The N terminus and transmembrane segment S1 of Kv1.5 can coassemble with the rest of the channel independently of the S1–S2 linkage

The voltage-gated potassium channel Kv1.5 belongs to the Shaker superfamily. Kv1.5 is composed of four subunits, each comprising 613 amino acids, which make up the N terminus, six transmembrane segments (S1–S6), and the C terminus. We recently demonstrated that, in HEK cells, extracellularly applied proteinase K (PK) cleaves Kv1.5 channels at a single site in the S1–S2 linker. This cleavage separates Kv1.5 into an N-fragment (N terminus to S1) and a C-fragment (S2 to C terminus). Interestingly, the cleavage does not impair channel function. Here, we investigated the role of the N terminus and S1 in Kv1.5 expression and function by creating plasmids encoding various fragments, including those that mimic PK-cleaved products. Our results disclosed that although expression of the pore-containing fragment (Frag(304–613)) alone could not produce current, coexpression with Frag(1–303) generated a functional channel. Immunofluorescence and biotinylation analyses uncovered that Frag(1–303) was required for Frag(304–613) to traffic to the plasma membrane. Biochemical analysis revealed that the two fragments interacted throughout channel trafficking and maturation. In Frag(1–303)+(304–613)-coassembled channels, which lack a covalent linkage between S1 and S2, amino acid residues 1–209 were important for association with Frag(304–613), and residues 210–303 were necessary for mediating trafficking of coassembled channels to the plasma membrane. We conclude that the N terminus and S1 of Kv1.5 can attract and coassemble with the rest of the channel (i.e. Frag(304–613)) to form a functional channel independently of the S1–S2 linkage.

superfamily of Kv channels. They are expressed in various tissues (1). In the heart, Kv1.5 is mainly expressed in the atria and conducts the ultra-rapidly activating delayed rectifier K ϩ current (I Kur ), which is important for atrial repolarization (2,3). In pulmonary artery smooth muscle cells, Kv1.5 plays an important role in regulating membrane potential and vascular tone (4,5). Elucidating the biochemical and biophysical properties of Kv channels is important and represents an active area of research. In particular, it is fundamental to understand how each part of ion channel structure contributes to assembly, trafficking, function, and regulation.
A single Kv channel is composed of four ␣-subunits, often identical, that assemble to form a tetramer (6,7). Each ␣-subunit contains six transmembrane segments, S1-S6, with N and C termini located intracellularly. Each structural component contributes distinctively to channel function. For example, the S4 segment acts as the voltage sensor (8). The linker of S5 and S6 contributes to forming the K ϩ -selective permeation pore (9). However, the roles of other structural components such as the N terminus and S1 segment in expression and function are less well understood.
We recently demonstrated that Kv1.5 channels can be cleaved at a single site in the S1-S2 linker by extracellularly applied proteinase K (PK) into an N-fragment that contains the N terminus and S1 and a C-fragment that contains S2 to the C terminus. However, PK-cleaved Kv1.5 still generates robust current (10). In the current study, we created plasmids encoding Frag(1-303) (fragment containing amino acid residues 1-303, which make up the N terminus and S1 segment) and the pore-containing Frag(304 -613) (fragment containing amino acid residues 304 -613, which make up S2-S6 and the C terminus) as well as further defined smaller fragments to investigate the role of the N terminus and S1 in Kv1.5 trafficking and function. Our results showed for the first time that even when the linkage within the S1-S2 linker was absent, Frag(1-303) was able to attract and assemble with Frag(304 -613) to form a functional channel. Specifically, residues 1-209 of Frag  interacted with the independent Frag(304 -613). Residues 210 -303 (the remaining N terminus and S1 of Frag(1-303)) mediated channel trafficking to the plasma membrane. Further analysis indicated that residues 210 -240 of the N terminus were required for the expression of S1-S2 linker-intact Kv1.5 channels in the plasma membrane. These results revealed novel roles of the N terminus and S1 in Kv1.5 channel trafficking and assembly.

