A high-resolution structure of the EF-hand domain of human polycystin-2

Autosomal dominant polycystic kidney disease (ADPKD) affects over 1:1000 of the worldwide population and is caused by mutations in two genes, PKD1 and PKD2. PKD2 encodes a 968-amino acid membrane spanning protein, Polycystin-2 (PC-2), which is a member of the TRP ion channel family. The C-terminal cytoplasmic tail contains an EF-hand motif followed by a short coiled-coil domain. We have determined the structure of the EF-hand region of PC-2 using NMR spectroscopy. The use of different boundaries, compared with those used in previous studies, have enabled us to determine a high resolution structure and show that the EF hand motif forms a standard calcium-binding pocket. The affinity of this pocket for calcium has been measured and mutants that both decrease and increase its affinity for the metal ion have been created.


Introduction
Autosomal dominant polycystic kidney disease (ADPKD) is a common inherited renal disease, with a worldwide prevalence of 1:1000 and is frequently complicated by cardiovascular disease including hypertension and cerebral aneurysm. It typically causes end-stage renal failure (ESRF) in the 6th to 8th decades of life requiring renal replacement ther-apy such as dialysis or transplantation. 1 About 85% of cases of ADPKD are caused by mutations in PKD1 with the remainder due to mutations in PKD2. 2 PKD1 encodes polycystin-1 (PC-1), a large multi-membrane spanning glycoprotein of 4302 residues and mass >460 kDa. It has a short 198 amino acid intracellular C-terminus that contains a coiledcoil domain that interacts with the C-terminus of polycystin-2 (PC-2), the protein product of PKD2 gene. 3 Mutations of PKD2 that result in the loss of functional polycystin-2 in renal tubular and biliary epithelial cells cause these cells to proliferate, lose their normal structure within tissue architecture, and go on to form cysts.
PC-2 is the prototypic member of the TRP (transient receptor potential) channel family, 4,5 and is expressed in most tissues. The PC-1:PC-2 complex functions as a mechanosensitive Ca 21 ion channel that localizes to primary cilia. 6 PC-2 consists of six transmembrane helices, with both the N-and Ctermini being cytoplasmic. The C-terminal cytoplasmic portion of PC-2 consists of three regions: an EF-hand motif (711-787), a flexible linker 787-827, and a coiled-coil domain that mediates the interaction with PC-1 (828-895). 7 EF hands motifs can bind calcium and often mediate responses to changes in calcium concentration. The EF hand motif in PC-2 has been shown to play an important role in modulating channel gating. This region of PC-2 has previously been investigated using NMR spectroscopy but these studies have given conflicting results. 8,9 As part of a long-standing study of the structure and function of the polycystins, 10 we have determined a high resolution NMR structure of the PC-2 EF-hand domain. This has allowed us to determine how the domain binds calcium and make a detailed comparison to the structures of other EF-hand domains.

