZ2 and CFTR remains unresolved.
To explore specificity of Glu 43 and His 27 on the interaction between PDZ1 and N we generated a PDZ1 variant where Glu 43 and His 27 were replaced by Asp and As spectively. As anticipated, PDZ1 with the Glu 43 Asp/His 27 Asn mutations eliminates ing between PDZ1 and NPT2A. The relevance of Glu 43 and His 27 on NPT2A-depe PTH-sensitive phosphate transport was validated by measuring phosphate upta OKH cells expressing the Glu 43 Asp/His 27 Asn-NHERF1 variant. As we presu Glu 43 Asp/His 27 Asn-NHERF1 blocks basal phosphate transport and is refractory to [44]. Notably, when Asp 183 and Asn 167 in PDZ2 were replaced respectively by Glu and the corresponding residues in PDZ1, the Asp 183 Glu/Asn 167 His rescue variant boun NPT2A C-terminal -TRL motif ( Figure 3) [45]. However, Asp 183 Glu/Asn 167 His-NH did not support NPT2A-dependent PTH-sensitive phosphate uptake (unpublished o vations). As expected, mutation of Glu 43 to Gly in PDZ1 of NHERF1 blocked phos uptake ( Figure 4) [44]. Experimental study confirms that PDZ1 is required for form of the NPT2A complex, and regulation of NPT2A-mediated hormone-sensitive phos transport. This finding strongly supports the concept that the specificity of the PDZ main is not determined by the conserved subset of residues (-23 GYGF 26 -and His 7 rather by the "outside the box" determinant Glu 43 . Support for this conclusion also c from thermodynamic parameters determined by isothermal microcalorimetry [45] stitution of Glu 43 by Asp increases enthalpy (ΔH ο ) and provides a jump in entropy making the interaction unfavorable. Mutation of His 27 to Asn has a minor effect. A change in the free energy (ΔΔG ο ), 0.8 and 0.4 kcal/mol for Glu 43 →Asp and His 27 → respectively, is attributed to the enthalpy-entropy compensation [45] illustrated in F 5.

The Role of the NPT2A Internal -TRL-Motif Remains to be Explored
In addition to its canonical C-terminal -T −2 RL 0 PDZ-binding motif, NPT2A po internal PDZ -494 T −2 RL 0 -recognition sequence that has not been characterized. Th minal motif is critical for interaction with NHERF1 [4]. Whether both PDZ motifs ute to or are required for proper localization and function of NPT2A and hormon is unknown. Two disease-associated mutations (Arg 495 His, Arg 495 Cys) have recen described in the putative internal PDZ motif (-494 TRL 496 -), while a third (Ser 585 Pr cated at the carboxy-terminal region. All are associated with elevated renal phosp cretion and consequent hypophosphatemia [10,11,61]. NPT2A-Arg 495 His and N Ser 585 Pro display different cell localization compared to wild-type (WT) NPT2A [ carboxy-terminal PDZ-binding motif (-T −2 R -1 L 0 ) [4,45,62] is required for cell me NPT2A localization and is necessary for hormone-regulated phosphate transpo The involvement of the putative internal NPT2A PDZ motif in association with N and its role in hormone action is not described. Pilot results indicate that mutatio internal PDZ motif interferes with PTH and FGF23 action and inhibits regulated ( Figure 6) despite normal binding to NHERF1 (unpublished work).

The Role of the NPT2A Internal -TRL-Motif Remains to be Explored
In addition to its canonical C-terminal -T −2 RL 0 PDZ-binding motif, NPT2A pos internal PDZ -494 T −2 RL 0 -recognition sequence that has not been characterized. Th minal motif is critical for interaction with NHERF1 [4]. Whether both PDZ motifs c ute to or are required for proper localization and function of NPT2A and hormon is unknown. Two disease-associated mutations (Arg 495 His, Arg 495 Cys) have recent described in the putative internal PDZ motif (-494 TRL 496 -), while a third (Ser 585 Pr cated at the carboxy-terminal region. All are associated with elevated renal phosph cretion and consequent hypophosphatemia [10,11,61]. NPT2A-Arg 495 His and N Ser 585 Pro display different cell localization compared to wild-type (WT) NPT2A [6 carboxy-terminal PDZ-binding motif (-T −2 R -1 L 0 ) [4,45,62] is required for cell me NPT2A localization and is necessary for hormone-regulated phosphate transpo The involvement of the putative internal NPT2A PDZ motif in association with N and its role in hormone action is not described. Pilot results indicate that mutatio internal PDZ motif interferes with PTH and FGF23 action and inhibits regulated ( Figure 6) despite normal binding to NHERF1 (unpublished work).

