journa l homepage: www.e lsev ier .com/ locate /ceca
he self-association of two N-terminal interaction domains plays an importantole in the tetramerization of TRPC4
ascale K. Lepage, Marc P. Lussier, Francois-Olivier McDuff, Pierre Lavigne, Guylain Boulay ∗
epartment of Pharmacology, Université de Sherbrooke, Sherbrooke, Quebec, Canada J1H 5N4
r t i c l e i n f o
rticle history:eceived 27 August 2008eceived in revised form 30 October 2008ccepted 3 November 2008
a b s t r a c t
Transient receptor potential canonical (TRPC) channels function as cation channels. In a previous study,we identified the molecular determinants involved in promoting TRPC subunit assembly. In the presentstudy, we used size-exclusion chromatography assays to show that the N-terminus of TRPC4 can self-associate and form a tetramer in cellulo. We further showed that the N-terminus of TRPC4 self-associates
vailable online 12 December 2008
eywords:RPCetramerizationalcium signalinghannel
via the ankyrin repeat domain and the region downstream from the coiled-coil domain. GST pull-down,yeast two-hybrid, and circular dichroism approaches demonstrated that both domains can self-associate.These findings indicated that the self-association of two distinct domains in the N-terminus of TRPC4 isinvolved in the assembly of the tetrameric channel.
Ca2+ signaling is involved in many cellular functions, includingell growth, differentiation, contraction, and secretion. The eleva-ion of [Ca2+]i occurs via two distinct phases upon activation ofGq protein-coupled receptor or a tyrosine kinase receptor. The
rst phase is mediated through the activation of phospholipase, which catalyzes the production of inositol 1,4,5-trisphosphate,hich in turn mobilizes Ca2+ from the endoplasmic reticulum. The
econd phase involves Ca2+ entry from the extracellular milieu,hich maintains [Ca2+]i higher than the basal level [1,2]. The tran-
ient receptor potential canonical (TRPC) protein family is involvedn Ca2+ entry. It is composed of seven members (TRPC1–TRPC7) thatre Ca2+-permeable cation channels [3,4]. Based on sequence sim-
larity, TRPCs can be further subdivided into four groups. Groups 1nd 2 contain the TRPC1 and TRPC2 isoforms, respectively. Group 3ncludes TRPC3, TRPC6, and TRPC7, whereas group 4 is composed ofRPC4 and TRPC5. Group 4 TRPCs are most closely related to group
Abbreviations: AD, Assembly domain; [Ca2+]i, intracellular Ca2+ concentration;D, circular dichroism; GAL4-AD, GAL4-activating domain; GAL4-BD, GAL4-bindingomain; GST, glutathione S-transferase; HA, hemagglutinin antigen; PBS, phosphateuffered salt; TRP, transient receptor potential; TRPC, TRP canonical; TRPV, TRPanilloid.∗ Corresponding author at: Department of Pharmacology, Faculty of Medicine
1. Functional TRPC channels are presumably formed by homote-tramers and heterotetramers [5,6]. Studies on the arrangement ofTRPCs in various systems have shown that the TRPC family can besubdivided into two distinct functional subgroups. All the mem-bers of one functional subgroup can selectively assemble togetherto form a specific channel [7,8]. One functional subgroup is com-posed of TRPC1, TRPC4, and TRPC5, whereas the other is composedof TRPC3, TRPC6, and TRPC7. TRPC1 and TRPC3, which belong todifferent functional subgroups, can also assemble together [9–11].In addition, TRPC4/5 can heteromultimerize with TRPC3/6/7 whenTRPC1 is present. These heteromers are found exclusively in embry-onic brain tissue [12].
TRPCs consist of a transmembrane domain composed of sixsegments with a putative pore between the fifth and the sixthtransmembrane segments. The amino and carboxy terminal tailsare cytoplasmic and each contains a coiled-coil domain [5]. The N-terminus also contains an ankyrin repeat domain. We previouslyreported that two domains are responsible for the subunit assem-bly of TRPC channels. One is in the N-terminus while the secondencompasses the pore region and the C-terminal tail. Using GSTpull-down assays, we further showed that the N-terminus containstwo minimal interaction domains in the third and fourth ankyrinrepeats and in the region downstream from the coiled-coil domain
[13].
The purpose of the present study was to determine whether theN-terminus of TRPC4 could oligomerize. We showed that the N-terminus of TRPC4 can self-associate and form a tetrameric complexin cellulo. We further demonstrated that the two N-terminal inter-
ction domains could exclusively self-associate. This suggests thathe N-termini of TRPCs play an important role in the tetramerizationf TRPC channels.
. Experimental procedures
.1. Materials
Cell culture media, serum, Hepes, trypsin, Opti-MEM I, Lipofec-amine 2000, and Zero Blunt Topo PCR cloning kit were purchasedrom Invitrogen (Burlington, ON, Canada). ALLN (calpain inhibitor) was from Calbiochem (San Diego, CA, USA). Rabbit polyclonal and
ouse monoclonal anti-hemagglutinin (HA)-specific antibodiesere from Covance (Berkeley, CA, USA). Rabbit polyclonal anti-
RPC4 N-terminal peptide (Tyr5–Arg17) antibody was a gift fromutz Birnbaumer (NIEHS, NIH, Research Triangle Park, NC, USA).ouse anti-c-myc 9B11 monoclonal antibody was from Cell Signal-
ng Technology (Danvers, MA, USA). Peroxidase-conjugated donkeynti-rabbit antibodies, peroxidase-conjugated sheep anti-mousentibodies, protein A-sepharose CL-4B, glutathione-sepharose, andhe protein calibration standards (thyroglobulin, ferritin, andldolase) were from GE Healthcare (Baie d’Urfé, QC, Canada).he bovine serum albumin was from Roche (Laval, QC, Canada).he Matchmaker Two-Hybrid-System 3 and Galacton-StarTM
hemiluminescent substrate were from Clontech (Palo Alto, CA,SA). The TNT® coupled reticulocyte lysate system was fromromega (Madison, WI, USA). Western Lightning Chemilumines-ence Reagent Plus and 0.2 �m nitrocellulose membranes wererom PerkinElmer Life Sciences (Woodbridge, ON, Canada). Allrimers and oligonucleotides were from Integrated DNA Tech-ologies (Coralville, IA, USA). Restriction enzymes were fromew England Biolab (Pickering, ON, Canada). The Phusion High-idelity DNA polymerase was from Finnzymes (Espoo, Finland).he TRPC4 (287ARLKLAIKYRQKEFVAQP304) and TRPC6 peptides363SRLKLAIKYEVKKFVAHP380) were synthesized by our peptideynthesis core facility, purified by reverse phase high-pressureiquid chromatography, and characterized by mass spectrome-ry. Unless otherwise stated, all other reagents were from SigmaOakville, ON, Canada) or Laboratoire MAT (Quebec City, QC,anada).
