Identification of a Membrane-targeting Domain of the Transient Receptor Potential Canonical (TRPC)4 Channel Unrelated to Its Formation of a Tetrameric Structure*

Background: Functional TRPC4/5 channels make tetrameric structures in the plasma membrane. Results: The 21–30 motif of TRPC4/5 regulates the membrane expression of the channels by interacting with PI(4,5)P2. Conclusion: The 21–30 motif of TRPC4/5 has an essential role in the membrane expression that does not involve the tetrameric structure. Significance: Our findings indicate the membrane expression regulating domain of TRP channels. Canonical transient receptor potential (TRPC) channels are Ca2+-permeable nonselective cation channels that are activated by a wide variety of stimuli, including G protein-coupled receptors (GPCRs). The TRPC4 channel is expressed in a punctate distribution in the membrane. To identify the regulating region of the channel trafficking to the membrane, we generated deletion mutants of the TRPC4 channel. We determined that when either region that was downstream of the 20 amino acids of the N terminus or the 700–730 amino acids was deleted, the mutants were retained in the endoplasmic reticulum. By coexpression of the wild-type TRPC4 with deletion mutants, we found that the 23–29 amino acids of the N terminus regulate a membrane trafficking. Additionally, by the fluorescence resonance energy transfer (FRET) method, we found that the regions downstream of the 99 amino acid region of the N terminus and upstream of the 730 amino acid region in the C terminus produce assembly of the TRPC4 tetramers. We inferred the candidate proteins that regulate or interact with the 23–29 domain of TRPC4.

ficking at the plasma membrane is limited, and they are retained in the endoplasmic reticulum (ER) or Golgi (1)(2)(3)(4). Among the TRPC channels, TRPC4 and TRPC5 channels are unique in the aspect of their activation by the G␣ i protein.
Additionally, the region that is responsible for the interaction with G␣ i protein is important for CaM binding. The deletion of this region causes the TRPC4 and TRPC5 channel to be nonfunctional (5,6). When Arg 718 , Lys 722 , and Arg 723 were substituted with alanine, mouse TRPC5 were destroyed and failed respond to stimulation by thrombin and bradykinin in both CHO and HEK293 cells. However, the mutant proteins were efficiently delivered to the plasma membrane when detected by surface biotinylation (6). In this report, we concluded that the calmodulin and inositol 1,4,5-triphosphate receptor-binding (CIRB) region is important for channel gating and/or channel activation by agonists rather than that there is a dysfunction in the trafficking and translocation of the channel protein. Similar results were obtained in the TRPC4 channel expressed in HEK cells (5). One point mutation within the region (N712R) did not show any typical doubly rectifying I-V curve, whereas R716A mutant showed an enhanced basal activity (5). This finding is in contrast to the finding that deletion of the CIRB site impaired the plasma membrane translocation of TRPC3 (3).
When TRPC4 channels were expressed in HEK cells, their distribution was not homogeneous at the plasma membrane, and instead, they existed as puncta or discrete microdomains. Several studies have shown that when the C terminus domain is deleted, the expression of the channels is limited because the deleted region is where the regulation of channel trafficking takes place (7,8). On the other hand, the importance of the N terminus and C terminus domain with regard to the formation of the TRPC4 multimer has also been studied (9 -11). The ankyrin repeat domains are especially important for the assembly of tetrameric TRPC4 structure. Recently, (12) showed the structure of TRPV1, which suggests that ARD was not involved in the assembly of TRPV1 tetramers but instead is involved in protein-protein interactions with other proteins.
In a previous study (5), we investigated the interaction of TRPC4 mutants with G␣ i proteins and observed their expression at the plasma membrane. Interestingly, the distribution of TRPC4 at the plasma membrane depended on the deletion regions. From the deletion mutants, we found that the 23-29 amino acid (aa) domain is important for membrane trafficking and that downstream of 99 aa of the N terminus and 700 -730 aa region is important for the assembly of TRPC4 tetramers.

