(Received for publication, November 11, 1996, and in revised form, February 14, 1997)
From the Department of Pharmacology II, ** Department
of Molecular Biology and Biochemistry, Faculty of Medicine, Osaka
University, Suita, Osaka 565, Japan, the § Department of
Cell Signaling, Yamagata University School of Medicine, Yamagata
993-23, Japan, the ¶ Department of Pharmacology, Akita University
School of Medicine, Akita 010, Japan, and the
Takai Biotimer
Project, ERATO, Japan Science and Technology Corporation, Kobe 651-22, Japan
An inwardly rectifying potassium channel predominantly expressed in glial cells, Kir4.1/KAB-2, has a sequence of Ser-Asn-Val in its carboxyl-terminal end, suggesting a possible interaction with an anchoring protein of the PSD-95 family. We examined the effects of PSD-95 on the distribution and function of Kir4.1 in a mammalian cell line. When Kir4.1 was expressed alone, the channel immunoreactivity was distributed homogeneously. In contrast, when co-expressed with PSD-95, prominent clustering of Kir4.1 in the cell membrane occurred. Kir4.1 was co-immunoprecipitated with PSD-95 in the co-expressed cells. Glutathione S-transferase-fusion protein of COOH terminus of Kir4.1 bound to PSD-95. These interactions disappeared when the Ser-Asn-Val motif was deleted. The magnitude of whole-cell Kir4.1 current was increased by 2-fold in cells co-expressing Kir4.1 and PSD-95 compared with cells expressing Kir4.1 alone. SAP97, another member of the PSD-95 family, showed similar effects on Kir4.1. Furthermore, we found that Kir4.1 as well as SAP97 distributed not diffusely but clustered in retinal glial cells. Therefore, PSD-95 family proteins may be a physiological regulator of the distribution and function of Kir4.1 in glial cells.
Distribution of ion channels in specialized areas of cell membrane is essential for various cell functions: for example, voltage-gated Na+ and K+ channels are concentrated at nodes of Ranvier for saltatory conduction in myelinated nerves (1). At the neuromuscular junction, the nicotinic acetylcholine receptor channels are concentrated at the postsynaptic membrane to initiate excitation-contraction coupling in skeletal muscle (2). Some molecular mechanisms responsible for the specific distribution of channel proteins have been characterized, e.g. agrin and rapsyn of the dystroglycan complex are involved in accumulation and clustering of nicotinic acetylcholine receptors (2). However, different molecular mechanisms may be responsible for clustering different channels. Recent studies using yeast two-hybrid method have revealed that clustering of voltage-gated K+ (3) and NMDA1 channels (4), was mediated by PSD-95 family proteins. PSD-95/SAP90 (5, 6) belongs to the membrane-associated anchoring proteins and is characterized with three PDZ domains in its amino terminus, an SH3 domain, and a carboxyl-terminal guanylate kinase homology domain. So far several members of the PSD-95 family have been cloned (7-9). The PDZ domains of PSD-95 and the Thr/Ser-X-Val motifs of the COOH termini of voltage-gated K+ and NMDA channels are supposed to interact for aggregation. Although co-localization of NMDA receptors and PSD-95 in cultured hippocampal neurons may indicate that these two proteins interact with each other in vivo (4), a physiological role for this interaction has not been fully elucidated.
Inwardly rectifying K+ channels comprise a family with more than ten members (10). These K+ channels play a pivotal role in determining resting membrane potential, in regulating action potential duration, and in transporting K+ ions. One inwardly rectifying K+ channel, Kir4.1/KAB-2, which we have cloned previously (11), has an amino acid motif of Ser-X-Val at its COOH terminus. The interaction of Kir4.1 and members of the PSD-95 family may affect the distribution and function of this channel. In this study, the interaction between Kir4.1 and PSD-95 members was studied. We show that PSD-95 and SAP97 (9) can interact with Kir4.1, resulting in clustering the channel proteins on the membrane and enhancing the Kir4.1 current in HEK293T cells. Because we also found that Kir4.1 and SAP97 clustered in retinal glial cells, the PSD-95 family of proteins may be a physiological regulator of the distribution and function of Kir4.1.
Rat Kir4.1 (11) was transfected with LipofectAMINE (Life Technologies, Inc.) into HEK 293T cells as described previously (12). Rat PSD-95 cDNA was tagged2 with 110 amino acid residues at the COOH terminus of biotin carboxylase carrier protein (BCCP) cDNA (13) and introduced into pCMV5 (kindly provided by Dr. D. E. Russell). For electrophysiological experiments, green fluorescent protein plasmid (14) was co-transfected with Kir4.1 cDNA.