The Role of the NPT2A Internal -TRL-Motif Remains to Be Explored
In addition to its canonical C-terminal -T −2 RL 0 PDZ-binding motif, NPT2A possess an internal PDZ -494 T −2 RL 0 -recognition sequence that has not been characterized. The C-terminal motif is critical for interaction with NHERF1 [4]. Whether both PDZ motifs contribute to or are required for proper localization and function of NPT2A and hormone action is unknown. Two disease-associated mutations (Arg 495 His, Arg 495 Cys) have recently been described in the putative internal PDZ motif (-494 TRL 496 -), while a third (Ser 585 Pro) is located at the carboxy-terminal region. All are associated with elevated renal phosphate excretion and consequent hypophosphatemia [10,11,61]. NPT2A-Arg 495 His and NPT2A-Ser 585 Pro display different cell localization compared to wild-type (WT) NPT2A [61]. The carboxy-terminal PDZ-binding motif (-T −2 R -1 L 0 ) [4,45,62] is required for cell membrane NPT2A localization and is necessary for hormone-regulated phosphate transport [4,6]. The involvement of the putative internal NPT2A PDZ motif in association with NHERF1 and its role in hormone action is not described. Pilot results indicate that mutation of the internal PDZ motif interferes with PTH and FGF23 action and inhibits regulated uptake ( Figure 6) despite normal binding to NHERF1 (unpublished work).

NPT2A-Dependent Hormone-Inhibitable Phosphate Transport Requires Asso tion between PDZ2 and GRK6A
G protein-coupled receptor kinase 6A (GRK6A), a natural partner of NHERF sesses a canonical PDZ ligand (-T −2 RL 0 ) at its C-terminus. GRK6A, like NPT2A, ass with NHERF1 PDZ domains through its C-terminal motif (-T −2 R −1 L 0 ). Knocking Grk6a by siRNA blocks Npt2a-dependent phosphate uptake in response to PT Thus, GRK6A is an essential regulatory component of NPT2A-dependent PTH-se phosphate transport and corroborates previous findings that GRK6A pharmacolog hibitors abolish PTH action [44]. Binding affinities (3-5 µM) for the C-terminal PDZ of GRK6A [63] or NPT2A (22 aa) [45] with NHERF1 are comparable and suggest t binding mechanism is presumable identical. NPT2A and GRK6A interact with PDZ a greater affinity than to PDZ2, thereby confirming that PDZ1 is naturally optim bind the -TRL sequence. Intriguingly, a minor interaction between GRK6A and P nonetheless critical for constitutive or PTH-induced phosphorylation of NHERF1 a [44,46], and for PTH-sensitive phosphate transport [44]. This finding highlights a b cal puzzle as to how PDZ2 becomes accessible to GRK6A and the role of PTH and in this process. Again, canonical PDZ recognition sites (-GYGF-and His) required binding of the C-terminal Leu 0 and Thr −2 of GRK6A cannot explain the observed s ity for PDZ2. Furthermore, mutation of the NHERF1 PDZ2 core-binding GYGF (N1P2-GAGA) decreased but did not abolish phosphate transport in response t (Figure 7). This finding strongly suggests that the PDZ2 domain retains its ability t act with the C-terminus of GRK6A in vivo. Clearly, binding determinants "outs box" control the formation of the PDZ2-GRK6A complex. Analysis of residues th contribute to the binding pointed to Ser 162 , known as a PKCα phosphorylation site man NHERF1 [43,49]. Notably, NHERF1 homologs (mouse, rabbit) harbor Asn at t responding position. PKCα action is unique for PDZ2 inasmuch as PDZ1 has Asn homologous location. The published docking structure of PDZ2 interacting with nate C-terminal peptide demonstrates that Ser 162 is masked by the C-terminal tail [4 in vivo significance of post-translational modification of Ser 162 is unknown. We in studies to elucidate the role of NHERF1 Ser 162 on NPT2A-dependent phosphate tra Ala replacement at Ser 162 (NHERF1 Ser 162 Ala) diminished PTH-inhibitable pho transport suggesting that Ser 162 is an essential regulator of hormone-sensitive pho

NPT2A-Dependent Hormone-Inhibitable Phosphate Transport Requires Association between PDZ2 and GRK6A