EXPERIMENTAL PROCEDURES
Cell Culture and Transient Transfection, cDNA Clones-Human embryonic kidney (HEK293) cells (ATCC, Manassas, VA) were maintained according to the supplier's recommendations. HEK293 cells were incubated in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% heat-inactivated FBS and penicillin (100 units/ml), streptomycin (100 g/ml) at 37°C in 5% CO 2 humidified incubator. Cells were seeded in confocal dish for recording FRET or 12-well plates for whole-cell patch clamp. The following day, transfection was performed with Fugene-6 according to the manufacturer's instructions. XFP (CFP or YFP)-tagged TRPC4, TRPC5, and TRPC1 were transfection in this way. The next day we performed electrophysiology or FRET experiments.
Electrophysiology-The cells were transferred onto a solution chamber on the stage of an invert microscope (IX70, Olympus, Japan). The whole cell configuration was used to measure TRPC channel current in HEK cells as described previously (13)(14)(15)(16)(17). Cells were left for 10 -15 min to attach to coverslips. Whole cell current were recorded using an Axopatch 200B amplifier (Axon instruments). Patch pipettes were made from bososilicate glass and had resistances of 3-5 M⍀ when filled with normal intracellular solutions. Bath solution was changed from Normal Tyrode (NT) to Cs ϩ rich external solution after whole cell recording system established. The NT contained 135 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 10 mM glucose, and 10 mM HEPES with a pH that was adjusted to 7.4 using NaOH. The Cs ϩ -rich external solution contained equimolar CsCl rather than NaCl and KCl. The internal solution contained 140 mM CsCl, 10 mM HEPES, 0.2 mM Tris-guanosine 5Ј-triphospate, 0.5 mM EGTA, and 3 mM Mg-adenosine 5Ј-triphosphate with a pH that was adjusted to 7.3 with CsOH. We used 0.2 mM GTP␥S that was purchased from Sigma. Voltage ramp pulse was applied from ϩ100 mV to Ϫ100 mV for 500 ms at Ϫ60 mV holding potential. Experiments were performed at room temperature (18 -22°C). The recording chamber was continuously perfused at a flow rate of 1-2 ml/min. Image Quantification and FRET Measurements-HEK293 cells were cultured in 35-mm coverslip bottom dish to obtain image and to measure FRET efficiency. To obtain the image and FRET efficiency of a cell, we used an inverted microscope (IX70, Olympus, Japan) with a 60ϫ oil objective lens and the 3 cube FRET calculation (18) controlled by MetaMorph 7.6 (Molecular Devices). We mainly used 3 cube FRET. 3 cube FRET efficiency (cube settings for CFP, YFP, and Raw FRET) were acquired from a pE-1 Main Unit to 3 cube FRET (excitation, dichroic mirror, filter) through a fixed collimator: CFP (ET 435/20 nm, ET CFP/YFP/mCherry beam splitter, ET 470/24 nm, chroma); YFP (ET500/20m, ET CFP/YFP/mCherry beam splitter, ET535/30 nm, chroma); and Raw FRET (ET435/20 nm, ET CFP/YFP/mCherry beam splitter, ET535/30 nm, Chroma). The excitation LED and filter were sequentially rotated, rota-tion period for each of filter cubes was ϳ0.5 s, and all images (three for CFP/YFP/Raw FRET, respectively) were obtained within 1.5 s. Each of the images was acquired on a cooled 10 MHz (14 bit) CCD camera (DR-328G-C01-SIL: Clara, ANDOR Technology) with an exposure time of 100 ms with 2 ϫ 2 or 3 ϫ 3 binning under the control of MetaMorph 7.6 software. Our FRET recording of the fluorophores was restricted in a range of CFP/YFP ratio being 0.5 to 2.0.