The Kv1.5 channel is cleaved in the S1-S2 linker by PK
Using Western blot analysis and whole-cell patch clamp method, we examined the effects of PK treatment on the expression and function of Kv1.5 channels. Kv1.5 channel proteins exist in two forms: the mature form, which is localized in the plasma membrane, and the immature form, which is localized intracellularly. It is the membrane-localized, mature protein that conducts Kv1.5 current (I Kv1.5 ) (10,11). Kv1.5 is synthesized as a 65-kDa protein. It undergoes core glycosylation to become a 68-kDa immature channel and further complex glycosylation to become a 75-kDa mature channel localized in the plasma membrane (Fig. 1A, lane 1). N-Glycosylation of Kv1.5 occurs at Asn-299 (GenBank accession number NM_002234.3) within the S1-S2 linker (10). Glycosylation is not a prerequisite for the trafficking of Kv1.5 channels to the plasma membrane (10). Thus, after inhibition of glycosylation with tunicamycin (Tuni; 10 g/ml, 48 h), both mature (plasma membrane) and immature (intracellular) channels display as a 65-kDa band (Fig. 1A, lane 2). PK, which cannot permeate the cell membrane, cleaves cell-surface protein when applied extracellularly (12). For normally glycosylated Kv1.5 channels, treatment of intact Kv1.5-expressing human embryonic kidney (HEK) (Kv1.5-HEK) cells with PK (200 g/ml, 20 min, 37°C) selectively abolished the 75-kDa Kv1.5 band without affecting the 68-kDa Kv1.5 band (Fig. 1A, lane 3). The disappearance of the 75-kDa band was accompanied by the appearance of a 42-kDa N-fragment (detected using an N terminus-targeting Kv1.5 antibody; Fig. 1A, left, lane 3) and a 33-kDa C-fragment (detected using a C terminus-targeting Kv1.5 antibody; Fig. 1A, right, lane 3). Resistance of the 68-kDa Kv1.5 protein to PK cleavage confirmed its intracellular location (Fig. 1A, lane 3). For unglycosylated channels (65-kDa; Tuni treatment), PK treatment cleaved the mature, plasma membrane-localized 65-kDa protein but not the immature, intracellularly localized 65-kDa protein (Fig. 1A, lane 4). As a result, PK treatment only decreased the intensity of the 65-kDa protein but did not abolish it. Cleavage of the unglycosylated plasma membrane Kv1.5 protein resulted in the appearance of a 32-kDa N-fragment (Fig.  1A, left, lane 4) and a 33-kDa C-fragment (Fig. 1A, right, lane 4). Thus, removal of glycosylation reduced the N-fragment from 42 to 32 kDa but did not affect the size of the C-fragment (Fig.  1A). These results indicate that the glycosylation site (Asn-299) is located within the N-fragment. Together with ExPASy ProtParam analysis based on the fragment sizes, these data indicated that PK likely cleaves Kv1.5 at a site between Asn-299 and the carboxyl end of the S1-S2 linker (Fig. 1B).
To determine whether the PK-generated N-and C-fragments associate with each other, we performed a coimmuno- Figure 1. PK cleaves Kv1.5 at the S1-S2 linker and increases I Kv1.5 . A, Western blots depicting Kv1.5 expression following PK treatment (200 g/ml, 20 min, 37°C) of Kv1.5-HEK cells cultured with or without Tuni (10 g/ml) for 48 h (n ϭ 6). Proteins were detected using anti-N-terminal (N-Ab) or anti-C-terminal (C-Ab) Kv1.5 antibody. Actin was used as a loading control. Molecular mass marker (Marker) is shown in the middle. B, schematic illustration of Kv1.5 PK cleavage. C, co-IP assay showing that the N-(N-FR) and C-fragments (C-FR) do not associate after PK cleavage. Whole-cell proteins were extracted from PK-treated WT Kv1.5-HEK cells, and an anti-N-terminal Kv1.5 antibody was added to precipitate the N-fragment and associated proteins. Although the N-fragment was detected in the precipitate, the C-fragment was not detected. Western blotting (WB) of uncleaved and PK-cleaved Kv1.5 is shown to indicate the fragments. GAPDH was used as the control. IP, immunoprecipitation; IB, immunoblotting. The same results were obtained from five co-IP experiments. D, I Kv1.5 in control (CTL) and PK-treated cells. The voltage protocol is shown above the current traces, and the summarized current-voltage (I-V) and g-V relationships are shown beneath the current traces (n ϭ 36 in control; n ϭ 35 in PK cleavage; **, p Ͻ 0.01 at 0 mV and above). Error bars represent S.E.