G protein-coupled receptor kinase 6A (GRK6A), a natural partner of NHERF1, possesses a canonical PDZ ligand (-T −2 RL 0 ) at its C-terminus. GRK6A, like NPT2A, associates with NHERF1 PDZ domains through its C-terminal motif (-T −2 R −1 L 0 ). Knocking down Grk6a by siRNA blocks Npt2a-dependent phosphate uptake in response to PTH [63]. Thus, GRK6A is an essential regulatory component of NPT2A-dependent PTH-sensitive phosphate transport and corroborates previous findings that GRK6A pharmacological inhibitors abolish PTH action [44]. Binding affinities (3-5 µM) for the C-terminal PDZ ligand of GRK6A [63] or NPT2A (22 aa) [45] with NHERF1 are comparable and suggest that the binding mechanism is presumable identical. NPT2A and GRK6A interact with PDZ1 with a greater affinity than to PDZ2, thereby confirming that PDZ1 is naturally optimized to bind the -TRL sequence. Intriguingly, a minor interaction between GRK6A and PDZ2 is nonetheless critical for constitutive or PTH-induced phosphorylation of NHERF1 at Ser 290 [44,46], and for PTH-sensitive phosphate transport [44]. This finding highlights a biological puzzle as to how PDZ2 becomes accessible to GRK6A and the role of PTH and FGF23 in this process. Again, canonical PDZ recognition sites (-GYGF-and His) required for the binding of the C-terminal Leu 0 and Thr −2 of GRK6A cannot explain the observed specificity for PDZ2. Furthermore, mutation of the NHERF1 PDZ2 core-binding GYGF motif (N1P2-GAGA) decreased but did not abolish phosphate transport in response to PTH (Figure 7). This finding strongly suggests that the PDZ2 domain retains its ability to interact with the C-terminus of GRK6A in vivo. Clearly, binding determinants "outside the box" control the formation of the PDZ2-GRK6A complex. Analysis of residues that may contribute to the binding pointed to Ser 162 , known as a PKCα phosphorylation site in human NHERF1 [43,49]. Notably, NHERF1 homologs (mouse, rabbit) harbor Asn at the corresponding position. PKCα action is unique for PDZ2 inasmuch as PDZ1 has Asn 22 at the homologous location. The published docking structure of PDZ2 interacting with its cognate C-terminal peptide demonstrates that Ser 162 is masked by the C-terminal tail [43]. The in vivo significance of post-translational modification of Ser 162 is unknown. We initiated studies to elucidate the role of NHERF1 Ser 162 on NPT2A-dependent phosphate transport. Ala replacement at Ser 162 (NHERF1 Ser 162 Ala) diminished PTH-inhibitable phosphate transport suggesting that Ser 162 is an essential regulator of hormone-sensitive phosphate uptake. Unexpectedly, phosphomimic Ser 162 Asp mutation in NHERF1 disrupts basal phosphate transport and blocks PTH action on phosphate uptake (Figure 8) [63]. Clearly, the discrepancy in charge and size between phosphate and carboxylate groups of phosphorylated Ser and Asp, respectively explains the observed difference. Since the phosphomimic mutation is structurally remote and cannot directly mediate the interaction between PDZ1 and NPT2A we assume that NHERF1-Ser 162 Asp exerts a large conformational change that affects the interaction between PDZ1 and NPT2A. There is no high-resolution structure of full-length NHERF1. We speculate that incorporation of Asp 162 with the negatively charge carboxylate group interrupts the self-inhibited NHERF1 conformation [38], releases the C-terminal tail, and increases conformational dynamics of NHERF1. The structurally disordered C-terminal tail may adopt a variety of conformations and screen the association between PDZ1 and NPT2A. Another possibility is that the C-terminal tail engages PDZ1 and interferes with NPT2A binding. In contrast to NHERF1-Ser 162 Asp, the double phosphomimic mutation at Ser 339 Asp/Ser 340 Asp increases the binding affinity of both PDZ1 and PDZ2 for CFTR through conformational changes in the linker regions [43] and long-range allosteric cooperativity in NHERF1 [36]. Of note, Ser 339 /Ser 340 are PKC phosphorylation sites [43,49] like Ser 162 [46]. It was suggested that Ser 339 /Ser 340 phosphorylation promotes phosphorylation of Ser 162 [43]. Whether this cooperativity exists in vivo remains to be established.
Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW mimic mutation is structurally remote and cannot directly mediate the interact tween PDZ1 and NPT2A we assume that NHERF1-Ser 162 Asp exerts a large conform change that affects the interaction between PDZ1 and NPT2A. There is no high-res structure of full-length NHERF1. We speculate that incorporation of Asp 162 with t atively charge carboxylate group interrupts the self-inhibited NHERF1 conformati releases the C-terminal tail, and increases conformational dynamics of NHERF structurally disordered C-terminal tail may adopt a variety of conformations and the association between PDZ1 and NPT2A. Another possibility is that the C-term engages PDZ1 and interferes with NPT2A binding. In contrast to NHERF1-Ser 162 A double phosphomimic mutation at Ser 339 Asp/Ser 340 Asp increases the binding aff both PDZ1 and PDZ2 for CFTR through conformational changes in the linker regio and long-range allosteric cooperativity in NHERF1 [36]. Of note, Ser 339 /Ser 340 a phosphorylation sites [43,49] like Ser 162 [46]. It was suggested that Ser 339 /Ser 340 phos lation promotes phosphorylation of Ser 162 [43]. Whether this cooperativity exists remains to be established.  To explore the biochemical role of phospho-Ser 162 (pSer) and its impact on th  . N1P2-GAGA-NHERF1 and WT-NHERF1 support PTH-inhibitable phosphate uptake, whereas N1P1-GAGA does not. Results report the mean ± SEM (n = 4, **** p < 0.0001, ANOVA) [63]. mimic mutation is structurally remote and cannot directly mediate the interacti tween PDZ1 and NPT2A we assume that NHERF1-Ser 162 Asp exerts a large conform change that affects the interaction between PDZ1 and NPT2A. There is no high-reso structure of full-length NHERF1. We speculate that incorporation of Asp 162 with th atively charge carboxylate group interrupts the self-inhibited NHERF1 conformatio releases the C-terminal tail, and increases conformational dynamics of NHERF structurally disordered C-terminal tail may adopt a variety of conformations and the association between PDZ1 and NPT2A. Another possibility is that the C-termi engages PDZ1 and interferes with NPT2A binding. In contrast to NHERF1-Ser 162 A double phosphomimic mutation at Ser 339 Asp/Ser 340 Asp increases the binding affi both PDZ1 and PDZ2 for CFTR through conformational changes in the linker regio and long-range allosteric cooperativity in NHERF1 [36]. Of note, Ser 339 /Ser 340 ar phosphorylation sites [43,49] like Ser 162 [46]. It was suggested that Ser 339 /Ser 340 phos lation promotes phosphorylation of Ser 162 [43]. Whether this cooperativity exists remains to be established.  To explore the biochemical role of phospho-Ser 162 (pSer) and its impact on th action between NHERF1 PDZ2 and GRK6A, pSer 162 was genetically introduced in Figure 8. Ser 162 is essential for PTH-inhibitable phosphate uptake. OKH cells were transiently transfected with WT-NHERF1 or with Ser 162 Ala-NHERF1 or Ser 162 Asp-NHERF1. Cells were treated with vehicle or with 100 nM PTH . Results report the mean ± SEM (n = 4, **** p < 0.0001, ANOVA) [63].
