The Cytoplasmic C-terminal Fragment of Polycystin-1 Regulates a Ca 2 1 -permeable Cation Channel*

The cytoplasmic C-terminal portion of the polycys-tin-1 polypeptide (PKD1(1–226)) regulates several important cell signaling pathways, and its deletion suffices to cause autosomal dominant polycystic kidney disease. However, a functional link between PKD1 and the ion transport processes required to drive renal cyst enlarge-ment has remained elusive. We report here that expression at the Xenopus oocyte surface of a transmembrane fusion protein encoding the C-terminal portion of the PKD1 cytoplasmic tail, PKD1(115–226), but not the N-terminal portion, induced a large, Ca 2 1 -permeable cation current, which shifted oocyte reversal potential ( E rev ) by 1 33 mV. Whole cell currents were sensitive to inhibition by La 3 1 , Gd 3 1 , and Zn 2 1 , and partially inhibited by SKF96365 and amiloride. Currents were not activated by bath hypertonicity, but were inhibited by acid pH. Outside-out patches pulled from PKD1(115– 226)-expressing oocytes exhibited a 5.1-fold increased NP o of endogenous 20-picosiemens cation channels of linear conductance. PKD1(115–226)-injected oocytes also exhibited elevated NP o of unitary calcium currents in outside-out and

The cytoplasmic C-terminal portion of the polycystin-1 polypeptide (PKD1(1-226)) regulates several important cell signaling pathways, and its deletion suffices to cause autosomal dominant polycystic kidney disease. However, a functional link between PKD1 and the ion transport processes required to drive renal cyst enlargement has remained elusive. We report here that expression at the Xenopus oocyte surface of a transmembrane fusion protein encoding the C-terminal portion of the PKD1 cytoplasmic tail, PKD1(115-226), but not the Nterminal portion, induced a large, Ca 2؉ -permeable cation current, which shifted oocyte reversal potential (E rev ) by ؉33 mV. Whole cell currents were sensitive to inhibition by La 3؉ , Gd 3؉ , and Zn 2؉ , and partially inhibited by SKF96365 and amiloride. Currents were not activated by bath hypertonicity, but were inhibited by acid pH. Outside-out patches pulled from PKD1(115-226)-expressing oocytes exhibited a 5.1-fold increased NP o of endogenous 20-picosiemens cation channels of linear conductance. PKD1(115-226)-injected oocytes also exhibited elevated NP o of unitary calcium currents in outside-out and cell-attached patches, and elevated calcium permeability documented by fluorescence ratio and 45 Ca 2؉ flux experiments. Both Ca 2؉ conductance and influx were inhibited by La 3؉ . Mutation of candidate phosphorylation sites within PKD1(115-226) abolished the cation current. We conclude that the C-terminal cytoplasmic tail of PKD1 up-regulates inward current that includes a major contribution from Ca 2؉permeable nonspecific cation channels. Dysregulation of these or similar channels in autosomal dominant polycystic kidney disease may contribute to cyst formation or expansion.
ϳ85% of autosomal dominant polycystic kidney disease (AD-PKD) 1 is secondary to mutations in the polycystin-1 (PKD1) gene (1). The 4303-aa PKD1 polypeptide has a large N-terminal extracellular domain region encompassing leucine-rich repeats, a C-type lectin domain, a low density lipoprotein A-like domain (2), 16 PKD domain repeats, and a single sperm receptor for egg jelly domain (3). Following a complex polytopic transmembrane domain is a 226-aa C-terminal cytoplasmic tail.
The hallmark pathology of ADPKD is the two-hit generation and gradual expansion of renal cysts. These cysts compress and, ultimately, render nonfunctional adjacent, apparently normal renal tissue (4). Cyst expansion is associated with the conversion of tubular epithelial cells from the normal phenotype of net solute reabsorption to the cystic phenotype of net secretion. This secretory phenotype may involve the function of the apically localized chloride channel, CFTR (5), but a direct link between the ADPKD disease genes for PKD1 and PKD2 and the ion transport processes required to mediate cyst expansion has remained elusive, despite the creation of mouse lines genetically deficient in PKD1 (6) or PKD2 (7). PKD disease mutations have been found throughout the PKD1 coding region (8 -10), but absence of the C-terminal cytoplamic tail suffices to cause ADPKD (11).
