| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* MRC Clinical Sciences Centre, Faculty of Medicine, Hammersmith Hospital Campus, Imperial College, London, UK; and
Gene Targeting Unit, Department of Cellular and Molecular Neurosciences, Division of Neuroscience and Psychological Medicine, Faculty of Medicine, Charing Cross Campus, Imperial College, London, UK
3Correspondence: MRC Clinical Sciences Centre, Imperial College Faculty of Medicine, Hammersmith Hospital Campus, Du Cane Rd., London W12 0NN, UK. E-mail: a.sardini{at}csc.mrc.ac.uk
ABSTRACT
Despite considerable similarity in their amino acid sequences and structural features, the mammalian members of the CLC chloride channel/transporter family have different subcellular locations. The subcellular location and function of one of these members, hClC-4, is controversial. To characterize its cellular function, we investigated its tissue distribution and subcellular location. Expression was high in excitable tissues such as the nervous system and skeletal muscle. When heterologously expressed in HEK293 cells and in skeletal muscle fibers, hClC-4 localizes to the endoplasmic/sarcoplasmic reticulum (ER/SR) membranes, in contrast to hClC-3, which localizes to vesicular structures. This location was confirmed by identification of endogenous ClC-4 in membrane fractions from mouse brain homogenate enriched for the sarco-endoplasmic reticulum ATPase SERCA2, an ER/SR marker. To identify the motif responsible for ER localization of hClC-4, we generated hClC-4 truncations and chimeras between hClC-4 and hClC-3 or the unrelated plasma membrane protein Ly49E. A stretch of amino acids, residues 14–63, at the N-terminus constitutes a novel motif both necessary and sufficient for targeting hClC-4 and other membrane proteins to the ER.—Okkenhaug, H., Weylandt, K.-H., Carmena, D., Wells, D. J., Higgins, C. F., Sardini, A. The human ClC-4 protein, a member of the CLC chloride channel/transporter family, is localized to the endoplasmic reticulum by its N-terminus.
Key Words: CLC family proteins • targeting motif • sarcoplasmic reticulum
THE MAMMALIAN MEMBERS of the CLC chloride channel/transporter family serve a wide variety of cellular functions such as stabilization of membrane potential in skeletal muscle fibers, regulation of the electrical excitability of neurons, transepithelial salt transport in the thick ascending limb of Henle’s loop in the kidney, and the stria vascularis of the inner ear, and in generating the anion shunt current necessary for effective acidification of endosomes, synaptic vesicles, and the reabsorption lacuna of osteoclasts (1)
. Human diseases such as osteopetrosis, Dent’s disease, Bartter’s syndrome, idiopathic generalized epilepsy, and myotonia congenita are due to mutations in ClC-7 (2)
, ClC-5 (3)
, ClC-Kb (4)
, ClC-2 (5)
, and ClC-1 (6)
, respectively. Not unexpectedly, therefore, the different CLC proteins have very different tissue distributions and subcellular locations. ClC-1, ClC-2, ClC-Ka, and ClC-Kb reside in the plasma membrane, whereas ClC-3 and ClC-5 are localized in endosomes and ClC-7 in late endosomes and lysosomes.
ClC-4 is one of the less characterized members of the CLC family. No disease-causing mutations have been identified in the human ClC-4 (hClC-4) gene (7
, 8)
. Outwardly rectifying plasma membrane chloride currents associated with heterologous expression of ClC-4 in overexpressing systems have been recorded (9
10
11)
. Recently, these currents have been shown to be associated with proton transport (12
, 13)
. The authors concluded that ClC-4 as well as ClC-5 behave as chloride/proton antiporters, similarly to the behavior of bacterial CLC proteins (14)
. However, the localization of ClC-4 to the plasma membrane in these systems is likely to be a consequence of overexpression since ClC-5 and ClC-3, which are known to be intracellular, also generate currents in the plasma membrane when overexpressed (9
, 15
, 16)
. In native cells, data are contradictory. ClC-4 has been reported to colocalize with the cystic fibrosis transmembrane conductance regulator (CFTR) in the brush border membrane of epithelial intestinal cells (17)
, with ClC-5 in early endosomes of COS-7 cells (18)
, and with the copper transporter ATP7B in the trans-Golgi network (TGN) of hepatocytes (19)
.
To characterize its cellular function, we investigated the endogenous distribution of ClC-4 in murine tissues (mClC-4). Expression was high in excitable tissues such as the nervous system and skeletal muscle. The subcellular location of hClC-4, when heterologously expressed in HEK293 cells and in skeletal muscle fibers transfected in vivo, was shown to be the endoplasmic/sarcoplasmic reticulum (ER/SR). This location was confirmed by identification of endogenous ClC-4 in ER-enriched membrane fraction from mouse brain homogenate. Generation of chimeras and mutations in hClC-4 identified a novel N-terminal sequence both necessary and sufficient for its ER localization. The identified segment is a generalizable signal for ER retention or retrieval since it is able to redirect the vesicular channel ClC-3 and the unrelated plasma membrane protein Ly49E to the ER.
MATERIALS AND METHODS
Cell culture and cell transfection
HEK293 (human embryonic kidney cells) cells were maintained at 37°C in a 5% CO2 vapor-saturated incubator in Dulbecco’s modified Eagle medium (DMEM; GIBCO, UK) supplemented with 10% FBS. HEK293 cells were transiently transfected with the appropriate plasmids by calcium phosphate DNA precipitation (CalPhosTM Mammalian Transfection Kit, BD Biosciences Clontech, Palo Alto, CA, USA). In brief, 12 to 24 h before transfection cells were plated onto poly-L-lysine coated glass coverslips, placed on 35 mm well, and transfected at 
50% confluency with 1 µg of DNA per well. After 8 h incubation at 37°C, the DNA precipitate was removed from the cells and the medium was replaced. Cells were fixed and processed for immunocytochemistry 36 to 48 h post-transfection. Permanently transfected clones were generated by electroporation with linearized plasmids as described previously (16)
. Clones were selected for resistance to 800 µg/ml of G418 and by fluorescence-activated cell sorting. Clones were maintained in DMEM, supplemented with 10%FBS and 500 µg/ml G418.