.2. Molecular biology and generation of chimeras
Standard molecular biology techniques were used for DNA iso-ation, analysis and cloning [14,15]. Depending on their intendedse, cDNAs encoding for mouse TRPC4 (NP 058680) and mouseRPC6 (NP 038866) were cloned into the mammalian cell expres-ion vector pcDNA3.1 (Invitrogen), the yeast expression vectorsGADT7 and pGBKT7 (Clontech), or the bacterial expression vec-or pGEX-4T-1 (GE Healthcare). Amino acids 1–304 of TRPC4ere cloned in-frame with the HA-epitope of pcDNA3.0-HA
a generous gift from Dr. Jean-Luc Parent, Université de Sher-rooke) [16]. Diverse N-terminal fragments of TRPC4 and TRPC6M1-P304T4, E87-H172T4, D254-P304T4, M1-H172T4/D249-P380T6 and64-P380T6) were cloned in-frame to the activating (pGADT7) orinding (pGBKT7) domains of GAL4. cDNA encoding [AnkT4]T6M1-H172 TRPC4/D249-R930 TRPC6) [13] was used as template toenerate the chimeric proteins M1-H172T4/D249-P380T6. The T64-248T6/E173-P304T chimera was generated by overlapping PCR.
he Q3-S183T4 and Q3-T401T6 constructs in pGBKT7 have alreadyeen used in our lab [17,18]. All PCR products were cloned intoCRII-Blunt-TOPO and transformed into an Escherichia coli hosttrain according to the manufacturer’s instructions. All constructsere confirmed using the dideoxynucleotide termination method
19].
m 45 (2009) 251–259
2.3. Cell culture and transfection
HEK293T cells were maintained at subconfluence in Dulbecco’smodified Eagle’s medium supplemented with 10% fetal bovineserum, penicillin (50 U/ml), and streptomycin (50 �g/ml) at 37 ◦C ina humidified atmosphere containing 5% CO2. The cells were tran-siently transfected using Lipofectamine 2000 transfection reagent.Briefly, 6-well plates were treated with poly-l-lysine for 30 min,rinsed with phosphate-buffered saline (PBS) (137 mM NaCl, 3.5 mMKCl, 10 mM sodium phosphate buffer, pH 7.4), and dried. Plas-mid DNA (1.6 �g) diluted in 250 �l of Opti-MEM I was added toeach well before adding 3 �l of Lipofectamine 2000 diluted in250 �l of Opti-MEM I. The mixture was incubated for 20 min atroom temperature. HEK293T cells (8.0 × 105) diluted in 1.5 ml ofculture medium without antibiotic were then added to the DNA-Lipofectamine 2000 complex and incubated for 16 h at 37 ◦C in ahumidified atmosphere containing 5% CO2. The medium was thenreplaced by complete culture medium and the cells were incubatedfor a further 24 h.
2.4. Immunoprecipitation assay
Transfected cells were treated for 3.5 h at 37 ◦C with a pro-teasome inhibitor (ALLN, 30 �M) and rinsed twice with PBSsupplemented with 0.9 mM CaCl2 and 1 mM MgCl2. The cellswere then incubated for 30 min at 4 ◦C with 0.5 ml of immuno-precipitation lysis buffer (150 mM NaCl, 1% Nonidet P-40, 0.5%deoxycholate, 5 mM EDTA, 20 mM Tris–HCl, pH 8.0, 1 �g/ml ofsoybean trypsin inhibitor, 5 �g/ml of leupeptin, and 100 �Mphenylmethylsulfonyl fluoride) with gentle agitation, followed byfive passages through a 25-gauge needle. The solubilized mate-rial was cleared by centrifugation for 15 min at 13,000 rpm at4 ◦C. The supernatant was mixed with 1.6 �l of anti-HA anti-body and 30 �l of a slurry of protein A-sepharose CL-4B and wasincubated overnight at 4 ◦C. Before use, the protein A-sepharoseCL-4B was washed three times with immunoprecipitation lysisbuffer. Samples were centrifuged for 3 min at 5000 rpm at 4 ◦Cand then washed three times with 1 ml of ice-cold immunopre-cipitation lysis buffer. Immunoprecipitated proteins were dissolvedin 20 �l of 2× Laemmli buffer and incubated for 30 min at60 ◦C. They were then separated on an SDS-polyacrylamide geland transferred to a nitrocellulose membrane for immunoblot-ting.
2.5. GST fusion proteins and GST pull-down assays
Amino acids 87-172 of TRPC4 were cloned in-frame into thepGEX-4T-1 plasmid to express GST fusion proteins in E. coli BL21.The expression of GST fusion proteins was induced with 0.2 mMisopropyl-d-thiogalactoside for 2 h at 30 ◦C. The cells were subse-quently collected by centrifugation at 2500 × g for 15 min, sonicatedin GST lysis buffer (20 mM Tris–HCl, pH 7.5, 1.0% Triton X-100,100 mM NaCl, 5 mM EDTA, 1 mM dithiothreitol, 1 �g/ml of soybeantrypsin inhibitor, 100 �M phenylmethylsulfonyl fluoride) on ice,and clarified by centrifugation for 15 min at 15,000 × g at 4 ◦C. Theclarified lysate was incubated with glutathione-sepharose beadsfor 1 h at room temperature (RT) and then washed six times withice-cold GST lysis buffer.
The N-terminal fragments of TRPC4 (E87-H172 and Q3-S183) thathad been subcloned in pGBKT7, which contains c-myc epitope tag,were translated using the TNT® coupled reticulocyte lysate system.
