KCNE4 can co-associate with the I Ks (KCNQ1–KCNE1) channel complex Lauren J. Manderfield 1 and Alfred L. George Jr 1,2 1 Department of Pharmacology, Vanderbilt University, Nashville, TN, USA 2 Department of Medicine, Vanderbilt University, Nashville, TN, USA Voltage-gated potassium (K V ) channels are essential for a variety of physiological processes, including the control of membrane potential, electrical excitability and solute transport. Many K V channels are hetero- multimeric protein complexes consisting of pore-form- ing subunits, encoded by a large number of distinct potassium channel gene subfamilies, and accessory proteins. At least four classes of K V accessory subunit have been identified, including K V b [1–4], KChIP [5,6], KChAP [7] and the KCNE proteins [8]. Accessory proteins provide an important mechanism for achiev- ing functional diversity amongst potassium channels. KCNE proteins are small, single transmembrane domain subunits that function to control or modulate K V channels in the heart, cochlea, small intestine and other tissues. KCNE1, originally named minK, was the first identified member of this family [9], and its expression has been demonstrated in several tissues, including the kidney, heart and uterus [10–12]. More than a decade later, the paralogous minK-related Keywords accessory subunits; KCNE4; KCNQ1; K V 7.1; potassium channel Correspondence A. L. George Jr, 529 Light Hall, 2215 Garland Avenue, Nashville, TN 37232-0275, USA Fax: +1 615 936 2661 Tel: +1 615 936 2660 E-mail: al.george@vanderbilt.edu (Received 24 August 2007, revised 11 December 2007, accepted 15 January 2008) doi:10.1111/j.1742-4658.2008.06294.x Voltage-gated potassium (K V ) channels can form heteromultimeric com- plexes with a variety of accessory subunits, including KCNE proteins. Het- erologous expression studies have demonstrated diverse functional effects of KCNE subunits on several K V channels, including KCNQ1 (K V 7.1) that, together with KCNE1, generates the slow-delayed rectifier current (I Ks ) important for cardiac repolarization. In particular, KCNE4 exerts a strong inhibitory effect on KCNQ1 and other K V channels, raising the pos- sibility that this accessory subunit is an important potassium current modu- lator. A polyclonal KCNE4 antibody was developed to determine the human tissue expression pattern and to investigate the biochemical associa- tions of this protein with KCNQ1. We found that KCNE4 is widely and variably expressed in several human tissues, with greatest abundance in brain, liver and testis. In heterologous expression experiments, immunopre- cipitation followed by immunoblotting was used to establish that KCNE4 directly associates with KCNQ1, and can co-associate together with KCNE1 in the same KCNQ1 complex to form a ‘triple subunit’ complex (KCNE1–KCNQ1–KCNE4). We also used cell surface biotinylation to demonstrate that KCNE4 does not impair plasma membrane expression of either KCNQ1 or the triple subunit complex, indicating that biophysical mechanisms probably underlie the inhibitory effects of KCNE4. The obser- vation that multiple KCNE proteins can co-associate with and modulate KCNQ1 channels to produce biochemically diverse channel complexes has important implications for understanding K V channel regulation in human physiology. Abbreviations GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HA, haemagglutinin; I Ks , cardiac slow-delayed rectifier current; I to , cardiac transient outward current; K V channel, voltage-gated potassium channel. 1336 FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors Journal compilation ª 2008 FEBS peptides encoded by human genes KCNE2, KCNE3, KCNE4 and KCNE5 were identified [13,14]. Although different KCNE proteins functionally interact with a variety of K V channels, all KCNE proteins have been shown to modulate heterologously expressed KCNQ1 (K V 7.1) with distinct effects [15–20]. The co-expres- sion of KCNE1 with KCNQ1 reconstitutes I Ks ,a potassium current important for myocardial repo- larization and the most well-studied physiological phenomenon mediated by a KCNE subunit [15,16]. Biophysical and biochemical experiments have demon- strated that two KCNE1 subunits associate with each tetramer of KCNQ1 [21]. All other KCNE proteins exert functional effects on KCNQ1 ranging from potentiation (KCNE3) [18] to suppression (KCNE4, KCNE5) [19,20] of channel activity. Given the varied KCNQ1 phenotypes generated by different KCNE proteins, and the overlapping expression patterns of these subunits [22], there may be multiple and diverse KCNE–KCNQ1 interactions within the same cells or tissues. One of the least characterized, but biophysically potent, members of this family is KCNE4. When expressed in heterologous systems, KCNE4 exerts dra- matic functional effects on KCNQ1 channels. Grunnet et al. [19] first demonstrated complete suppression of KCNQ1 activity by KCNE4 in both oocytes and Chi- nese hamster ovary cells. In addition to KCNQ1, other K V channels, including K V 1.1 and K V 1.3, are also inhibited by KCNE4 [23]. KCNE4 can also exert func- tional inhibition on K V channels even in the presence of other accessory subunits. For example, KCNE4 can inhibit I Ks stably expressed in Chinese hamster ovary cells [24], as well as the transient outward current (I to ) reconstituted in heterologous systems by the co-expres- sion of K V 4.3 with KChIP2 [25]. KCNE4 inhibition of heterologously expressed KCNQ1 in the presence or absence of KCNE1, the overlapping mRNA expression patterns of KCNE and KCNQ1 genes, and the observation that the KCNQ1 tetramer can accommodate at least two KCNE sub- units has raised the possibility that multiple accessory subunits can