FR and FRET Efficiency Computation (18)-FRET Ratio (FR) is equal to the fractional increase in YFP emission due to FRET and was calculated as in Equation 1, where E is the intrinsic FRET efficiency when fluorophoretagged molecules are associated with each other, A b is the fraction of YFP tagged molecules that are associated with CFPtagged molecules, and the bracketed term is the ratio of YFP and CFP molar extinction coefficients scaled for the FRET cube excitation filter (21). We determined this ratio to be 0.094 based on maximal extinction coefficients for ECFP and EYFP (19) and excitation spectra measured in our laboratory. FRET was also assessed by measuring dequenching of donor emission following nearly complete acceptor photodestruction (20) by 30 min of strong illumination through a 540AF30 excitation filter. This spared the CFP chromophore in control experiments. Here, the effective FRET efficiency is calculated as Equation 3, where S CFP (DA) before and S CFP (DA) after are CFP emission prior to and following YFP photobleaching, and Db is the fraction of CFP-tagged molecules that are associated with YFP-tagged molecules (21). Western Blotting Analyses, Co-IP, and Surface Biotinylation-For Western blotting, cells were seeded in 6-well plates. On the next day, 0.5-2 g/well of TRPC4␤ cDNA was transfected into cells using the transfection reagent Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. After transfection for 24 h, the cells were harvested as follows. Lysates were prepared in lysis buffer (0.5% Triton X-100, 50 Tris-Cl, 150 NaCl, 1 EDTA, pH 7.5, (in mM)) by being passed through a 26-gauge needle 7-10 times after sonication. Lysates were centrifuged at 13,000 ϫ g for 10 min at 4°C, and the protein concentration in the supernatants was determined. The proteins extracted in sample buffer were loaded onto 8% Tris-glycine SDS-PAGE gels, and then subsequently transferred onto a PVDF membrane. The proteins were probed with GFP (Invitrogen) or ␤-Actin (GeneTex) antibodies for GFP-tagged or housekeeping protein as indicated.
Surface Biotinylation-PBS-washed cells were incubated in 0.5 mg/ml sulfo-NHS-LC-biotin (Pierce) in PBS for 30 min on ice. Afterward, the biotin was quenched by the addition of 100 mM glycine in PBS. The cells were then processed as described above to make cell extract. Forty microliters of 1:1 slurry of immobilized avidin beads (Pierce) were added to 300 l of cell lysates (500 g of protein). After incubation for 1 h at room temperature, beads were washed three times with 0.5% Triton-X-100 in PBS, and proteins were extracted in sample buffer. Collected proteins were then analyzed by Western blot. The experiment was performed as previously described in detail (22).
Measurement of Colocalization and Quantification in Images-To determine colocalization of two image in membrane of cell, we made a binary mask image using the "image threshold" routine of MetaMorph 7.6. A CFP mask and a YFP mask were made from a CFP image and a YFP image, respectively. We defined the overlapping area of the CFP mask and the YFP mask as the colocalization region. Finally, the colocalization was calculated by dividing the spatially area of colocalization region over the CFP masked area.
To determine the quantification of membrane expression in image, we substrate background intensity from original image using the "Background subtraction" routine of MetaMorph 7.6. We obtained plasma membrane region from the PH-YFP image expressing membrane region. Then we obtained intensity of membrane region defined by PH-YFP image and total intensity whole cell region. Finally, the membrane/total ration was calculated by dividing the membrane intensity over the total intensity.

Membrane Expression of TRPC4 Deletion
Mutants-It has been suggested that the TRPC channels assemble into tetramers, and the expression of the wild type (WT) TRPC4 (TRPC4-WT) channel shows a punctate distribution at the plasma membrane (PM) (23). However, some of the deletion mutants of the TRP channels, namely, TRPV4 (4), TRPC3 (3), TRPV5, 6 (2), TRPM4 (1), and TRPV5 (24), have restricted translocation to the PM and appeared to be predominantly located in the intracellular compartments, for example, the endoplasmic reticulum. Previous reports have shown that because the C terminus region of TRPC4 regulates the insertion of the channel into the PM, the deletion of certain C terminus regions restricts trafficking the channel proteins at the PM (7, 25). We generated deletion mutants of TRPC4␤ to find regions that regulate the expression of the channel at the PM and its function (Fig. 1A).