For the GST pull-down experiments, the N-terminal fragments ofTRPC4 were incubated for 1 h at RT with sepharose bead-bound GSTfusion proteins. After washing with binding buffer (20 mM Tris–HCl,pH 7.5, 1% Nonidet P-40, 100 mM NaCl, 5 mM EDTA, 1 �g/ml of soy-bean trypsin inhibitor, 100 �M phenylmethylsulfonyl fluoride), the
eads were resuspended in 2× Laemmli buffer. The proteins wereeparated on a 16.5% Tris-tricine polyacrylamide gel and transferredo a nitrocellulose membrane. They were then detected by stain-ng with Ponceau S or by immunoblotting with anti-c-myc epitope
ouse monoclonal antibodies.
.6. Immunoblots
Cell lysate and immunoprecipitated proteins were separatedy SDS-PAGE and transferred to a 0.2 �m nitrocellulose mem-rane (PerkinElmer) in 150 mM glycine, 20 mM Tris-base, and0% methanol (350 mA for 3 h or 100 mA overnight at 4 ◦C). Thelots were stained with Ponceau S to visualize the marker pro-eins, destained with TBST (20 mM Tris–HCl, pH 7.5, 137 mMaCl, 0.3% Tween 20), and blocked for either 1 h at RT orvernight at 4 ◦C in TBST containing 7% (w/v) nonfat dry milk.he blots were then incubated for 3 h at RT or overnight at 4 ◦Cith rabbit anti-HA (dilution 1:1000), mouse anti-HA (dilution
:1000), rabbit anti-TRPC4 (dilution 1:2000), or mouse anti-c-yc (dilution 1:1000). After three washes with TBST, the blotsere incubated with peroxidase-conjugated donkey anti-rabbit-
gG (1:50,000) or peroxidase-conjugated sheep anti-mouse-IgG1:25,000) for 1 h at RT in TBST. The blots were washed threeimes with TBST and the immune complexes were visualized usinghe Western Lightning Chemiluminescence Reagent Plus detectionystem.
.7. Yeast two-hybrid and ˇ-galactosidase assay
Protein interactions in the two-hybrid system can be moni-ored both qualitatively and quantitatively. cDNAs encoding aminocids 1–304, 3–183, and 254–304 of TRPC4 or amino acids 3–401f TRPC6 were used as bait. They were cloned in-frame with theAL4 DNA-binding domain in pGBKT7-BD and transformed intoeast strain AH109. cDNAs encoding amino acids 1–304, 87–172,nd 254–304 of TRPC4, amino acids 64–380 of TRPC6, and the M1-172T4/D249-P380T6 and T64-H248T6/E173-P304T4, chimeras weresed as prey. They were cloned in-frame with the GAL4 DNA-ctivating domain in pGADT7-AD and transformed into yeast strain187. Yeast strain AH109 expressing the bait was then mated witheast strain Y187 expressing the prey, according to the manufac-urer’s instructions. Mated yeast cells were first grown on lowtringency selection plates (-Leu, -Trp) and then on high strin-ency selection plates (-Leu, -Trp, -His, -Ade) in the presence of-�-gal.
A liquid culture �-galactosidase assay was used to quantify thenteractions according to the method described in the Clontecheast Protocols handbook using the Galacton-StarTM chemilumi-escent substrate (Clontech). Briefly, the diploid cells were grownvernight at 30 ◦C in 5 ml of low stringency medium with agi-ation (250 rpm). The overnight culture (1 ml) was transferred toPDA medium (8 ml) and grown at 30 ◦C with agitation (250 rpm)o an optical density (OD600) between 0.4 and 0.6. The cells werearvested by centrifugation at 13,000 rpm for 30 s and resus-ended in 1.5 ml of Z buffer (60 mM Na2HPO4, 40 mM NaH2PO4,H 7, 10 mM KCl, 1 mM MgSO4). They were harvested by cen-rifugation, resuspended in 300 �l of Z buffer, and subjected tohree freeze-thaw cycles in liquid nitrogen. The cell lysate (25 �l)as incubated with Galacton-StarTM reaction mixture (200 �l) at
T for 60 min and then centrifuged at 13,000 rpm for 1 min at◦C. The supernatants were transferred to luminometer tubes and
ight emission was recorded at 5-s intervals using a Mini LumatB 9506 luminometer (EG&G Berthold, Bad Wildbad, Germany).-Galactosidase activity was normalized to the correspondingD600.
m 45 (2009) 251–259 253
2.8. Size-exclusion chromatography
Transfected cells were treated for 3.5 h at 37 ◦C with a pro-teasome inhibitor (ALLN, 30 �M) and rinsed twice with PBSsupplemented with 0.9 mM CaCl2 and 1 mM MgCl2. They were thenscraped into 750 �l of PBS supplemented with 0.9 mM CaCl2, 1 mMMgCl2, 1 �g/ml of soybean trypsin inhibitor, 5 �g/ml of leupeptin,and 100 �M phenylmethylsulfonyl fluoride, and homogenized in a1.5 ml tube using a pellet pestle motor homogenizer (Kontes). Thehomogenate was centrifuged at 13,000 rpm for 30 min at 4 ◦C andthe supernatant containing the cytosolic proteins was collected,concentrated using an Amicon Ultra-4 10K filter (Millipore), andseparated by size-exclusion chromatography on a Superdex 200(10/300) (GE Healthcare) column, mounted on an ÄKTAprimeTM
(GE Healtcare) at RT. Samples were eluted (0.5 ml/min) and the elu-tion profile was monitored at 280 nm. The proteins in the fractionswere separated on 12% SDS-PAGE gels and detected by immunoblot-ting with anti-HA mouse monoclonal antibodies or by stainingwith Coomassie blue. Band densities were quantified using a meanintensity value within a specific area, and the background from aneighboring empty location was subtracted. The same surface areawas used for all the quantification calculations.