interact simultaneously with KCNQ1 channels. In this study, the expression of KCNE4 pro- tein was demonstrated in human tissues. We further show that KCNE4 physically interacts with KCNQ1, but does not suppress channel activity by impairing the cell surface expression of this K V channel. Finally, we demonstrated that KCNE1 and KCNE4 can simul- taneously associate with KCNQ1 to form KCNE1– KCNQ1–KCNE4 channel complexes expressed at the plasma membrane. Together, our findings contribute to the understanding of the role of KCNE4 as a potentially important regulator of KCNQ1 and other K V channels. Results Characterization of KCNE4 antibody A rabbit polyclonal antibody raised against a C-termi- nal epitope of human KCNE4 was characterized. The antibody (anti-KCNE4) recognized a single band of approximately 28 kDa on immunoblots of proteins from cells transfected with an epitope (haemagglutinin, HA)-tagged KCNE4 cDNA, but did not recognize spe- cific bands in non-transfected cells or when excess anti- genic peptide was present to block immunodetection (Fig. 1A). An identical band was observed when the immunoblots were probed with anti-HA, but not when the immunoblots were probed with pre-immune rabbit serum. In separate experiments designed to demon- strate specificity, anti-KCNE4 recognized a band of approximately 25 kDa only in cells transfected with untagged KCNE4, and did not exhibit cross-reactivity with other human KCNE proteins (Fig. 1B). The observed mass of the native KCNE4 protein ( 25 kDa) is slightly larger than that predicted from the ORF ( 18 kDa), and we speculate that this dis- crepancy may be the result of anomalous electropho- retic migration of KCNE4 on SDS-PAGE, as observed with other small, highly acidic proteins [26,27]. The molecular mass difference between tagged and untag- ged KCNE4 ( 28 versus  25 kDa) is very consistent with the predicted mass of the epitope tag ( 3 kDa). All subsequent biochemical experiments utilized untag- ged KCNE4 unless otherwise stated. Expression of KCNE4 in human tissues Anti-KCNE4 was utilized to probe immunoblots pre- pared with a panel of 16 human tissues to determine the expression pattern of this protein (Fig. 1C). KCNE4 exhibited the highest levels of expression in the brain, liver and testis. By contrast, colon, lung, placenta and prostate had little or no KCNE4 expres- sion. Many of the tissues examined had been studied previously by real-time quantitative RT-PCR [22] and, in most tissues, mRNA levels were concordant with protein levels. Interestingly, brain and liver, two of the tissues with high levels of KCNE4 protein expression, had low KCNE4 mRNA expression [22]. Conversely, placenta and spleen, two of the tissues with the highest KCNE4 mRNA expression, had low or no KCNE4 protein expression [22]. We inferred from these data that post-transcriptional mechanisms contribute to the L. J. Manderfield and A. L. George Jr KCNE4 co-association with KCNQ1–KCNE1 channels FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors Journal compilation ª 2008 FEBS 1337 steady-state level of KCNE4 protein in certain tissues. A similar lack of correlation between mRNA levels and protein expression has also been observed for KCNE1 throughout regions of the heart [28]. KCNE4 interacts with KCNQ1 We and others have demonstrated that KCNE4 inhib- its KCNQ1 function in vitro [19,24]. We hypothesized that this effect is a result of a direct interaction of KCNE4 with KCNQ1. This hypothesis was tested by examining whether KCNE4 forms protein complexes with KCNQ1. KCNE4 and KCNQ1 were transiently co-expressed in COS-M6 cells, the protein complexes were immunoprecipitated from cellular lysates using a KCNQ1 antibody (anti-KCNQ1), and the immuno- blots were probed with anti-KCNE4. The results indi- cated that KCNE4 interacts with KCNQ1 (Fig. 2). The specificity of this interaction was demonstrated by several control experiments. Pre-incubation of anti- KCNQ1 with antigenic peptide prevented the immuno- precipitation of KCNQ1 or KCNE4 (Fig. 2, lane 3). When cell lysates from cells expressing only KCNQ1 or KCNE4 were mixed, interaction was not observed, thus excluding a post-lysis artefact (Fig. 2, lane 4). Neither KCNE4 nor KCNQ1 was immunoprecipitated with Protein-G Sepharose beads alone (Fig. 2, lane 5) or pre-immune serum matched to the species origin of anti-KCNQ1 (Fig. 2, lane 6). When KCNQ1 and KCNE4 were expressed alone (Fig. 2, lanes 7 and 8), no cross-reactivity was observed between the respective antibodies. These experiments offer conclusive evidence that KCNE4 forms channel complexes with KCNQ1 in vitro. The suppression of I Ks by KCNE4 could poten- tially be explained by displacement or sequestra- tion of KCNE1 by KCNE4. The possibility that KCNE4 can displace KCNE1 from KCNQ1 was + KCNE4 + KCNE4 + antigenic peptide + Rabbit pre- immune serum HAIB: + –– 50 kDa 30 25 37 kDa IB:GAPDH 50 kDa 30 25 IB:KCNE4 NT KCNE1 KCNE2 KCNE3 IB:KCNE4 IB:GAPDH 50 kDa 30 25 35 kDa KCNE4 KCNE5 – – Brain Heart Colon Ileum Kidney Liver Lung Ovary Pancreas Palcenta Prostae Muscle Spleen Testicle Thymus Uterus A B C Fig. 1. Specificity of anti-KCNE4. (A) Whole cell lysates from COS-M6 cells transfected with HA epitope-tagged KCNE4 (+) or non-transfect- ed cells ()) were subjected to SDS-PAGE and western blotting with the indicated immunoreagent. A specific protein with a molecular mass of approximately 28 kDa was identified by immunoblotting with either anti-HA or anti-KCNE4. (B) Western blot of lysates derived from non- transfected cells (NT) or cells expressing each individual KCNE protein probed for KCNE4. All lysates were also probed for GAPDH in order to demonstrate protein expression. (C) Western blot of lysates