The N terminus and S1 in Kv1.5 expression and function precipitation (co-IP) assay. Our results indicated that the PKgenerated N-fragment (42 kDa) and C-fragment (33 kDa) of Kv1.5 did not associate (Fig. 1C) (10). Interestingly, PK cleavage of Kv1.5 did not impair channel function but instead led to an increase in I Kv1.5 without affecting the activation-voltage (g-V) relationships (Fig. 1D). The g-V relationships were fit to the Boltzmann function. The half-maximal activation voltage and the slope factor were Ϫ6.95 Ϯ 0.91 mV and 5.55 Ϯ 0.34 in control (n ϭ 36) and Ϫ5.82 Ϯ 0.99 mV and 6.12 Ϯ 0.30 in the PK cleavage group (n ϭ 35, p Ͼ 0.05).

Frag(304 -613) requires Frag(1-303) to traffic to the membrane and form functional channels
Western blot analysis of whole-cell proteins extracted from Frag(1-303)-and/or Frag(304 -613)-expressing cells showed that the two fragments are robustly expressed regardless of whether they are independently expressed or coexpressed (Fig.  3A, left). However, in cells coexpressed with both fragments, a full-length Kv1.5 protein (75 kDa glycosylated or 65 kDa unglycosylated) was not observed (Fig. 3A, left), indicating that the current is generated by a Frag(1-303)ϩ(304 -613)-coassembled channel. Frag(304 -613) contains the pore region but did not generate current when expressed alone. To determine whether Frag(304 -613) requires Frag(1-303) for trafficking to The N terminus and S1 in Kv1.5 expression and function the plasma membrane, we isolated membrane protein using a biotinylation method. Frag(304 -613) was barely detected in the membrane protein when it was independently expressed (Fig. 3A, right). In contrast, when both fragments were coexpressed, Frag(304 -613) was robustly detected in the isolated membrane protein (Fig. 3A, right). In contrast, Frag(1-303) was detected in the membrane protein regardless of whether it was independently expressed or coexpressed with Frag(304 -613) (Fig. 3A, right). These results indicate that Frag(1-303) mediates the trafficking of Frag(304 -613) to the plasma membrane. This notion was confirmed by data from immunofluorescence microscopy experiments. As shown in Fig. 3B, Frag(1-303) was able to traffic to the plasma membrane when expressed alone or coexpressed with Frag(304 -613). However, Frag(304 -613) was retained intracellularly when expressed alone but was present on the plasma membrane when it was coexpressed with Frag(1-303) (Fig. 3B).
The T1-S1 linker and/or S1 segment are important for plasma membrane expression of WT Kv1. 5
The N terminus and S1 in Kv1. 5

The biophysical properties of Frag(1-303)؉(304 -613)coassembled Kv1.5 channels are different from WT channels
Our results showed that the N terminus interacts with other regions of Kv1.5 channels (Figs. 2-6), which is in line with previous studies demonstrating that the N terminus affects K ϩ channel gating (18). Consistently, although WT Kv1.5 was inactivated to a similar extent (flat) upon 5-s depolarizing voltages between 0 and 90 mV, the N-terminal truncation Kv1.5 mutant ⌬N209 displayed a U-shaped voltage dependence of inactiva-tion (Fig. 8A), an observation that was also reported by Kurata et al. (14,15). Interestingly, although Frag(1-303)ϩ(304 -613)coassembled Kv1.5 channels contained the N terminus, they displayed a U-shaped voltage dependence of inactivation, similar to ⌬N209 (Fig. 8A). The U-shaped voltage dependence of inactivation has been proposed to be a result of inactivation occurring from late closed states of the channel (19). Repeated depolarizations with short periods of negative voltage facilitate inactivation from late closed states, leading to excessive cumulative inactivation (14,19). As shown in Fig. 8B, shuttling between 60 mV for 90 ms and Ϫ80 mV for 10-ms steps resulted in more cumulative inactivation in both ⌬N209 and Frag(1-303)ϩ(304 -613)-coassembled channels than in WT Kv1.5 channels.