Multiple potential signaling and binding functions of the 226-aa PKD1 C-terminal cytoplasmic tail have been defined through study of overexpressed transmembrane fusion proteins encompassing all or portions of this C-terminal region (12)(13)(14). In this fusion protein context, the PKD1 C-terminal cytoplasmic tail activates PKC␣ and (in a process requiring active Rac or Cdc-42) c-Jun N-terminal kinase. Both pathways are required for downstream activation of AP-1-dependent transcriptional events (12,13).
The varied signaling and binding functions of the PKD1 C-terminal cytoplasmic tail led us to hypothesize that it might regulate ion transport processes that mediate or regulate cyst expansion. We now report that the distal fragment of the cytoplasmic C-terminal tail of PKD1 (PKD1(115-226)), expressed as a transmembrane fusion protein, up-regulates inward current and a Ca 2ϩ -permeable nonspecific cation channel of Xenopus oocytes.
Preparation of cRNA-SalI-or NotI-linearized recombinant pXT7 2 In most experiments, CD16.7-PKD1(1-92) was expressed as either of two variants which included C-terminal extensions beyond PKD1 cytoplasmic domain residue 92, originating from polylinker sequence. One form added a single C-terminal Ser residue. The other form added the 7-residue C-terminal sequence Ser-Ala-Ala-Ala-Arg-Glu-Ile. Both of these fusion proteins were expressed at the oocyte surface, but neither increased oocyte currents. The experiment shown in Fig. 6D utilized a form of CD16.7-PKD(1-92) devoid of any C-terminal extension. 3 The construct CD16.7 control provides a novel, PKD1(1-92)-unrelated cytoplasmic C-terminal amino acid sequence of the same length as PKD1(1-92). The sequence is: ESWHLSPLLCVGLWALRLWGALRLGA-VILRWRYHALRGELYRPAWEPQDYEMVELFLRRLRLWMGLSKVKES. This sequence is encoded by a frameshifted PKD1(1-92) cDNA (due to a single nucleotide deletion, GenBank NM000296 nucleotide 12446).  1-92). A, current traces in normal bath (ND96) generated in a two-microelectrode voltage clamp experiment on a CD16.7 control-injected oocyte. The voltage pulse protocol generated 20-mV voltage steps from Ϫ100 mV to ϩ20 mV, from a holding potential of Ϫ30 mV. B, representative current traces from a PKD1(115-226)-injected oocyte. C, composite current-voltage relationship of CD16.7 control-injected oocytes and CD16.7-PKD1(115-226)-injected oocytes. Slope conductances were 2.1 and 8.7 S, and reversal potentials were Ϫ31 and ϩ2 mV in the control and PKD1(115-226) groups, respectively. D, analysis of the indicated deletion constructs implicates a region of the cytoplasmic tail encompassing residues 115-189 as required for expression of increased current. La 3ϩ -sensitive current (2 mM La 3ϩ ) was assayed at Ϫ100 mV during two microelectrode studies. *, p Ͻ 0.05; **, p Ͻ 0.01, by post-test versus control. Oocytes expressing CD16.7-PKD1(115-226) (shaded bar) were studied in a separate experiment (inset, I-V relationship of La 3ϩ -sensitive difference current).
templates were transcribed with T7 RNA polymerase. Xenopus oocyte isolation, culture, and microinjection were performed using standard techniques (28,29). Oocytes injected with 12-25 ng of cRNA in a volume of 50 nl maintained integrity up to 18 days after injection. Generally, however, oocytes were subjected to electrical recording studies 2-3 days after microinjection.
