Trafficking motifs present in the intracellular regions of ion channels affect their subcellular location within neurons. The mechanisms that control trafficking to dendrites of central neurons have been identified, but it is not fully understood how channels are localized to the soma. We have now identified a motif within the calcium-activated potassium channel KCa2.1 (SK1) that results in somatic localization. Transfection of hippocampal neurons with KCa2.1 subunits causes expression of functional channels in only the soma and proximal processes. By contrast, expressed KCa2.3 subunits are located throughout the processes of transfected neurons. Point mutation of KCa2.1 within this novel motif to mimic a sequence present in the C-terminus of KCa2.3 causes expression of KCa2.1 subunits throughout the processes. We also demonstrate that blocking of clathrin-mediated endocytosis causes KCa2.1 subunit expression to mimic that of the mutated subunit. The role of this novel motif is therefore not to directly target trafficking of the channel to subcellular compartments, but to regulate channel location by subjecting it to rapid clathrin-mediated endocytosis.
Neuronal excitability is regulated by activation of both voltage- and calcium (Ca2+)-dependent potassium (K+) currents. Investigation into the subcellular location of voltage-dependent K+ channels has shown them to be located within specific regions of the cell (Sheng et al., 1992; Hoffman et al., 1997; Misonou et al., 2004; Gu et al., 2006). However, less is known about the subcellular location of Ca2+-dependent K+ channels (BK and SK) the activation of which regulates action potential firing by underlying the generation of Ca2+-dependent afterhyperpolarization(s) (AHP) that hyperpolarize the membrane potential away from the threshold for action potential initiation. Large conductance Ca2+-activated (BK) potassium channels are located in the strata oriens and lucidum of the hippocampus, suggesting that they are located within axons and basal dendrites of pyramidal cells and mossy fibres of granule cells from the dentate gyrus (Misonou et al., 2006; Sausbier et al., 2006).
Three subtypes of SK channel have been cloned (KCa2.1-2.3), with KCa2.2 and 2.3 being sensitive to the bee venom toxin apamin (Köhler et al., 1996) (see Grunnet et al., 2001) and KCa2.1 originally being reported to be insensitive. All three KCa2 subunits are present in the hippocampus. For example, KCa2.3 (SK3) is present across the hippocampus, with clear localization in mossy fibre terminals (Sailer et al., 2002). By contrast, KCa2.1 (SK1) subunits are located in the soma of pyramidal neurons (Bowden et al., 2001; Sailer et al., 2002). The mechanisms controlling trafficking of some voltage-dependent potassium channels are beginning to be understood, with specific amino acid motifs being responsible for targeting the channel subunit to a particular subcellular location (Lim et al., 2000; Rivera et al., 2003). The role of intracellular transport proteins such as kinesins in delivering vesicle-bound voltage-dependent K+ (Kv) channels has been described in some cases (Chu et al., 2006; Gu et al., 2006). More recently, the mechanism of directing kinesin-powered vesicles to distinct membrane compartments has been investigated. For example, Kv4.2-containing vesicles are directed to dendrites by interaction with the plus-end-directed motor myosin Va, which permits vesicles to associate with the dendritically localized kinesin Kif17 to direct the dendritic targeting of Kv4.2 channel subunits (Rivera et al., 2003; Lewis et al., 2009). By contrast, little is known about how KCa2 channel subunits are trafficked in neurons and what would determine the different subcellular locations of KCa2.1 and 2.3 subunits.
We have used expression of epitope-tagged KCa2.1 and 2.3 channels in cultured rat hippocampal neurons to circumvent a lack of suitable primary antibodies. This approach has enabled the study of the subcellular location and trafficking of these subunits in neurons, demonstrating that the restricted somatic location of KCa2.1 is achieved by general insertion into the plasma membrane followed by rapid endocytosis.