derived from specified human tissues probed for KCNE4. Brain lysates were derived from the cerebellum. Colon lysates were derived from the descending colon. Heart lysates were derived from the left ventricle. Muscle lysates were derived from skeletal muscle (quadriceps). Supplementary Table S1 provides age and sex information for the tissue donors. All lysates were also probed for GAPDH in order to demonstrate protein expression. KCNE4 co-association with KCNQ1–KCNE1 channels L. J. Manderfield and A. L. George Jr 1338 FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors Journal compilation ª 2008 FEBS first examined by testing whether KCNE1 remained associated with KCNQ1 even in the presence of KCNE4. Cells were transiently transfected with KCNE4, KCNE1 3FLAG and KCNQ1, and the cell lysates were subjected to immunoprecipitation with anti-KCNQ1. In these experiments, KCNE4 and KCNE1 interactions with KCNQ1 were detected by immunoblot using anti-KCNE4 or anti-FLAG. Figure 3 illustrates that, in cells transfected with all three channel subunits, anti-KCNQ1 immunoprecipi- tates both KCNE1 and KCNE4. This interaction was specific for the KCNQ1 antibody, did not occur during processing of the cell lysates, and could not be attributed to antibody cross-reactivity or non-spe- cific interactions with Protein-G Sepharose. In the immunoblots in Fig. 3, KCNE4 appears as a dou- blet, which may be a result of post-translational pro- cessing. These experiments demonstrate that KCNQ1 can associate with both KCNE1 and KCNE4 in the same population of cells, providing evidence that dis- tinct KCNQ1–KCNE1 and KCNQ1–KCNE4 com- plexes are formed. We next examined the hypothesis that KCNE4 directly binds and sequesters KCNE1 was examined as an explanation of why KCNE4 functionally suppresses I Ks [24]. KCNE4 and an epitope-tagged KCNE1 (KCNE1 3FLAG ) were co-expressed and immunoprecipi- tated with anti-FLAG, followed by immunoblotting using anti-KCNE4. There was no evidence of KCNE1–KCNE4 interaction when both subunits were co-expressed (Fig. 4A). Furthermore, both KCNE subunits were expressed at the plasma membrane (Fig. 4B), and this observation rules out intracellular degradation as an explanation for a lack of KCNE1– KCNE4 interaction [29]. The apparent decrease in KCNE1 at the plasma membrane in the presence of KCNE4 is not sufficient to explain the dominant effect of KCNE4 on I Ks . The multiple molecular mass bands ranging from approximately 15 to 25 kDa observed in the immunoblots probed for KCNE1 3FLAG represent differentially glycosylated forms of this protein that have been described previously [30]. KCNE1 and KCNE4 co-assemble with KCNQ1 The existence of KCNQ1–KCNE1 complexes in the experiment described above would be expected to con- tribute some level of I Ks expression. However, this was not observed in previous electrophysiological studies when KCNQ1, KCNE1 and KCNE4 were co- expressed [24]. One possible explanation is that all three subunits form a triple subunit complex (i.e. KCNE1–KCNQ1–KCNE4) in which KCNE4 exerts a dominant inhibitory effect. To probe for the existence of KCNE1–KCNQ1–KCNE4, we examined whether both KCNE1 and KCNE4 could be incorporated into the same KCNQ1 complex. As stated above, we estab- lished that these two different KCNE subtypes did not interact with each other in the absence of KCNQ1 (Fig. 3) and that both subtypes bound KCNQ1 when co-expressed in the same cell population (Fig. 4). To detect KCNE1–KCNQ1–KCNE4 complexes, cells were transfected with all three channel subunits, the cell lysates were immunoprecipitated using anti-FLAG (recognizes KCNE1), and KCNE4 was immunodetect- ed. Figure 5 illustrates that anti-FLAG was indeed able to co-immunoprecipitate both KCNQ1 and KCNE4, thus providing evidence for the existence of KCNE1–KCNQ1–KCNE4 complexes. These interac- tions were specific, as demonstrated by the absence of co-immunoprecipitation of KCNE4 and KCNQ1 in any of the control conditions. Therefore, these data represent biochemical evidence for a KCNQ1 chan- nel complex incorporating two different KCNE subunits. +–+++++– –++++++– KCNQ1 KCNE4 30 50 kDa 25 IP:KCNQ1 IB:KCNE4 75 kDa IB:KCNQ1 25 30 kDa IB:KCNE4 IP:KCNQ1 IB:KCNQ1 75 kDa 50 12 456783 Fig. 2. KCNE4 interacts with KCNQ1. Whole cell lysates were immunoprecipitated with anti-KCNQ1 and then subjected to SDS- PAGE and western blot analysis. Lane 1, non-transfected COS-M6 cells. Lane 2, cells transfected with KCNQ1 and KCNE4. Lane 3, cells transfected with KCNQ1 and KCNE4, but anti-KCNQ1 used for immunoprecipitation was pre-incubated with antigenic peptide. Lane 4, mixture of lysates from cells expressing either KCNQ1 or KCNE4 only combined prior to immunoprecipitation. Lane 5, KCNQ1 and KCNE4 transfected cells immunoprecipitated with Pro- tein-G Sepharose. Lane 6, KCNQ1 and KCNE4 transfected cells immunoprecipitated with goat pre-immune serum. Lane 7, cells expressing KCNQ1 only. Lane 8, cells expressing KCNE4 only. The first immunoblot shows samples immunoprecipitated with anti- KCNQ1 and immunoblotted for KCNQ1. The second image shows a KCNQ1 immunoblot of the initial lysates demonstrating KCNQ1 expression. The third immunoblot shows the anti-KCNQ1 immuno- precipitated samples which were probed with anti-KCNE4. The final image shows a KCNE4 immunoblot of the initial lysates demon- strating KCNE4 expression. L. J. Manderfield and A. L. George Jr KCNE4 co-association with KCNQ1–KCNE1 channels FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors Journal compilation ª 2008 FEBS 1339 KCNE4 does not inhibit KCNQ1 trafficking One potential mechanism by which KCNE4 could sup- press I Ks is by impairing KCNQ1 cell surface expres- sion. This possibility