Report

TRP-PLIK, a Bifunctional Protein with Kinase and Ion Channel Activities

See allHide authors and affiliations

Science  09 Feb 2001:
Vol. 291, Issue 5506, pp. 1043-1047
DOI: 10.1126/science.1058519

Abstract

We cloned and characterized a protein kinase and ion channel, TRP-PLIK. As part of the long transient receptor potential channel subfamily implicated in control of cell division, it is a protein that is both an ion channel and a protein kinase. TRP-PLIK phosphorylated itself, displayed a wide tissue distribution, and, when expressed in CHO-K1 cells, constituted a nonselective, calcium-permeant, 105-picosiemen, steeply outwardly rectifying conductance. The zinc finger containing α-kinase domain was functional. Inactivation of the kinase activity by site-directed mutagenesis and the channel's dependence on intracellular adenosine triphosphate (ATP) demonstrated that the channel's kinase activity is essential for channel function.

Phototransduction in Drosophila invokes phospholipase C (PLC)–mediated activation of transient receptor potential (TRP) channels, leading to membrane depolarization (1,2). The mammalian TRP channel family may be divided by sequence similarity into short, long, and osm 9–like subfamilies [reviewed in (3)]. Receptor-mediated stimulation of PLC activates many members of the short TRP channels, and physical or chemical stimuli activate isoforms of the osm 9–like TRP channel. Long TRP channels (LTRPC), such as melastatin, MTR1, and TRP-PLIK, are distinguished by their long coding sequences. Melastatin expression correlates with melanocytic tumor progression, whereas MTR1 is associated with Beckwith-Wiedemann syndrome and a predisposition to neoplasias (3). The gating mechanisms of the LTRPC group are unknown.

A yeast two-hybrid (Y2H) screen of a rat brain library with the C2 domain–containing COOH-terminus of PLC-β1 as bait identified a potential interacting partner with similarity to Dictyostelium myosin heavy chain kinase B (MHCK B) and eukaryotic elongation factor 2 kinase (eEF-2 kinase). Rapid amplification of cDNA ends (RACE) experiments based on the rat Y2H clones revealed one open reading frame encoding a putative kinase with 347 amino acids and a predicted molecular mass of 39.6 kD (4). A BLAST search with the sequence of the Y2H clone against the NCBI nonredundant nucleotide database revealed that the sequence is part of a larger open reading frame within a 7105–base pair (bp) transcript (accession number: AF149013; ChaK) encoding an 1863–amino acid protein with a predicted molecular mass of 212.4 kD (Fig. 1A). We cloned the predicted larger protein directly by polymerase chain reaction (PCR) from mouse brain cDNA (5). Comparison of the deduced amino acid sequence with those in the nucleotide and protein databanks demonstrated substantial similarity of ChaK to TRP family members, with greatest similarity to a member of the LTRPC group, melastatin. We designated the smaller protein PLIK, for “phospholipase C interacting kinase,” and the larger protein TRP-PLIK. The interaction of PLIK with PLC-β1 was confirmed by coimmunoprecipitation of expressed proteins in CHO-K1 cells and by glutathione S-transferase (GST) pull-down purification (6).

Figure 1

Sequence of TRP-PLIK and assessment of expression. (A) The deduced amino acid sequence of TRP-PLIK (13). Secondary structure algorithms and Kyte-Doolittle analysis of the amino acid sequence predict an integral membrane protein with at least six transmembrane domains (underlined). The TRP family amino acid motif “EWKFAR” that follows the last predicted transmembrane domain is modified in TRP-PLIK (VWKYQR) and is underscored by asterisks. The putative kinase domain containing a region homologous to MHCK B and eEF-2 kinase is boxed in red. The “nucleotide binding” motif (GXGXXG) is boxed and the zinc-finger motif is underlined. Residues targeted for mutational analysis are designated by solid black circles. (B) Northern blot analysis of TRP-PLIK. (C) Expression of full-length TRP-PLIK-HA was assessed by transient transfection of TRP-PLIK-HA (lane 1) or vector (lane 2) into CHO-K1 cells by immunoprecipitation and Western blotting. (D) Confocal microscope images of TRP-PLIK-HA expressed in HM1 cells.