Electropulse injection into mouse skeletal muscle
In vivo transfection of murine skeletal muscle fibers was based on the method of McMahon et al. (20)
. In brief, 6-wk-old C57BL/10 anesthetized male mice were injected in the anterior tibial muscle with 25 µl ClC-4-GFP plasmid DNA (1 µg/µl), 2 h after pretreatment of the same muscle with 10 U of bovine hyaluronidase (H-4272; Sigma). An electrical field was then applied to the muscle. The ClC-4-GFP injected muscles were dissected from sacrificed animals 2–5 days after the plasmid electrotransfer, embedded in Cryo-M-Bed (Bright, Huntingdon, UK), and 10 µm sections were obtained on a cryostat. The expression of hClC-4-GFP was ascertained by confocal microscopy.
Immunocytochemistry and confocal imaging
Cells grown on poly-L-lysine (Sigma, St. Louis, MO, USA) -coated glass coverslips were fixed with a solution of 4% formaldehyde in 50% PBS containing 4% sucrose and 2.5 mM CaCl2 for 20 min at room temperature. After quenching the reactive aldehyde groups with 50 mM NH4Cl in PBS for 10 min, cells were permeabilized by exposure to 0.1% TritonX-100 in PBS for 4 min and treated with 0.2% fish skin gelatin (Sigma) in PBS for 2 h to reduce nonspecific staining. Cells were then incubated with primary antibody (Ab) at 4°C overnight, followed by incubation with appropriate secondary Ab at room temperature for 30 min. Coverslips were mounted in Vectashield (Vector Burlingame, CA, USA). The following antibodies were used: rabbit polyclonal antibody (pAb) anti-ryanodine receptor Ab (sc-13942, Santa Cruz Technology, Santa Cruz, CA, USA); mouse monoclonal anti-SERCA2 ATPase Ab (MA3–910, Affinity Bioreagents); goat polyclonal anti-SERCA2 Ab (sc-8094 Santa Cruz); mouse monoclonal anti-Na+/K+ATPase 
-1 Ab (clone C464.6, Upstate Biotechnology, Lake Placid, NY, USA); mouse monoclonal anti-EEA-1 Ab (PA1–063, Affinity BioReagents, Neshanic Station, NJ, USA). Primary antibodies were detected using goat anti-mouse, goat anti-rabbit, and donkey anti-goat antibodies conjugated with Alexa568 (Molecular Probes, Leiden, the Netherlands). Nuclei were visualized by staining with propidium iodide (PI), diamidinophenyl-indole (4',6'-diam idino-2-phenylidole [DAPI]), or TOTO-3 (Molecular Probes). Nuclear staining was generally performed after Ab staining. For PI staining, cells were fixed and permeabilized as described above, then incubated sequentially with RNase (100 µg/ml) for 30 min at 37°C to and PI (40 µg/ml) for 30 min at 37°C. For DAPI and TOTO-3 nuclear stains, cells were washed briefly in 0.05% Tween20/PBS and incubated with DAPI reagent (300 nM in PBS) or TOTO-3 reagent (0.5 µM in PBS) for 30 min at room temperature. The actin-myosin skeletal muscle fibers were labeled with Texas red-X phalloidin (Molecular Probes). ER in live cells was labeled with ER-tracker® Blue-White DPX (Molecular Probes), a membrane-permeable fluorescent dapoxyl dye that specifically labels ER membrane lipids. In brief, cells were incubated with 500 nM ER-tracker® in culture medium for 30 min at 37°C. After washing with PBS, cells were fixed in formaldehyde, as described above, for 20 min at room temperature, mounted in Vectashield, and visualized by the UV laser line. Images were acquired using a Leica SP confocal microscope mounted with a PlanApochromat 63 x 1.4NA oil immersion objective. The GFP (green fluorescent protein) fluorophore was excited with the 488 nm line of an Argon laser; Alexa 568, PI, and Texas red-X phalloidin were excited with the 568 nm line of a Krypton laser and TOTO-3 with a 633 nm line of a diode laser. The emitted signals were detected through a 488/568/633 dichroic mirror using emission filters with a band width of 498–538 nm, 574–623 nm, or 646–728 nm, respectively. DAPI and ER tracker Blue-White DPX were excited with a 351 line of an UV laser and the emitted fluorescence collected through an emission filter of 400–493 band width. If the preparation was stained with different fluorophores, to prevent overlap of the emission signals in the acquisition channels, image acquisition was performed sequentially, acquiring the UV image last to minimize bleaching of the other fluorescent labels. Stacks of confocal sections separated by 0.5 µm increments were taken and images analyzed with Metamorph 5.0v1 software (Universal Imaging, Downington, PA, USA). Figures were assembled for publication with Adobe Photoshop 6.0 software (Adobe Systems, San Jose, CA, USA).