was examined previously by Grunnet et al. [19], where it was demonstrated that KCNE4 did not decrease KCNQ1 plasma membrane expression in Xenopus oocytes assayed by cell surface biotinylation. Here we investigated whether KCNE4 expression affected KCNQ1 plasma membrane expres- sion in mammalian cells, and whether KCNQ1, KCNE1 and KCNE4 reached the cell surface when co-expressed. We examined KCNQ1 trafficking when KCNQ1 was either expressed alone, with KCNE1 or with KCNE4. KCNQ1 co-expression with KCNE1 served as a control for KCNQ1 trafficking, as it is presumed that KCNE1–KCNQ1 complexes reach the plasma membrane to enable functional I Ks . Total protein, non-biotinylated and biotinylated fractions from cells expressing KCNQ1 only, KCNQ1 with KCNE1 3FLAG and KCNQ1 with KCNE4 3HA were collected and probed with anti-KCNQ1. Figure 6 illustrates that KCNQ1 cell surface expression was not inhibited by the expression of either KCNE1 or KCNE4. KCNQ1 was specifically detected in all protein fractions under all three conditions. KCNQ1 was not immunodetected in any fraction from non-transfected cells (data not shown). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) protein was immunodetected in only the total protein and non-biotinylated fractions, demon- strating clean separation of plasma membrane +++++++– +++++++– +++++++– –+– +–– ––+ KCNQ1 KCNE4 KCNE1 3FLAG 75 kDa 50 IP:KCNQ1 IB:KCNQ1 75 kDa IB:KCNQ1 25 30 kDa IB:KCNE4 25 kDa 15 IB:FLAG 50 kDa 30 25 IP:KCNQ1 IB:KCNE4 30 kDa 25 15 IP:KCNQ1 IB:FLAG 87654321 91011 Fig. 3. KCNE4 interacts with KCNQ1 in the presence of KCNE1. Whole cell lysates were immunoprecipitated with anti-KCNQ1 and then subjected to SDS-PAGE and western blot analysis. Lane 1, non-transfected COS-M6 cells. Lane 2, cells transfected with KCNQ1, KCNE4 and KCNE1 3FLAG . Lane 3, cells transfected with KCNQ1, KCNE4 and KCNE1 3FLAG , but anti-KCNQ1 used for immunoprecipitation was pre- incubated with an antigenic peptide. Lane 4, mixture of lysates from cells expressing KCNQ1, KCNE4 or KCNE1 3FLAG only were combined prior to immunoprecipitation. Lane 5, mixture of lysates from cells expressing either KCNQ1 and KCNE4 or KCNE1 3FLAG only combined prior to immunoprecipitation. Lane 6, mixture of lysates from cells expressing either KCNQ1 and KCNE1 3FLAG or KCNE4 only combined prior to immunoprecipitation. Lane 7, KCNQ1, KCNE4 and KCNE1 3FLAG transfected cells immunoprecipitated with Protein-G Sepharose. Lane 8, KCNQ1, KCNE4 and KCNE1 3FLAG transfected cells immunoprecipitated with goat pre-immune serum. Lane 9, cells expressing KCNQ1 only. Lane 10, cells expressing KCNE1 3FLAG only. Lane 11, cells expressing KCNE4 only. The first row of immunoblots shows samples immuno- precipitated with anti-KCNQ1 and immunoblotted for KCNQ1. The second row of immunoblots shows the initial lysates confirming KCNQ1 expression. The third row of immunoblots shows the anti-KCNQ1 immunoprecipitated samples that were probed with the KCNE4 antibody. The fourth row of immunoblots shows the initial lysates confirming KCNE4 expression. The fifth row of immunoblots shows the anti-KCNQ1 immunoprecipitated samples which were probed with the FLAG antibody. The sixth row of immunoblots shows the initial lysates confirming KCNE1 expression. KCNE4 co-association with KCNQ1–KCNE1 channels L. J. Manderfield and A. L. George Jr 1340 FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors Journal compilation ª 2008 FEBS (biotinylated fraction) and cytosolic proteins (non-bio- tinylated fraction). Similarly, calnexin was only immu- nodetected in the total protein and non-biotinylated fractions (data not shown). The percentage of KCNQ1 protein present at the cell surface was not significantly different between the three conditions (KCNQ1 alone, 49.8 ± 8.4%; KCNQ1 plus KCNE1, 40.1 ± 7.6%; KCNQ1 plus KCNE4, 31.0 ± 3.4%; mean ± SEM; n = 3 each; Fig. 6), indicating that impaired KCNQ1 cell surface expression cannot explain the suppression of I Ks by KCNE4. Next, we examined the ability of KCNE proteins to traffic to the plasma membrane in the presence of KCNQ1. Total protein, non-biotinylated and biotiny- lated fractions from cells transfected with KCNQ1 plus KCNE1 and KCNQ1 plus KCNE4 were collected and probed for KCNE1 (anti-FLAG) or KCNE4 (anti-HA). Figure 7A illustrates that KCNE1 reaches the cell surface in the presence of KCNQ1. Multiple bands representing the differentially glycosylated forms of KCNE1 were detected, indicating normal matura- tion of the protein. Figure 7B illustrates that KCNE4 also traffics to the cell surface in the presence of KCNQ1. Finally, we examined if KCNE1–KCNQ1–KCNE4 complexes exist at the cell surface. The three channel subunits were co-expressed, and total protein, non-bio- tinylated and biotinylated fractions were collected. Figure 7C illustrates qualitatively that KCNQ1, KCNE1 and KCNE4 were all detected in the biotiny- lated fraction, suggesting expression of the KCNE1– KCNQ1–KCNE4 complex at the surface plasma mem- brane (Fig. 7C). Discussion In this study, we demonstrated the expression pattern of KCNE4 protein in human tissues, and provided in vitro biochemical evidence that KCNE4 interacts with KCNQ1. We also determined that KCNE1 and KCNE4 can simultaneously co-associate with KCNQ1 to form KCNE1–KCNQ1–KCNE4 ‘triple’ subunit complexes, and that the inhibitory effect of KCNE4 cannot be explained by impaired cell surface 30 kDa 75 kDa IB:Transferrin 75 kDa IB:Transferrin 50 kDa IB:KCNE4 Biotinylated Fractions 25 15 E1NT E1+E4 IB:FLAG 30 25 NT E4 E1+E4 + – + + + + + – – + + + + + + – KCNE1 3FLAG KCNE4 IB:FLAG IP:FLAG IB:KCNE4 IB:KCNE4 IP:FLAG IB:KCNE1 30 kDa 25 15 25 kDa 15 30 50 kDa 25 30 kDa 25 1 8 7 6 5 4 3 2 A B Fig. 