To observe the subcellular localization of TRPC4␤ deletion mutants, the constructs of CFP-or YFP-tagged PLC␦ PH domain as a membrane marker were co-expressed in HEK293 cells (Fig. 1B). Red and green represent CFP and YFP in all of the images. The PH domain of PLC␦ binds specifically to PIP 2 at the PM (26). The TRPC4-WT channel is at the PM in the manner of a punctate distribution. To find the domain that is crucial for the expression of the TRPC4 channel at the PM, we checked a deletion mutant distribution in the HEK293 cells. Fig. 1B shows that the TRPC4-WT channel and N terminus deletion mutants TRPC4-⌬1-10 and TRPC4-⌬11-20 manifested a robust punctate distribution at the cell surface, whereas TRPC4-⌬21-30, TRPC4-⌬1-30, TRPC4-⌬11-30, TRPC4-⌬1-98, and TRPC4-⌬1-124 showed retention in the cytosol. When the C terminus deletion mutants were observed, the CIRB domain and SESTD1-deleted mutants (TRPC4-⌬700 -728, TRPC4-⌬700 -710, and TRPC4-⌬710 -720) were not expressed at the PM and were instead retained in the cytosol. Interestingly, the deletion mutant, in which 150 aa were deleted in the C terminus (TRPC4-⌬720 -870), was expressed at the plasma membrane at a similar expression level as TRPC4-WT. A line scan across the image showed similar results (Fig. 1B). To quantify the surface expression of the mutants, the intensity of surface-expressed fluorescence in the image divided by the total intensity of the whole cell is shown (Fig. 1C).
To confirm the surface expression quantitatively with a biochemical experiment, we used cell surface biotinylation (Fig.  2B). TRPC4 and deletion mutant expression in whole cell lysate and the biotinylated fraction were determined using immunoblotting. Strikingly, both TRPC4␤-WT and TRPC4␤-⌬21-30 were present at the plasma membrane, although the latter is significantly reduced (30.76 Ϯ 10.23%, n ϭ 3) at the PM. This finding was confirmed by quantification of the protein levels (Fig. 2B).The activity of the deletion mutants was measured by the TRPC4 current in the deletion mutants in response to GTP␥S, which is a well-known activator of TRPC4 (Fig. 2C). TRPC4␤-WT, TRPC4␤-⌬1-10, TRPC4␤-⌬11-20, and TRPC4␤-⌬720 -870 showed the typical TRPC4 current with a double rectifying shape that is a general shape of the I/V shape of TRPC4. On the other hand, other mutants did not show typical TRPC4 currents. We also generated a mutant that lacks the last 4 C terminus aa (TTRL) that comprise a PDZ domain-binding motif. This motif binds to the PDZ domain of the scaffolding proteins EBP50 and ZO-1 and regulates the surface expression of TRPC4 (7,25). In contrast to the previous results, we could not find a difference between the membrane expression and the function of TRPC4␤-⌬TTRL and TRPC4-WT (Fig. 2D).
These results suggest that there is a correlation between the whole cell current and punctate distribution at the PM of the TRPC4 channel and between trafficking into the plasma membrane (surface biotinylation) and the punctate distribution at the PM of the TRPC4 channel, except for TRPC4-⌬1-10.