Discussion
Elucidating structure-function relationships is an essential step toward understanding the properties and regulation of Kv channels. Functional reconstitution from two nonoverlapping contiguous fragments represents a useful strategy to study the structure-function relationships for several transmembrane proteins such as ␤ 2 -adrenergic receptors, M1/M2 muscarinic acetylcholine receptors, lactose permease enzymes (20 -23), and voltage-gated chloride channels (24 -26). Upon functional reconstitution of transmembrane proteins from independent fragments, the contribution and behavior of different regions of the protein with regard to membrane insertion, protein interaction, assembly, trafficking, and function are further understood.
We previously demonstrated that, among various K ϩ channels, PK selectively cleaves hERG at a site within the S5-pore linker into N-and C-fragments, leading to a complete loss of hERG function (27). In contrast, we found that PK also cleaves Kv1.5 at a site within the S1-S2 linker into N-and C-fragments but does not affect channel function (10). In this recently published work, we used a Kv1.5 stable cell line that displayed very large currents (ϳ900 pA/pF upon 50-mV depolarization) to detect any loss of function. Due to the large currents, some cells displayed currents that exceeded the scale of the patch clamp amplifier (Axopatch 200B). As a result, the PK cleavageinduced increase in Kv1.5 current was overlooked in the previous study (10). To focus on the effects of PK cleavage on the functionality of Kv1.5 channels, we created a new stable Kv1.5-HEK cell line that expresses whole-cell current of 150 -300 pA/pF upon 50-mV depolarization. Our data showed that Kv1.5 current increased upon PK cleavage (Fig. 1D). The mechanisms of PK cleavage-induced current increase are currently under investigation but are not due to the altered channel gating. PK cleavage did not affect the voltage dependence of activation (Fig. 1D). PK cleavage neither resulted in U-shaped voltage dependence of inactivation nor diminished Kv1.5 inactivation (data not shown). The present study focused on how Frag(1-303) and Frag(304 -613) work together to form a functional channel. The mechanism underlying the formation of a functional channel after coexpression of the two DNA fragments remains to be fully elucidated. Our results showed that the two fragments are successfully coimmunoprecipitated (Fig.  4) unlike the fragments produced by PK treatment of full- The N terminus and S1 in Kv1. 5   The N terminus and S1 in Kv1.5 expression and function length Kv1.5 (Fig. 1C). We analyzed the involvement of various regions within Frag(1-303) to further understand Kv1.5 trafficking and function. Our results reveal a novel role of the N terminus in channel assembly: the N terminus can attract and interact with the independent Frag(304 -613) to form a functional channel.

expression and function
A structural element, termed the T1 (tetramerization) domain, within the N terminus of Shaker channels has been suggested to play a central role in the interaction and assembly of channels into homotetrameric or heterotetrameric proteins (28 -30). The T1 domain is a module of ϳ120 amino acids that adopts a "hanging gondola" structure, and its conformation is highly conserved among Kv channel subfamilies (31,32). However, several reports have demonstrated that Kv channels can still form functional channels upon deletion of the T1 domain (33)(34)(35). Nonetheless, the T1 domain has been suggested to impart stability on assembled channels by interactions within the T1 domain or with other channel elements (30). Our data showed that ⌬N209 Kv1.5, a naturally occurring isoform of Kv1.5 that lacks most of the T1 domain, is able to generate current (Fig. 7B), an observation that is consistent with previous reports (14,15,36). However, the present study revealed an important role of amino acid residues 1-209 in channel assembly in the absence of a covalent linkage within the S1-S2 linker. Our data show that Frag(1-209) interacted and associated with Frag(304 -613) (Fig. 5E). However, this interaction is not sufficient to form a functional channel as Frag(1-209) did not traffic to the membrane (Fig. 5D). In contrast, Frag(210 -303) facilitated trafficking to the plasma