Measurement of cell surface of Ly49E by flow cytometry
HEK293 cells transiently transfected by calcium phosphate DNA precipitation (see above) with plasmids coding for Ly49E or a chimera of Ly49E with ClC-4, were harvested by exposure to 2 mM EDTA in PBS and suspended in HEPES-buffered, phenol red-free DMEM (Sigma) containing 0.2% fish skin gelatin. Cells were stained for cell surface Ly49E by 2 h incubation on ice with culture supernatant containing a rat antiLy49E monoclonal antibody (mAb), CM4, directed against an external epitope of the protein (21)
, followed by incubation for 45 min on ice with a 1:20 dilution of FITC-conjugated rabbit antirat Ab (DAKO, Carpinteria, CA, USA). Cells were analyzed by flow cytometry on a FACScalibur machine using CellQuest software (Becton Dickinson, Franklin Lakes, NJ, USA). Live cells were identified by PI exclusion. FITC excited with the 488 nm line of an Argon laser, and emitted fluorescence collected on the FL-1 channel (band pass filter 530±15 nm). Data were elaborated off-line by FlowJo software (Tree Star, Inc., Ashland, OR, USA). Frequency distributions of fluorescence intensity were compared by the Kolmogorov-Smirnov (K-S) algorithm; statistical significance was accepted for P 
0.05.
Western blot analysis
Tissues dissected from adult mice were snap-frozen in liquid nitrogen, pulverized and homogenized with an Ultra Turrax on ice in 1.5 ml of homogenization buffer containing 50 mM Tris-HCl pH7.4, 50 mM mannitol, 2 mM EDTA, and complete protease inhibitors (Roche, Nutley, NJ, USA). Cell debris and nuclei were removed by centrifugation at 500 g for 10 min. The supernatant was layered on 3.2 ml buffer containing 300 mM mannitol, 50 mM Tris-HCl pH7.4, 2 mM EDTA and centrifuged for 45 min at 40,000 rpm. The membrane pellet was resuspended in 200 µl of homogenization buffer containing 1% SDS. Samples were stored at –20°C until used. Samples from either from cell or tissue homogenates were treated with benzonuclease (Sigma; 5U/µl) for 30 min at room temperature, then dissolved in Laemmli sample buffer (4 x sample buffer: 200 mM Tris-HCl pH6.8, 40% glycerol, 20% ß-mercaptoethanol, 0.2% bromophenol blue, 8% SDS), heated for 5 min at 37°C, and loaded onto 8% SDS-PAGE gels. Unless indicated otherwise, 50 µg of protein sample was loaded per lane. Proteins were separated by electrophoresis for 1 h at 100–150V in a Bio-Rad system (Hercules, CA, USA), then transferred onto PVDF membranes (immobilion-P, Millipore, Billerica, MA, USA) by electroblotting at 10W for 1.5 h in a semi-dry blotting system (Anachem, Allen, TX, USA) using a transfer buffer containing 25 mM Tris-HCl (pH 8.3), 192 mM glycine, 20% methanol. The PVDF membrane was rinsed three times in PBS, and incubated in 100 ml blocking buffer (0.02% Tween 20, 4% dried skimmed milk in PBS) for 2 h at room temperature. For Western blot analysis, the membrane was incubated sequentially with primary Ab (5–10 ml of a 1:1000 dilution in blocking buffer) overnight at 4°C, then with an appropriate horseradish peroxidase (HRP) -conjugated secondary Ab (1:2000 dilution in blocking buffer) for 1 h at room temperature. Protein bands were detected by enhanced chemiluminescence (Amersham, Arlington Heights, IL, USA). For sequential analysis of multiple antigenic proteins, all antibodies were removed from the blot by a 30 min incubation in stripping buffer (0.01% ß-mercaptoethanol, 1% SDS, 10 mM Tris-HCl in dH20) at 65°C prior to reprobing the blot with the next primary Ab. Two different antibodies against ClC-4 were used. A rabbit pAb against human ClC-4 raised against the peptide LREKSRDTDRHRKITSKSKESI (HO2), corresponding to amino acids 36–57 of hClC-4 and an Ab against ClC-4 provided by C. E. Bear (Toronto, Canada) raised against the N-terminus of ClC-4 (17
, 18)
. Both anti-ClC-4 antibodies were assessed for specificity by Western blot analysis of HEK293 parental cell line and of the HEK293 cell lines permanently expressing hClC-3-GFP, hClC-5-GFP, and hClC-4-GFP. To raise rabbit polyclonal antibodies against the peptide HO2, the peptide was conjugated to keyhole limpet hemocyanin (KLH) via an extra-N-terminal cysteine. Antiserum against peptide HO2 was generated by injection into rabbits following a standard immunization protocol (Harlan SERA-LAB, Hillcrest, UK) and affinity-purified using N-hydroxysuccinimide-activated Sepharose affinity resin (Amersham) conjugated with HO2 peptide. Eluates were analyzed by SDS-PAGE and Coomassie brilliant blue staining and by ELISA. The Ab-containing fractions were then concentrated using spin columns (Millipore) with a cutoff size of 70 kDa. All purified antibodies were tested for specificity by cross-reactivity with an hClC-3 peptide (MTNGGSINSSTHLLD) in ELISA.