Expression of KCa2.1 subunits in hippocampal neurons produces functional somatic channels
Immunovisualization of KCa2.1 subunits in rat hippocampal slices (Sailer et al., 2002) and acutely dissociated CA1 pyramidal neurons (Bowden et al., 2001) has shown the presence of endogenous protein, suggesting that these neurons can process the subunit correctly. However, with the exception of one example (Shah et al., 2006), there have been no reports of cultured hippocampal neurons expressing endogenous functional KCa2 channels, despite numerous examples reporting the presence of these channels in the hippocampal slice preparation. Transfection of cultured hippocampal neurons with a plasmid encoding EGFP alone produces an expression pattern throughout the soma and processes, in excess of 150 μm from the cell body (Fig. 1A). By contrast, expressed rat KCa2.1 subunits have a discrete subcellular distribution, with protein localized to the soma and proximal processes (Fig. 1B). The observed subcellular location of expressed KCa2.1 subunits within the soma and proximal processes of hippocampal neurons is remarkably similar to that observed for endogenous subunits in dissociated hippocampal CA1 neurons (Bowden et al., 2001). The restricted location for this subunit is apparent in all cells tested and immunostaining is only present within 60 μm from the cell body (see later). The same somatic location of expressed KCa2.1 subunits is seen using a channel construct with a Myc tag inserted into the S1-S2 extracellular loop (Fig. 1C). This construct displays the cell-surface expression of subunits that are clearly restricted to the cell soma. Whole-cell recording from neurons transfected with the FLAG-KCa2.1 subunit shows the presence of a large Ca2+-dependent, voltage-independent potassium conductance that is absent in cells expressing EGFP alone (Fig. 1D-F). Expressed KCa2.1-mediated current is insensitive to apamin (100 nM; data not shown) (see D'hoedt et al., 2004) and the properties of the recorded current are consistent with those expected for KCa2-mediated whole-cell currents (Strøbaek et al., 2000; Grunnet et al., 2001; Dale et al., 2002; Benton et al., 2003; D'hoedt et al., 2004). Therefore, our data demonstrate the expression of functional somatic channels in transfected neurons.
Somatic localization of KCa2.1 subunits results from rapid internalization
The localization of P2X2 receptors is suggested to result from the presence of a stabilization motif that might render the receptor insensitive to internalization (Chaumont et al., 2004). We used live-cell antibody uptake of the extracellular Myc-tagged KCa2.1 subunit to test whether wild-type KCa2.1 is retained in the soma and proximal processes by rapid internalization. This approach shows that rat KCa2.1 subunits are internalized within the soma of transfected neurons (Fig. 2A). This labelling is specific for KCa2.1 subunits that are internalized during the 20 minutes incubation at 37°C, as no labelling is observed when internalization is prevented by maintaining cells at 4°C (Fig. 2B) or when cells are incubated at 37°C for 20 minutes but not permeabilized (Fig. 2C). These data, together with the lack of labelling of intracellular KCa2.1 subunits in the processes of permeabilized cells (FLAG-KCa2.1; Fig. 1B), indicate that channel subunits are mainly internalized in the soma. This suggests that targeted internalization within the soma prevents KCa2.1 subunits from being located throughout the processes.
Internalization of KCa2.1 subunits is mediated by clathrin-dependent endocytosis
Cells use various mechanisms to internalize plasma membrane proteins, including clathrin-mediated and caveolae-dependent endocytosis and phagocytosis. These internalization mechanisms utilize the large GTPase dynamin for scission of vesicles (Praefcke and McMahon, 2004). Dynamin is present in hippocampal neurons, being located in both the soma and processes (Fig. 3A). We have used dynamin 1(K44A) as a dominant negative mutant dynamin (dyn-DNM) (Damke et al., 1994) to determine whether dynamin-dependent internalization affects the subcellular location of expressed KCa2.1 subunits. Expression of dyn-DNM in hippocampal neurons blocks the internalization of fluorescently labelled transferrin (data not shown), indicating that this dominant negative construct inhibits dynamin-dependent internalization in hippocampal neurons. Overexpression of wild-type human dynamin 1 in hippocampal neurons (Fig. 3Bi) has no effect on the subcellular location of expressed KCa2.1 subunits, with channel subunits still being located in the soma and proximal processes (Fig. 3Bii). By contrast, coexpression of dyn-DNM and KCa2.1 subunits changed the subcellular location of channel subunits to throughout the soma and distal processes (Fig. 3Cii). Block of dynamin-dependent internalization was confirmed to result in the appearance of cell-surface KCa2.1 subunits throughout the soma and distal processes, by using the extracellular Myc-tagged KCa2.1 subunit. Coexpression of dyn-DNM and Myc-tagged KCa2.1 produced cell-surface channel subunits throughout the distal processes, suggesting that blocking of dynamin-dependent endocytosis permits channel subunits to diffuse laterally from the soma to distal processes (Fig. 3Ciii). These data indicate that the somatic location of expressed KCa2.1 subunits is regulated by dynamin-dependent internalization.