membrane (Fig. 5D) but did not interact with Frag(304 -613) (Fig. 5F). Thus, both Frag(1-209) and (210 -303) are required for Frag(304 -613) to form a functional channel (Fig. 2). Furthermore, residues 1-209 and 210 -303 must be conjoined as a single entity (Frag(1-303)) to assemble with Frag(304 -613) and form a functional channel because simultaneous expression of Frag(1-209) and Frag (210 -303) with Frag(304 -613) did not generate any current (Fig. 5B). These observations indicate that coexpressed Frag(1-209) and Frag(210 -303) could not interact and assemble properly with Frag(304 -613) to form a functional channel. This notion is consistent with our data obtained in the dominantnegative suppression experiments (Fig. 6). Frag(1-303) disrupted WT Kv1.5 function. One possible interpretation of this result is that Frag(1-303) is able to assemble with WT subunits to form heterotetrameric channels. Because Frag(1-303) lacks the pore region, its inclusion results in a dominant-negative suppression of WT channels (Fig. 6). In contrast, neither Frag(1-209) nor Frag(210 -303) produced a dominant-negative effect on WT Kv1.5 channels (Fig. 6). The lack of dominantnegative suppression of WT Kv1.5 with Frag(210 -303) is obvious because it did not interact with Frag(304 -613) (Fig. 5F). However, although Frag(1-209) interacted with Frag(304 -613) (Fig. 5E), it did not disrupt WT channels either. We reasoned that, because WT subunits traffic normally, but Frag (1-209) is trafficking-deficient (Fig. 5D), they eventually disassociate during maturation and trafficking.
Understanding the dominant-negative effects of Kv1.5 fragments is of biological and clinical importance. Previous reports have demonstrated that overexpression of Kv1.1 polypeptides containing the N terminus and the first transmembrane segment results in a dominant-negative effect on the WT Kv1.5 potassium channel, resulting in arrhythmias (37). A naturally occurring Kv1.5 truncation mutant with a stop codon in amino acid position 375 (conserving the N terminus and S1-S3 segments) was also shown to produce a dominant-negative effect on WT Kv1.5 channels, resulting in atrial fibrillation (17). Our results show that coexpression of Frag(1-303), but not Frag(304 -613) (data not shown), with WT Kv1.5 resulted in a dominant-negative suppression of Kv1.5 current (Fig. 6). These observations are consistent with the idea that the T1 domain of the N terminus mediates assembly of Kv channels (28 -30). We also observed that the dominant-negative suppression induced by Frag(1-303) was selective to the Kv1.5 channel as the functions of hERG (Kv11.1) and Kv4.3 were unaffected (data not shown). These data support the notion that the presence of the unique N-terminal domain can prevent coassembly between subfamilies (35).
The S1 segment has been shown to play a role in assembly, trafficking, and insertion of Kv channels into the membrane (16,28,38,39). Our results show that Frag(210 -303) was able to traffic to the plasma membrane (Fig. 5D). Because residues 210 -303 contain the S1 segment (residues 248 -269), this result may indicate the importance of the S1 segment in channel assembly, trafficking, and membrane expression. Although deletion of the N terminus up to residue 209 (⌬N209) did not disrupt Kv1.5 function, further deletion of amino acid residues 210 -240, which removes a region known as the T1-S1 linker of the N terminus, disrupted plasma membrane expression and function (Fig. 7). The T1-S1 linker possesses an ␣-helical conformation that separates the T1 domain from the S1 transmembrane segment (31,40). Deletion of this region may have changed the conformation of the S1 segment and thus its hydrophobic nature, preventing membrane expression. Additionally, although both Frag(210 -303) and the del(2-240) Kv1.5 channel can traffic through the Golgi apparatus (Figs. 5C and 7C), only Frag(210 -303) was able to insert into the plasma membrane, supporting the notion that the conformation of S1 is disrupted by the del(2-240) mutation. In brief, the finding that Frag(304 -613) alone does not reach the plasma membrane (Figs. 2 and 3) but ⌬N209 (residues 210 -613) does (Fig.  7, B and D) implies that the T1-S1 linker and/or the S1 segment is important for insertion of Kv1.5 channels into the plasma membrane.