4. KCNE4 does not interact with KCNE1. (A) Whole cell lysates were immunoprecipitated with anti-FLAG and then subjected to SDS-PAGE and western blot analysis. Lane 1, non-transfected COS-M6 cells. Lane 2, cells transfected with KCNE1 3FLAG and KCNE4. Lane 3, cells transfected with KCNE1 3FLAG and KCNE4, but anti-FLAG used for immunoprecipitation was pre-incubated with an antigenic peptide. Lane 4, mixture of lysates from cells expressing either KCNE4 or KCNE1 3FLAG only combined prior to immunoprecipi- tation. Lane 5, KCNE1 3FLAG and KCNE4 transfected cells immuno- precipitated with Protein-G Sepharose. Lane 6, KCNE1 3FLAG and KCNE4 transfected cells immunoprecipitated with mouse pre- immune serum. Lane 7, cells expressing KCNE1 3FLAG only. Lane 8, cells expressing KCNE4 only. The first immunoblot shows samples immunoprecipitated with anti-FLAG and immunoblotted for KCNE1. The second blot shows a FLAG immunoblot of the initial lysates confirming KCNE1 expression. The third immunoblot shows the anti-FLAG immunoprecipitated samples which were probed with the KCNE4 antibody. The fourth image shows a KCNE4 immunoblot of the initial lysates confirming KCNE4 expression. (B) Representa- tive western blots examining KCNE1 and KCNE4 protein trafficking to the plasma membrane. The protein lysate composition of each lane is denoted as NT for non-transfected, E1 for KCNE1 3FLAG ,E4 for KCNE4 and E1 + E4 for KCNE1 3FLAG +KCNE4. Only the bio- tinylated fractions are illustrated. Lysates were probed with anti- FLAG to demonstrate the presence of KCNE1, or anti-KCNE4 to demonstrate the presence of KCNE4. All lysates were also probed with an antibody against transferrin to demonstrate complete sepa- ration of biotinylated proteins. L. J. Manderfield and A. L. George Jr KCNE4 co-association with KCNQ1–KCNE1 channels FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors Journal compilation ª 2008 FEBS 1341 expression. The observation that multiple KCNE pro- teins can associate with and modulate KCNQ1 chan- nels at the plasma membrane to produce biochemically diverse channel complexes has important implications for understanding physiologically relevant channel reg- ulation. +++++++– +++++++– +++++++– +–– ––+ –+– KCNQ1 KCNE4 KCNE1 3FLAG 87654321 9 10 11 30 kDa 25 IB:KCNE4 30 50 kDa 25 IP:FLAG IB:KCNE4 25 kDa 15 IB:FLAG 30 kDa 25 15 IP:FLAG IB:KCNE1 75 kDa 50 IP:FLAG IB:KCNQ1 IB:KCNQ1 75 kDa Fig. 5. KCNE1 and KCNE4 co-assemble with KCNQ1. Whole cell lysates were immunoprecipitated with anti-FLAG and then subjected to SDS-PAGE and western blot analysis. All lane compositions are the same as defined in Fig. 3. The first row of immunoblots shows samples immunoprecipitated with anti-FLAG and immunoblotted for KCNE1. The second row shows FLAG immunoblots of the initial lysates confirm- ing KCNE1 expression. The third row of immunoblots shows the anti-FLAG immunoprecipitated samples which were probed with KCNQ1 antibody. The fourth row of immunoblots shows the initial lysates confirming KCNQ1 expression. The fifth row of immunoblots shows the anti-FLAG immunoprecipitated samples that were probed with the KCNE4 antibody. The sixth row of immunoblots shows the initial lysates confirming KCNE4 expression. 35 kDa 75 kDa 50 Total protein KCNQ1 alone KCNQ1 + KCNE1 IB:KCNQ1 IB:GAPDH KCNQ1 + KCNE4 0 10 20 30 40 50 60 70 NS Non-biotinylated Biotinylated Total protein Non-biotinylated Biotinylated Total protein Non-biotinylated Biotinylated % Surface expression KCNQ1 + KCNE4 KCNQ1 + KCNE1 KCNQ1 alone Fig. 6. KCNE4 does not inhibit KCNQ1 cell surface expression. Representative western blots examining KCNQ1 trafficking to the plasma membrane when expressed alone, with KCNE1 or with KCNE4. The protein lysate composition of each lane is denoted as total protein, non-biotinylated or biotinylated for the three conditions examined. A bar graph illustrates the relative proportions of surface- expressed KCNQ1 as a percentage of total protein for the three conditions (NS, non-significant). All lysates were probed with anti-KCNQ1 to demonstrate KCNQ1 expression and with anti-GAPDH in order to demonstrate complete separation of biotinylated and non-biotinylated proteins. KCNE4 co-association with KCNQ1–KCNE1 channels L. J. Manderfield and A. L. George Jr 1342 FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors Journal compilation ª 2008 FEBS Diversity of KCNE4 expression KCNE4 protein is expressed widely in both excitable and non-excitable human tissues, suggesting that this subunit could impact a wide array of cell types and physiological functions. Excitable tissues that express KCNE4 include the brain, heart and skeletal muscle. The expression of KCNE4 in brain, coupled with the previously demonstrated inhibitory effect of this sub- unit on K V 1.1 and K V 1.3 channels, raises the possibil- ity of important physiological effects on neuronal excitability, synaptic neurotransmission and impulse conduction [23]. In the heart, we speculate that KCNE4 exerts a suppressive effect on I Ks and may be critical for the regulation of cardiac repolarization. The fact that I Ks has been detected in cardiac myocytes suggests that KCNE4 does not associate with all avail- able KCNQ1 channels, possibly because of excess KCNE1, the most highly expressed KCNE mRNA in heart [24]. We previously showed significant changes in KCNE4 mRNA expression in the setting of end-stage cardiomyopathy [24] and there have been recent sug- gestions of an influence of KCNE4 polymorphisms on the susceptibility to atrial arrhythmias [31]. There are no data available on the role of KCNE4 in skeletal muscle. KCNE4 also exhibits robust expression in epithelial tissues, including the pancreas and kidney. Several studies