Preparation of ER vesicles by discontinuous sucrose gradient centrifugation
Subcellular membrane fractionation of homogenates from cultured cells or mouse brain was performed by ultracentrifugation on a discontinuous sucrose gradient using the following procedure. Cells were homogenized in a glass Dounce homogenizer in 3.5 ml homogenization medium. Cell debris, nuclei, and mitochondria were pelleted by centrifugation at 1000 g for 10 min (to remove cell debris and nuclei), then at 7700 g for 5 min (to remove mitochondria). A crude microsome fraction was obtained from the supernatant by ultracentrifugation for 1.5 h at 100,000 g in an SW41 rotor (Beckman, Fullerton, CA, USA): 3 ml supernatant was loaded onto a sucrose cushion of 2.8 ml 2M sucrose and 5.5 ml 0.35 M sucrose (both sucrose solutions in 1 mM EDTA, 5 mM Tris-HCl pH7.0). Crude microsomes were harvested at the 0.35–2 M interface. They were further purified by centrifugation on a sucrose gradient consisting of: 1 ml 45% sucrose, 1 ml 40% sucrose, 2 ml crude microsomes, 1 ml 35% sucrose, 1 ml 30% sucrose, 1 ml 25% sucrose, 1 ml 20% sucrose, overlaid with 5% sucrose (all sucrose solutions in 1 mM EDTA, 5 mM Tris-HCl pH7.0). After centrifugation overnight at 100,000 g in a Beckman SW41 rotor membrane, fractions were collected at the interfaces between the sucrose layers. All fractions were analyzed for protein content (Lowry assay) and refractive index, and stored at –80°C. Equal volumes of each fraction were analyzed by SDS-PAGE and immunoblotting using specific cellular marker proteins: Na+/K+ ATPase (detected by Mab clone C464.6, Upstate) for plasma membrane; SERCA2 (detected by the MAb MA3–910, Affinity Bioreagents) for ER; and an anti-GFP Ab (JL-8, BD Biosciences Clontech) to determine which fractions contained ClC-4-GFP. A similar protocol was followed for the analysis of microsomes derived from mouse brain homogenate. In this case a whole mouse brain snap-frozen in liquid nitrogen was pulverized, dissolved in 1.5 ml of homogenization buffer and homogenized with an Ultra Turrax. A crude microsome preparation was obtained as above described and subjected to ultracentrifugation on discontinuous sucrose gradient as previously specified. Microsome fractions were tested negative for 58K protein, lamp-1, and EEA-1 as markers for Golgi membranes, lysosomes, and endosomes, respectively.
Plasmids
The hClC-4-GFP/pCIneo mammalian expression plasmid, and the pNSNP1-Ly49E vector were kind gifts from Thomas Jentsch (Hamburg, Germany) and Colin Brooks (Newcastle, UK), respectively. All plasmids for expression in mammalian cells were generated on a pCIneo backbone (Promega, Madison, WI, USA). TOPO Blunt PCR4 (Invitrogen, Carlsbad, CA, USA) was used for rapid subcloning of polymerase chain reaction (PCR) fragments; their identity was confirmed by DNA sequencing using T3 or T7 sequencing primers (Invitrogen). Inserts on the pCIneo backbone were verified alternatively by sequencing with an oligonucleotide primer complementary to a stretch of bases located just 5' of the multiple cloning site (primer F0:ccacaggtgtccactcccag).
Generation of exchange chimeras
hClC-4 exchange chimeras were generated in a two-step overlap extension PCR reaction using overlapping oligonucleotide primers (22)
(see Table 1
for a list of all the primer sequences used in this study). All chimeric cDNA fragments were first subcloned in the TOPO pCR®4 Blunt cloning vector (Invitrogen), sequenced, and inserted into the pCIneo mammalian expression vector for expression in HEK293 cells. The chimeras ClC-4(551–760)xClC3-GFP, ClC-4(1–71)xClC3-GFP, and ClC-4(72-760)xClC3-GFP were generated. In addition, the chimera ClC-4(1–63)xLy49E-GFP was generated by replacing the cytoplasmic N-terminus of the type II plasma membrane protein Ly49E with the first 63 amino acids of hClC-4. Plasmids coding for full-length hClC-4-GFP, hClC-3-GFP, and Ly49E-GFP were transfected as controls.
|
Generation of hClC-4 truncations
In these chimeric proteins, part of the N-or C-terminal tail of hClC-4 was deleted by shifting the Kozak sequence and START codon (gccaccATG) or by introduction of an inframe XbaI restriction site (CTAGAC) at the truncation point. All truncations were fused at the C-terminus to enhanced green fluorescent protein (eGFP, BD Biosciences, Bedford, MA, USA). All hClC-4 truncation fragments were produced by PCR using the hClC4-GFP plasmid as the PCR template and the respective oligonucleotide primer pairs, and subcloned in the TOPO pCR®4Blunt. For expression in HEK293 cells the chimeric cDNA was excised from the TOPO vector by restriction enzyme digestion, gel-purified and inserted 5' of and inframe with the eGFP ORF on a pCIneo backbone (Promega).
Mutagenesis
The QuickChange Site-Directed Mutagenesis kit (Stratagene, San Diego, CA, USA) was used according to the manufacturer’s instructions, to generate point-mutations in miniprep plasmid DNA. The presence of the desired mutation, and the absence of any additional mutations, was verified by sequencing.
RESULTS
mClC-4 is present mainly in nervous and skeletal muscle tissues
The tissue distribution of mClC-4 was analyzed by Western blotting using membrane preparations from homogenates of adult murine tissues and a pAb raised against a KHL-conjugated antigenic peptide corresponding to amino acids 35–57 (amino acids 23–34 of mClC-4). The anti-ClC-4 Ab was demonstrated to be specific for ClC-4, and not to cross-react with ClC-3 or ClC-5 (Fig. 1
A). mClC-4 protein was predominantly present in excitable tissues such as the nervous system (brainstem, cerebellum, and cerebrum) and skeletal muscle (Fig. 1B
). This is consistent with Northern blot data (23
, 24
, 25)
showing high levels of ClC-4 mRNA in skeletal muscle and brain. However, in contrast to the Northern blot studies, mClC-4 was not detected in heart (Fig. 1B
). This tissue distribution for mClC-4 protein was confirmed by Western blotting using a different Ab against ClC-4 provided by C. E. Bear (17
, 18)
.