Endocytosis of membrane proteins is achieved by two main mechanisms, one clathrin mediated and the other cavaeolae dependent (Praefcke and McMahon, 2004). The internalization of clathrin-coated vesicles is a process that is commonly used to facilitate the endocytosis of specific cargoes such as receptors and transporters (Marsh and McMahon, 1999). We tested whether the dynamin-dependent internalization of KCa2.1 subunits occurs by clathrin-mediated endocytosis. AP180 is a major component of clathrin coats, binding to clathrin and stimulating clathrin cage assembly to cause endocytosis (Ford et al., 2001). Expression of a dominant negative C-terminal construct of AP180 (AP180-C) in COS-7 cells inhibited endocytosis of transferrin by blocking clathrin-coated pit formation (Ford et al., 2001). Consistent with this result, expression of AP180-C in hippocampal neurons blocks endocytosis of fluorescently labelled transferrin, indicating clathrin-mediated endocytosis is inhibited (data not shown). Coexpression of pFLAG-KCa2.1 and AP180-C in hippocampal neurons causes a dramatic change in the subcellular location of the Ca2+-activated potassium channel. Instead of the channel exhibiting a somatic and proximal process location (Fig. 3D), co-expression of KCa2.1 and AP180-C causes the channel subunit to be distributed in the soma and throughout the processes (Fig. 3E). This distribution pattern is very similar to the subcellular location of KCa2.1 co-expressed with dyn-DNM (Fig. 3Cii). These data demonstrate that expressed functional KCa2.1 channels are restricted to the soma and proximal processes of cultured hippocampal neurons by rapid internalization resulting from clathrin-mediated endocytosis.
Subcellular location of expressed KCa2.1 subunits is determined by a novel motif
Various endocytotic-sorting signals have been identified, including a tyrosine-based motif (Owen and Evans, 1998). For example, a non-canonical tyrosine-based endocytotic motif with the sequence YxxGΦ (where Φ indicates any amino acid with a hydrophobic side chain and x any amino acid) is present within the C-terminus of P2X4 receptors (Royle et al., 2002), enabling the receptor to undergo clathrin-mediated endocytosis (Royle et al., 2005). In addition, it has been reported that the subcellular location of P2X2 receptors expressed in hippocampal neuronal cultures is changed on mutation of either a tyrosine (Y) or lysine (K) within a YxxxK motif present in the C-terminus of the receptor (Chaumont et al., 2004). A YxxxK sequence is also present in the C-terminus of KCa2.3, but is seen as YxxxR406 in KCa2.1 (Fig. 4A). To investigate if this sequence could mediate the restricted location of expressed KCa2.1, we mutated arginine (R) 406 in KCa2.1 to a lysine (K) residue. KCa2.1(R406 to K) [hereafter referred to as KCa2.1(R406K)] produces a very different subcellular expression pattern from KCa2.1 when expressed in hippocampal neurons. Expressed KCa2.1(R406K) subunits are distributed throughout the soma and processes (Fig. 4B), in contrast to the soma and proximal process location of expressed wild-type KCa2.1 (Fig. 1B,C, Fig. 2A, Fig. 3D). The subcellular location of expressed KCa2.1(R406K) subunits is similar to that of FLAG-tagged KCa2.3 subunits expressed in hippocampal neurons (Fig. 4C). Punctate labelling is observed with both KCa2.1(R406K) and KCa2.3 subunits (boxed regions in Fig. 4B,C). Measurement of averaged pixel intensity within the soma and processes shows that labelling of both KCa2.1(R406K) and KCa2.3 subunits extends beyond 150 μm within five processes of a single neuron (Fig. 4E) and a randomly selected process of six to nine neurons (Fig. 4F). Like KCa2.1, KCa2.1(R406K) produces a Ca2+-dependent, voltage-independent potassium current (Fig. 4Di,ii) that is insensitive to the bee venom toxin apamin (data not shown). Comparison of whole-cell current amplitude at –80 mV normalized to cell size for both the wild-type (–254.9±58.4 pA/pF, n=9) and mutant (–352.4±77.7 pA/pF, n=8) rKCa2.1-mediated current showed that a larger current was observed with rKCa2.1(R406K) (ANOVA, P<0.1). The observed tendency of a larger current suggests that the YxxxK motif stabilizes functional channels in the plasma membrane, whereas channels with the YxxxR motif are more sensitive to internalization.