To explore the biochemical role of phospho-Ser 162 (pSer) and its impact on the interaction between NHERF1 PDZ2 and GRK6A, pSer 162 was genetically introduced in recombinant PDZ2 (133-300 aa) using amber codon suppression [64]. Previously, semi-synthesis was effectively applied to generate site-specific phosphorylated PDZ domains [33]. Here, we used amber codon suppression to genetically encode pSer at position 162 [63], which in the future will allow to generate phosphorylated full-length proteins comprising both PDZ domains. Replacement of Ser 162 by pSer 162 in recombinant PDZ2 permits estimating the binding affinity between pSer 162 -PDZ2 and the C-terminal PDZ-ligand of GRK6A. Fluorescence anisotropy (FA) was applied to measure dissociation constants (K D 's) between pSer 162 -PDZ2, Ser 162 Ala-PDZ2, wild type PDZ2 (133-300 aa) and FITC-labeled GRK6A (22 aa). The results demonstrate that the K D values for the interaction between WT PDZ2 or Ser 162 Ala-PDZ2 and GRK6A were comparable (51.3 ± 0.4, 43.4 ± 0.4 µM) and double that for pSer 162 -PDZ2 (26.1 ± 0.8 µM), thus, demonstrating that the binding affinity of PDZ2 is regulated by phosphorylation of Ser 162 . Site-specific incorporation of pSer 162 applied for NHERF1 PDZ2 introduces an analogous but not identical functional group compared to phosphomimetic mutagenesis. The phosphate group has a -2 negative charge compared to the single negative charge of the Asp carboxylate group. The increased size of the sidechain may importantly perturb the local protein conformation [65]. MD simulations provide structural details associated with incorporation of pSer 162 (Figure 9). Three main observations come from the simulation studies. First, pSer 162 forms an electrostatic interaction with the positively charged side chain of Arg −1 of the C-terminal TRL motif of GRK6A and therefore has a significant impact on dynamics and conformational flexibility of the C-terminus of GRK6A. Second, pSer 162 promotes the formation of the electrostatic network (pSer 162 -Arg −1 -Asp 183 ) wherein the sidechain of Asp 183 changes its orientation and moves toward the sidechain of Arg −1 . Third, the GYGF/GAGA substitution in the carboxylatebinding site of pSer 162-PDZ2 does not impede the interaction with GRK6A. Consequently, NHERF1 with the modified core-binding motif (N1P2-GAGA) supports PTH-sensitive phosphate uptake (Figure 7). Thus, a strong stabilizing influence of pSer 162 underscores the limitation and potential hazard of using phosphomimetics to draw conclusions about phosphorylation and demonstrates the strength of introducing site-specific pSer using experimental and computational methods.

Phosphorylation of Ser290 Controls NHERF1 Conformation and Interactions with NPT2A
Recent combined NMR and small-angle neutron scattering (SANS) experiments revealed that full-length NHERF1 cannot be characterized by a single conformation. Rather, NHERF1 represents an ensemble of diverse PDZ configurations connected by flexible linkers with the C-terminal unstructured tail [66]. Notably, the flexible linker connecting PDZ2 and the C-terminus has 17 Ser/Thr residues. Compared to structurally determined rigid domains, intrinsically disordered regions typically contain a high density of phosphorylation sites [67]. Site-specific phosphorylation within these regions promotes structurally relevant conformational transitions that affect protein function [67,68]. Reversible post-translational modification of Ser/Thr residues within this region may regulate NHERF1 activity and signaling. We showed that dephosphorylation-phosphorylation of Ser 290 located in this flexible