2.9. Circular dichroism spectroscopy measurements
Circular dichroism (CD) measurements were performed in0.1-cm path length quartz cuvettes using a Jasco J-810 spectropo-larimeter (JASCO Inc.) equipped with a Peltier-type thermostat.Spectra were recorded at 37 ◦C from 300 to 190 nm. The averageof three scans at 0.1 nm intervals is presented. All the CD spec-tra were corrected by subtracting the spectrum obtained withbuffer alone. Proteins were diluted to the desired concentrationsin PBS supplemented with 0.9 mM CaCl2 and 1 mM MgCl2. Theraw m◦ values were transformed into mean residue molar ellip-ticities (deg cm2 dmol−1) using the following equation: [�]222 = CDsignal (deg) × MRW/concentration (g/ml) × l × 10, where MRW isthe mean residue weight and l is the path length of the CD cell. CDmeasurements were obtained in three independent experimentsfor each protein.
3. Results
3.1. TRPC4 N-termini tetramerize
A previous study showed that the N-terminus of TRPC4 pos-sesses several molecular determinants that are involved in theassembly of the tetrameric channel [13]. To determine the roleof the N-terminus of TRPC4 in the oligomerization of the chan-nel, we first wanted to confirm that it interacts with intact TRPC4.HEK293T cells were co-transfected with HA-tagged TRPC4 and theN-terminal fragment of TRPC4 (M1-P304T4) to confirm that it inter-acts with intact TRPC4. HA-tagged TRPC4 was immunoprecipitatedusing an anti-HA antibody, and the presence of the M1-P304T4 frag-ment in the immunoprecipitate was assessed by immunoblottingwith a specific antibody directed against the N-terminus of TRPC4.As shown in Fig. 1A, upper panel, we detected the M1-P304T4 frag-ment, migrating with a predicted molecular mass of 35 kDa, in theimmunoprecipitate of the HA-tagged TRPC4. The lower panel ofFig. 1A shows that the HA-tagged TRPC4 migrated with a molecularmass of 113 kDa. These results confirmed our previous observationthat the N-terminus of TRPC4 contributes to the assembly of the
tetrameric channel. We then looked at whether the N-terminusof TRPC4 could self-associate. HEK293T cells were co-transfectedwith the HA-tagged M1-P304T4 fragment and the untagged M1-P304T4 fragment. Following an immunoprecipitation step using ananti-HA antibody, the presence of the M1-P304T4 fragment in the
254 P.K. Lepage et al. / Cell Calciu
Fig. 1. TRPC4 M1-P304T4 fragment self-associates in HEK293T cells. (A) cDNAs cod-ing for amino acids M1-P304 of TRPC4 (1–304T4), full-length HA-tagged TRPC4(T4HA) or the combination of both were transiently transfected into HEK293T cells.The TRPC4 proteins were immunoprecipitated with a mouse monoclonal anti-HAantibody. The immunoprecipitated complexes were separated by SDS-PAGE andanalyzed by immunoblotting with an anti-TRPC4 (upper panel) or with a rabbit poly-clonal anti-HA antibodies (lower panel). (B) cDNAs coding for amino acids 1–304 ofTRPC4 (1–304T4) and HA-tagged 1–304 of TRPC4 (HA1–304), alone or in combi-n 1 304
pTi
iciwcnMPdd
cco
ation, were transiently transfected in HEK293T cells. HA-tagged TRPC4 M -Proteins were immunoprecipitated with a mouse monoclonal anti-HA antibodies.he immunoprecipitated complexes were separated by SDS-PAGE and analyzed bymmunoblotting with a rabbit polyclonal anti-TRPC4 antibody.
mmunoprecipitate was assessed by immunoblotting using a spe-ific antibody directed against the N-terminus of TRPC4. As shownn Fig. 1B, we detected the untagged M1-P304T4 fragment, migrating
ith a predicted molecular mass of 35 kDa, in the immunopre-ipitate of the HA-tagged M1-P304T4 fragment. It is important toote that, due to the presence of the HA-epitope, the HA-tagged1-P304T4 fragment migrated slightly above the untagged M1-
304T4 fragment, with a molecular mass of 37 kDa. These resultsemonstrate that the M1-P304T4 fragment possesses the essential
eterminants to self-associate.
The HA-tagged M1-P304T4 fragment was expressed in HEK293Tells and the cytosolic proteins were separated by size-exclusionhromatography as described in Section 2 to determine theligomerization state of the M1-P304T4 fragment. The proteins
m 45 (2009) 251–259
in the fractions were detected by immunoblotting using ananti-HA antibody (Fig. 2A), which was analyzed by densitom-etry (Fig. 2B), and by Coomassie blue staining (Fig. 2C). Wedetected the HA-tagged M1-P304T4 fragment in the fractions 15–35.The densitometric analysis revealed four peaks corresponding to44.7 ± 2.8 kDa, 76.2 ± 2.3 kDa, 109.1 ± 6.8 kDa, and 161.2 ± 2.6 kDaproteins (Fig. 2B), which is consistent with monomeric, dimeric,trimeric and tetrameric protein complexes, respectively (Fig. 2B).These results suggested that the N-terminus of TRPC4 can tetramer-ize in cellulo.
3.2. Importance of each interaction domain within TRPC’sN-terminus
In a previous study [13], we identified two regions in theN-terminus of TRPC4 involved in channel assembly using a GST-pulldown assay. The first region involves the third and fourthankyrin repeats (E87-H172) while the second region is downstreamfrom the coiled-coil domain (D254-P304). To further investigate theimportance of each of these regions in the assembly of TRPC4,we generated two chimeras by swapping similar regions betweenthe N-termini of TRPC4 and TRPC6 as illustrated in Fig. 3A. It isimportant to note that TRPC4 and TRPC6 belong to different func-tional subgroups of TRPCs and are thus not expected to interact.The interaction of the chimeras with the N-terminus of TRPC4(Gal4 BD-fused M1-P304T4 fragment) using the yeast two-hybridapproach was quantified by measuring the �-galactosidase activ-ity in diploid yeast cells. As shown in Fig. 3B, the �-galactosidaseactivity of diploid yeast cells expressing the Gal4 BD-fused M1-P304T4 and the Gal4 AD-fused M1-H172T4/D249-P380T6 chimera was65 ± 5% of the �-galactosidase activity of diploid yeast cells express-ing the Gal4 BD-fused M1-P304T4 and the Gal4 AD-fused M1-P304T4.The �-galactosidase activity of diploid yeast cells expressing theGal4 BD-fused M1-P304T4 and the Gal4 AD-fused T64-H248T6/E173-P304T4 chimera was only 32 ± 10% of the best combination. Thisvalue was similar to the �-galactosidase activity of diploid yeastcells expressing the Gal4 BD-fused M1-P304T4 and the Gal4 AD-fused T64-P380T6, which are not expected to interact in mammaliancells. This value was thus considered to represent a non-selectiveassociation in the yeast two-hybrid assay. The same experimen-tal approach was used to study the interaction between thesechimeras with the N-terminus of TRPC6 (Gal4 BD-fused Q3-T401T6).The �-galactosidase activity of diploid yeast cells expressing theGal4 BD-fused Q3-T401T6 and the Gal4 AD-fused T64-H248T6/E173-P304T4 chimeras was 110 ± 38% of the �-galactosidase activityof diploid yeast cells expressing the Gal4 BD-fused Q3-T401T6and the Gal4 AD-fused T64-P380T6 chimeras (Fig. 3C). The �-galactosidase activity of diploid yeast cells expressing the Gal4BD-fused Q3-T401T6 and the Gal4 AD-fused M1-H172T4/D249-P380T6chimeras was 84 ± 4% of the best combination. These results con-firmed that both ankyrin repeats and coiled-coil regions wererequired for stable interactions between the N-terminus interac-tion domains of TRPC4. However, these results also indicated that,in the case of TRPC6, the presence of either the ankyrin repeatsor the coiled-coil region was sufficient to ensure a stable interac-tion.