|
hClC-4-GFP expressed in HEK293 cells localizes in the endoplasmic reticulum
To identify the subcellular location of human ClC-4, clonal HEK293 cell lines permanently expressing hClC-4-GFP were generated. Confocal sections of two fields of HEK293 cells expressing hClC-4-GFP show that the GFP fluorescent signal is localized in a reticular network pervading the entire volume of the cell from the nucleus to the periphery, exhibiting continuity with the outer nuclear membrane and with poor or absent definition at the cell borders (Fig. 2
A). This is the typical pattern of ER proteins. To exclude the possibility that the GFP tag affects hClC-4 localization, the C-terminal GFP-tag was replaced with the much smaller C-terminal FLAG-tag (DYKDDDDK). hClC-4-Flag showed a similar subcellular location to hClC-4-GFP as shown by the colocalization of the two signals in a cell simultaneously transfected with hClC-4-Flag and hClC-4-GFP (Fig. 2C
). Membrane fractionation, obtained by discontinuous sucrose gradient centrifugation, provided further evidence of an ER location for hClC-4-GFP (Fig. 2B
). Only microsome fractions enriched for ER membranes, as evidenced by the presence of the sarco/endoplasmic reticulum calcium ATPase type 2 (SERCA2), contained hClC-4GFP. A plasma membrane marker, the -subunit of the Na+/K+ ATPase, was found predominately in different fractions from hClC-4.
|
To confirm an ER location, cells were costained for two specific ER markers, SERCA2 (Fig. 3
A) and the ER-tracker Blue-White DPX, a fluorescent probe that specifically binds to ER membrane lipids (Fig. 3B
). Colocalization of hClC-4-GFP with SERCA2 and ER-tracker was observed. No colocalization with EEA-1, a marker of early endosomes, was seen (Fig. 3C
).
|
hClC-4-GFP expressed in vivo in mouse skeletal muscle fibers localizes to the sarcoplasmic reticulum
It is possible that heterologous expression of hClC-4-GFP in HEK293 cells might induce mislocalization of the protein, since this cell line does not express endogenous ClC-4. Skeletal muscle fibers endogenously express high levels of ClC-4, as shown by our Western blot analysis. Preliminary data obtained in the murine myoblast cell line H2K showed that hClC-4-GFP localizes to the ER (data not shown). hClC-4-GFP was therefore expressed in vivo in mouse skeletal muscle fibers by plasmid electrotransfer into the anterior tibial muscle. Muscles were dissected 2–5 days after transfection and cryosections imaged by confocal microscopy. hClC-4-GFP fluorescence was in a precise reticular pattern throughout the muscle fibers (Fig. 4
A), consistent with ES/SR localization. The signal did not overlap with actin-myosin skeletal fibers visualized by labeling with Texas red-X phalloidin (Fig. 4B
). The hClC-4-GFP signal colocalized with two typical ER/SR resident proteins: SERCA2 (Fig. 4C
) and the ryanodine receptor (Fig. 4D
). The incomplete overlay of the GFP signal with the SERCA2 and ryanodine receptor was a consequence of the decreasing hClC-4-GFP expression radiating from the transfected nucleus present near the fiber surface. These data confirm that hClC-4 expressed in vivo in skeletal muscle fibers also localizes to the ER/SR.
|
Endogenous mClC-4 is identified in membrane fractions from mouse brain homogenate enriched for endoplasmic reticulum calcium ATPase
To investigate the intracellular location of endogenous ClC-4, and to obviate the lack of tissue staining with Ab against ClC-4, microsome membranes obtained from mouse brain homogenates were separated by discontinuous sucrose gradient centrifugation (Fig. 5
). ClC-4 signal peaked in fractions 15, 16 and 17; these fractions resulted particularly enriched as well as in SERCA2. Na+/K+ATPase, a marker of plasma membrane, although presents also in fractions 15, 16 and 17, showed a more consistent enrichment in fractions 12 and 13. We consider fractions 15, 16 and 17 as bona fide fractions enriched for endoplasmic reticulum membranes; the presence of Na+/K+ATPase in these fractions can be attributed to the to the fact that Na+/K+ATPase traffics to the plasma membrane through the endoplasmic reticulum as well as to the technical difficulty of a complete separation in distinct fractions of plasma membranes from endoplasmic reticulum membranes. We cannot exclude the presence of mClC-4 on the plasma membrane since fractions 11, 12, and 13, characterized by a strong Na+/K+ATPase signal, showed the presence of ClC-4 as well as a SERCA2 signal trough. In conclusion, these data suggest that endogenous mClC-4 is predominantly located in the ER of cells of the nervous system.
|
Identification of the ER targeting motif of hClC-4
ER residency of membrane proteins is frequently due to cytoplasmic double-lysine or double-arginine dibasic motifs (26)
. The amino acid sequence of hClC-4 has two cytoplasmic double-lysine motifs (KK739–740 and KSK52–54), in the C- and N-terminal tails, respectively, and seven double-arginine motifs RHR45–47, RR252–253, RRR358–360, RPRR601–604, RR651–652, RQR662–664 and RR695–696 of which the two double-arginine motifs RR252–253 and RRR358–360, in its transmembrane domains. However, with the exception of RQR662–664 and RR695–696, these motifs are also present in hClC-3 and hClC-5, which are not retained in the ER (16
, 27
, 28)
, (Fig. 6
). To determine whether any of these motifs are responsible for ER retention or to identify an alternative ER localization motif in hClC-4, three sets of modified GFP-tagged hClC-4 proteins were generated (Fig. 7
): A) exchange chimeras, in which amino acids of hClC-4 were exchanged with the corresponding segments of the highly homologous hClC-3, or the unrelated type II plasma membrane protein Ly49E; B) truncations, in which parts of the N- and/or C-terminal tail of hClC-4 were deleted; C) point mutations, in which single amino acid changes were introduced into a motif common to hClC-4 and the ER resident protein Sec63. The subcellular distribution of these altered hClC-4-GFP proteins following transient transfection into HEK293 cells was analyzed by confocal microscopy. Residence in the ER was evaluated by colocalization with a SERCA2.