Regions of TRPC4 for the Punctate Distribution at the Plasma Membrane-In the experiments mentioned above, the mutants might be retained in the ER due to either a lack of the region for membrane anchoring or impaired tetrameric assembly. A functional TRPC channel complex is believed to be composed of a tetrameric structure of TRPC proteins. Thus, we can infer that TRPC4-WT might rescue TRPC4 deletion mutants that do not have the membrane anchoring region by making tetramers. To test whether TRPC4-WT could rescue deletion mutants that are retained in the endoplasmic reticulum, we co-expressed TRPC4␤-WT-YFP and CFP-fused deletion mutants of TRPC4␤ in HEK293 cells and visualized the subcellular distribution by microscopy. As shown in Fig. 3A, the TRPC4␤-WT-YFP at the plasma membrane overlapped only in cells that expressed TRPC4␤-WT-CFP, TRPC4-⌬1-10-CFP, ⌬11-20-CFP, ⌬21-30-CFP, ⌬1-30-CFP, ⌬11-30-CFP, ⌬1-98-CFP, and ⌬720 -870-CFP, whereas retention in the cytosol was detected in the other mutants. We observed that TRPC4-WT is capable of ushering TRPC4␤-⌬21-30, TRPC4␤-⌬1-30, TRPC4␤-⌬11-30, and TRPC4␤-⌬1-98 into the plasma membrane, most likely by forming mixed tetramers. A line scan across the image showed similar results (Fig. 3A). We quantified the PM distribution of mutants when co-expressed with the wild type channel or when expressed alone (Fig. 3B), and we measured the colocalization of the wild type TRPC4 with deletion mutants at the plasma membrane (Fig. 3C). Fig. 3C shows the overlapping region of wild type TRPC4 with deletion mutants at the PM of the cell. This result indicates that some deletion mutants can be inserted into the PM when co-expressed with wild type TRPC4, whereas when expressed alone, they are retained in the endoplasmic reticulum. This result also FIGURE 1. Membrane expression of TRPC4 deletion mutants. A, diagram of deletion mutants of TRPC4␤. The TRPC4 protein has four ankyrin-like repeats and a coiled-coil domain in their N terminus strands followed by six transmembrane domains and a C terminus that contains the SESD1, CIRB, and PDZ binding domain. B, CFP-tagged deletion mutants of TRPC4␤ (green) were coexpressed with the PH-YFP (red), selectively binding to PI(4,5)P 2 , as an indicator of plasma membrane in HEK293 cells. Wild type TRPC4␤ channel, ⌬1-10, ⌬11-20, and ⌬720 -870 deletion mutants were observed in the plasma membrane of the cells. The scale bar represents 10 m. The line scan shows CFP-tagged mutants and PH-YFP intensity followed by a white dashed line. C, quantification of the membrane expression of the mutants. This graph shows the fluorescence intensity ratio of the membrane expression of deletion mutants to the whole cell expression. Error bars indicate S.E. DECEMBER 12, 2014 • VOLUME 289 • NUMBER 50

JOURNAL OF BIOLOGICAL CHEMISTRY 34993
shows that the deleted regions are important for the insertion of the channels into the membrane. Consequently, when independently expressed, the ⌬21-30, ⌬1-30, ⌬11-30, and ⌬1-98 deletion mutants cannot be trafficked to the PM, but when co-expressed with WT, they are trafficked to the PM. Knowing that these mutants require the aid of WT to be trafficked to the PM, we can reasonably infer that the deleted regions of the deletion mutants, especially 21-30 aa, are the crucial regions for membrane insertion and punctate distribution at the PM. To confirm that TRPC4␤ (⌬21-30) could be expressed at the plasma membrane by forming tetrameric structures with TRPC4␤-WT, we used surface biotinylation. To accomplish this goal, the cells were transfected with TRPC4␤-Flag and TRPC4␤-⌬21-30 or TRPC4␤-⌬21-30 alone. A portion of the total lysate was retained to determine the expression levels of the transfected constructs. Biotinylated surface proteins were then recovered by incubation with immobilized avidinagarose. As shown in Fig. 4A, although the cell total fraction of the band for the TRPC4␤-⌬21-30 that co-expressed with TRPC4␤-WT-Flag appeared to be less dense than that derived from cells transfected with TRPC4␤-⌬21-30 alone, we found that the level of TRPC4␤-⌬21-30 co-expressed with TRPC4␤-WT-Flag increased in the cell surface compared with TRPC4␤-⌬21-30 alone. A similar result was obtained for TRPC4␤-⌬11-30 (Fig. 4A). However, TRPC4␤-WT did not increase the expression of TRPC4␤-⌬1-124 at the cell surface. Nevertheless, all of the cotransfected cells showed the typical TRPC4 current with a double rectifying shape (Fig. 4B).
As observed in the TRPC4 channel, when TRPC5-⌬21-30 is expressed in the HEK293 cell, the mutants were retained in the endoplasmic reticulum (Fig. 4C). However, when TRPC5-⌬21-30 is co-expressed with TRPC5-WT with the aid of the wild type, TRPC5-⌬21-30 is successfully expressed in the membrane (Fig. 4D). These results suggest that the 21-30 domain of the TRPC5 and TRPC4 channels was responsible for the membrane trafficking and punctate distribution but not for channel tetramerization.

