Impaired endocytosis of the ion channel TRPM4 is associated with human progressive familial heart block type I

PFHBI disease is linked to a mutation in TRPM4.

We extended the clinical and genetic basis of our previous mapping and linkage study in the Afrikaner pedigree with PFHBI, traced to an ancestral founder who immigrated from Portugal to South Africa in 1696 (5). Of 71 mutation carriers identified, 48 had PMs implanted. For some family members with PMs, we had access to ECGs prior to the date of PM implantation in which they still had normal atrioventricular conduction. Of the combined individuals not in need of PMs and patients with normal atrioventricular conduction prior to implantation, 19 patients had RBBB, 8 had RBBB with left anterior or left posterior hemiblock, and 7 had no ECG abnormalities (Supplemental Table 1; supplemental material available online with this article; doi: 10.1172/JCI38292DS1). For the 19 patients with RBBB, heart rates, mean P-wave duration, and mean PR interval were in the normal range (Supplemental Table 1). Atrial premature activity and extra systoles were not seen. None of the characterized patients showed a Brugada syndrome type I pattern in ECG analysis, and syncope caused by Torsades de pointes was not observed in these families. It should be noted that with bundle branch blocks, QTc may be prolonged because of late depolarization, and consequent late repolarization, of one ventricle.

With additional microsatellite markers, we narrowed the PFHBI locus to an interval between markers D19S1059 and D19S604 corresponding to a region of approximately 0.5 Mb in chromosome 19q13.33 with approximately 25 genes (Supplemental Figure 1). Of these 25 genes, 5 are expressed in cardiac tissue: U1 small nuclear ribonucleoprotein (SNRP70); histidine-rich calcium-binding protein (HRC); transient receptor potential cation channel, subfamily M, member 4 (TRPM4); transcriptional enhancer 4 (TEAD2); and potassium voltage-gated channel, Shaker-related subfamily, member7 (KCNA7) (11, 14–18). Previously, KCNA7 was excluded as a possible candidate gene (16). Here, we focused on TRPM4 (11), which was prominently expressed in human Purkinje fibers compared with septum, atrium, and right and left ventricles (Figure 2A). We sequenced the 25 exons of TRPM4 (Supplemental Table 2). DNA samples of 23 affected and 35 unaffected members of the Afrikaner pedigree with PFHBI were available for analysis. We detected exclusively in DNA samples of affected family members a heterozygous G→A mutation at nucleotide 19 in exon 1 (c.19G→A; Figure 2B). The mutation generated a new recognition site for the restriction enzyme MboII, allowing for independent identification of the c.19G→A TRPM4 mutation by MboII DNA digestion (Supplemental Figure 2). In 2 unaffected control populations (230 ancestry-matched, unrelated Afrikaner and 389 unrelated individuals of mixed European descent), we did not detect the 19A allele. The c.19G→A mutation in TRPM4 predicts the substitution of TRPM4 glutamic acid at position 7 to lysine (p.E7K) within the TRPM4 N terminus (referred to herein as TRPM4^E7K). Glu7 is part of an N-terminal TRPM4 sequence motif evolutionary conserved among TRPM4 orthologs across phyla (Figure 2C). The motif is likely to play an important function in TRPM4 channel activity unrelated to conserved sequence domains previously reported to affect TRPM4 channel assembly or activity (Figure 2D and ref. 19).

The PFHBI-associated mutation increases TRPM4 current density.

To understand the functional significance of the PFHBI-associated mutation, we expressed WT and mutant TRPM4 channels in HEK 293 cells and measured TRPM4 current properties in response to a variety of stimuli known to influence TRPM4 channel gating (11, 19). WT and mutant TRPM4 channels responded in an essentially identical manner to both change in membrane voltage and rise in intracellular Ca²⁺ (Figure 3, A and B). At +80 mV, half-maximal current activation was observed, with 0.44 ± 0.03 and 0.40 ± 0.04 μM Ca²⁺ for WT and mutant TRPM4 channels, respectively (n = 7 per group; Figure 3B), in agreement with data previously reported for the TRPM4 channel expressed in HEK 293 cells (11). WT and mutant TRPM4 channel exhibited essentially identical sensitivities to block by intracellular ATP in inside-out membrane patches. Bath application of 500 μM adenylyl imidodiphosphate tetralithium salt (AMP-PNP), a nonhydrolizable ATP analog, inhibited WT and mutant TRPM4 channels with similar potency (TRPM4, 72.5% ± 2.5% block; TRPM4^E7K, 71.3% ± 3.9% block, n = 3; Figure 3C, Supplemental Figure 3A, and ref. 20). Furthermore, application of phosphatidylinositol-4,5-bisphosphate [PtdIns(4,5)P₂] to inside-out patches stimulated TRPM4 and TRPM4^E7K current amplitudes to a similar degree (TRPM4, 232% ± 12%, TRPM4^E7K, 238% ± 15%, n = 3; Supplemental Figure 3, B and C). These data indicated that PtdIns(4,5)P₂ sensitivity of TRPM4^E7K channels was unchanged compared with WT channels (21).