Results
Structural studies of the C-terminal EF-hand domain of PC-2 A series of fragments of the C-terminal cytoplasmic region of PC-2 were cloned, expressed and their suitability for NMR studies determined. Constructs containing the C-terminal coiled-coil were found to precipitate upon the addition of calcium and were not investigated further. Focusing on the region around the EF hand motif, we found that sequence alignment and secondary structure prediction [ Fig.  1(a)] suggested that it was embedded with a larger structured region (residues 680-792). NMR spectra were recorded of a series of constructs encompassing this region in absence and presence of 5 mM calcium and the peptide corresponding to residues 717-792 was found to be most suitable for NMR studies. Additional residues at the N-terminus caused the domain to be less stable and hence these constructs were not used. These findings were in agreement with those of Schumann et al. who used a clone from 680 to 796 and observed the region from 680 to 705 to be characterized by somewhat broad lines, chemical shifts close to those observed for random coil amino acids and NMR relaxation data consistent with a weakly structured region. 8 Furthermore, Schumann et al. were unable to detect resonance for residues from 707 to 724, suggesting this region to be in intermediate exchange. It is possible that the region from 680 to 717 can become structured upon interacting with either the membrane or a functionally important protein, but under the conditions used the region was not sufficiently ordered to obtain structural information. Interestingly, two structures of ion-channels identified by PHYRE to have homology to polycystin-2 were shown to have helical motifs in this region, 11,12 suggesting that the ion channel may be linked to the EF-hand motif by a helical extension. 15 N-1 H HSQC measurements of PC-2 EF-hand indicated that this region of the protein has the characteristics of a molten-globule in the absence of calcium [ Fig. 2(a)], but upon addition of calcium the spectrum shows dispersion typical of a folded domain [ Fig. 2 Fig. S2) from the chemical shift and RDC measurements. The structures calculated with and without RDCs were essentially the same with the lowest energy structures having an r.m.s.d. of 1.01 Å over residues 717-785. CNS was subsequently used to determine a high-resolution solution structure of the domain using NOE, dihedral angle, hydrogen-bond and residual dipolar coupling (RDC) constraints. Structures were initially determined with a calcium ion not included in the structure calculations to prevent bias. Once an unambiguous calcium-binding pocket was seen (Supporting Information Fig. S3), a calcium ion was incorporated into structure calculation based on the coordination observed in other EF hands using established protocols. A summary of all conformational constraints and statistics is presented in Table I. A comparison of our structure with those determined previously (Fig. 4) reveals similarity to that determined by Petri 9 and colleagues with an r.m.s.d. of 4.0 Å for the backbone residues (720-785) of 2KQ6. The construct used by Petri was shorter at the N-terminus than the one used here probably as a result of the use of limited proteolysis to determine the domain boundaries. We observe several long range NOEs from residues missing in their construct, in particular from L717. The removal of these residues may have also affected the stability of their protein, and consequently the quality of the spectra, resulting in a low number of long-range NOEs in their calculations (41 for the 78 residues construct). Superimposition of the domain using residues 734-785 reveals a greater degree of agreement with an r.m.s.d. of 2.21 Å . There are, however, significant differences in the position of the side-chains of residues known to be important for calcium binding [ Fig. 4(d,f)]. The large number of NOEs determined for our structure enabled the side-chains in the EFhand to be constrained in positions consistent with calcium binding (Supporting Information Fig. 3).
Our structure shows less similarity to those determined by Schumnann et al. 8 with an r.m.s.d. of 4.28 and 6.02 Å for 2KLD and 2KLE, respectively. The 2KLD structures were determined using CNS whilst 2KLE was determined using a protocol involving AUREMOL-ISIC which uses known structures to assist in refinement. The construct used for the 2KLD and 2KLE structures were considerably longer than those of Petri (680-796) but unfortunately the spectra did not allow for the assignments for residues 707-724. As such, the N-terminus of the structures determined by Schumann et al. is undefined and this possibly contributed to the lack of definition at the C-terminus due to the loss of long-range NOEs between these regions. Whilst there are more NOEs used in the calculation of 2KLD and 2KLE structures, the EF-hand region of the protein is also somewhat undefined [ Fig. 4(e)] with several of the side-chains adopting conformations inconsistent with calcium binding. Our structure provides the first detailed information on the structure of the EF-hand motif of the PC-2 domain and thus provides a framework for further analysis of the mechanism of calcium sensing. Proteins usually contain at least two EF hand motifs, typically with two motifs interacting to form a compact abaaba fold in which the calcium ions are bound by residues in the loops between the first and second and the third and four helices. The PC-2 domain has a similar fold although in this case only the loop between the third and fourth helices binds a calcium ion. The loop between the first two helices that would normally bind a second calcium ion is truncated in this domain. Although it still forms a short stretch of sheet with residues in the canonical EF hand loop, a key feature of EF hand proteins, several of the residues that would normally bind a calcium ion are missing [ Fig. 1(b)]. A glutamate residue is present at the start of helix 3 in what would normally be the final position of an EF hand motif but mutation of this residue to glutamine has no effect on calcium binding.
The calcium affinity of the PC-2 EF hand domain is rather low for this class of protein (Table II). We noted that residue Q768, at position 6 in the canonical EF-hand motif, adopts an unusual phi angle not typically associated with nonglycine residues. Given that position 6 of an EF-hand is most often a glycine we made a Q768G mutant to observe the effect on calcium binding. Strikingly the Q768G mutant was able to bind calcium 10-fold tighter than the wild-type EF-hand domain (12 6 0.2 mM and 122 6 6.1 mM, respectively). This is possibly a result of the linkage in this domain between folding and binding as it is more energetically favorable for glycine to adopt this backbone conformation.
Members of the EF-hand superfamily are typically divided into three conformations: open, closed, and semi-open, depending on the relative orientation of the helices. 13 Calcium-free EF-hand motifs are invariably found in the "closed" conformation. Upon binding calcium, a rearrangement of the helices in the motif can occur that leads to an "open" conformation in which a large hydrophobic surface is exposed that has the potential to bind to target proteins or ligands. The "semi-open" conformation has to date only been found in myosins. 14 Calciuminduced conformation changes are common in EFhand proteins, including calmodulin, that have a role in signal transduction proteins. Not all EFhand containing proteins; however, undergo large conformational change upon binding calcium and these are thought to simply buffer calcium concentration. 15 The final three helices of the PC-2 domain are very similar in structure to the calcium-bound form of calmodulin 16 and troponin C 17 (Supporting Information Fig. 4). The orientation of the first helix is, however, very different from that normally observed in "open" EF-hand domains possibly, in part, because of the truncation of the loop between the first two helices. In the PC-2 domain the first helix is orthogonal to the equivalent helix in regular open EF hand domain structures and makes extensive contacts to the C-terminal helix. As such, it can be seen that the N-terminal helix appears to fill the hydrophobic patch used to typically bind other proteins or cognate partners. 17