have indicated that pancreatic acinar cells gen- erate a slowly activating potassium current resembling I Ks , and that this current promotes a driving force for efficient chloride secretion [32,33]. Conceivably, KCNE4 could modulate pancreatic exocrine secretion through attenuation of the I Ks -like current. In the kid- ney, KCNE4 may interact with KCNQ1, which is localized to the lumenal membranes of the mid to late proximal tubule [15,34,35]. Evidence from studies of knockout mice has revealed that KCNQ1 is important for proximal tubule repolarization and the mainte- nance of the electrical driving force for Na + reabsorp- tion under conditions of enhanced electrogenic reabsorption [36]. We speculate that renal expression of KCNE4 may modulate this channel activity and affect reabsorption, but additional studies are needed to demonstrate the subcellular location of this protein in the proximal tubule. Multiple KCNE subunits can co-associate with KCNQ1 There is substantial evidence for the functional diver- sity of potassium channels as a result of the differential assembly of channel subunits. Functional diversity can B 35 kDa IB:GAPDH 50 kDa T NB B KCNQ1 + KCNE4 30 25 IB:HA C IB:KCNQ1 IB:GAPDH IB:FLAG KCNQ1 + KCNE1 + KCNE4 B 75 kDa 50 35 kDa 30 kDa 25 15 30 25 50 kDa IB:HA A IB:GAPDH 35 kDa KCNQ1 + KCNE1 IB:FLAG NB B 30 kDa 25 15 T T NB Fig. 7. KCNE trafficking in the presence of KCNQ1, and KCNQ1 trafficking in the presence of multiple KCNE proteins. Representa- tive western blots examining KCNE protein trafficking to the plasma membrane when expressed with KCNQ1. The protein lysate composition of each lane is denoted as T for total protein, NB for non-biotinylated proteins and B for biotinylated proteins. All lysates were probed with anti-GAPDH in order to demonstrate complete separation of biotinylated and non-biotinylated proteins. (A) KCNE1 traffics to the plasma membrane in the presence of KCNQ1. All lysates were probed with anti-FLAG to demonstrate the presence of KCNE1. (B) KCNE4 traffics to the plasma mem- brane in the presence of KCNQ1. All lysates were probed with anti- HA to demonstrate the presence of KCNE4. (C) KCNQ1 traffics to the plasma membrane in the presence of KCNE1 and KCNE4. The top immunoblot was probed with anti-KCNQ1 to examine KCNQ1 expression. The second immunoblot was probed with anti-FLAG to examine KCNE1 expression. The third immunoblot was probed with anti-HA to examine KCNE4 expression. L. J. Manderfield and A. L. George Jr KCNE4 co-association with KCNQ1–KCNE1 channels FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors Journal compilation ª 2008 FEBS 1343 be achieved through either the assembly of different pore-forming subunits, as illustrated by the generation of the neuronal M-current through the co-assembly of KCNQ2 and KCNQ3 (K V 7.2 and K V 7.3) [37], or by the association of channels with different accessory subunits. Conceivably, the variety of channel com- plexes can be expanded further by mechanisms com- bining pore-forming subunits with multiple different types of accessory subunits. In considering this possi- bility with regard to the KCNE family, we were inspired by the well-established heteromultimeric nat- ure of neuronal voltage-gated sodium channels which comprise a single a-subunit combined with two distinct accessory b-subunits. This precedent led us to investi- gate the possibility that more than one type of KCNE protein could simultaneously co-associate with KCNQ1. We first proposed that KCNQ1 could associate with two different KCNE proteins based on our find- ing that the transient expression of KCNE4 in a cell line stably expressing KCNQ1 and KCNE1 (I Ks cells) suppressed I Ks [24]. This observation suggested that either KCNE1 was displaced from KCNQ1 com- plexes, or that KCNE4 and KCNE1 associated jointly with KCNQ1, but the inhibitory effect of KCNE4 was dominant. In our study, biochemical strategies were applied to determine that the latter explanation was most likely. Since our initial biophysical study, other groups have examined KCNQ1 co-association with KCNE1 and KCNE2 [38,39]. One of these stud- ies used a similar biochemical strategy to demonstrate the presence of a KCNE1–KCNQ1–KCNE2 channel complex [39]; however, these data are somewhat difficult to interpret because of the evidence for KCNQ1 protein aggregation (aberrant molecular mass of KCNQ1 monomer) and the absence of control experiments to exclude artefactual subunit interac- tions. In Caenorhabditis elegans, an A-type K + chan- nel, KVS-1, has been shown to biophysically associate with MPS-2 and MPS-3, two KCNE-related subunits [40]. The association of both MPS proteins with KVS-1 generates potassium currents which are distinct from those generated when KVS-1 is expressed with either MPS-2 or MPS-3 alone [40]. We speculate that other K V channel complexes can be modulated dis- tinctly by the incorporation of multiple KCNE subunits. KCNE4 does not impair KCNQ1 membrane expression Finally, it was tested whether impairment of cell surface expression might explain the inhibition of KCNQ1 by KCNE4. Certain classes of potassium channel accessory subunit (i.e. K V b) have been shown to increase membrane expression of K V channel a-sub- units [41,42], and it was hypothesized that other types of accessory subunit could have the opposite effect. Indeed, a missense KCNE1 mutant (L51H) associated with congenital long-QT syndrome causes retention of both KCNE1 and KCNQ1 in the endoplasmic reticu- lum [43,44]. One previous study examined KCNQ1 and K V 1.1 trafficking and found that KCNE4 did not diminish the cell surface expression of either K V chan- nel [19,23]. We confirmed this finding related to KCNQ1, but also