3.3. Self-association of the third and fourth ankyrin repeatsdomain and the region downstream from the coiled-coil domainof TPRC4
To determine whether these regions could self-associate orinteract, we performed GST pull-down and yeast two-hybrid assays.Fig. 4A shows the various TRPC4 N-terminus fragments used in theGST pull-down and yeast two-hybrid assays and indicates the loca-tion of the various domains. To determine whether the E87-H172T4
Fig. 2. TRPC4 M1-P304 fragments oligomerize as a tetramer. HA-tagged TRPC4 M1-P304 cDNA was transiently transfected into HEK293T cells. The cells were lysed into PBSw parat( am). Ti ed by
fStgAwrtv
wafaAPdfA
ith a mechanical homogenizer and the membrane and cytosolic proteins were seMillipore) and separated by size-exclusion chromatography (Superdex 200, Amershmmunoblotted using anti-HA mouse monoclonal antibodies (A), which was analyz
ragment self-associates, we incubated the in vitro translated Q3-183T4, which includes the first to fourth ankyrin repeats, andhe in vitro translated E87-H172T4 with E87-H172T4 immobilized onlutathione-Sepharose 4B affinity beads (GST-fused E87-H172T4).s shown in Fig. 4B and C, no significant binding was observedith GST alone, while E87-H172T4 and Q3-S183T4 were efficiently
etained by GST-fused E87-H172T4. These results demonstrated thathe third and fourth ankyrin repeats of TRPC4 can self-associate initro.
We also performed a yeast two-hybrid assay to verify in vivohether both N-terminus interaction domains could self-associate
nd/or interact. M1-P304T4, E87-H172T4, and D254-P304T4 wereused to the Gal4-activating domain and transformed individu-lly into Y187. The transformed Y187 yeast was mated with the
H109 yeast strain expressing M1-P304T4, Q3-S183T4, or D254-304T4 fused to the Gal4-binding domain. As shown in Table 1,iploid yeast cells expressing the N-terminal fragment M1-P304T4used to Gal4-AD and Gal4-BD, the ankyrin repeats fused to Gal4-D and Gal4-BD, and the region downstream from the coiled-coil
ed by centrifugation. Cytosolic proteins were concentrated using an Amicon filterhe proteins from the different fractions obtained were separated by a 12% SDS-PAGE,densitometry (B), or stained with Coomassie blue (C).
domain fused to Gal4-AD and Gal4-BD were able to grow on highstringency medium. Diploid yeast cells expressing the region down-stream from the coiled-coil domain and the ankyrin repeats wereunable to grow on high stringency medium. We performed a �-galactosidase assay to quantify the stability of the interactions.The �-galactosidase activity of diploid yeast cells expressing theankyrin repeats fused to Gal4-AD and Gal4-BD was 63.5 ± 14.1% thatof diploid yeast cells expressing M1-P304T4 fused to Gal4-AD andGal4-BD. The �-galactosidase activity of diploid yeast cells express-ing the region downstream from the coiled-coil domain fused toGal4-AD and Gal4-BD was 164.9 ± 28.7% of that of diploid yeast cellsexpressing M1-P304T4 fused to Gal4-AD and Gal4-BD. These resultsshowed that the third and fourth ankyrin repeats self-associate,that the region downstream from the coiled-coil domain also self-
associates, and that the two domains do not interact with eachother.
To further confirm that the region downstream from thecoiled-coil domain self-associates, we synthesized the pep-tide 287ARLKLAIKYRQKEFVAQP304 of TRPC4 and the peptide
Fig. 3. Both interaction domains are required for stable interactions with the N-terminus of TRPC4 while either interaction domains of TRPC6 are sufficient to ensurea stable interaction. (A) Schematic representation of the chimeras used for the liq-uid culture �-galactosidase assay. The TRPC4 amino acid sequence is indicated by awhite box, and the TRPC6 amino acid sequence is indicated by a black box. (B and C)S. cerevisiae Y187 pretransformed with the GAL4-activating domain (pGADT7) con-structs was mated with S. cerevisiae AH109 pretransformed with the GAL4-bindingdomain (pGBKT7) constructs. The histogram illustrates the relative strength of two-hybrid interactions as determined by a chemiluminescent assay for lacZ reportergene expression. Liquid cultures were grown to log phase in SD/-Trp, -Leu mediumand assayed for �-galactosidase expression as described in Section 2. The resultsac(
3
rfTtopltatihatpam
Fig. 4. Self-association of the third and fourth ankyrin repeats domain of TRPC4 invitro. (A) Representation of the various TRPC4 N-terminus domains used in the GSTpull-down and in the yeast two-hybrid assays (Table 1). These constructions wereintroduced into the vectors pGADT7, pGBKT7 and/or pGEX. The fragments TRPC4E87-H172 (B) and TRPC4 Q3-S183 (C), translated in vitro, were incubated for 1 h at
ity ([�]) around 222 nm. This points out to a situation where themonomeric TRPC peptides are mostly unfolded when not oligomer-ized. On the other hand, if increasing the total concentration ofthe TRPC peptides leads to the formation of a quaternary structure(self-assembly), this should be accompanied by the induction of a
Table 1Self-association of the ankyrin repeats and the region downstream of the coiled-coildomain of TRPC4. S. cerevisiae AH109 pretransformed with the GAL4-binding domain(GAL4-BD) constructs were mated with the S. cerevisiae Y187 strain pretransformedwith the GAL4-activating domain (GAL4-AD) constructs. The mated clones that grewon a high stringency selective medium (-Leu, -Trp, -Ade, -His, +X-�-gal) at 30 ◦C areindicated by a positive sign and mated clones that did not grow on high stringencyselective medium are indicated by a negative sign. The �-galactosidase activity,which represents the relative strength of two-hybrid interactions, was determinedby a chemiluminescent assay for lacZ reporter gene expression as described in Sec-tion 2. The results (mean ± S.D.) are expressed as a percentage of the self-associationof M1-P304T4 and are obtained from three independent experiments.