linker regulates the association between NHERF1 PDZ1 and NPT2A. Preventing Ser 290 phosphorylation with a phosphoresistant mutation (Ser 290 Ala) or pharmacologically inhibiting the action of GRK6A kinase decreases the binding of NTP2A to NHERF1 and reduces PTH-sensitive phosphate transport [44]. Evidently, Ser 290 determines regulated stability of the NHERF1-NPT2A complex through long range allosteric communication. Hydrogen deuterium exchange mass spectrometry (HDX-MS) analysis was used to analyze the region in NHERF1 undergoing conformational changes along the dephosphorylationphosphorylation cycle of Ser 290 [44]. In addition to the linker region between PDZ2 and EBD, substantial conformational changes were found in PDZ1 near Glu 43 , a critical determinant for NPT2A binding [44,45]. Thus, dephosphorylation-phosphorylation of Ser 290 allosterically regulates the conformation of the side chain of Glu 43 and switches the interaction on and off with Arg −1 of the C-terminal -TRL motif of NPT2A and thereby controls PTH-sensitive phosphate uptake.  Recent combined NMR and small-angle neutron scattering (SANS) experiments revealed that full-length NHERF1 cannot be characterized by a single conformation. Rather, NHERF1 represents an ensemble of diverse PDZ configurations connected by flexible linkers with the C-terminal unstructured tail [66]. Notably, the flexible linker connecting  Another example of allosteric regulation within NHERF1 relates to Leu 110 Val, Glu 225 Lys and Arg 153 Gln mutations identified in patients with hypophosphatemia [9]. The diseaseassociated mutations decrease or abolish the interaction between PDZ1 and NPT2A and block PTH-sensitive phosphate transport [4]. It was suggested that Arg 153 Gln and Glu 225 Lys, located in PDZ2, stabilize a self-inhibited conformation in which the NHERF1 C-terminus -FS −2 NL 0 , a PDZ recognition motif itself, is engaged in the PDZ2 domain, masking PDZ1 and thereby disrupting the interaction with NPT2A [4]. Remarkably, when the C-terminal Leu 0 of the NHERF1 PDZ ligand, required for engagement with PDZ2, was exchanged for Ala (NHERF1 Leu 0 Ala) and was paired with the Arg 153 Gln (NHERF1 Arg 153 Gln/Leu 358 Ala) disease mutation, the interaction with NPT2A was rescued and PTH-sensitive phosphate transport was restored and indistinguishable from that of wild-type NHERF1 [4], notably in the ongoing presence of the naturally occurring Arg 153 Gln mutation. Leu 110 Val, a similar dysfunctional mutant, locates in the α4 helix of the carboxy-terminal helix-turn-helix extension (α3-loop-α4) of PDZ1 [36]. This subdomain of 30 residues allosterically regulates the binding affinity between NHERF1 PDZ domains and their cellular targets [69]. NMR studies suggested that an extensive hydrophobic network between PDZ1 and the α3-loop-α4 extension stabilizes the structure with Leu 110 serving as a part of this hydrophobic network (64). It was proposed that mutation of Leu 110 to Val could potentially disrupt the packing interactions and reduce the stability and affinity of the PDZ1 domain [36]. Replacement of Leu 110 , even with a shorter hydrophobic side chain of Val, will likely disturb the hydrophobic network inside the α3-loop-α4 extension and promote its rearrangement. In addition, conformational changes associated with such an Leu 110 Val mutation may limit phosphorylation of Thr 95 , a PKC phosphorylation site located near Leu 110 , and thereby disrupt cooperativity between Thr 95 and Ser 77 phosphorylation events required for normal NPT2A-dependent PTH-inhibitable phosphate transport [70]. These conjectures require experimental verification.