Immunocytochemistry-Oocytes injected 3, 5, or 10 days previously with cRNA were incubated with purified 3G8 mouse anti-human CD16 monoclonal antibody (gift of O. Mandelstam, J. Strominger, and J. Unkeless; or purchased from Meditech) at a concentration of 2 g/ml in ND96 at room temperature for 3-4 h, then rinsed several times in ND96. Oocytes were fixed either in absolute methanol or in 3% paraformaldehyde in 140 mM NaCl, 20 mM sodium phosphate, pH 7.4 (PBS). Paraformaldehyde-fixed oocytes were rinsed in PBS, quenched with several washes in PBS plus 50 mM glycine, then in PBS. Methanolfixed oocytes were rehydrated into PBS. Oocytes were then incubated in secondary Cy3-conjugated donkey anti-mouse IgG (Jackson Immunoresearch Laboratories, West Grove, PA) diluted 1/500 in PBS, rinsed several times in PBS, then dehydrated in methanol, and cleared in benzyl benzoate/benzyl alcohol (2:1) prior to imaging with a Bio-Rad MRC 1024 confocal microscope. Some oocytes were incubated with both primary and secondary antibody prior to fixation, with indistinguishable results.
Two-electrode Voltage Clamp-Oocytes were placed in a 1-ml chamber (model RC-11, Warner Instruments, Hamden, CT) on the stage of a dissecting microscope and impaled with microelectrodes under direct view. Electrodes were pulled from borosilicate glass with a Narashige puller, filled with 3 M KCl, and had resistances of 2-3 megohms. Currents were measured with a Geneclamp 500 amplifier (Axon Instruments, Burlingame, CA) interfaced to a Hewlett Packard computer with a Digidata 1200 interface (Axon Instruments). Data acquisition and analysis utilized pCLAMP 6.0.3 software (Axon Instruments). Voltage pulse protocols consisting of 720-ms 20-mV steps between Ϫ100 and ϩ20 mV were generated by the Clampex subroutine. Bath resistance was minimized by the use of agar bridges filled with 3 M KCl, and a virtual ground circuit clamped bath potential to zero.
Standard bath solution was 96 mM NaCl, 2 mM KCl, 5 mM HEPES, 1.8 mM CaCl 2 , and 1 mM MgCl 2 , and holding potential was Ϫ30 mV. All solutions were of pH 7.40. Cation/anion permeability determinations were by equimolar replacement of bath Na ϩ with N-methyl-D-glucamine (NMDG), and of bath Cl Ϫ with gluconate. Determination of relative monovalent cation permeabilities was by equimolar substitution of bath Na ϩ with K ϩ , Li ϩ , and NH 4 A, PKD1(115-226)-associated currents were significantly reduced and the reversal potential was significantly hyperpolarized when bath sodium was replaced by NMDG. Bath chloride subsitution with gluconate had no effect on reversal potential. B, significant La 3ϩ -sensitive currents persisted after 16 h chloride depletion and subsequent recording in a sodium methanesulfonate bath after injection of EGTA. Current at Ϫ100 mV was Ϫ421 Ϯ 52 nA and Ϫ270 Ϯ 34 nA in control and La 3ϩ -containing bath, respectively; t ϭ 5.7, p ϭ 0.0001, by paired t test.
Patch Clamp Recording of Cation Currents-Patch pipettes were pulled to a resistance of 5-8 megohms, and oocytes were devitellinized as described previously (29). Gigaseals were first attained in the cellattached mode. Outside-out patches were formed by breaking the oocyte plasma membrane with a pulse and withdrawing the pipette slowly. Currents were measured with an Axopatch 1D amplifier (Axon Instruments) interfaced via a Digidata 1200 AD/DA board to a HP Vectra computer. Data was acquired at 1 kHz and digitized at 5 kHz. For outside-out experiments, pipette solutions contained 128 mM cesium aspartate, 12 mM cesium EGTA, and 10 mM HEPES and the bath solution was 140 mM sodium methanesulfonate, 10 mM HEDTA, 10 mM HEPES. In cell-attached experiments, the pipette contained 100 mM CaCl 2 , 10 mM HEPES, and 10 mM HEDTA.