Gal4-BD Gal4-AD Growth �-Galactosidase activity
re expressed as a percentage of the self-association of M1-P304T4 (A) or the asso-iation of T64-P380T6 with Q3-T401T6 (B). The histogram shows the average valuesmean ± S.D.) obtained from three independent experiments.
63SRLKLAIKYEVKKFVAHP380 of TRPC6. These two peptides rep-esent the last 18 amino acid residues of the region downstreamrom the coiled-coil domain, which is a very similar region betweenRPC4 and TRPC6. We hypothesized that the regions involved inhe channel assembly should be structurally very similar through-ut the TRPC family. To determine whether the TRPC peptidesossess the ability to self-associate, we also used CD with the fol-
owing proviso in mind. According to the law of mass action, ifhe peptides exist in equilibrium between their monomeric statend an oligomeric state, increasing the total concentration of pep-ides should tip the balance in favor of the oligomeric state and thenduction of a secondary structure. However, short peptides cannotave stable secondary or tertiary structures without being part ofstable quaternary structure. Fig. 5A shows the far-UV CD spec-
ra representative of three independent experiments of syntheticeptides corresponding to 287ARLKLAIKYRQKEFVAQP304 of TRPC4nd 363SRLKLAIKYEVKKFVAHP380 of TRPC6. These peptides wereostly unfolded (random coils) with a residual secondary structure
room temperature with TRPC4 E87-H172 fused to GST and adsorbed to glutathione-Sepharose beads. GST pull-down complexes were analyzed by 16.5% Tris-tricine gelelectrophoresis, stained with Ponceau S and immunoblotted with anti-c-myc mousemonoclonal antibodies.
(e.g., �-helix) as indicated by a typical negative molar elliptic-
Fig. 5. Self-association of A287-P304T4 or S363-P380T6 peptides. (A) Far-UV CDspectra of A287-P304TRPC4 peptide (closed circles) and S363-P380TRPC6 (open cir-cles) recorded at neutral pH and 37 ◦C. The concentration of the peptides was100 × 10−6 M. (B) Concentration dependence of the molar ellipticity at 222 nm(cit
sAtheo
4
bdc(btoifiaadttp
[�]222) of A287-P304TRPC4 peptide (closed circles) and S363-P380TRPC6 (open cir-les) at neutral pH and 37 ◦C. The molar ellipticity [�] was calculated as describedn Section 2. The spectra shown in (A) and the graphics in (B) are representative ofhree independent experiments.
econdary structure or a more negative molar ellipticity at 222 nm.s shown in Fig. 5B, the secondary structure content of both pep-
ides increased in parallel with their concentration. This result is theallmark of an auto-association process and supports our hypoth-sis that the TRPC peptides are part of the self-assembly domainsf the TRPC.
. Discussion
In a previous study, we identified two domains that are responsi-le for the oligomerization of TRPC channels [13]. The first assemblyomain (AD1) overlaps the N-terminal ankyrin repeats and theoiled-coil domain (M1-P304T4). The second assembly domainAD2) encompasses the putative pore region, including the linkeretween the fourth and fifth transmembrane segments, and the C-erminal tail (I516-L974T4). In the present study, we showed that AD1f TRPC4 forms tetramers. The present study also identified twonter-subunit interaction domains in the N-terminus of TRPC. Therst interaction domain was in the third and fourth ankyrin repeatsnd corresponded to residues E87-H172 of TRPC4. The second inter-
ction domain was downstream from the predicted coiled-coilomain and corresponded to residues D254-P304 of TRPC4. Thosewo interacting domains are also present in TRPC6. The implica-ion of the N-termini in the assembly of the channel was alsoresent in other channels structurally related to TRPC. As exam-
m 45 (2009) 251–259 257
ple, voltage-dependent potassium channels (Kv) [20–22], for whichtetramerization has been intensively studied. The crystal structuresof Shaker channels, which are members of the Kv channel fam-ily, have also been published [23,24]. An N-terminal cytoplasmicdomain (T1, ∼130 amino acids) recognizes members of the samefamily and prevents co-assembly with members of other Kv fami-lies. The structure of T1 contains four identical subunits arrangedin fourfold symmetry around a centrally located pore as well asseveral interaction domains between the four subunits [24,25]. TheN-termini of TRPV4, 5, and 6 play an essential role in self-assembly[26–28]. Mutations of conserved amino acids and deletion of apart of the ankyrin repeats abolish assembly and result in non-functional channels. When overexpressed in E. coli, the ankyrinrepeat domains of TRPV1, TRPV2, TRPV5, and TRPV6 do not assem-ble to form a tetramer and remain monomeric in solution [29–32].However, when the ankyrin repeat domain of TRPV6 in mammaliancells is overexpressed, it can oligomerize and has dominant nega-tive properties [33]. It is likely that the assembly of the ankyrinrepeat domains of the TRPV channels requires some cellular factorsor chaperones that are absent in E. coli [30]. However, further stud-ies are needed to identify these cellular factors or chaperones andtheir possible role in the TRPC assembly.