|
|
Exchange chimeras with the highly homologous hClC-3 suggest that the N-terminus of hClC-4 is responsible for its ER localization
Although hClC-3 and hClC-4 share 78% amino acid sequence identity, they reside in different intracellular compartments. Transient transfection of HEK293 cells showed that hClC-4-GFP is localized in the ER (Fig. 8
A) while hClC-3-GFP is mainly localized in intracellular vesicles (Fig. 8B
), as previously reported (29
, 27
, 16)
. In contrast to hClC-4, no colocalization of hClC-3 was observed with SERCA2. This confirmed the results for permanently transfected cell lines described above. Chimeras of hClC-3-GFP and hClC-4-GFP were generated, based on the premise that they would reside in the ER if they contain the putative ER localization motif of hClC4 or in intracellular vesicles when the ER localization motif is replaced by the corresponding sequence for hClC-3. The following nomenclature has been used for the exchange chimeras: ClC-4(a-b)xP-GFP, where amino acids a to b of hClC-4 have been substituted to the corresponding region of protein P.
|
Since the C-terminus of hClC-4 contains five putative ER localization motifs (RPRR601–604, RR651–652, RQR662–664, RR695–696 and KK739–740), the exchange chimera ClC-4(551–760)xClC3-GFP in which the cytoplasmic C-terminus of hClC-3 was replaced with amino acids 551–760 of hClC-4, was generated. This chimera is localized in small intracellular vesicles, indistinguishable from hClC-3-GFP, and showed no colocalization with SERCA2 (Fig. 8C
). These data suggest that the putative ER retention motifs in the cytoplasmic C-terminus of hClC-4 are not responsible for ER localization, and any ER signal likely resides within the N-terminal 550 amino acids of hClC-4 (four putative ER motifs are identified in this region: RHR45–47, KSK52–54, RR252–253, and RRR358–360).
To study the role of the N-terminus of hClC-4, two chimeras were generated: ClC-4(1–71)xClC3-GFP in which the first 71 amino acids at the N-terminus of hClC-4 containing the putative ER localization motifs RHR45–47 and KSK52–54 replaced the cytoplasmic N-terminus of hClC-3; and the complementary chimera ClC-4(72-760)xClC3-GFP, including the first 71 amino acids of hClC-3. These mirror chimeras provide controls to exclude the possibility that ER targeting is a consequence of the retention of misfolded protein in the ER, as combining segments from different proteins to generate chimeras can increase the frequency of protein misfolding and induce ER retention by the protein quality control machinery (30)
. If the chimeras are not simply retained by the quality control apparatus of the ER, one chimeric protein of this mirror image pair should be localized in the ER, whereas the other should be targeted to a vesicular compartment. Indeed, ClC-4(72–760)xClC3-GFP expressed in HEK293 cells was localized in small intracellular vesicles (Fig. 8D
), as for hClC-3-GFP. In contrast, the subcellular localization of ClC-4(1–71)xClC3-GFP closely resembled that of hClC-4-GFP and colocalized with SERCA2, suggesting that the chimera resides in the ER (Fig. 8E
). In conclusion, the cytoplasmic N-terminus of ClC-4 appears to contain essential targeting information to direct hClC-4 and, when substituted to the N-terminus of hClC-3, hClC-3 to the ER. hClC-3 is targeted in neuronal and non-neuronal cells to a vesicular compartment by an adaptor protein AP-3-dependent mechanism (31)
. AP-3 proteins recognize dileucine-type motifs that are present throughout the sequence of both hClC-3 and hClC-4. The N-terminus of hClC-3 when substituted to the N-terminus of hClC-4 may direct hClC-4 to the vesicular compartment by removal of a dominant targeting signal presents in the ClC-4 N-terminus.
N-terminal truncations of hClC-4 localize to the ER
To ascertain whether the motif contained in the N-terminus of hClC-4 is sufficient for ER localization and does not require additional sorting signals in other segments of the protein, the hClC-4 truncation ClC-4(1–112)t-GFP was generated. This truncation consists of the first 112 amino acids of hClC-4 and includes the N-terminus and the first transmembrane domain. ClC-4(1–112)t-GFP showed a distribution comparable to hClC-4-GFP and largely overlapping with SERCA2 (Fig. 9
A).
|
The chimera ClC-4(1–63)t-GFP, consisting of the first 63 amino acids of hClC4, was also generated to ascertain whether the N-terminal transmembrane -helix of hClC-4 is necessary for the localization to the ER. Since this chimera does not contain a transmembrane -helix, it was predicted to be a soluble protein unless the cytoplasmic N-terminus of hClC-4 was able to interact per se with the ER retrieval/retention machinery. The ClC-4(1–63)t-GFP truncation chimera had an expression pattern identical to that of the cytoplasmic eGFP (data not shown).
Since the principal sequence differences between hClC-4 and murine ClC-4 (mClC-4) are within the cytoplasmic N-terminus of the protein, where mClC-4 is shorter by 13 amino acids, it seems unlikely that the retention motif of hClC-4 is within the first 13 amino acids. This was confirmed by ER localization of the human truncation chimera ClC-4(14–112)t-GFP that mimics a truncation of mClC-4 (Fig. 9B
).