Domain architecture of polycystin-2
The domain architecture of the C-terminus of PC-2 was analyzed using the programs JPRED 18 and Phyre 19 to identify regions likely to have a discrete fold. Sequence alignments were generated using CLUSTALW.

Expression and purification of PC-2 EF-hand domain
The C-terminus of PC-2 and mutant protein domains were cloned into a modified pRSETa 20 vector containing a thrombin cleavage site to remove the poly-histidine tag and expressed in C41 cells grown in 2xTY medium. Isotopically-labeled protein was prepared by growing cells in K-MOPS minimal media 21 containing 15 NH 4 Cl and/or [ 13 C]-glucose.
Protein was purified using Ni-NTA superflow resin and affinity chromatography. Following thrombin cleavage and further Ni-NTA affinity chromatography unbound protein was applied to a Superdex 75 gel-filtration column. Double de-ionised water was used to make the buffer solutions. The buffer solutions were then passed through ChelexV R 100 resin, analytical grade (Bio-Rad Laboratories, Hemel Hempstead HP2 7DX, UK). This was to remove any residual calcium in the buffer solutions. EDTA treated and thoroughly washed plastic containers were always used.

Isothermal titration calorimetry
All ITC measurements were carried out at 25 C on a VP-ITC system (MicroCal). All solutions were prepared with Milli-Q V R water and degassed before 1 mM CaC1 2 solution (in the syringe) was titrated into 10-20 mM protein in the sample cell (1.44 mL), using 4 or 10 mL of titrant per injection over 12 or 30 s. The protein and calcium solutions (1 mM) were in 20 mM HEPES buffer at pH 7.5. Time intervals of 420-600 s were used between injections to allow complete thermal equilibration. The data were fitted to a one site model using a nonlinear least-squares algorithm packaged with the OriginV R software, provided with the VP-ITC system, The stoichiometry (N), association constant (K a ), and enthalpy change (DH) were each obtained directly from the titration data. The dissociation constant (K d ) was calculated (Supporting Information Fig. S1).