"Outside the Box" Determinants Are Involved in the Interaction between NHERF1 PDZ Domains and PTHR
PTHR, a Family B GPCR, plays a key role in mineral-ion metabolism and bone physiology [2,22]. The C-terminal sequence of PTHR represents a typical PDZ binding motif (-T −2 -V-M 0 ) with upstream Glu residues at positions -3, -5, and -6 (-E −6 -E −5 -W-E −3 -T −2 -V-M 0 ). Although crystallographic or NMR structures for the complex between the PDZ domains of NHERF1 and PTHR have not been solved, molecular determinants of the interaction were characterized by different approaches. Biochemical studies [18,71] and Molecular Dynamics simulations [62] predict that both PDZ domains engage in PDZ-ligand interactions with C-terminal Met 0 and Thr −2 of PTHR. The promiscuous -T −2 -V-M 0 motif permits PTHR to bind PDZ1 or PDZ2, or both through canonical PDZ-ligand interactions established between the conserved GYGF motif of PDZ1/PDZ2 and the His 72 /His 212 residue at the top of the α2 helix. Although these canonical interactions are essential components of the binding, determinants outside the PDZ-ligand pocket enhance formation of the binary [NHERF1-PTHR] complex [18,29]. MD simulations on a~70-ns time scale provide clear evidence that Arg 40 and His 27 or Arg 180 of PDZ1 and Asn 169 in PDZ2 are involved in forming stable interactions with Glu −3 of PTHR [29,62]. Fluorescence anisotropy (FA), and isothermal titration calorimetry (ITC) measurements experimentally validated and extended computational predictions by showing that Glu −3 to Ala substitution destabilizes the complex [29]. This finding corroborates X-ray crystallographic [23,24] and NMR studies [36] that demonstrated the formation of electrostatic interactions between Arg 40 /Arg 180 of PDZ1/PDZ2 and Asp −3 of the C-terminal PDZ-binding motif of CFTR (-DTRL). Similar electrostatic interactions between NHERF1 PDZ1/PDZ2 and Asp −3 from the C-terminal PDZ-binding motif of the β-AR (-ESLL) were confirmed by X-ray crystallography [23,24]. The determined binding affinities in the nM-µM range for PDZ1 or PDZ2 domains bound the C-terminal peptide ligands of CFTR or PTHR support the formation of favorable electrostatic interactions [29,36,43,72].
In addition to the Arg 40 /Arg 180 -Glu −3 pair, residues from the β2-β3 loop of PDZ1/PDZ2 and the negatively charged Glu residues at positions -5 and -6 of the PTHR C-terminus are also involved in the interactions [29]. The β2-β3 loop is flexible. This explains why the impact of the residues from the loop on the binding with the upstream residues of the PTHR (E −6 E −5 -WETVM) is challenging to estimate by MD simulation [29]. The affinities determined by a plate-binding assay demonstrate that mutation of Glu 585 (Glu −5 ) to Ala or double substitution of Glu 585 Ala/Glu 586 Ala decreases binding to NHERF1 [18]. Thermodynamic parameters of the binding between NHERF1 PDZ domains and the C-terminal peptide of PTHR (8 aa) with a single or double Ala substitution were evaluated by ITC [29]. The loss of the binding strength following Glu-to-Ala substitution is attributed to unfavorable changes in enthalpy (∆H o ) and entropy (∆S o ). Mutation of Glu −5 in combination with Glu −6 exerts a synergistic effect on ∆H o . Again, a corresponding change in the free energy of binding (∆∆G o ) is not observed or is only minor due to the enthalpyentropy compensation and rather define the PDZ-ligand specificity [29]. The ∆S o values for PDZ1 and PDZ2 domains bound to the ensemble of Ala variants of the C-terminal motif of PTHR were plotted against ∆H o . A strong correlation between ∆S o and ∆H o is well-matched to PDZ1-NPT2A mutant variants ( Figure 5). A linear plot for PDZ1-NPT2A and PDZ1/PDZ2-PTHR confirmed that the binding mechanism is similarly conserved.

Concluding Remarks
NHERF1 PDZ domain specificity for NPT2A and PTHR is modulated by residues outside the binding pocket and by allosteric long-range communication. Despite significant progress, crystallization or cryo-electron microscopy to solve the structure of NHERF1 PDZ domains with the C-terminal -TRL motif of NPT2A will help distinguish binding determinants and improve our understanding of the origins of NHERF1 and other PDZ protein functional specificity. The recent finding that an internal PDZ region in NPT2A may regulate phosphate transport raises new questions about the structural determinant of this interaction and its impact on NHERF1. The influence of post-translational, site-specific phosphorylation on NHERF1 binding specificity and regulation of NPT2A-mediated phosphate uptake in health and disease remains uncharacterized. A combination of structural analysis, protein engineering and cell biology will be required to address these gaps in our understanding.