We may argue that the ER location of the truncations described above is not simply due to their misfolding. Small polypeptide fragments of ClC-1 are able autonomously to properly fold and insert into membranes, since when expressed independently they reconstitute a full functional channel protein (32)
. Furthermore, it has been reported that the fluorescence of the C-terminal GFP tag of fusion proteins is a good indicator of proper folding of the upstream protein (33
, 34)
. Our truncations are conjugated at the C-terminus with a GFP tag, and their intense and readily detectable fluorescence would suggest their proper folding.
In conclusion, the cytosolic N-terminus of hClC-4 is necessary for ER retention although it requires the first transmembrane domain to direct hClC-4 to the ER. Since the chimera ClC-4(1–71)xClC3-GFP that contains only the N-terminus of hClC-4 is localized to the ER, it seems unlikely that the first transmembrane domain of hClC-4 contains an ER localization motif, but instead is simply required for initial sorting to the membrane compartment.
The N-terminus of hClC-4 directs the unrelated type II plasma membrane protein Ly49E to the ER
To confirm that the N-terminus of hClC-4 contains an ER targeting motif, this sequence was used to replace the N-terminus of Ly49E, a type II plasma membrane protein. Although the T-lymphocyte surface receptors CD4 and CD8 are regularly used as plasma membrane reporter proteins to identify targeting motifs in ER membrane proteins (35
, 36)
, these are both type I plasma membrane proteins with a single transmembrane -helix and an extracellular N-terminus, and therefore are not suitable for exchange with hClC-4, which has a cytoplasmic N-terminus. The type II plasma membrane protein, Ly49E, found in murine natural killer cells was therefore used as reporter protein. Ly49E consists of a short (44 amino acids) cytoplasmic N-terminus, a single transmembrane -helix, and a large extracellular C-terminal region (37)
. When transfected into HEK293 cells, Ly49E-GFP was primarily observed at the periphery of the cell and in small intracellular vesicles close to the plasma membrane (Fig. 10
A). Replacing the N-terminus of Ly49E with the first 63 amino acids of hClC-4, generating the chimera ClC-4(1–63)xLy49E-GFP, resulted in a redistribution of the protein (Fig. 10B
). Although some fluorescence was still observed on the plasma membrane and in small intracellular vesicles, the majority of protein colocalized with SERCA2 in the ER. Quantification of surface expression of Ly49E by flow cytometry using an Ab recognizing an extracellular antigen of Ly49E (Fig. 10C
) showed that surface expression of ClC-4(1–63)xLy49E (red trace) was reduced in comparison with Ly49E (green trace) (P<0.001, K-S test). Thus, the first 63 N-terminal amino acids of hClC-4 are both necessary and sufficient to confer ER residency on the plasma membrane protein Ly49E and displace it from its normal physiological compartment.
|
Effect of point mutations in the N-terminus of hClC-4
The N-terminal 63 amino acids of hClC-4 were analyzed by the pattern discovery program PRATT (website: http://www.ebi.ac.uk/pratt/; ref. 38
), searching for motifs absent from hClC-3 and hClC-5 but common to hClC-4 and a number of unrelated ER resident membrane proteins such as SERCA, IP3-receptor, RyR, P450 isoforms, and Sec63. The sequence [DE]-[TS]-[DE]-R-[EH] was identified in hClC-4 (DT43DRH) as well as in the translocation pore component Sec63 (DSDRE and ESDRE), but not in hClC-3 (DRERH) or hClC-5 (DRDRH). A similar sequence was also found in Sec12 (DTETR), an ER protein required for the formation of transport vesicles from the ER, and ClC-6 (RSDRD), another CLC chloride channel reported to be localized to the ER (39)
. To evaluate the involvement of the DT43DRH motif in ER retention/retrieval of hClC-4, point mutations were introduced into ClC-4(1–122)t. The ClC-4(1–112)/T43R protein was generated, by mutating Thr43 to arginine, equivalent to the corresponding amino acid residue in both hClC-3 and hClC-5. RDTDRH is also a consensus sequence for phosphorylation by PKC, PKA, and casein kinase I (CK-I), and phosphorylation of the cytoplasmic N-terminus may influence the subcellular trafficking, as for the potassium channel KCNK3 (40)
and the lip35 major histocompatibility antigen class II-associated invariant chain (41)
. Thr43 therefore was also mutated to aspartic acid (ClC-4(1–112)/T43D) and alanine (ClC-4(1–112)/T43A), mimicking constitutively phosphorylated and nonphosphorylated motifs, respectively. Expression of ClC-4(1112)/T43R, ClC-4(1–112)/T43D, and ClC-4(1–112)/T43A resulted in a fluorescence distribution pattern comparable to that of hClC-4GFP and ClC-4(1–112)t-GFP, and colocalization with SERCA2 (Fig. 11
). Thus, the DT43DRH motif, despite being conserved in Sec63 and ClC-4 and absent from hClC-3 and hClC-5, is not responsible for ER localization of hClC-4.
|
DISCUSSION
A few studies of the function and subcellular location of ClC-4 have appeared recently. ClC-4 has been reported to colocalize with CFTR on the plasma membrane (17)
, with ClC-5 in endosomes where it might contribute to their acidification (18)
, or with the copper pump ATP7B in the TGN to facilitate transport of copper by generation of a countercurrent (19)
. In this study we show that ClC-4 is highly expressed in excitable tissues such as the nervous system and skeletal muscle and that it may reside mainly in their sarco/endoplasmic membranes. Indeed, heterologous transfection of hClC-4-GFP in skeletal muscle fibers showed that hClC-4 localized in sarco/endoplasmic reticulum membranes; and endogenous mClC-4 was identified predominantly in ER-enriched fractions of mouse brain homogenate. Furthermore, heterologous expression of hClC-4-GFP in HEK293 cells showed a location in the ER and allowed us to identify a stretch of amino acids, residues 14–63, at the N-terminus that is both necessary and sufficient for targeting hClC-4 and other membrane proteins to the ER. This segment does not contain known ER targeting motifs that are not shared by the non ER-resident proteins hClC-3 and hClC-5. We propose that residues 14–63 of hClC-4 constitute an extended structural motif able to retain hClC-4 and other membrane proteins to the ER.