However, comparison of TRPM4 channel densities showed that HEK 293 cells expressing TRPM4^E7K consistently exhibited approximately 2-fold TRPM4 current density (43.8 ± 6.3 pA/pF at +80 mV, n = 13) compared with that of cells expressing WT channel (18.8 ± 3.7 pA/pF at +80 mV, n = 10). We obtained identical results using CHO cells (Supplemental Figure 4). Furthermore, HEK 293 cells transfected with TRPM4^E7K cDNA or cotransfected with equal amounts of TRPM4 and TRPM4^E7K cDNA (43.9 ± 5.2 pA/pF at +80 mV, n = 6) showed essentially the same increase in current density (Figure 3D). The dominant mutational effect on TRPM4 current amplitude apparently reflects the dominant inheritance of the PFHBI phenotype.

Macroscopic current amplitude I is the product of 3 parameters: single-channel conductance (γ), open probability (P_o), and number of channels (N) expressed at the cell surface. First, we investigated in the inside-out patch-clamp configuration parameters γ and P_o, biophysical parameters intrinsic to the channel. They were essentially identical for WT (γ, 19.1 ± 0.8 pS; P_o, 0.25 ± 0.09, n = 8) and TRPM4^E7K channels (γ, 18.8 ± 0.4 pS; P_o, 0.26 ± 0.02, n = 9; Figure 3, E–G). The data indicate that the observed mutational effect on TRPM4 current amplitude was most likely the result of a change in N.

The PFHBI-associated mutation increases TRPM4 channel density.

We addressed this hypothesis directly and evaluated the number of WT and mutant TRPM4 channels residing in the plasma membrane. In order to evaluate N, we introduced a Myc tag into an extracellular site of TRPM4 between the first and second transmembrane-spanning segments, according to the predicted membrane topology of TRPM4 (Supplemental Figure 5 and ref. 19). Functional tests showed that Myc-tagged WT and mutant TRPM4 channels mediated currents indistinguishable from the untagged channel (Supplemental Figure 6, A and B). Next, we labeled nonpermeabilized HEK 293 cells expressing Myc-tagged WT or TRPM4^E7K channels with FITC-coupled anti-Myc antibody and quantitated N using fluorescence-activated cell sorting (FACS; Figure 4A). HEK 293 cells expressing untagged WT channel were used for background control. In agreement with our electrophysiological results, the FACS results showed approximately 2.5-fold larger N for Myc-tagged TRPM4^E7K than for WT TRPM (Figure 4, A and B).

A likely explanation for increased TRPM4^E7K channel density in the plasma membrane was a mutational influence on TRPM4 protein stability. To address this hypothesis, we investigated the steady-state concentration of TRPM4 WT and mutant channel protein in transiently transfected cells by Western blot and immunofluorescence microscopy. We used FLAG-tagged channels with current density similar to that of untagged controls (Supplemental Figure 6, C and D) to analyze TRPM4 protein concentration in transfected cells. Western blot analysis of respective cell lysates revealed a more intense signal for FLAG-tagged TRPM4^E7K than for WT (n = 3; Figure 4E). Fluorescence image quantitation (see Methods) revealed significantly larger immunofluorescence intensity for TRPM4^E7K than for WT (Figure 4, C and D). Thus, the effect of the PFHBI mutation on TRPM4 channel density correlated with an elevated steady-state level of TRPM4 protein.

The PFHBI-associated mutation affects TRPM4 channel endocytosis.

The observed TRPM4^E7K concentration increase in the plasma membrane indicated an altered balance between the anterograde and retrograde TRPM4 trafficking pathways that determine steady-state protein expression at the cell surface. Because mutations mostly obstruct protein function (22–27), we considered a default in retrograde (i.e., endocytotic) TRPM4 trafficking as a likely explanation for the PFHBI mutational effect. Dynein-based endocytotic trafficking, a common pathway in ion channel endocytosis (28–30), is effectively inhibited in cells that overexpress dynamitin (31, 32). Overexpression of dynamitin is an established method to disrupt the dynein motor system and to interfere with ion channel endocytosis. Thus, we investigated the effect of dynamitin overexpression on TRPM4 channel density. Cotransfection of HEK 293 cells with dynamitin and TRPM4 channel cDNAs resulted in significantly increased TRPM4 channel density (39.3 ± 7.9 pA/pF, n = 9; Figure 5A). Conversely, TRPM4^E7K current density was dynamitin insensitive (42.7 ± 8.7 pA/pF, n = 7; Figure 5A). We conclude that the PFHBI mutation impaired dynein-based TRPM4 channel endocytosis.