The intracellular location of expressed KCa2.1(R406K) subunits is very similar to that seen after blocking dynamin-dependent endocytosis (Fig. 3C,E). Expression of either wild-type dynamin (Fig. 5A) or dyn-DNM (Fig. 5B) had no effect on the location of coexpressed Myc-tagged KCa2.1(R406K), with surface labelling using the extracellular Myc tag revealing channel subunits distributed throughout the soma and processes. These data confirm that expressed KCa2.1(R406K) subunits are resistant to dynamin-dependent internalization and suggest that clathrin-mediated endocytosis of wild-type KCa2.1 is dependent on the intact YxxxR motif.
The YxxxR motif is a novel non-canonical endocytotic motif
Our data indicate that the C-terminal arginine residue in the YxxxR sequence allows the subunit to be a substrate for clathrin-mediated endocytosis. However, to clarify whether this sequence can be classified as a novel non-canonical endocytotic motif, we mutated the initial tyrosine to a sterically similar phenylalanine (F) residue within the YxxxR sequence of KCa2.1 [KCa2.1(Y402F)]. This mutation produces subunit expression throughout the processes, similar to that seen with expression of KCa2.1(R406K) subunits (Fig. 6A). This indicates a primary requirement for the initial tyrosine residue within this sequence to enable clathrin-meditated endocytosis.
Replacement of the C-terminal R with K prevents the subunit from being a substrate for endocytosis (Fig. 4B). However, it is not clear whether the loss of R or the gain of K enables the KCa2.1 subunit to become resistant to clathrin-mediated endocytosis. Mutating the basic R to an alanine (A) produces a subunit KCa2.1(R406 to A) [KCa2.1(R406A)] that is expressed throughout the processes of hippocampal neurons (Fig. 6B) similar to KCa2.1(R406K). These data demonstrate that it is the loss of the C-terminal R that produces a subunit that cannot be rapidly internalized by endocytosis. Taken together, we propose that the YxxxR motif within the C-terminus of KCa2.1 is a functional non-canonical endocytotic motif.
We propose that clathrin-mediated endocytosis from the soma restricts the location of KCa2.1 subunits, but it is important to understand whether the subcellular location of KCa2.1 subunits is also regulated by additional targeting sequences. This has to be achieved using the KCa2.1(R406K) mutant subunit, as wild-type KCa2.1 subunits are restricted to a somatic location. Co-labelling hippocampal neurons expressing KCa2.1(R406K) subunits with either MAP2 (to label dendrites; Fig. 6C) or Tau (to label axons; Fig. 6D) shows that these subunits are located in both subcellular compartments. The lack of localization to either dendrites or axons indicates that KCa2.1(R406K) (and by association KCa2.1) subunits do not possess an additional targeting sequence to direct the subunit to either axons or dendrites, with the wild-type KCa2.1 subunit being restricted to the cell soma by rapid internalization.
The YxxxR motif determines rapid endocytosis in another channel subunit
The dramatic difference in KCa2.1 subunit localization by the mutation R406 to K (Fig. 4B) suggests that the reverse mutation in KCa2.3 might produce a subunit that displays a somatic and proximal process location, by conferring sensitivity to rapid endocytosis through the YxxxR motif and thereby preventing migration into distal processes. Transfection of hippocampal neurons with KCa2.3(K587R) (the lysine that corresponds to R406 in KCa2.1) still produces expression of subunits throughout the soma and processes (Fig. 6E), but this expression is clearly different from that observed with wild-type KCa2.3 (Fig. 4C). Expression of the KCa2.3(K587R) subunit produces smooth labelling of processes (Fig. 6E), whereas expressed KCa2.3 subunits are distributed throughout the processes with punctuate labelling (Fig. 4C). Block of clathrin-mediated endocytosis by coexpression of KCa2.3(K587R) with AP180-C restores the punctate labelling (Fig. 6F). These data indicate that introduction of the YxxxR motif into KCa2.3 can induce its rapid internalization by endocytosis. However, it also suggests that the KCa2.3 subunit possesses additional trafficking motif(s) that are lacking in KCa2.1 (Fig. 6C,D), enabling KCa2.3 subunits to be targeted to the processes.