Calcium Fluorescence Measurements-Oocytes from both groups were injected with 70-kDa Calcium Green dextran (Molecular Probes) to a final concentration of 3.5-7 M, then subjected to bath [Ca 2ϩ ] change from 0 to 10 mM. Relative change in intracellular calcium concentration ([Ca 2ϩ ] i ) was determined by exciting the fluorophore at 490 nm and imaging at 530 nm as described (29). Additional oocytes were loaded for 45 min with 8 M Fura2-AM, then subjected to bath [Ca 2ϩ ] changes from 1.8 mM to nominal 0 mM to 10 mM. In these oocytes [Ca 2ϩ ] i was determined by alternately exciting the fluorophore at 340 and 380 nm. 510-nm emission images were acquired and analyzed as described (29), using IMAGE1 and IMAGE1-FL software (Universal Imaging). 45 Ca 2ϩ Influx Studies-The uptake procedure was as reported (30), with modifications. Oocytes were incubated 1 h in ND-96 modified to contain 0.13 mM CaCl 2 with 2.5 Ci/l 45 Ca 2ϩ , washed in ND-96, then counted in a liquid scintillation counter (Packard 2200CA).
The increased inward current and positive voltage-shifted E rev elicited by CD16.7-PKD1(115-226) but not by CD16.7 control was not explained by differences in oocyte surface expression of the different fusion proteins. As shown in Fig. 2, immunostaining of intact, unfixed oocytes with anti-CD16 antibody revealed comparable intensities of CD16 signal at the oocyte surface whether the epitope was borne on CD16.7- PKD1(115-226) or by CD16.7 control. Surface staining intensity increased at longer times after cRNA injection and in oocytes coexpressing both cRNAs (data not shown). The inactive CD16.7-PKD1(1-92) constructs were also expressed at the surface of 57% of injected oocytes at fluorescence intensities comparable to that of the active construct CD16.7-PKD1(115-226). Surface staining was always absent in water-injected oocytes (data not shown) and in oocytes expressing CFTR, which lacks the CD16 epitope (Fig. 2d).
The CD16.7-PKD1(115-226)-induced Current Is Associated with Elevated Ca 2ϩ Permeability-The hypothesis that the PKD1(115-226)-associated current might conduct Ca 2ϩ was tested by subjecting oocytes to voltage ramps first in the absence and then in the presence of 10 mM bath Ca 2ϩ (Fig. 4A). The difference current at Ϫ100 mV attributable to Ca 2ϩ was Ϫ436 Ϯ 95 nA in CD16.7-PKD1(115-226)-expressing oocytes compared with ϩ21 Ϯ 27 nA in oocytes previously injected with water (p ϭ 0.004, two-tailed t test). Thus, the nonspecific cation current activated in oocytes by expression of PKD1(115-226) also exhibited substantial permeability to Ca 2ϩ . The time course of holding currents was studied with calcium bath substitution (see Fig. 4B). The holding potential chosen was Ϫ30 mV, close to the chloride equilibrium potential, to minimize potential contribution of endogenous chloride currents. The upper trace from an oocyte expressing CD16.7-PKD1(115-226) reveals decreased inward current upon bath Na ϩ substitution with NMDG, which increased upon subsequent addition to the NMDG of 10 mM Ca 2ϩ (representative of four experiments). The lower trace from an oocyte expressing CD16.7 control shows that current remained unaltered during the same bath change protocol (representative of four experiments), as expected from the dominant endogenous chloride conductance. Fig. 4C shows current voltage relationships of oocytes subjected to the protocol of Fig. 4B. The calcium difference current (solid line) was significant, and the E rev was consistent with that expected of a cation current. 45 Ca 2ϩ influx into oocytes expressing PKD1(115-226) was 3.94 Ϯ 0.14 pmol/oocyte⅐h (Fig. 4D, n ϭ 24), and greatly exceeded the 0.85 Ϯ 0.26 pmol/oocyte⅐h in oocytes expressing CD16.7-PKD1(1-92) (n ϭ 9, p Ͻ 0.00001). 