The N terminus and T1 domain also play a role in channel gating. Crystallographic experiments have demonstrated that the T1 domain can assume different conformations, which are responsible for altering voltage-gating characteristics of Kv channels as it aligns with the pore and S4 transmembrane segment (18,31,41). In line with this, our results demonstrate that the N-terminal truncation mutant ⌬N209 displays a U-shaped voltage dependence of inactivation (Fig. 8), an observation that has been reported previously (14,15). The whole N terminus is present in Frag(1-303). The fact that Frag(1-303)ϩ(304 -613)-coassembled channels display U-shaped voltage dependence of inactivation suggests that that the impact of the N terminus on channel inactivation is lost in Frag(1-303)ϩ(304 -613)-coassembled channels (Fig. 8). PK cleavage did not result The N terminus and S1 in Kv1.5 expression and function in U-shaped voltage dependence of inactivation. Also, although the two PK-produced fragments of WT Kv1.5 did not associate (Fig. 1C), the coexpressed Frag(1-303) and Frag(304 -613) did associate (Fig. 4), and the N terminus is critical for the association (Fig. 5E). Thus, the N terminus of Frag(1-303)ϩ(304 -613)-coassembled channels is localized differently than WT channel regardless of PK cleavage, leading to the U-shaped voltage dependence of inactivation (Fig. 8).
In summary, by reconstituting functional Kv1.5 channels using Frag  and Frag(304 -613) (separated at the S1-S2 linker), our study revealed a novel role of the N terminus in channel maturation: the N terminus is powerful enough to attract and interact with the rest of the channel (i.e. Frag(304 -613)) to form a functional channel independently of the linkage between S1 and S2. Because the S1-S2 linker is usually intact in WT channels, the contribution of the N terminus to channel assembly through targeting the C-terminal side of the channel found in the present study is entirely novel. In addition, our results revealed the importance of the T1-S1 linker and/or S1 segment in plasma membrane expression. were established using G418 for selection (1 mg/ml) and maintenance (0.4 mg/ml). The WT Kv1.5-HEK cell line in the present study displays whole-cell currents in the range of 2000 -4000 pA (150 -300 pA/pF). These stable cell lines and otherwise indicated transient expressions were used throughout the study. HEK cells were cultured in minimum essential medium (MEM) supplemented with 10% fetal bovine serum, 1 mM sodium pyruvate, and 1ϫ nonessential amino acids (Thermo Fisher Scientific, Waltham, MA). Twenty-four hours after transfection, cells were collected for biochemical and patch clamp experiments. GFP-positive cells were used for whole-cell patch clamp recordings.

Western blot analysis and coimmunoprecipitation
Western blot analysis was used to detect protein expression of WT, mutants, and fragments of Kv1.5 channels. Cells were washed with ice-cold 1ϫ PBS, collected, and centrifuged at 100 ϫ g for 4 min. Next, the cell pellets were lysed using highfrequency sonification in ice-cold lysis buffer containing 1% PMSF and 1% protease inhibitor mixture. The lysates were centrifuged at 10,000 ϫ g for 10 min. The protein concentrations were determined using a protein assay kit. Appropriate amounts of double-distilled water and loading buffer containing 5% ␤-mercaptoethanol were added to the protein to make 0.3 g/l samples. Protein samples (15 g) were loaded and separated on 8, 10, or 15% SDS-polyacrylamide gels and transferred onto polyvinylidene difluoride membranes. BLUeye Prestained Protein Ladder was used to identify the mass of proteins. The membranes were blocked to prevent nonspecific protein interactions with 5% nonfat skim milk and 0.1% Tween 20 in Tris-buffered saline for 1 h at room temperature. Next, membranes were incubated with either 1:1000 N terminus-, 1:500 S1-S2 linker-, or 1:1000 C terminus-specific polyclonal rabbit anti-Kv1.5 primary antibody for 1 h at room temperature. Actin was detected as a loading control using 1:2000 monoclonal mouse anti-actin primary antibody. The membranes were then incubated with goat anti-rabbit (for Kv1.5 detection) or horse anti-mouse (for actin detection) horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 h at room temperature. An enhanced chemiluminescence detection kit was used to visualize the protein bands on X-ray films.