The ER not only assists folding and delivery of proteins to the secretory pathway, but also retains and eliminates misfolded proteins (30)
. This dual role complicates the study of protein localization to the ER. To ascertain that ER localization of hClC-4 was not simply a consequence of misfolding, several controls were undertaken. ER localization was not due to overexpression since no difference was observed using either transient or permanent transfection of hClC-4-GFP expressing different protein levels. Furthermore, comparable levels of expression of hClC-3 and hClC-5 showed a vesicular localization (16)
. Similarly, ER localization was not due to the absence of appropriate targeting mechanisms in the recipient cells: transfection into skeletal muscle fibers, a tissue that naturally expresses ClC-4, showed ES/SR localization. Identification of endogenous ClC-4 in a fraction enriched in ER membranes derived from mouse brain confirmed that ClC-4 is predominantly located to the ER/SR. This study, however, did not rule out that a lesser amount of ClC-4 is possibly located in the plasma membrane.
The stretch of amino acids, residues 14 –63, in the N-terminal cytoplasmic tail of ClC-4 that is responsible for ER targeting was identified by a chimeric protein strategy by exchanging segments between ClC-4 and the highly homologous ClC-3 that does not reside in the ER but in a vesicular compartment. A similar approach has been utilized to identify targeting motives involved in differential localization of SERCA and the plasma membrane calcium ATPase (PMCA) (42
43
44)
. This approach obviates the criticism that ER location of ClC-4 is simply a consequence of its misfolding. No obvious motifs are present in this N-terminal segment, residues 14–63, which are absent from the equivalent segments of hClC-3 and hClC-5, suggesting that an extended structural motif within amino acids 14–63 may form the targeting motif responsible for localization of hClC-4 to the ER. There are precedents for extended structural motifs for targeting such as that at the C-terminus of hemolysin required for its export (45)
. Furthermore, the segment so identified is both necessary and sufficient for ER localization; when linked to an unrelated plasma membrane membrane protein as Ly49E, it is able to displace it from its natural location. We conclude that the segment identified must contain a generalazible signal for ER retention or retrieval.
Different locations of ClC-4 have been reported in intestinal cells, kidney, and hepatocytes (17
18
19)
. In these tissues, where ClC-4 is not the main CLC channel, interaction of ClC-4 with other CLC proteins, or with ancillary proteins such as barttin (that functions as a ß-subunit for ClC-Ka and ClC-Kb) (46)
, could mask the ER localization signal present on the N-terminus of ClC-4. This mechanism has been shown to regulate, for example, the ER exit of KATP channels: KATP channels are retained in the ER by a localization signal, and their exit from the ER is dependent on the interaction with the sulfonylurea receptor SUR1 that masks their localization signal from the cellular machinery that recognizes them (47)
. A recent paper shows heterodimerization of ClC-3, ClC-4, and ClC-5 when heterologously expressed in the same expression system (48)
. It remains to be determined whether in vivo ClC-4 is also able to form heterodimer and to what extent. To date the only evidence for in vivo heterodimerization of CLC proteins it is between ClC-4 and ClC-5 in kidney tissue (18)
, where ClC-5 is the main expressed intracellular CLC protein. We have shown that ClC-4 is mainly expressed in excitable tissues where ClC-3 is also abundantly expressed (16)
. However, colocalization and functional complementation of ClC-4 and ClC-3 is unlikely since the ClC-3 knock-out mouse shows an extensive degeneration of the nervous system and no up-regulation of ClC-4 (27)
.
What could be the physiological role of ClC-4 in the ER/SR membranes? Excitable cells such as neurons and skeletal muscle fibers mobilize calcium from the SR/ER to the cytosol during activity. The cellular homeostatic control restores the cytosol calcium concentration by extruding it from the cell or by pumping it back into the SR/ER compartments via SERCA activity. By providing an influx of chloride anions into and an efflux of protons from the SR/ER, ClC-4 might contribute to dissipate the electrical gradient generated by the SERCA activity and facilitate calcium reuptake into the SR/ER. The role of ClC-4 could be accompanied by a concomitant H+ and K+ efflux from the SR/ER since both conductances are known to be present in SR/ER membrane.
If ClC-4 plays such a role in the ion homeostasis of the SR/ER, a screening for mutations in ClC-4 could prove useful for the identification of the etiology of diseases of excitable tissues such as skeletal muscle diseases and epilepsy.
ACKNOWLEDGMENTS
We thank Thomas Jentsch (Hamburg University, Germany) for the hClC-4GFP/pCIneo mammalian expression plasmid; Julian Dyson (CSC, London, UK) and Colin Brooks (Newcastle University, UK) for the pNSNP1-Ly49E vector and the Ab CM4 against Ly49E; Christine Bear (University of Toronto, Canada) for the Ab against ClC-4; Jenny Morgan (CSC, London, UK) for the H2K myoblast cell line.
FOOTNOTES
1 Present address: Inositide Laboratory, The Babraham Institute, Babraham, Cambridge CB2 4AT, UK. 
2 Present address: Medizinische Klinik m. S., Gastroenterologie und Hepatologie, Charité, Campus Virchow-Klinikum, Augustenburger Platz 1, 13353 Berlin, Germany.
Received for publication February 8, 2006. Accepted for publication June 12, 2006.
REFERENCES
| ||||||||||||