Electrophysiological RecordingsWhole-cell and single-channel currents of HEK cells were measured as described (12).
Solutions and ChemicalsIn whole-cell experiments, the bathing solution contained (in mM): 40 KCl, 100 NaCl, 1.8 CaCl2, 0.53 MgCl2, 5.5 glucose, and 5.5 HEPES-KOH, pH 7.4. The pipette solution contained (in mM): 140 KCl, 5 K2 ATP, 1 MgCl2, 5 EGTA, and 5 HEPES-KOH, pH 7.3. In single-channel recordings, the pipette solution contained (in mM): 140 KCl, 1 CaCl2, 1 MgCl2 and 5 HEPES-KOH, pH 7.4. The bath was perfused with the solution composed of (in mM): 140 KCl, 5 EGTA, and 5 HEPES-KOH, pH 7.3, with 2 MgCl2.
AntibodyAnti-KAB-2A1 antibody and anti-KAB-2C2 antibody were raised in rabbits against synthetic peptides corresponding to amino acids 13-26 in the amino-terminal region and amino acids 366-379 in the COOH terminus of rat Kir4.1, respectively. Both antibodies were purified with antigenic peptide-coupled Sulfolink resin (Pierce) (15).
ImmunocytochemistryFor immunocytochemical experiments, the anti-KAB-2C2 antibody was used. For identification of Müller cells, monoclonal anti-vimentin antibody (Zymed Laboratory, San Francisco, CA) was used. Efficiencies of PSD-95 and PSD-95-BCCP on clustering Kir4.1 immunoreactivity were essentially the same. Anti-SAP97 antibody was raised in rabbit.2 Cells were examined with a confocal microscopy (MRC-1024, Bio-Rad, Hertfordshire, United Kingdom).
Biochemical AnalysisFusion proteins with glutathione
S-transferase (GST) of Kir 4.1 amino acids 165-376
(GST-Kir4.1C) or amino acids 165-379 (GST-Kir4.1C) were expressed
in Escherichia coli and purified on glutathione-Sepharose.
HEK cells transfected with PSD-95 were homogenized and solubilized in a
lysis buffer (40 mM Tris-HCl, pH 7.4, 0.15 M
NaCl, 10 mM EDTA, 0.2 mg/ml benzamidine, 30 kilounits/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, and 1%
Triton X-100) and then centrifuged at 100,000 × g for
30 min at 4 °C. Solubilized homogenates were incubated for 2 h
with 2 µg of purified fusion protein bound to 20 µl of
glutathione-Sepharose. Samples were washed four times with
phosphate-buffered saline (PBS) containing 0.4 M NaCl,
resolved by SDS-10% polyacrylamide gel electrophoresis (SDS-PAGE), and
then analyzed by Western blot using anti-PSD-95 antibody (Transduction
Laboratories, Lexington, KY). For immunoprecipitation, Kir4.1 (amino
acids 1-379) or Kir4.1
(amino acids 1-376), which lacks the
Ser-Asn-Val motif at the COOH-terminal end, were expressed with PSD-95
in HEK cells. Solubilized proteins of the cells (1 × 107 cells) were incubated with anti-KAB-2A1
antibody (4 µg/ml) overnight at 4 °C. 50 µl of protein
A-Sepharose was added to the solution. Protein A pellets were washed
four times with PBS containing 1% Triton X-100. Immunoprecipitated
proteins were resolved by SDS-PAGE and probed with anti-PSD-95
antibody.
Müller cells were isolated from retinas of Wistar rats (Nippon Doubutsu, Kyoto, Japan) (16).
To investigate the interaction between Kir4.1 and PSD-95, we
transiently expressed Kir4.1 and PSD-95-BCCP in HEK cells and examined
the distribution immunocytochemically. The anti-Kir4.1 antibody
(anti-KAB-2C2) (15) was used. The distribution of Kir4.1 immunoreactivity was homogenous in most cells transfected with Kir4.1
alone (green in Fig. 1A).
Prominent clustering of Kir4.1 occurred when PSD-95 was co-expressed
(arrows in Fig. 1B ). In the cells showing
clustered distribution of Kir4.1, the PSD-95 protein was always
co-expressed (red in Fig. 1C). Double staining showed that Kir4.1 and PSD-95 co-localized in close vicinity
(yellow in Fig. 1D). Because Kir4.1 and PSD-95
were transiently expressed, some cells expressed only Kir4.1. In Fig.