For co-IP, 0.5 mg of whole-cell protein in 0.25 ml of lysis buffer was incubated with appropriate primary antibody overnight at 4°C. GAPDH was precipitated with rabbit anti-GAPDH primary antibody as a negative control. Protein A/G Plus-agarose beads were then added to the protein complexes for 4 h at 4°C prior to precipitation by centrifugation at 10,000 ϫ g for 1 min. Next, the beads were washed five times with cold lysis buffer before being resuspended in 2ϫ sample buffer. The samples were boiled for 5 min and centrifuged at 20,000 ϫ g for 5 min. Finally, the supernatants were loaded into SDS-polyacrylamide gels and subjected to Western blot analysis.

Electrophysiological recordings
A whole-cell patch clamp method was used to record I Kv1.5 . Cells were settled on the bottom of a 0.5-ml perfusion chamber in the bath solution. Patch glass pipettes were pulled using thinwalled borosilicate glass (World Precision Instruments, Sarasota, FL). The pipettes had inner diameters of 1.5 m and resistances of 2 megaohms when filled with solution. An Axopatch 200B amplifier and pCLAMP10 (Molecular Devices, San Jose, CA) were used for data acquisition and analysis. Data were sampled at 20 kHz and filtered at 5 kHz. Series resistance was compensated by 80%, and leak subtraction was not used. I Kv1.5 was elicited from a holding potential of Ϫ80 mV by depolarizing steps to voltages between Ϫ70 and ϩ70 mV in 10-mV increments for 200 ms. A repolarizing step to either Ϫ50 or Ϫ20 mV was applied for 250 ms before returning to the holding potential. To measure the voltage dependence of the channel inacti-The N terminus and S1 in Kv1.5 expression and function vation, a series of 5-s conditioning pulses to various potentials were followed by a 1-s test pulse to ϩ60 mV. The current amplitudes during the test pulse were plotted against voltages of the conditioning pulses to construct voltage-inactivation relationships. In addition, 100 repetitive depolarization pulses to ϩ60 mV for 90 ms and repolarization to Ϫ80 mV for 10 ms were applied continuously to observe excessive cumulative inactivation. The bath solution contained 135 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 10 mM glucose, and 10 mM HEPES (pH 7.4 with NaOH). The pipette solution consisted of 135 mM KCl, 5 mM MgATP, 5 mM EGTA, and 10 mM HEPES (pH 7.2 with KOH). Patch clamp experiments were performed at room temperature (22 Ϯ 1°C).

Isolation of cell-surface proteins
To determine the cell-surface localization of the N-and C-fragments, a cell surface protein isolation kit was used to extract membrane proteins from HEK cells expressing the fragments. Cells were grown to 90% confluence in 100-mm culture dishes. Membrane proteins were labeled with a membrane-impermeant thiol-cleavable amine-reactive biotinylation reagent, Sulfo-NHS-SS-biotin, at 250 g/ml for 30 min at 4°C. Quenching solution was then added to cease the labeling reaction. Cells with biotinylated surface proteins were collected and lysed with lysis buffer containing 1% protease inhibitor mixture. After centrifugation at 10,000 ϫ g for 2 min at 4°C, biotin-tagged proteins were isolated using NeutrAvidin-agarose columns and eluted with SDS-polyacrylamide sample buffer containing DTT. The isolated cell-surface proteins were analyzed via Western blot analysis. As a loading control, Na ϩ /K ϩ -ATPase expression was detected using mouse anti-Na ϩ /K ϩ -ATPase ␣1 antibody and horse anti-mouse HRP-conjugated secondary antibody.

Immunofluorescence microscopy
HEK cells expressing WT Kv1.5, Kv1.5 mutants, or fragments were grown on glass coverslips. Live cell membranes were stained with Oregon Green 488 wheat germ agglutinin (5 g/ml) for 1 min in Hanks' balanced salt solution. The cells were then fixed using 4% ice-cold paraformaldehyde in PBS for 15 min, permeabilized with 0.1% Triton X-100 for 10 min, and blocked with 5% BSA in PBS for 1 h. WT and mutants of Kv1.5 were labeled with N terminus-, S1-S2 linker-, or C terminusspecific rabbit anti-Kv1.5 primary antibody and Alexa Fluor 594 -conjugated donkey anti-rabbit secondary antibody. The coverslips were mounted onto glass slides using Prolong Gold antifade reagent. Images were obtained using a Leica TCS SP2 multiphoton confocal microscope (Leica, Heidelberg, Germany).