JOURNAL OF BIOLOGICAL CHEMISTRY 34995
TRPC4␤ (23-29, C1) overlaps with the ER resident chaperone calnexin. Co-localization of TRPC4␤ (23-29, C1) with the Golgi marker, FAPP1, was not detected. TRPC4 chimera were trafficked when 27-29, 25-29, or 23-26 aa of the TRPC4␤ channel were substituted with the relevant part of TRPC1, whereas independently expressed TRPC4␤ (23-29, C1) could not be located at the PM. When the chimera mutants co-expressed with WT, however, all four chimera were successfully located at the PM (Fig. 6B). The function of the chimera mutants was tested with a patch clamp technique (Fig. 6C). Except for TRPC4␤ (23-29, C1) chimera, other chimera mutants showed I/V curves that had the general TRPC4 double rectifying shape in response to GTP␥S. These findings, up to this point, demonstrate that N terminus 23-29 aa of TRPC4 is a critical region for insertion into the PM.
The Important Region for the Formation of the TRPC4 Tetrameric Structure-Collecting these data, we hypothesized that deletion mutants that impaired proper folding and could not insert into the PM were not expressed alone and were not coexpressed with the wild type TRPC4. Fig. 3A showed that TRPC4␤-⌬1-124, TRPC4␤-⌬700 -728, TRPC4␤-⌬700 -710, and TRPC4␤-⌬710 -720 failed to reach the plasma membrane and showed ER retention even if co-expressed with TRPC4␤-WT. In other words, the preceding 4 deletion mutants were not properly folded or assembled with the wild type. From these results, we assumed that there are two regions that are responsible for the tetrameric assembly of TRPC4 channels. First, the N terminus region from the 98th aa to the 124th aa is one of the regions for making a tetrameric structure. The reason is that the mutants that were deleted further than the 98th aa could not be trafficked into the PM even with the aid of TRPC4-WT. Second, the deleted SESTD1 and CIRB domain showed the same phenomenon by deletion. For the C terminus deletion mutants TRPC4␤-⌬700-728, TRPC4␤-⌬700-710, and TRPC4␤-⌬710 -720, the channels could not be trafficked to the PM as well as when co-expressed with WT (9 -11). The SESTD1 (SEC14-like and spectrin-type domain 1) binding domain of TRPC4 and TRPC5 is crucial for proper working and regulation (5,28). This sequence is highly conserved in TRPC5 and largely overlaps with the CIRB domain.
Interestingly, the FRET signal between TRPC4␤-WT-CFP and TRPC4␤-⌬720 -870-YFP (14.32 Ϯ 1.32%, n ϭ 5) was lower than the positive control. This finding of reduced FRET efficiency appears to be due to the loss of the 150 aa in the C terminus of TRPC4␤. Consequently, CFP and YFP were separated further compared with the C terminus-tagged fluorescence proteins and TRPC4␤-WT.

JOURNAL OF BIOLOGICAL CHEMISTRY 34997
have significantly reduced the FRET efficiency but have greater FRET efficiency than that at the PM. TRPC4␤-⌬720 -870-YFP (14.32 Ϯ 1.32%, n ϭ 5) has a similar FRET efficiency to that at the PM.