Two mechanisms have been proposed for the polarized distribution of membrane proteins within neurons: selective retention or selective delivery. For example, the sodium channel Nav1.2 is selectively retained within axons of hippocampal neurons. This channel possesses a dileucine motif that directs it to the axonal domain by selective endocytosis from the somatodendritic domain (Garrido et al., 2001). By contrast, selective delivery occurs by the polarized delivery of membrane protein-containing vesicles and insertion by targeted exocytosis (Burack et al., 2000; Horton and Ehlers, 2003; Sampo et al., 2003). For example, there is a high density of Kv4.2 A-type potassium channels in the dendrites of hippocampal CA1 pyramidal neurons, which serve to limit dendritic action potential firing (Sheng et al., 1992; Hoffman et al., 1997). The selective delivery of this channel is mediated by a 16 amino acid, dileucine-containing motif that is present in the C-terminus of the channel (Rivera et al., 2003).
In contrast to the polarized sorting of membrane molecules, the mechanism(s) determining somatic location of ion channels is poorly understood. For example, endogenous and over-expressed Kv2.1 subunits are restricted to the soma and proximal dendrites of hippocampal neurons where they contribute to the repolarization of the somatic action potential (Misonou et al., 2004). Somatic localization occurs by subunits residing within large clusters that limits lateral diffusion (Misonou et al., 2004; O'Connell et al., 2006). By contrast, surface-expressed KCa2.1 subunits do not reside in clusters (Fig. 1C) and therefore would be free to diffuse within the membrane. This is clearly shown when clathrin-mediated endocytosis of KCa2.1 subunits is blocked by coexpression of dyn-DNM (Fig. 3C) or AP180-C (Fig. 3E). Subunit migration by diffusion has been observed for Kv1.4 and Kv1.3 subunits expressed in HEK293 cells (Burke et al., 1999; O'Connell and Tamkun, 2005), and β2 adrenergic receptors expressed in hippocampal neurons (Yudowski et al., 2006). In addition, Kv2.1 subunits undergo targeted insertion into somatic clusters and diffuse into distal processes after actin disruption, implying that these subunits can also diffuse laterally in the membrane (O'Connell et al., 2006). The distances implied for lateral diffusion of KCa2.1 subunits are mirrored by those observed for diffusion of the NMDA (N-methyl D-aspartate) and AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) glutamate receptors within hippocampal neurons, where both receptor types are inserted into the somatic plasma membrane and diffuse laterally to synapses within dendritic spines (Bresler et al., 2004; Adesnik et al., 2005). The rate of lateral diffusion of K+ channels within the membrane has been estimated to be approximately 5×10–11 cm2/second (Weis et al., 1986; Burke et al., 1999). This diffusion coefficient means that a channel will require 3-6 hours to diffuse across an area the size of the soma of a hippocampal neuron. The half-lifetime of different ion channels in the membrane has been estimated to be approximately 5-20 hours (Passafero et al., 1992; Takimoto et al., 1993). It is clear that a channel could laterally diffuse a significant distance in that time, exiting the soma and migrating up dendritic processes. If a somatic location of an ion channel is required to regulate excitability within the cell body, an ion channel could be corralled to prevent migration (Kv2.1) or selectively internalized after a time to prevent channel migration up processes. It is proposed that it is this latter mechanism that is used to maintain somatically located KCa2.1 channel subunits.
Expressed KCa2.1 subunits are observed in the soma and proximal processes, with no labelling being apparent in distal processes. The lack of distal labelling indicates that subunits are not present in the cytosol of processes, because intracellular subunits would be identified in permeabilized cells using the intracellular FLAG epitope. This lack of distal labelling (Fig. 1B,C, Fig. 2A, Fig. 3Bii,D and Fig. 4E,F) indicates that the KCa2.1 subunit does not possess a motif to traffic subunits to the dendrites or axons, which is supported by the distribution of KCa2.1(R406K) subunits in both MAP2- and Tau-positive processes (Fig. 6C,D). By contrast, expression of KCa2.3(K587R) subunits in processes of hippocampal neurons (Fig. 6E) suggests that this subunit possesses an unidentified motif that targets the subunits to neuronal processes. KCa2.3 displays a dileucine motif formed from two overlapping leucine zipper sequences in the N-terminus (Jones et al., 2004) and three dileucine motifs in its C-terminus. By contrast, KCa2.1 does not have a dileucine motif in either terminal. For example, KCa2.3 possesses a dileucine motif in the C-terminus that is seen as a leucine-valine in KCa2.1 (Fig. 4A). We generated the double mutant KCa2.3(K587 to R, L589 to V), which replicated the sequence found in KCa2.1, to determine whether a subunit that lacked this dileucine motif but could undergo rapid endocytosis, would be expressed only in the soma and proximal processes. This was not observed, with KCa2.3(K587R,L589V) expression being identical to KCa2.3(K587R) (data not shown). This data is consistent with the finding that trafficking motifs in both the N- and C-termini must be instrumental in determining the dendritic location of KCa2.3, because deletion of both termini was required to change the location of subunits in processes (Decimo et al., 2006).