demonstrated cell surface expression of KCNE4 protein and the triple KCNE1–KCNQ1– KCNE4 complex. Mechanisms other than impaired plasma membrane expression must explain the impaired KCNQ1 func- tion in the presence of KCNE4. For example, KCNE4 may cause a strong shift in the voltage dependence of activation, or lock the channel in a closed state by immobilizing the activation gate or voltage sensor. A strong depolarizing shift in activa- tion appears to explain the suppression of KCNQ1 by KCNE5 [20]. There are many examples of K V channel gating modulation by accessory subunits, and this provides another potential mechanism for KCNE4 effects. For example, heterologous expression of K V 1.5 generates a non-inactivating current when expressed alone, but becomes a rapidly inactivating outward current when co-expressed with K V b1 [45]. The ability of K V b1 to dramatically alter the K V 1.5 current has been attributed to a specific structure within the N-terminus of the protein that is similar to the inactivating N-terminal peptide in A-type K V channels [1,46]. This structure allows for rapid inacti- vation of the channel through blocking of the internal pore following depolarization of the membrane. The KCNE4 protein possesses a large cytoplasmic C-ter- minal tail. We speculate that this structure may func- tion in a similar, albeit voltage-independent, manner to block the internal pore of KCNQ1. The C-termi- nus of KCNE4 might also stabilize the channel in another non-conducting state. There have been no investigations into the structural determinants of KCNE4 inhibition. Conclusions KCNE4 is a widely expressed K V accessory subunit implicated in the assembly of biophysically diverse channel complexes in both excitable and non-excitable tissues. The inhibitory actions of KCNE4 are exerted at the plasma membrane, but the precise functional KCNE4 co-association with KCNQ1–KCNE1 channels L. J. Manderfield and A. L. George Jr 1344 FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors Journal compilation ª 2008 FEBS mechanism remains unknown. KCNE4 can co-associ- ate with KCNE1 and KCNQ1 to form a heteromulti- meric complex that is non-functional at the cell membrane. These findings indicate that KCNE4 is a physiologically relevant K V channel modulator. Experimental procedures Generation of KCNE4 polyclonal antibody A polyclonal rabbit antibody, targetted to a unique sequence in the KCNE4 C-terminus (residues 73–94, YKDEERLWGEAMKPLPVVSGLR), was generated by Proteintech Group, Inc. (Chicago, IL, USA). A cysteine residue was added to the N-terminus of the peptide to facil- itate KLH conjugation. Sera were screened using ELISA and the final antibody preparations were affinity purified against the antigenic peptide. Construction of epitope-tagged KCNE1 and KCNE4 KCNE1 and KCNE4 were subcloned into pIRES2-EGFP (BD Biosciences-Clontech, Mountain View, CA, USA), as described previously [24]. A triple FLAG epitope (DY- KDHDGDYKDHDIDYKDDDDK) was introduced by recombinant PCR into the KCNE1 cDNA, immediately upstream of the stop codon. The FLAG sequence was PCR amplified from p3XFLAG-CMVÔ-13 (Sigma-Aldrich, St Louis, MO, USA). Similarly, a triple HA epitope (YP- YDVPDYAGYPYDVPDYAGSYPYDVPDYA) was intro- duced into the KCNE4 cDNA, immediately upstream of the stop codon. The HA sequence was PCR amplified from a plasmid provided by Sabina Kuperschmidt (Vanderbilt University, Nashville, TN, USA). All constructs were veri- fied by complete sequencing of the coding regions. The addition of epitope tags did not affect the electrophysiologi- cal effects of KCNE1 or KCNE4 (supplementary Figs S1 and S2). Cell culture and transfection COS-M6 cells were grown at 37 °Cin5%CO 2 in Dul- becco’s modified Eagle’s medium (DMEM; Life Technolo- gies, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (ATLANTA Biologicals, Norcross, GA, USA), penicillin (50 unitsÆ mL )1 ), streptomycin (50 lgÆmL )1 ) (Life Technologies) and 20 mm HEPES. COS-M6 cells were transiently transfected using FuGene-6 (Roche Applied Science, Indianapolis, IN, USA). Full-length KCNQ1 was expressed from the pcDNA5 ⁄ FRT vector (Invitrogen, San Diego, CA, USA), whereas all KCNE cDNAs were constructed in pIRES2-EGFP. Cells were harvested 48 h post-transfection. Preparation of cell lysates Two 100 mm dishes of COS-M6 cells were transfected per condition, and two dishes of non-transfected COS-M6 cells were used in parallel as a control. Forty-eight hours post- transfection, cells were placed on ice and washed twice with ice-cold phosphate buffered saline (PBS) (137 mm NaCl, 2.7 mm KCl, 10 mm Na 2 HPO 4 ,2mm KH 2 PO 4 , pH 7.4). The cells from one dish were lysed with 1 mL of ice-cold NP-40 buffer (1% NP-40, 150 mm NaCl, 50 mm Tris, pH 8.0) supplemented with a Complete miniprotease inhibi- tor tablet (Roche Applied Science) for 3 min. Cells were then scraped and incubated on ice for another 3 min. The lysate was transferred to a 1.5 mL microfuge tube and rocked for 15 min at 4 °C, followed by centrifugation at 14 000 g for 10 min. The supernatant was collected and centrifuged again under the same conditions. Prior to immunoprecipitation experiments, aliquots of the final supernatant were incubated with Protein-G SepharoseÔ 4 Fast Flow (GE Healthcare Life Sciences, Piscataway, NJ, USA) to pre-clear non-specific protein binding to the Sepharose beads. Final pre-cleared lysates were quantified using the Bradford reagent (Bio-Rad Laboratories, Hercu- les, CA, USA), and equal amounts of proteins were used in the immunoprecipitation experiments. Preparation of tissue lysates Human autopsy tissues were obtained from the NICHD Brain and Tissue Bank for Developmental Disorders under contracts N01-HD-4-3368 