Candidates for Regulators of TRPC4 Channel Trafficking-
We investigated which candidate molecules interact with the 23-29 aa of the TRPC4 channel and, consequently, regulate the channel trafficking and punctate distribution at the PM. Previ- Middle column: CFP-tagged chimera mutants were coexpressed with YFP-tagged FAPP1-PH, to be a Golgi marker. Right column: CFP-tagged chimera mutants were coexpressed with YFP-tagged calnexin, to be an ER marker. Chimera mutants expressed a punctate distribution in the membrane except for TRPC4␤ (23-29, C1). The scale bar represents 10 m. B, chimera mutants (green) were expressed with the wild type TRPC4␤ (red). All of the chimera mutants expressed in the plasma membrane involved TRPC4␤ (23-29, C1), which was expressed in the ER when transfected alone. The scale bar represents 10 m. C, whole-cell patch clamp recordings reveal that cells that expressed chimera mutants show robust GTP␥-induced currents, except for mTRPC4␤ (23-29, C1) (n ϭ 3-10). The white column was measured with ϩ100 mV, and the black column with Ϫ60 mV. Error bars indicate S.E. ous studies have shown that the following candidates, for example, caveolin (29), PI(4,5)P 2 (30), small G protein (31), cytoskeleton, SESTD1 (28), and protein 4.1 (32), interacted with TRPC4 and could regulate membrane trafficking and docking.
Finally, the Val 57 -Thr 58 -Val 59 tripeptide at the N terminus of human concentrative nucleoside transporter 3 (hCNT3) appears to be the core of the endoplasmic reticulum export signal (47). Our preceding results have demonstrated that the 23 VRAETEL 29 aa residues require TRPC4 expression at the cell surface. This domain of TRPC4 and TRPC5 also contains 26 EXE 28 , which is very similar to the di-acidic ER export motifs that have been identified in several plasma membrane proteins (48 -52). However, TRPC4␤ (E26A/E28A) but not the (⌬23, 24) channel were expressed at the plasma membrane (Fig. 8F).

DISCUSSION
The present studies showed that 1) N terminus 23-29 aa of TRPC4 are a crucial domain for membrane insertion of channel proteins, 2) modifying this domain could cause trafficking to the plasma membrane by making proper folding or heterotetrameric assembly with wild type TRPC4, 3) N or C terminus interact with each other to make a tetrameric structure, 4) PI(4,5)P 2 is a possible candidate for interaction with the membrane-targeting domain at the N terminus of TRPC4, and 5) N terminus 98 -124 aa and C terminus 700 -728 aa are critical for the tetrameric assembly of TRPC4.
By the imaging approach, we could find a membrane-targeting domain of the channels independent of the domains that are critical for the tetrameric or dimeric structure. The 23 VRAETEL 29 aa residues are required for TRPC4 expression at the cell surface (Fig. 6). This domain is not related to ER export. The N terminus of the TRPC3 channel acts as a half-domain of the PH domain with the PLC␥ PH domain (45,(53)(54)(55). The report also showed that the TRPC4 N terminus has a similar amino acid sequence and acts as a half domain of the PH domain.
There was a correlation between the whole cell current and the punctate distribution at the PM of the TRPC4 channel (Figs. 1 and 2) and between trafficking into the plasma membrane (surface biotinylation) and the punctate distribution at the PM of the TRPC4 channel, except for TRPC4-⌬1-10 (Fig. 2). This exception might occur due to the specific definition of PM expression. We defined PM expression of the TRPC4 channel to occur when the TRPC4 channel distributes as a patch or puncta rather than a random distribution at the PM. In the case of the TRP channels, to distinguish the subplasmalemmal distribution from the PM distribution was difficult. The TRPC4-⌬1-10 deletion mutant might localize uniquely at a specific area and function in a more efficient way even though the total protein at the cell surface is lower. A similar result was obtained in TRPV5 (24).
The patch clamp technique is useful for measuring the current from the functioning ion channels. In our hands, this technique has been very sensitive and reliable for the TRPC4 channels (5, 14 -16, 22, 30). On the other hand, this technique was not good for the rescue experiment of deletion mutants with TRPC4-WT or the co-expression experiment of more than 2 constructs (Figs. 2C and 4B). In general, making a dimer or tetramer construct takes a long time, and dimer or tetramer constructs usually do not work well due to the linker length or the flexibility of the linker (22). To combine the imaging methods and the FRET methods was useful for finding a membranetargeting domain or tetrameric assembly domain.
In the present study, we found a novel membrane-targeting sequence ( 23 VRAETEL 29 ) of TRPC4 channels. The N terminus region (98 -124 aa) and the C terminus region (700 -728 aa) are critical for tetrameric assembly.