Rat KCa2.1 subunits form functional channels when expressed in hippocampal neurons, something not observed when these subunits are expressed in various cell lines (Bowden et al., 2001; Benton et al., 2003; D'hoedt et al., 2004). The mutant KCa2.1(R406K) subunit does not provide functional channels when expressed in HEK293 cells (data not shown), which indicates that this motif is not a prerequisite for functional expression. Instead, this motif permits the channel to be a substrate for clathrin-mediated endocytosis (Fig. 3). The YxxxR motif is clearly a novel non-canonical endocytotic motif, as mutation of either the leading tyrosine or the tail arginine caused the loss of somatic localization. The arginine residue is critical in determining whether the subunit is a substrate for clathrin-mediated endocytosis, with the extremely conservative substitution of a lysine residue to mimic the sequence found in KCa2.3, making the subunit resistant to endocytosis (Fig. 4B). This substitution is mimicked by mutation of the C-terminal arginine to alanine, which confirms that YxxxR is a non-canonical endocytotic motif and it is the loss of the tail R rather than the gain of a K residue that makes the subunit resistant to clathrin-mediated endocytosis. The motif is clearly different from tyrosine-based sorting sequences that are of the form YxxΦ. These motifs direct the protein to be endocytosed by interaction of both the tyrosine and the hydrophobic residues with the μ2 subunit of the AP2 clathrin adaptor protein complex (Bonifacino and Traub, 2003). A similar non-canonical tyrosine-based sequence has been found in the C-terminus of P2X4 receptors, which has the form of YxxGΦ (Royle et al., 2005). This motif also dictates clathrin-mediated endocytosis by interaction of the tyrosine and hydrophobic residues with the μ2 subunit of AP2 (Royle et al., 2005). By contrast, the novel motif identified in KCa2.1 has a basic arginine residue in place of the hydrophobic leucine that is present in P2X4 receptors, suggesting a different mechanism.
Clear correlation is apparent between the subcellular location of channels and the presence of the YxxxR motif in the C-terminus of the subunit. The voltage-dependent potassium channel subunits KCNQ3 and -5 are expressed in the soma of hippocampal neurons (Yus-nájera et al., 2003; Geiger et al., 2006) and both possess YxxxR in their C-termini. The L-type Ca2+ channel subunit CaV1.3, also has the C-terminal YxxxR motif and is expressed only in the soma and proximal dendrites of hippocampal neurons (Bowden et al., 2001). Although a detailed mutational analysis for these channels has not been performed, our results with the KCa2.1 channel suggest that targeted endocytosis might also restrict these channels to their specific subcellular location. By contrast, Kv1.4 channels are located in dendrites of hippocampal neurons (Rivera et al., 2003) and have a C-terminal YxxxK motif. Finally, both KCNQ2 and KCa2.2 subunits have the YxxxK motif and are distributed in both dendrites and dendritic spines (Martire et al., 2004; Ngo-Anh et al., 2005; Weber et al., 2006). These observations, together with the subcellular distribution of expressed KCa2.1 and -2.3 subunits, suggest that general insertion and rapid endocytosis might be a common mechanism to provide a restricted somatic location.