1B, diffuse staining of Kir4.1 was detected in two cells
(arrowheads), where the expression of PSD-95 was not
detected (arrowheads in Fig. 1C). Furthermore, when PSD-95 was expressed alone, it distributed diffusely (Fig. 1G). These results suggest that there is a reciprocal
relationship between Kir4.1 and PSD-95 on the cell membrane.
SAP97, another member of the PSD-95 family, when co-transfected with Kir4.1 clustered the channel proteins (Fig. 1E). In these cells, SAP97 itself was also clustered (Fig. 1F), although it distributed diffusely when transfected alone (Fig. 1H). Because both of the antibodies against Kir4.1 and SAP97 were raised in rabbit, double staining of Kir4.1 and SAP97 could not be performed. These results, however, strongly suggest that SAP97 also co-localized with Kir4.1 and clustered the channel protein on the cell membrane.
To examine the molecular mechanism responsible for interaction between
Kir4.1 and PSD-95, a mutant COOH terminus of Kir4.1 (amino acids
165-376), which lacks Ser-Asn-Val motif, and the normal COOH-terminal
region of Kir4.1 (amino acids 165-379) were expressed in E. coli with GST as fusion proteins (Fig. 2A,
GST-Kir4.1C and GST-Kir4.1C, respectively). The
fusion protein was bound to glutathione-Sepharose and incubated with
the lysate of HEK cells, which had been transfected with PSD-95. The
bound material was resolved by SDS-PAGE and subjected to Western blot
analysis using anti-PSD-95 antibody. PSD-95 in the lysate did not
associate with GST-Kir4.1C
(Fig. 2B, GST-Kir4.1C
PSD-95) but did so with GST-Kir4.1C (Fig. 2B, GST-Kir4.1C
PSD-95). Such binding was not detected when the lysate of
mock-transfected cells was used (Fig. 2B, GST-Kir4.1C control). To determine whether the interaction of Kir4.1 with PSD-95 occurs in cells, co-immunoprecipitation experiments were performed. Kir4.1
(amino acids 1-376) or the control Kir4.1 was expressed with PSD-95 in HEK cells. Expression of the channel proteins
in these cells could be confirmed by anti-KAB-2A1 antibody (Fig. 2A, Kir4.1
and Kir4.1).
Immunoprecipitants from these cells using anti-KAB-2A1 were
separated by SDS-PAGE and immunoblotted with anti-PSD-95 antibody. The
immunoprecipitant from the cells expressing Kir4.1 plus PSD-95 (Fig.
2B, Kir4.1+PSD-95 IP), but not from those expressing
Kir4.1
plus PSD-95 (Fig. 2B, Kir4.1
+PSD-95 IP), contained the PSD-95-band. These results indicate that
PSD-95 may form a macroprotein complex with Kir4.1 and that the
Ser-Asn-Val motif of Kir4.1 is indispensable for the interaction.
PSD-95 may also affect the function of Kir4.1. Using the patch clamp
technique, we compared Kir4.1 channel activity in HEK cells
co-expressed with or without PSD-95 (Fig. 3). In the
whole-cell clamp mode, the membrane currents elicited by command steps
were compared (Fig. 3A). Both cells expressed K+ currents
which were blocked by 1 mM Ba2+ (Fig.
3B). In the control cells or those transfected with only PSD-95, the Ba2+-sensitive K+ current was not
detected (n = 10 for each, data not shown). The inwardly rectifying K+ current in the cells transfected
with both Kir4.1 and PSD-95 was more than two times greater than that
observed in those transfected with Kir4.1 alone. At 120 mV, the
Ba2+-sensitive K+ current was 325.6 ± 56.3 pA/pF in those transfected with both Kir4.1 and PSD-95
(n = 6), and 134.3 ± 50.2 pA/pF in the cells expressing Kir4.1 alone (n = 11, p < 0.001, Fig. 3C). When Kir4.1 was co-transfected with SAP97,
the K+ current was more than three times larger than that
observed with Kir4.1 alone (426.9 ± 64.7 pA/pF, n = 7, p < 0.001, Fig. 3C).
The single channel properties of Kir4.1 in the cells transfected with
both Kir4.1 and PSD-95 and in those with Kir4.1 alone did not differ
from each other (Fig. 3, D and E). The unitary channel conductances of Kir4.1 alone, Kir4.1 plus PSD-95 and Kir4.1 plus SAP97 were 22.1 ± 1.4 pS (n = 3), 20.6 ± 1.0 pS (n = 3), 21.7 ± 0.8 pS
(n = 3), respectively. Open probabilities of the Kir4.1
channel with and without PSD-95 family proteins were constant ~0.9
between 40 and
100 mV (data not shown). These results suggest that
the larger current in the cells co-expressing Kir4.1 and PSD-95 or
SAP97 has resulted from an increase in the number of functional Kir4.1
channels in the membrane.