Endocytosis is the initial step in retrograde movement of membrane protein. Subsequently, internalized proteins can follow multiple routes with different outcomes. One well-recognized fate of internalized protein is targeting for proteasomal degradation (33). Therefore, we reasoned that WT TRPM4 is more sensitive to proteasomal degradation than is TRPM4^E7K, for which endocytosis is disrupted. To test this hypothesis, we investigated the effect of the proteasome inhibitor MG132 (34) on cell surface expression of WT and mutant Myc-tagged TRPM4 channels. Quantitative FACS analysis of Myc-specific antibody–labeled transfected HEK 293 cells showed that MG132 produced a significant, approximately 3-fold increase in channel density for WT (310% ± 30%, n = 3), but not TRPM4^E7K (Figure 5B). This finding concurs with our observation that the endocytotic TRPM4 trafficking was impaired by PFHBI mutation, possibly by its interference with a regulatory pathway.

The TRPM4^E7K channel is less sensitive to SUMOylation.

The regulatory pathway presumably involves posttranslational TRPM4 modification, for example, by protein phosphorylation and/or dephosphorylation. In keeping with this idea, we applied 8-Brc-AMP to TRPM4-expressing HEK 293 cells to stimulate protein kinase A and then studied the effects of the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine and the protein phosphatase inhibitor okadaic acid on TRPM4 current density. Both WT and TRPM4^E7K current density was insensitive to this treatment (Supplemental Figure 7). As an alternative posttranslational modification, one that may regulate ion channel density in the plasma membrane, we considered SUMOylation (33). First, we directly investigated SUMOylation of FLAG-tagged TRPM4. FLAG-tagged WT or mutant TRPM4 protein was pulled down with immobilized FLAG-specific antibodies from cellular lysates and immunostained with small ubiquitin-like modifier–specific (SUMO-specific) antibodies in Western blots. We found that TRMP4 protein was directly SUMOylated, whereas vector controls showed no detectable SUMOylation signal; moreover, TRPM4^E7K protein displayed markedly stronger immunostaining than did WT protein (n = 3 per group; Figure 6A).

Next, we investigated the effect of the SUMO ligase ubiquitin-conjugating enzyme 9 (Ubc9; ref. 35) and of the SUMO protease SUMO1/sentrin specific peptidase 1 (SENP1; ref. 36) on TRPM4 current density. Coexpression of WT TRPM4 with Ubc9 produced a significant increase of TRPM4 current density (43.3 ± 12.7 pA/pF, n = 9, P = 0.04). Conversely, coexpression of WT TRPM4 with SENP1 decreased TRPM4 current density (4.5 ± 1.0 pA/pF, n = 8) to background levels measured on mock-transfected cells (3.5 ± 0.4 pA/pF, n = 7, P = 0.39), whereas coexpression with the inactive SENP1 mutant SENP1-C603S (37) showed no effect on WT TRPM4 current density (21.6 ± 3.8 pA/pF, n = 9, P = 0.79; Figure 6B). In another control experiment, we investigated the effect of SENP1 protease on human ether-a-go-go–related gene (HERG) channel activity, which plays an important role in ventricular action potential repolarization (38). HERG current density was insensitive to SENP1 (absence, 53.7 ± 10.6 pA/pF, n = 6; presence, 54.8 ± 11.2 pA/pF, n = 7). In contrast to WT TRPM4, coexpression of TRPM4^E7K with Ubc9 (63.3 ± 7.1 pA/pF, n = 17, P = 0.46) or SENP1 (64.5 ± 17.2 pA/pF, n = 6, P = 0.65; Figure 6B) neither significantly increased nor significantly decreased current density. Quantitative FACS analyses in which we investigated surface expression of Myc-tagged TRPM4^E7K after coexpression with Ubc9 or SENP1 yielded similar results (Figure 6C). These results imply that deSUMOylation serves as a signal for TRPM4 internalization and that, accordingly, constitutive SUMOylation protects TRPM4^E7K against endocytosis and proteasomal degradation.

Clinical evaluation and genetic mapping.