and N01-HD-4-3383. All tissues had previously been frozen and stored at )80 °C. Prior to homogenization, tissues were ground with a mortar and pestle. Three millilitres of ice-cold NP-40 buffer with pro- tease inhibitors was added to 1 g of ground tissue and homogenized for 30 s using a mechanical homogenizer (Tekmar Company, Cincinnati, OH, USA). Homogenates were rocked at 4 °C for 1 h, and then centrifuged at 14 000 g for 10 min. The supernatant was collected and centrifuged again under the same conditions. The Bradford reagent was used to quantify the protein concentration in the final lysates. Cross-linking of antibodies to Protein-G Sepharose Ten micrograms of antibody were combined with 750 lLof borate buffer (200 mm sodium tetraborate decahydrate, pH 9.0), added to 50 lL of Protein-G SepharoseÔ 4 Fast Flow and rocked at room temperature for 1 h. The beads were then washed twice with borate buffer. After the sec- ond wash, the beads were resuspended in 1 mL of borate buffer supplemented with 20 mm dimethyl pimelimidate di- hydrochloride, and rocked at room temperature for 30 min. L. J. Manderfield and A. L. George Jr KCNE4 co-association with KCNQ1–KCNE1 channels FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors Journal compilation ª 2008 FEBS 1345 [...]... washing twice with DPBS with 100 mm glycine The cells were then incubated in the same DPBS with glycine solution for 10 min and then washed twice with DPBS Cellular lysates were then prepared as described above, except that the cells were lysed with ice-cold RIPA buffer (150 mm NaCl, 50 mm Tris-Base, pH 7.5, 1% IGEPAL, 0.5% sodium deoxycholate, 0.1% SDS, supplemented with a Complete miniprotease inhibitor... part of the online article from http://www.blackwell-synergy.com KCNE4 co-association with KCNQ1–KCNE1 channels Please note: Blackwell Publishing are not responsible for the content or functionality of any supplementary materials supplied by the authors Any queries (other than missing material) should be directed to the corresponding author for the article FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors... at 4 °C The supernatant fraction was collected and an aliquot was retained as the total protein fraction The remaining supernatant was incubated with ImmunoPure Immobilized Streptavidin beads (Pierce Chemical Co.) overnight at 4 °C The samples were centrifuged for 1 min at 14 000 g, and the supernatant fraction was collected and retained as the non-biotinylated fraction The beads were washed with RIPA... three times for 5 min at 4 °C Then, the biotinylated proteins were eluted with Laemmli sample buffer (Bio-Rad Laboratories) for 30 min at room temperature, and this final elution was retained as the biotinylated fraction All fractions were then subjected to western blot analysis as described above, with the exception that all membranes were incubated overnight at 4 °C with the appropriate primary antibody... transfection, the cells were washed twice with ice-cold Dulbecco’s PBS (DPBS) supplemented with CaCl2 and MgCl2 (GIBCO ⁄ Invitrogen, Grand Island, NY, USA) After washing, the cells were incubated in DPBS plus 1.5 mgÆmL)1 sulfo-NHS-biotin (Pierce Chemical Co., Rockford, IL, USA) for 1 h on ice with shaking The biotinylation solution was removed and the reaction was quenched by washing twice with DPBS with 100.. .KCNE4 co-association with KCNQ1–KCNE1 channels L J Manderfield and A L George Jr The cross-linking reaction was quenched by the addition of ethanolamine buffer (200 mm ethanolamine, pH 8.0) at room temperature for 5 min The Sepharose beads were centrifuged and resuspended in 1 mL of ethanolamine buffer and rocked at room temperature for 1 h The beads were washed twice in PBS and then resuspended... temperature with a mouse transferrin monoclonal antibody (1 : 500, Zymed Laboratories ⁄ Invitrogen Corporation, Carlsbad, CA, USA) Proteins were quantified by densitometry using imagej software (National Institutes of Health, Bethesda, MD, USA) The percentage of protein at the membrane was calculated by dividing the biotinylated fraction value by the total protein fraction value, and is reported as the mean... research fellowship FEBS Journal 275 (2008) 1336–1349 ª 2008 The Authors Journal compilation ª 2008 FEBS L J Manderfield and A L George Jr award from the American Heart Association Southeast Affiliate We wish to thank Dr Carlos Vanoye for valuable comments on the manuscript We also acknowledge the Brain and Tissue Bank for Developmental Disorders at the University of Maryland for providing human tissue samples... resuspended in PBS and stored at 4 °C until use Co-immunoprecipitations Cellular lysates pre-cleared with Protein-G Sepharose were rocked with 50 lL of cross-linked antibody at 4 °C for 4 h Following incubation, the samples were washed three times with ice-cold NP-40 for 5 min at 4 °C Immune complexes were then eluted with SDS-PAGE sample buffer (1% SDS, 50 mm Tris, 1% glycerol, 100 mm dithiothreitol) at 50... proteins to form cardiac IKs potassium channel Nature 384, 80–83 16 Barhanin J, Lesage F, Guillemare E, Fink M, Lazdunski M & Romey G (1996) KvLQT1 and IsK (minK) proteins associate to form the IKs cardiac potassium current Nature 384, 78–80 17 Tinel N, Diochot S, Borsotto M, Lazdunski M & Barhanin J (2000) KCNE2 confers background current characteristics to the cardiac KCNQ1 potassium channel EMBO J 19, . establish that KCNE4 directly associates with KCNQ1, and can co-associate together with KCNE1 in the same KCNQ1 complex to form a ‘triple subunit’ complex (KCNE1–KCNQ1 KCNE4) and KCNE4 can simultaneously co-associate with KCNQ1 to form KCNE1–KCNQ1 KCNE4 ‘triple’ subunit complexes, and that the inhibitory effect of KCNE4 cannot