The interaction of PSD-95 family and Kir4.1 might also occur in
vivo. To examine this possibility, we studied cellular
distribution of Kir4.1 in retinal glial cells (Müller cells) (11,
17). Isolated Müller cells were stained with the anti-Kir4.1
antibody. The immunoreactivity of Kir4.1 (Fig.
4A) clustered on the membrane of cells, which
were stained with vimentin, a marker of Müller cells (Fig.
4B). Reverse transcriptase-polymerase chain reaction using
mRNA of an isolated single Müller cell and specific primers for the second PDZ domain of PSD-95 family showed that at least SAP97
was expressed in the cell. Thus, we stained Müller cells with
anti-SAP97 antibody (Figs. 4, D-F). SAP97 was expressed and clustered in Müller cells in a similar manner to Kir4.1. It is, therefore, strongly suggested that SAP97 might be responsible for
aggregation of Kir4.1 in Müller cells.
The major findings of this study are 1) Kir4.1 was shown to interact with PSD-95 through the Ser-Asn-Val motif; 2) the PSD-95 family proteins caused not only clustering of Kir4.1 on the cell membrane, but also increased the number of functional channels; 3) this interaction may occur in vivo and might be important for the physiological function of Kir4.1 channels.
The co-localization of Kir4.1 and PSD-95 in the cell membrane may result from direct interaction between these proteins, because PSD-95 bound to the GST-fusion protein of Kir4.1 C terminus and, moreover, was co-immunoprecipitated with Kir4.1. Because both interactions disappeared when the Ser-Asn-Val motif was deleted from Kir4.1, the motif is indispensable for Kir4.1-binding to PSD-95. The interaction is not, however, specific to PSD-95 itself, because SAP97 could also cause clustering of Kir4.1.
PDZ domains in various proteins can interact with each other, e.g. the PDZ domain of neuronal nitric-oxide synthase interacts with the second PDZ of PSD-95, and nitric-oxide synthase and skeletal muscle syntrophin also interact through their PDZ domains (18). Thus, PSD-95 proteins might self-aggregate through their PDZ domains. However, because PSD-95 and SAP97 diffusely distributed when transfected alone (Figs. 1, G and H), the formation of a protein complex with Kir4.1 might be essential for their aggregation.
Co-expression of Kir4.1 with PSD-95 or SAP97 prominently enhanced the whole-cell Kir4.1 currents. Because single-channel properties of Kir4.1 were not changed by the expression of PSD-95 or SAP97, an increase in the functional number of Kir4.1 channels should be mainly responsible for the increase. An increase in functional channel number can be achieved either by increasing the number of Kir4.1 proteins in the membrane and/or by facilitating formation of functional channels. hdlg, a member of the PSD-95 family, was reported to bind band 4.1 protein, which co-localizes with actin (8). Thus, Kir4.1 forming a macroprotein complex with PSD-95 might be linked to the cytoskeletal matrix through associated proteins and could be stabilized in the membrane. Another possibility is that the PSD-95 family might facilitate formation of functional Kir4.1 tetramers, because the tetrameric structure is essential for channel function (19). Through such a mechanism, even if the total number of monomer subunit proteins expressed was not increased, the number of functional tetramers could be increased. At present, we cannot discriminate these possibilities. To elucidate the molecular mechanism of enhancement of Kir4.1 channel activity by PSD-95 family proteins, further studies are needed.
We found clustered distribution of Kir4.1 on the membrane of retinal Müller glial cells. Thus, clustering of Kir4.1 occurs in vivo. Immunostaining also showed that SAP97 was clustered in Müller cells. SAP97 might, thus, be responsible for clustering of Kir4.1 in Müller cells. Because Kir 4.1 is the dominant Kir channel in glial cells, Kir4.1 is considered to play a major role in spatial buffering of K+ ions, which regulates extracellular K+ ions released by synaptic excitations in brain and retina (20). The clustering and enhanced function of Kir4.1 mediated by PSD-95 family proteins may be essential for rapid K+ exchange between glial cells and neurons.
We thank Dr. Ian Findlay (Université de Tours, Tours, France) for the critical reading of the manuscript.