The clinical status of members of the Afrikaner pedigree was ascertained at Tygerberg Hospital, University of Stellenbosch, and phenotypically characterized as previously published (5, 9). Clinical and genetic studies were carried out according to the institutional review boards of University of Stellenbosch and the Ethikkommission der τrztekammer Westfalen-Lippe und der Westfälischen Wilhelms—Universität Münster. Informed consent was given by all participating individuals. Detailed clinical and family histories, including cardiac symptoms, device implantation, and standard 12-lead ECGs, were obtained. Genomic DNA was isolated from venous blood samples, and 25 TRPM4 exons were PCR amplified using oligonucleotides located 60–100 bp from the splice sites (Supplemental Table 2). Unrelated healthy individuals (389 mixed European descent, 230 Afrikaner) served as controls for the observed genetic variation.

Cloning of expression plasmids.

For heterologous expression in tissue culture cells, the following cDNAs were cloned into pcDNA3 vector: human TRPM4 (provided by A. Guse, University Hospital Hamburg-Eppendorf, Hamburg, Germany; ref. 52), HERG (generated in our laboratory), dynamitin (provided by R. Vallee, Columbia University, New York, New York, USA; ref. 53), SENP1 (provided by P. O’Hare, Marie Curie Research Institute, Oxted, United Kingdom; ref. 36), and Ubc9 (provided by S. Müller, Max-Planck-Institute of Biochemistry, Munich, Germany; ref. 35). TRPM4^E7K cDNA was obtained by in vitro mutagenesis. For detection in Western blot or immunofluorescence experiments, N-terminal (after Met 1) or C-terminal (after Asp 1214) FLAG tags, respectively, were inserted into the coding sequence by PCR. For quantitative FACS analysis, a Myc-Tag was inserted after Leu 722; the fluorescence image quantitation program ImageJ (http://rsbweb.nih.gov/ij/) was used for quantitative analysis of transfected cells immunostained with anti-FLAG antibodies. Plasmid sequences were verified by sequence analysis.

Quantitative RT-PCR.

Heart tissue samples were a gift of U. Ravens and E. Wettwer (Technische Universität Dresden, Dresden, Germany). Tissue samples were prepared from nondiseased human hearts technically unusable for transplantation. All experimental protocols were approved by the Ethical Review Board of the University Hospital of Dresden. Total RNA was extracted from heart tissue samples using TRIzol (Invitrogen) according to the manufacturer’s instructions. The RNA was reverse transcribed into cDNA with Superscript II RT (Invitrogen) and subjected to real-time PCR. TRPM4 expression was monitored using a TaqMan Assay for TRPM4 (Hs00214167_m1; Applied Biosystems Inc.). For relative expression analysis, a RPLPO TaqMan Assay was used (4333761T; Applied Biosystems Inc.). Analysis was carried out as published previously (54).

Cell culture.

HEK 293, CHO, and African green monkey (COS-7) cells were cultured at 37°C and 5% CO₂ in DMEM (HEK 293 and COS-7) or MEM-α medium (CHO) supplemented with 10% FBS (Invitrogen) and 2 mM glutamine. Plasmid DNA was transfected into HEK 293, COS-7, and CHO cells with Lipofectamine 2000 (Invitrogen) according to the manufacturer’s specifications.

FACS analysis.

HEK 293 cells were transiently transfected with plasmids encoding for TRPM4 control, Myc-tagged TRPM4, or Myc-tagged TRPM4^E7K channels. At 24 hours after transfection, cells were harvested by Accutase (PAA Laboratories GmbH), washed once with PBS, and resuspended at a density of 1 × 10⁶ cells/ml in FACS buffer (PBS plus 0.5% BSA and 0.02% NaN₃). We stained 1 × 10⁵ cells with FITC-coupled Myc-specific antibody (Sigma-Aldrich) at a dilution of 1:50 for 30 minutes at 4°C. Subsequently, cells were washed with 1 ml FACS buffer, resuspended in 400 μl FACS buffer, and directly analyzed for FITC fluorescence.

Western blot.

Transiently transfected COS-7 cells expressing TRPM4 were harvested 24–48 hours after transfection. Cells were pelleted by centrifugation for 3 minutes at 300 g. Pellets were washed twice in PBS and resuspended in 1 ml H₂O. Cells were lysed by freeze-thaw cycles in liquid nitrogen and 37°C H₂O 3 times before the nuclear fraction was pelleted by centrifugation for 3 minutes at 2,000 g (4°C). Supernatant was removed, and the membrane-enriched fraction was pelleted for 20 minutes at 18,000 g (4°C), resuspended in 30–50 μl HEPES-lysis buffer (150 mM NaCl, 50 mM HEPES, and 0.5% Triton X-100, pH 7.4), and subjected to SDS-PAGE. Proteins were transferred onto PROTRAN nitrocellulose membrane (Whatman GmbH) and stained with antibodies against FLAG tag (diluted 1:500) for TRPM4 expression, SUMO-1 (diluted 1:2,000), or actin (diluted 1:2,000; all from Sigma-Aldrich).