NMR spectroscopy
Protein samples prepared for NMR spectroscopy experiments were typically 0.5 mM in 90% H 2 O, 10% D 2 O, containing 20 mM Tris, pH 7.0, 100 mM NaCl, and 5 mM b-mercaptoethanol. All spectra were acquired using either a Bruker DRX800 or DRX600 spectrometers equipped with pulsed field gradient triple resonance at 25 C, and referenced relative to external sodium 2,2-dimethyl-2-silapentane-5-sulfonate (DSS) for proton and carbon signals, or liquid ammonia for that of nitrogen. Assignments were obtained using standard NMR methods using 13 C/ 15 N-labeled, 15 N-labeled, 10% 13 C-labeled and unlabeled PC-2 EF-hand NMR samples. 22,23 Backbone assignments were obtained using the following   H-13 C HSQC. Additional assignments were made using 2D TOCSY and DQF-COSY spectra. A set of distance constraints were derived from 2D NOESY spectra recorded from a 1.0 mM sample with a mixing time of 100 ms. Hydrogen bond constraints were included for a number of backbone amide protons whose signals were still detected after 10 min in a 2D 1 H-15 N-HSQC spectrum recorded in D 2 O at 278 K (pH 5.0). Candidates for the acceptors were identified using the program HBPLUS for the hydrogen bond donors that were identified by the H-D exchange experiments. When two or more candidates of acceptors were found for the same donor in different structures, the most frequently occurring candidate was selected. For hydrogen bond partners, two distance constraints were used where the distance (D) H-O (A) corresponded to 1.5-2.5 Å and (D) N-O (A) to 2.5-3.5 Å . Torsional angle constraints were obtained from an analysis of C 0 , N, C a H a and C b chemical shifts using the program TALOS. 24 The stereospecific assignments of H b resonances determined from DQF-COSY and HNHB spectra were confirmed by analyzing the initial ensemble of structures. Stereospecific assignments of H g and H d resonances of Val and Leu residues, respectively, were assigned using a fractionally 13 C-labeled protein sample. 25 Residual dipolar couplings were measured for proteins aligned in 5% C12E5/1-hexanol. Initial modeling was performed using CS-Rosetta and included the measured RDC values. The three-dimensional structures of the PC-2 EF-hand domain were calculated using the standard torsion angle dynamics-simulated annealing protocol in the program CNS 1.2. 26 Alignment tensor values for the RDC constraints were calculated using SSIA and the RDC constraints were incorporated in the final round of structure calculations. Structures were accepted where no distance violation was greater than 0.25 Å and no dihedral angle violations >5 . The final coordinates have been deposited in the Protein Data Bank (PDB accession no. 2Y4Q).

Conclusions
In this study, we have determined a high-resolution structure of the human polycystein-2 EF-hand domain. Based on this, it would appear that this module is derived from a protein originally containing two EF motifs that has lost the first calciumbinding site. Loss of this site prevents the PC-2 domain from binding calcium cooperatively, a common feature of many EF hand proteins. The PC-2 domain also differs from the majority of EF hand containing proteins in its affinity for calcium apparently as a result of a conserved substitution in the EF hand loop. Recently, it has become clear that primary cilia are specialized calcium signaling organelles with a calcium concentration distinct from that of the cytosol. The polycystin channels role appears to be to control the calcium concentration within in primary cilia and thereby regulate signaling pathways linked to this organelle. It is likely that the atypical calcium binding properties of the EF hand domain from PC-2 may have evolved to allow the sensing of a difference range of calcium concentrations in the cilium, to that found in the cytoplasm. 27 Loss of the first calcium-binding site in the PC-2 domain has been accompanied by a rearrangement of the N-terminal helix that allows it to make more extensive contacts with the rest of the structure. This has important functional implications as the first helix now blocks the hydrophobic binding site used by other open conformation EF hands to interact with ligands. At present, it is unclear how the PC-2 EF hand domain transduces changes in calcium concentration. As, in this protein, calcium binding is coupled to folding many features of the folded conformation could in principle mediate the response to changes in calcium concentration, or potentially simple compaction of the chain due to the folding processes could juxtapose other elements of the protein or their binding partner closer to the channel. It has been shown that other regions of the cytoplasmic tail of PC-2 also bind calcium and the changes induce in PC-2 in response to changes in calcium concentration are probably complex. The structure reported here can provide a basis for detailed structure function studies of the role of this region in the regulation of polycystin that will lead to a clear understanding of the molecular basis of PKD.