1 mM La 3ϩ inhibited 45 Ca 2ϩ influx into oocytes expressing PKD1(115-226) by 69% to 1.24 Ϯ 0.17 pmol/oocyte⅐h (n ϭ 26, p Ͻ 0.00001), whereas La 3ϩ was without effect on 45 Ca 2ϩ influx into oocytes expressing PKD1(1-92) (n ϭ 4, Fig. 4D). Thus, 1 mM La 3ϩ inhibited 88% of the 45  vated Na ϩ Current- Fig. 5A shows unitary currents recorded at a holding potential of Ϫ50 mV from a representative outsideout patch with cesium aspartate in the pipette and sodium methanesulfonate in the bath. Addition to the bath of 1 mM La 3ϩ suppressed nearly completely the inward Na ϩ currents. The unitary Na ϩ current exhibited an amplitude of Ϫ1.1 pA at Ϫ50 mV (Fig. 5B), and a linear conductance of 20 pS (r 2 ϭ 0.99) with E rev ϳ0 mV (P Na /P Cs ϳ1; Fig. 5C). The mean NP o of 0.16 Ϯ 0.04 was 5.1-fold higher than the NP o (0.03 Ϯ 0.03) of indistinguishable channels detected in patches pulled from oocytes expressing CD16.7 control (Fig. 5D, n ϭ 5, p ϭ 0.03). Bath addition of 1 mM La 3ϩ to patches pulled from oocytes expressing CD16.7-PKD1(115-226) reduced mean NP o by 96% (Fig.  5A), from 0.47 Ϯ 0.19 to 0.02 Ϯ 0.01 (n ϭ 5, p ϭ 0.04, one-tailed t test). Bath addition of 10 M SKF96365 reduced NP o 82% to 0.03 Ϯ 0.02 (n ϭ 5, p ϭ 0.03). Both inhibitors were considerably more potent as antagonists of outside-out patch currents than of whole oocyte currents (Table I).
The effect of bath pH on the induced cation current was tested. Bath pH change from 7.4 to 6.0 decreased inward current measured at Ϫ100 mV by 33%, from Ϫ1041 Ϯ 71 nA to Ϫ697 Ϯ 69 nA (n ϭ 5, p ϭ 0.0002). In contrast, bath pH change to 8.0 was without effect on inward current.
Importance of Selected Serine and Tyrosine Residues for Cation Current Activation by CD16.7-PKD1(115-226)-Several amino acid residues have been shown to be phosphorylated within PKD1(115-226) fusion proteins in vitro by exogenously added kinases and in transiently transfected cells (23). Within this region, serine 175 (Ser-4252 of holo-PKD1) is a preferred substrate of cAMP-dependent protein kinase, and tyrosine 160 (Tyr-4237 of holo-PKD1) is a preferred substrate of c-Src. The adjacent serine residues 174 and 175 (4251/4252) were mutated in tandem either to alanine or to aspartate, since each is a candidate cAMP-dependent protein kinase site. Each double mutation abolished CD16.7-PKD1(115-226)-associated inward currents (Fig. 8A), despite unimpaired expression of the doubly mutant polypeptides at the oocyte surface (Fig. 8B). Similarly, mutation of tyrosine 160 to phenylalanine (Y4237F) abolished CD16.7-PKD1(115-226)-associated current (Fig. 8A) despite unimpaired surface expression of this mutant polypeptide (Fig. 8B). DISCUSSION We report that expression of the CD16.7-PKD1(115-226) fusion protein, encoding the 112 C-terminal residues of the PKD1 cytoplasmic C-terminal tail fused to the CD16.7 transmembrane chimera, up-regulates a Ca 2ϩ -permeable cation conductance in Xenopus oocytes. The cation conductance is nonselective among monovalent cations, is inhibited by La 3ϩ , Gd 3ϩ , SKF96365, amiloride, and Zn 2ϩ , but not by DIDS or niflumate, and is not activated by hypertonicity. Expression of CD16.7 control, or of CD16.7-PKD1(1-92), CD16.7-PKD1(1-155), or of CD16.7 without any intracellular tail failed to activate this cation conductance, despite normal surface expression of these inactive polypeptides. Activation of whole cell currents was accompanied by increased open probability (NP o ) of cation channels of 20 -22 pS linear conductance, which likely mediated or substantially contributed to the recorded whole cell currents.