Immunofluorescence.

COS-7 cells were plated onto glass coverslips and transiently transfected with expression plasmids for TRPM4 or TRPM4^E7K. At 48 hours after transfection, cells were fixed with 4% PFA and permeabilized with 0.2% Triton X-100. After blockade for unspecific binding with 10% horse serum, cells were stained for TRPM4 expression by mouse FLAG-specific antibody (Sigma-Aldrich) at a dilution of 1:500 at 22°C for 1 hour. After washing, cells were treated with a goat anti-mouse secondary antibody coupled to Alexa Fluor 546 (Invitrogen) for 20 minutes at 22°C in a 1:1,000 dilution. Stained cells were analyzed with a Fluoview FV1000 confocal microscope (Olympus GmbH).

Electrophysiology.

Whole-cell, inside-out, and single-channel patch-clamp recordings were performed with an EPC9 patch-clamp amplifier (HEKA Elektronik) at a sampling rate of 20 kHz. Patch electrodes had a DC resistance between 1 and 3 MΩ when filled with intracellular solution for whole-cell and inside-out patches, and between 6 and 10 MΩ for single-channel recordings. An Ag-AgCl wire was used as a reference electrode. Capacitance and access resistance were monitored continuously, and cell membrane capacitance values were used to calculate current densities. For whole-cell recordings on HEK 293 and CHO cells, the bath solution used was buffer A (156 mM NaCl, 5 mM CaCl₂, 10 mM glucose, and 10 mM HEPES) at pH 7.2. The pipette solution used was buffer B (156 mM CsCl, 1 mM MgCl₂, 10 mM EGTA, and 10 mM HEPES) at pH 7.4. Free Ca²⁺ in the pipette solution was 0.02 and 7.4 μM by addition of CaCl₂. Ca²⁺ concentrations were measured with the Ca²⁺-sensitive dye Fluo4FF (Invitrogen). Holding potential was –60 mV, and current traces were elicited by voltage ramps for 250 ms from –120 to +100 mV (52). Inside-out patch-clamp recordings, for determination of TRPM4 channel block induced by AMP-PNP (Sigma-Aldrich), were performed with buffer A at pH 7.4 as bath and pipette solution. For application of AMP-PNP, cells were superfused with buffer B plus 0.3 mM CaCl₂ at pH 7.2, with and without 500 μM AMP-PNP (20). Holding potential was 0 mV; channels were activated with a 250-ms pulse to +100 mV after a 500-ms pulse to –100 mV (20). For single-channel recordings on CHO cells, bath solution contained 140 mM NaCl, 4.8 mM KCl, 1.2 mM MgCl₂, 10 mM glucose, and 10 mM HEPES (pH 7.4). The pipette solution used was buffer C (145 mM NaCl, 1 mM CaCl₂, 1.2 mM MgCl₂, 10 mM glucose, and 10 mM HEPES) at pH 7.4. For activation of TRPM4 channels, patches were superfused with buffer C at pH 7.2 (47). Bath solution for recordings of HERG channels in HEK 293 cells contained 150 mM NaCl, 1.8 mM CaCl₂, 5 mM glucose, 4 mM KCl, 1 mM MgCl₂, and 10 mM HEPES (pH 7.4). Pipette solution was composed of 126 mM KCl, 4 mM Mg-ATP, 2 mM MgSO₄, 5 mM EGTA, 0.5 mM CaCl₂, and 25 mM HEPES (pH 7.3). We measured K currents by 3.5-second depolarizing test pulses from a holding potential of –80 mV to test potentials between –50 and +50 mV, followed by repolarization to –40 mV (55). Recordings were carried out at room temperature (22°C–25°C).

Accession numbers.

GenBank accession numbers are as follows: Homo sapiens TRPM4 cDNA, NM_017636; Homo sapiens TRPM4, NP_060106; Macaca fascicularis TRPM4, BAC41778; Rattus norvegicus TRPM4, XP_574447; Mus musculus TRPM4, AAH96475; Canis familiaris TRPM4, XP_541500; Ornithorhynchus anatinus TRPM4, XP_001508557; Danio rerio TRPM4, XP_001339981; Xenopus laevis TRPM4, AAH72955.

Statistics.

All data are presented as mean ± SEM. To test for statistical significance, unpaired 2-tailed Student’s t test was used. A P value less than 0.05 was considered significant.