Northern blot analysis of polyadenylated RNAs revealed an ∼8-kb transcript in brain and skeletal muscle, with stronger signals in kidney, heart, liver, and spleen, consistent with the 7105-bp transcript size for TRP-PLIK (Fig. 1B) (7). SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and protein immunoblot analysis of transiently expressed TRP-PLIK modified at the COOH-terminus to contain the hemagglutinin (HA) epitope (TRP-PLIK-HA) revealed a signal migrating at ∼220 kD, in accord with the predicted molecular mass of TRP-PLIK (212.4 kD; Fig. 1C) (8). Consistent with TRP-PLIK being an integral membrane protein, TRP-PLIK-HA transiently expressed in HM1 cells [human embryonic kidney (HEK)–293 cells stably transfected with the muscarinic type 1 receptor] revealed both punctate membrane and cytoplasmic staining (Fig. 1D) (9).

TRP-PLIK and PLIK contain a region with similarity to the catalytic domains of MHCK B and eEF-2 kinase. MHCK B and eEF-2 kinase belong to the atypical α-kinase family (10–12). A Clustal W alignment of the catalytic regions of the three proteins revealed that TRP-PLIK contains the first and last of the three conserved glycine residues within the “GXGXXG” nucleotide binding motif that is found in many adenosine triphosphate (ATP)–binding proteins. Further sequence analysis of the proposed kinase domain with the program BLOCKS (www.blocks.fhcrc.org) revealed that residues in TRP-PLIK are similar to the “(R/K)(R/K)HHCR motif” (13) in FYVE zinc fingers (14). One of the cysteines within this motif is conserved among the three kinases. Mutation of a conserved pair of cysteine residues COOH-terminal to the GXGXXG sequence (Cys313 and Cys317 in rat eEF-2K) completely inactivated eEF-2K (15).

We tested whether TRP-PLIK and PLIK could encode functional protein kinases and probed for possible “dominant-negative” mutations for functional studies by constructing recombinant fusion proteins of the kinase domain and two mutant forms attached to the COOH-terminus of GST. ATP-mut is the kinase domain substitution of aspartate exchanged for the final conserved glycine in the putative ATP-binding motif of TRP-PLIK (Gly1796). Zn-mut is the kinase domain substitution of two cysteines (Cys1809 and Cys1812) by alanine within the FYVE zinc finger homology domain from TRP-PLIK (Fig. 1A). GST fusion proteins of the wild-type kinase domain (GST-WT) and kinase domain mutants (GST-ATP-mut and GST-Zn-mut) were expressed in Escherichia coli as soluble proteins and purified by glutathione-agarose chromatography. The isolated fusion proteins were used for in vitro kinase assays with myelin basic protein (MBP) as a test substrate (Fig. 2, A and B) (16). The incorporation of 32P into MBP catalyzed by GST-WT, but not by GST or MBP alone, indicated that the kinase domain of TRP-PLIK exhibited protein kinase activity. The incorporation of 32P into GST-WT itself is presumably due to autophosphorylation, a common feature of protein kinases (17). Indeed, Fig. 2C shows that the full-length TRP-PLIK was autophosphorylated. The incorporation of32P into MBP catalyzed by GST-ATP-mut was estimated by densitometry to be 0.005 of that for GST-WT, supporting the identification of “GPANLG” (residues 1791 to 1796) (13) as the ATP-binding site. No kinase activity by GST-Zn-mut could be detected. Because FYVE zinc fingers are stabilized by interactions of conserved cysteines with Zn2+, the tertiary structure of the enzyme mutant within this region could be destabilized (18).

Figure 2

Kinase activity of TRP-PLIK (16). (A) Coomassie-stained gel after SDS–gel electrophoresis of phosphorylation reactions containing GST-kinase fusion proteins with or without MBP. lane M, molecular mass marker; lane 1, 10 μg of GST; lane 2, 500 ng of GST-WT; lane 3, 500 ng of GST-Zn-mut; lane 4, 500 ng of GST-ATP-mut; lane 5, 1 μg of MBP; lane 6, GST and 1 μg of MBP; lane 7, GST-WT and 1 μg of MBP; lane 8, GST-Zn-mut and 1 μg of MBP; and lane 9, GST-ATP-mut and 1 μg of MBP. (B) Autoradiogram of same sample gel. (C) Autophosphorylation of immunoprecipitated HA-tagged TRP-PLIK upon incubation with [γ-32P]ATP.

TRP-PLIK was functionally characterized by whole-cell and single-channel recordings of CHO-K1 cells transfected with TRP-PLIK (19). In transfected cells, a large outwardly rectifying current was elicited by a voltage ramp and voltage steps ranging from –100 to +100 mV, whereas mock-transfected cells exhibited only a small linear background current (Fig. 3, A and B). Because the large outward current above +50 mV was outside the physiologically relevant range, we focused on the inward currents (see inset in Fig. 3A). Mean inward current density was 15.4 ± 2.6 pA/pF at –100 mV (Fig. 3C).