We demonstrated by yeast two-hybrid and circular dichroismthat the region downstream from the coiled-coil domain (D254-P304T4) can self-associate (Table 1 and Fig. 5B). However, we couldnot detect a specific interaction between M1-P304T4 and the T64-H248T6/E173-P304T4 chimera (Fig. 3B). The presence, in the chimera,of the ankyrin repeats domain of TRPC6 could explain these dis-crepant results. The main difference between the ankyrin repeatsdomain of TRPC4 and the ankyrin repeats domain of TRPC6 residesin the loop between the ankyrin repeats 3 and 4, which is longerin TRPC6. Thus, the ankyrin repeats domain of TRPC6 might causean unfavorable conformation or steric constraints that could desta-bilize the interaction between two coiled-coil domains of TRPC4.This difference might also explain the selectivity in the assemblybetween TRPCs.
A recent study by Schindl et al. [34] showed that the expres-sion of F59-S137 of TRPC4 and the equivalent region of TRPC5 havea dominant negative effect on TRPC4 and TRPC5 channel activity,respectively, by reducing the formation of the TRPC channel com-plex. Based on our sequence alignment for the ankyrin repeats ofTRPC [17], F59-S137TRPC4 is the second and third ankyrin repeat inTRPC4. The third ankyrin repeat is a major molecular determinantfor TRPC channel assembly. However, we cannot rule out the pos-sibility that other determinants in the second and fourth ankyrinrepeats are needed to support multimerization. The presence of twointeracting domains has also been suggested as being necessary forthe assembly of TRPV6 [28]. Erler et al. proposed a model in whichthe third ankyrin repeat initiates a molecular zippering process thatproceeds past the fifth ankyrin repeat and creates an intracellu-lar anchor that is necessary for functional subunit assembly. Sincewe showed that the N-terminus of TRPC4 tetramerizes through theself-association of the two interaction domains, the mechanism bywhich the interaction domains of TRPCs interact to form a tetramermight also involve a zippering process. In this model, each of theinteracting domains has the capacity to tetramerize (Fig. 6A). Theresults do not rule out the possibility of a second model, in which thefirst interaction domain is responsible for the interaction betweentwo subunits leading to the formation of a dimer, while the sec-ond interaction domain is responsible for the association of the twoother subunits with the dimer (Fig. 6B). However, further studies are
required to validate these models.
In summary, we showed that the N-terminus of TRPC4 forms atetramer via the self-association of two interaction domains in thethird and fourth ankyrin repeats and the region downstream froma predicted coiled-coil domain.
258 P.K. Lepage et al. / Cell Calciu
Fig. 6. Hypothetic models for the self-association of the two interaction domainswithin the TRPC’s N-termini. (A) In the first model, each interaction domain hasthe capacity to tetramerize. This model proposes that the third and fourth ankyrinrepeats domain initiate a molecular zippering process that proceeds past the regiondownstream of the coiled-coil domain. (B) In the second model, the first interac-tfa
A
oaefP
R
[
[
[
[
[
[
[
[
[
[28] I. Erler, D. Hirnet, U. Wissenbach, V. Flockerzi, B.A. Niemeyer, Ca2+-selective tran-
ion domain is responsible for the interaction between two subunits leading to theormation of a dimer, while the second interaction domain is responsible for thessociation of the two other subunits with the dimer.
cknowledgments
This work was supported by grants from the Canadian Institutesf Health Research and the Quebec Heart & Stroke Foundation. G.B.nd P.L. is the recipient of a scholarship from the Fonds de Recherchen Santé du Québec. M.P.L. is the recipient of a Ph.D. studentshiprom the Canadian Heart & Stroke Foundation. This work is part of.K.L.’s PhD thesis at the Université de Sherbrooke.
eferences
[1] M.J. Berridge, M.D. Bootman, P. Lipp, Calcium—a life and death signal, Nature395 (1998) 645–648.
[2] M.J. Berridge, P. Lipp, M.D. Bootman, The versatility and universality of calciumsignalling, Nature Reviews Molecular Cell Biology 1 (2000) 11–21.
[
m 45 (2009) 251–259
[3] D.E. Clapham, D. Julius, C. Montell, G. Schultz, International Union ofPharmacology. XLIX. Nomenclature and structure-function relationships oftransient receptor potential channels, Pharmacological Reviews 57 (2005) 427–450.
[4] C. Montell, L. Birnbaumer, V. Flockerzi, R.J. Bindels, E.A. Brudorf, M.J. Caterina,D.E. Clapham, C. Harteneck, S. Heller, D. Julius, I. Kojima, Y. Mori, R. Penner, D.Prawitt, A.M. Scharenberg, G. Schultz, N. Shimizu, M.X. Zhu, A unified nomen-clature for the superfamily of TRP cation channels, Molecular Cell 9 (2002)229–231.
[5] L. Birnbaumer, X. Zhu, M.S. Jiang, G. Boulay, M. Peyton, B. Vannier, D. Brown, D.Platano, H. Sadeghi, E. Stefani, M. Birnbaumer, On the molecular basis and regu-lation of cellular capacitative calcium entry: roles for Trp proteins, Proceedingsof the National Academy of Sciences of the United States of America 93 (1996)15195–15202.
[6] K. Mio, T. Ogura, Y. Hara, Y. Mori, C. Sato, The non-selective cation-permeablechannel TRPC3 is a tetrahedron with a cap on the large cytoplasmic end, Bio-chemical and Biophysical Research Communications 333 (2005) 768–777.
[7] T. Hofmann, M. Schaefer, G. Schultz, T. Gudermann, Subunit composition ofmammalian transient receptor potential channels in living cells, Proceedingsof the National Academy of Sciences of the United States of America 99 (2002)7461–7466.
[8] M. Goel, W.G. Sinkins, W.P. Schilling, Selective association of TRPC channel sub-units in rat brain synaptosomes, Journal of Biological Chemistry 277 (2002)48303–48310.
[9] X.Z.S. Xu, H.S. Li, W.B. Guggino, C. Montell, Coassembly of TRP and TRPL producesa distinct store-operated conductance, Cell 89 (1997) 1155–1164.