Materials and Methods
The N-terminal pFLAG-tagged rat KCa2.1 has been described previously (Fletcher et al., 2003). Len Kaczmarek (Yale University, School of Medicine, New Haven, CT) kindly provided human KCa2.3, which was subcloned into pFLAG-CMV2 to provide an N-terminal pFLAG-tagged KCa2.3. A Myc-tagged AP180-C was provided by Harvey McMahon (MRC Laboratory of Molecular Biology, Cambridge, UK). Single and double point mutations within the C-termini of pFLAG-KCa2.1 and -KCa2.3 were introduced by site-directed mutagenesis and confirmed by dye-termination sequencing. To provide a channel subunit with an extracellular epitope [either wild-type pFLAG-KCa2.1 or -KCa2.1(R406K)], the Myc sequence was introduced into the first extracellular loop of pFLAG-KCa2.1 by creation of a silent mutation at amino acid Y128 (TAC to TAT), which created a BstZ17I restriction site. The mutagenesis was performed using the QuickChange mutagenesis method (Stratagene) and two complementary oligonucleotides harbouring the necessary mutation (5′-GTCCTGGGGTGTGTATACCAAGGAGTCTCTG-3′ and 5′-CAGAGACTCCTTGGTATACACACCCCAGGAC-3′). The resulting plasmid pFLAG-rKCa2.1-BstZ17I was then linearized with BstZ17I and the 5′ ends dephosphorylated with calf intestine alkaline phosphatase. Two complementary oligonucleotides were then synthesized that contained the Myc sequence (5′-CGAACAAAAGCTTATTTCTGAAGAAGACCTGGG-3′ and 5′-CCCAGGTCTTCTTCAGAAATAAGCTTTTGTTCG-3′). The oligonucleotides were phosphorylated with T4 polynucleotide kinase and hybridized by heating to 100°C and slowly cooled to room temperature. The phosphorylated, double-stranded oligonucleotide was then ligated into the linearized pFLAG-rKCa2.1-BstZ17I plasmid. This procedure generated the plasmid pFLAG-Myc-rKCa2.1, with the Myc sequence EQKLISEEDL (plus an additional glycine at the C-terminal side to maintain the correct reading frame) inserted between amino acids Y128 and T129. The integrity of the insertion was verified by sequencing the entire open reading frame of the mutated channel subunit.
Cell culture and transfections
Hippocampal cultures were prepared as described previously (Corrêa et al., 2004). The cells were plated onto 22-mm glass coverslips coated with poly-L-lysine and transfected 15-18 days after plating using Lipofectamine 2000 (Invitrogen). Neurons were used 24 hours after transfection. Only approximately 3% of the cells were transfected with this method. This did not hamper electrophysiological recording, because cells were co-transfected with EGFP as a marker.
Whole-cell voltage-clamp recordings were made from cultured hippocampal neurons at room temperature (20-24°C) 24 hours post-transfection using an Axopatch 200A amplifier (Axon Instruments). Hippocampal neurons were bathed in a high K+ solution described previously (Bowden et al., 2001). Transfected cells were identified by visualization of co-transfected EGFP. Fire-polished electrodes (3-5 MΩ) pulled from borosilicate glass contained 120 mM KMeSO4, 20 mM KCl, 1.31 mM MgCl2, 10 mM HEPES, 10 mM EGTA, 3 mM Na2ATP, CaCl2 (calculated free [Ca2+] 60 nM or 1 μM) (pH 7.4). Hippocampal neurons were held at 0 mV. Expressed KCa2.1-mediated currents were only accepted when voltage-dependent K+ currents (delayed rectifier and A-current) were evoked by depolarizing voltage steps (10 mV increments) from a holding potential –90 mV. Capacitance and series resistance compensation (>90%) was used throughout, with currents filtered at 1 kHz (8-pole low pass Bessel filter; Frequency Devices, CT) and sampled at 10 kHz using Pulse (HEKA, Lambrech, Germany). KCa2.1-mediated currents were revealed by 10-mV-increment voltage steps of 1-second duration and measured at the end of the 1-second pulse.
All antibodies were acquired commercially: FLAG M2 [monoclonal antibody (mAb); Sigma], FLAG [monoclonal antibody (pAb); Sigma], dynamin (mAb Hudy 1; Upstate), MAP2 (Clone HM-2 mAb; Sigma), Tau-1 (MAB 3420 mAb; Chemicon), Myc (9E10 mAb; Santa Cruz Biotechnology).
Immunocytochemistry, microscopy and data analysis
To visualize expressed pFLAG-tagged rKCa2.1 and hKCa2.3 proteins and endogenous levels of dynamin, TAU and MAP2, transfected hippocampal neurons were grown on poly-L-lysine-coated coverslips and then washed three times with warm HEPES-buffered saline (HBS) of composition: 119 mM NaCl, 5 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 25 mM HEPES, 30 mM glucose (pH 7.2), fixed in warm 4% paraformaldehyde (PFA, pH 7.4) in 0.1 M phosphate-buffered saline (PBS) for 10 minutes at room temperature and permeabilized for 5 minutes in 0.1% Triton X-100. The cells were then blocked with 10% BSA (Sigma) and incubated for 80 minutes at room temperature with primary antibodies diluted in 10% BSA. Alexa Fluor secondary antibodies (1:200; Molecular Probes) to the appropriate species were diluted in 10% BSA and incubated at 4°C for 1 hour. Non-specific labelling was determined by incubating neurons with secondary antibody alone. Coverslips were mounted on slides with Mowiol. To visualize only surface-expressed Myc-KCa2.1 channels, hippocampal neurons were transfected with pFLAG-Myc-KCa2.1. Hippocampal neurons overexpressing pFLAG-Myc-KCa2.1 subunits were washed and briefly fixed using 4% PFA and incubated with rabbit anti-Myc antibody (10 μg/ml; Santa Cruz Biotechnology) at room temperature for 1 hour. Cells were subsequently washed with HBS and incubated with anti-rabbit Alexa Fluor 568 secondary antibody (Molecular Probes) to visualize the extracellular Myc tag of FLAG-Myc-KCa2.1 subunits. For each experimental condition, 10-15 cells were imaged and analyzed. Each experimental condition was repeated three times using hippocampal cultures from different preparations.