Figure 3

TRP-PLIK currents in transfected CHO-K1 cells. (A) Representative recordings of TRP-PLIK currents evoked by a 500-ms voltage ramp ranging from –100 to +100 mV in TRP-PLIK-transfected (a) and mock-transfected CHO-K1 cells (b). Holding potential = 0 mV. Inset shows current traces obtained during ramps from –100 to 0 mV performed on the same cells but amplified to provide a more detailed view of the inward currents. pA/pF, picoamperes/picofarad. (B) Currents elicited by voltage steps ranging from –100 to +100 mV in a TRP-PLIK–expressing cell show that the current steeply rectifies and is time invariant during the duration of the step. (C) Averaged I-Vrelation of the TRP-PLIK current (n = 10). (D) Blockade of TRP-PLIK by 2 mM La3+ (outward currents, upward bars; inward currents, downward bars; ± SEM;n = 5). (E) Single-channel currents of TRP-PLIK at various test potentials obtained in outside-out patches. Dashed lines represent the closed state. (F) TRP-PLIK single-channel I-V relation. Single-channel current amplitude was determined by measuring amplitude histograms at each potential. A linear regression fit from +40 to +100 mV (solid line; ±SEM) yielded a slope conductance of 105 ± 8 pS (n = 4).

TRP-PLIK currents were not altered when NaCl was substituted by CH3SO3Na, indicating that Cl does not permeate TRP-PLIK channels (6). The relative permeability of cations relative to that of Cs+ was 1.1, 0.97, and 0.34 for K+, Na+, and Ca2+, respectively. TRP-PLIK was not blocked by 1 mM Ba2+, 1 mM TEA, or 0.2 mM Zn2+. La3+ (2 mM) blocked inward and outward TRP-PLIK currents by 97% (P < 0.01, n = 5) and 37% (P < 0.05, n = 5), respectively. Single-channel currents were measured at positive voltages in outside-out patches, and their net activity agreed with that predicted from the whole-cell currents. The slope conductance of the single-channel currents was 105 ± 8 pS (Fig. 3, E and F).

We tested the effects of mutations that alter kinase activity on channel function. Whole-cell current amplitudes of ATP-mut and Zn-mut TRP-PLIK were markedly decreased compared with those of the nonmutated TRP-PLIK, suggesting that kinase activity was required for TRP-PLIK channel function (Fig. 4, A and B). This was supported by experiments in which current amplitudes in cells dialyzed with an ATP-containing pipette solution (5 mM ATP, 1 mM Mg2+) initially increased, followed by a slow decrease over several minutes. Currents in cells dialyzed with 0 mM ATP (1 mM Mg2+) pipette solution were significantly smaller and did not vary during recordings (Fig. 4, C and D).

Figure 4

TRP-PLIK channels are controlled by kinase activity. (A) Representative currents recorded in TRP-PLIK–expressing cells (top) and currents elicited by voltage ramps in Zn-mut, ATP-mut, and mock (vector)–transfected CHO-K1 cells (bottom). (B) Mean current amplitude (±SEM,n = 8) of TRP-PLIK, Zn-mut, ATP-mut, and control (vector). Upward bars, outward current; downward bars, inward current. * indicates P <0.01 compared with TRP-PLIK. (C) Representative outward (top) and inward TRP-PLIK currents (bottom) recorded at 0, 200, and 800 s after initiation of whole-cell dialysis. (D) Time-dependent changes in TRP-PLIK current amplitude in cells dialyzed with either 0 or 5 mM ATP (1 mM Mg2+) in normal intracellular solution (±SEM,n = 5).

TRP-PLIK is a protein that is both an ion channel and a kinase. As a channel, it conducts calcium and monovalent cations to depolarize cells and increase intracellular calcium. As a kinase, it is capable of phosphorylating TRP-PLIK and other substrates. The kinase activity is necessary for channel function, as shown by its dependence on intracellular ATP and by the kinase mutants. Although kinases have long been known to modulate ion channels (20), TRP-PLIK is unusual in that the channel has its own kinase. The presence of the kinase domain adjacent to the sixth transmembrane segment (S6) supports the hypothesis that it plays an important role in channel gating, because S6 appears to be commonly involved in the gating of ion channels (21). We postulate that TRP-PLIK is controlled by intracellular ATP levels and may be linked to a signal transduction cascade that modulates the channel's kinase activity. This member of the LTRPC family may link calcium-dependent processes in cells, such as cell division and apoptosis, to receptor and plasma membrane–associated signal transduction events.

  • * To whom correspondence should be addressed. E-mail: clapham{at}rascal.med.harvard.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article