The CD16.7-PKD1(115-226) fusion protein spans the lipid bilayer only once. The transmembrane domain derived from CD7 within the CD16.7 fusion proteins is not expected itself to form an ion channel. Indeed, similar cation channel activity appeared present at 5-8-fold lower activity in oocytes previously injected with water or with cRNA encoding either CD16.7 control or CD16.7-PKD1(1-92). Taken together, the data suggest that the PKD1(115-226) domain activates a cation channel native to oocytes. This hypothesis suggests that CD16.7-PKD1(115-226) either triggers a signaling pathway that tonically activates cation channel activity, or that the PKD1 fragment can serve as a chaperonin or a functional subunit of an endogenous pore-forming polypeptide or polypeptide complex.
Which signaling pathways might mediate cation channel activation by CD16.7-PKD1(115-226)? This portion of the PKD1 C-terminal cytoplasmic tail activates the siamois promoter via inhibition of GKS-3␤, ␤-catenin stabilization, and up-regulation of TCF-mediated transcriptional signaling (13). Xwnt8-mediated activation of the siamois promoter was also enhanced by PKD1(115-226). Although itself incapable of activating transcriptional signaling mediated by AP-1 elements, this PKD2-binding fragment of PKD1 enhanced activation by PKD2 of AP-1-mediated transcription (21). PKD1(115-226) also dorsalized and prevented posterior development in zebrafish embryos (13), and bound and stabilized RGS7 (15). PKD1(115-226) also harbors a major in vitro cAMP-dependent protein kinase phosphorylation site (23,24). We tested by mutagenesis the functional importance of this site for cation channel activation. Mutation of the tandem serine residues 4251/4252 to either alanine or to aspartate nearly abolished CD16.7-PKD1(115-226)-associated cation currents, despite undiminished expression of the mutant polypeptides at the oocyte surface. Mutation of the putative tyrosine kinase site Tyr-4237 (23) to Phe also reduced cation current without decreasing surface expression of polypeptide. Therefore, these residues play important roles in cation channel activation. Whether this occurs by direct phosphorylation or by another mechanism remains to be determined.
Tsiokas et al. (18) showed that PKD2 through distinct binding domains complexes with the Ca 2ϩ -permeable cation channel hTRPC1, but not with hTRPC3. Akaike and colleagues (37) showed that degenerate hTRP-1 oligonucleotides reduced thapsigargin-stimulated currents in oocytes injected with a rat TRP4 homolog. More recently, Berridge and colleagues have cloned Xtrp1 (38), but its function remains unreported. Our preliminary experiments failed to demonstrate reduction in CD16.7-PKD1(115-226)-associated cation channel activity after coinjection of Xtrp1 antisense oligonucleotides (data not shown).
In summary, expression in Xenopus oocytes of a CD16.7 transmembrane fusion protein containing the PKD1 C-terminal cytoplasmic tail aa 115-226 activated Ca 2ϩ -permeable nonspecific cation currents in Xenopus oocytes likely mediated by channels of 20-pS linear conductance. Preliminary experiments (data not shown) indicate that similar channels were activated in 293T cells transiently transfected with CD16.7-PKD1(115-226). In oocytes, the elevated cation conductance correlated with increased Ca 2ϩ entry and elevation of [Ca 2ϩ ] i . These changes in ion channel activity are the first directly linked to the expression of any portion of the PKD1 polypeptide. We propose that changes in intracellular Ca 2ϩ signaling resulting from mutations in PKD1 contribute to the creation and/or maintenance of the secretory and/or proliferative phenotype of renal cysts in ADPKD1, and to the resultant cyst expansion that ultimately leads to renal failure.