Immunofluorescent staining was observed using a 63× oil immersion lens on a Zeiss LSM510 confocal microscope (Oberkochen, Germany). Fluorophores were excited with 488 or 568 nm wavelengths and emission from a single confocal plane was detected through 505-530 nm band-pass and 560 nm long-pass filters. Images comprising a single confocal plane or Z-stack were processed using Adobe Photoshop 6.0 (Adobe) and CorelDraw 12.0. The quantification of the pixel intensity was performed with ImageJ software (NIH, http:/rsb.info.nih.gov/ij/index.html). The images of neurons expressing either pFLAG-KCa2.1, pFLAG-KCa2.1(R406K) or pFLAG-KCa2.3 were first calibrated by a two-stage process. First, the image was scaled according to the physical parameters by which the image was acquired. Second, a region of diffuse fluorescence near the cell body was selected as background and the threshold was set at twice this background level. The cell was analyzed by drawing a line from approximately the centre of the cell body to the end of a randomly selected process, and the profile of the pixel intensity was measured. Neurons expressing pFLAG-KCa2.1, where staining was observed only in the cell body, used the corresponding transmission image as a reference. The listed measurements obtained from the ImageJ software were then transferred to an Excel work sheet and the pixel intensity was averaged into 10 μm bins and then normalized to the pixel intensity of the cell body. The normalized data for all the cells pFLAG-KCa2.1 (n=9), pFLAG-KCa2.1(R406K) (n=6) and pFLAG-KCa2.3 (n=5) were then averaged and analyzed using SPSS 15.0. Statistical significance was determined by comparison of the means of the 10 μm bins for pFLAG-KCa2.1, pFLAG-KCa2.1(R406K) and pFLAG-KCa2.3 using one-way ANOVA. All of the values reported are means ± s.e.m.
For channel internalization assays, live hippocampal neurons overexpressing pFLAG-Myc-KCa2.1 subunits were incubated with anti-Myc (pAb, 10 μg/ml; Santa Cruz Biotechnology) for 20 minutes at 4°C. The unbound excess of Myc antibody was quickly washed off with cold HBS and cells were incubated for another 20 minutes at 37°C to allow the Myc-labelled channels to be internalized, or for 20 minutes at 4°C to prevent channel internalization. Neurons were briefly fixed with 4% PFA after three washes with HBS, and the remaining surface-expressed pFLAG-Myc-KCa2.1 subunits were blocked by a 90-minute incubation with unlabelled anti-rabbit secondary antibody (1:50 dilution; Sigma). Neurons were then permeabilized with 0.1% Triton X-100 for 10 minutes, washed and incubated with Alexa Fluor 568 anti-rabbit antibody (Molecular Probes) for 1 hour to visualize internalized pFLAG-Myc-KCa2.1 subunits. Parallel samples of cells expressing pFLAG-Myc-KCa2.1 subunits, incubated at 37°C, were also blocked with unlabelled anti-rabbit secondary antibody (1:50 dilution, Sigma) and directly labelled with Alexa Fluor 568 without permeabilization.
We wish to thank David Prole and David Stephens for helpful discussions. We are indebted to Atsushi Nishimune for help with generation of the point mutants. Finally, we thank Jeremy Henley, George Banting, John Crabtree and Dawn Shepherd for critical reading of the manuscript. This work was supported by the Medical Research Council (UK). Deposited in PMC for release after 6 months.
↵* Present address: Department of Biological Sciences, University of Warwick, Coventry, CV4 7AL, UK
- Accepted September 9, 2009.
- © The Company of Biologists Limited 2009