[10] B. Lintschinger, M. Balzer-Geldsetzer, T. Baskaran, W.F. Graier, C. Romanin,M.X. Zhu, K. Groschner, Coassembly of Trp1 and Trp3 proteins generatesdiacylglycerol- and Ca2+-sensitive cation channels, Journal of Biological Chem-istry 275 (2000) 27799–27805.
[11] X.B. Liu, B.C. Bandyopadhyay, B.B. Singh, K. Groschner, I.S. Ambudkar, Molecularanalysis of a store-operated and 2-acetyl-sn-glycerol-sensitive non-selectivecation channel, Journal of Biological Chemistry 280 (2005) 21600–21606.
12] C. Strubing, G. Krapivinsky, L. Krapivinsky, D.E. Clapham, Formation of novelTRPC channels by complex subunit interactions in embryonic brain, Journal ofBiological Chemistry 278 (2003) 39014–39019.
[13] P.K. Lepage, M.P. Lussier, H. Barajas-Martinez, S.M. Bousquet, A.P. Blanchard,N. Francoeur, R. Dumaine, G. Boulay, Identification of two domains involvedin the assembly of transient receptor potential canonical channels, Journal ofBiological Chemistry 281 (2006) 30356–30364.
[14] J. Sambrook, D.W. Russell, Molecular Cloning: A Laboratory Manual, Cold SpringHarbor, New York, 2001.
[15] F.M. Ausubel, R. Brent, R.E. Kingston, D.D. Moore, J.G. Seidman, J.A. Smith, K.Struhl, Current Protocols in Molecular Biology, Green Publishing Associates andWiley Interscience, New York, 2002.
[16] J.L. Parent, P. Labrecque, M.D. Rochdi, J.L. Benovic, Role of the differen-tially spliced carboxyl terminus in thromboxane A(2) receptor trafficking—identifcation of a distinct motif for tonic internalization, Journal of BiologicalChemistry 276 (2001) 7079–7085.
[17] M.P. Lussier, S. Cayouette, P.K. Lepage, C.L. Bernier, N. Francoeur, M. St Hilaire, M.Pinard, G. Boulay, MxA, a member of the dynamin superfamily, interacts withthe ankyrin-like repeat domain of TRPC, Journal of Biological Chemistry 280(2005) 19393–19400.
[18] M.P. Lussier, P.K. Lepage, S.M. Bousquet, G. Boulay, RNF24, a new TRPC interact-ing protein, causes the intracellular retention of TRPC, Cell Calcium 43 (2008)432–443.
[19] F. Sanger, S. Nicklen, A. Coulson, DNA sequencing with chain-terminatinginhibitors, Proceedings of the National Academy of Sciences of the United Statesof America 74 (1977) 5463–5467.
20] B.A. Niemeyer, Structure-function analysis of TRPV channels, Naunyn-Schmiedebergs Archives of Pharmacology 371 (2005) 285–294.
21] M. Schaefer, Homo- and heteromeric assembly of TRP channel subunits,Pflugers Archiv-European Journal of Physiology 451 (2005) 35–42.
22] I.S. Ambudkar, Ca2+ signaling microdomains: platforms for the assembly andregulation of TRPC channels, Trends in Pharmacological Sciences 27 (2006)25–32.
23] S.B. Long, E.B. Campbell, R. MacKinnon, Crystal structure of a mammalianvoltage-dependent Shaker family K+ channel, Science 309 (2005) 897–903.
24] A. Kreusch, P.J. Pfaffinger, C.F. Stevens, S. Choe, Crystal structure of the tetramer-ization domain of the Shaker potassium channel, Nature 392 (1998) 945–948.
25] N.V. Shen, P.J. Pfaffinger, Molecular recognition and assembly sequencesinvolved in the subfamily-specific assembly of voltage-gated k+ channel sub-unit proteins, Neuron 14 (1995) 625–633.
26] M. Arniges, J.M. Fernandez-Fernandez, N. Albrecht, M. Schaefer, M.A. Valverde,Human TRPV4 channel splice variants revealed a key role of ankyrin domainsin multimerization and trafficking, Journal of Biological Chemistry 281 (2006)1580–1586.
27] Q. Chang, E. Gyftogianni, S.F.J. van de Graaf, S. Hoefs, F.A. Weidema, R.J.M.Bindels, J.G.J. Hoenderop, Molecular determinants in TRPV5 channel assembly,Journal of Biological Chemistry 279 (2004) 54304–54311.
sient receptor potential V channel architecture and function require a specificankyrin repeat, Journal of Biological Chemistry 279 (2004) 34456–34463.
29] X.S. Jin, J. Touhey, R. Gaudet, Structure of the N-terminal ankyrin repeatdomain of the TRPV2 ion channel, Journal of Biological Chemistry 281 (2006)25006–25010.
Calciu
[
[
[
[
knock-down strategies fail to affect CRAC channels in mucosal-type mast cells,
P.K. Lepage et al. / Cell
30] C.B. Phelps, R.J. Huang, P.V. Lishko, R.R. Wang, R. Gaudet, Structural analyses ofthe ankyrin repeat domain of TRPV6 and related TRPV ion channels, Biochem-istry 47 (2008) 2476–2484.
31] P.V. Lishko, E. Procko, X. Jin, C.B. Phelps, R. Gaudet, The ankyrin repeats of TRPV1bind multiple ligands and modulate channel sensitivity, Neuron 54 (2007)905–918.
32] C.J. McCleverty, E. Koesema, A. Patapoutian, S.A. Lesley, A. Kreusch, Crystal struc-ture of the human TRPV2 channel ankyrin repeat domain, Protein Science 15(2006) 2201–2206.
[
m 45 (2009) 251–259 259
33] H. Kahr, R. Schindl, R. Fritsch, B. Heinze, M. Hofbauer, M.E. Hack, M.A. Mortel-maier, K. Groschner, J.B. Peng, H. Takanaga, M.A. Hediger, C. Romanin, CaT1
Journal of Physiology-London 557 (2004) 121–132.34] R. Schindl, I. Frischauf, H. Kahr, R. Fritsch, M. Krenn, A. Derndl, E. Vales, M. Muik,
I. Derler, K. Groschner, C. Romanin, The first ankyrin-like repeat is the minimumindispensable key structure for functional assembly of homo- and heteromericTRPC4/TRPC5 channels, Cell Calcium 43 (2008) 260–269.
