Muscarinic Acetylcholine Receptor Regulation of TRP6 Ca 2 1 Channel Isoforms

In this study, we report the molecular cloning of cDNAs encoding three distinct isoforms of rat (r) TRP6 Ca 2 1 channels. The longest isoform, rTRP6A, contains 930 amino acid residues; rTRP6B lacks 54 amino acids (3–56) at the N terminus, and rTRP6C is missing an additional 68 amino acids near the C terminus. Transient transfection of COS cells with expression vectors encoding rTRP6A or rTRP6B increased Ca 2 1 influx and gave rise to a novel Ba 2 1 influx after activation of M 5 muscarinic acetylcholine receptors. By contrast, passive depletion of intracellular Ca 2 1 stores with thapsigargin did not induce Ba 2 1 influx in cells expressing rTRP6 isoforms. Ba 2 1 influx was also stimulated in rTRP6A-expressing cells after exposure to the diacylglycerol an-alog, 1-oleoyl-2-acetyl- sn -glycerol (OAG), but rTRP6B-expressing cells failed to show OAG-induced Ba 2 1 influx. Expression of a rTRP6 N-terminal fragment of rTRP6B or rTRP6A antisense RNA blocked M 5 musca- rinic acetylcholine receptor-dependent Ba 2 1 influx in COS cells that were transfected with rTRP6 cDNAs. Together these results suggest that rTRP6 participates in the formation of Ca 2 1 channels that are regulated by a G-protein-coupled receptor, but not by intracellular Ca 2 1 stores. In contrast to the results we obtained with 0.01% H 2 O 2 , 0.04% NiCl 2 , and 50 m M Tris-Cl, pH 7.5. Deglycosylation— Cell lysates were treated with endoglycosidase H (0.05 units, Roche Molecular Biochemicals) or peptide N -glycosidase F (5 units, Roche Molecular Biochemicals) overnight at 4 °C with rotation. The samples were then mixed with an equal volume of 2 3 SDS sample buffer and resolved by SDS-polyacrylamide gel electrophoresis and Western blots as described above.


Lei Zhang ‡ and David Saffen
From the Department of Neurochemistry, Faculty of Medicine, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113-0033, Japan In this study, we report the molecular cloning of cDNAs encoding three distinct isoforms of rat (r) TRP6 Ca 2؉ channels. The longest isoform, rTRP6A, contains 930 amino acid residues; rTRP6B lacks 54 amino acids  at the N terminus, and rTRP6C is missing an additional 68 amino acids near the C terminus. Transient transfection of COS cells with expression vectors encoding rTRP6A or rTRP6B increased Ca 2؉ influx and gave rise to a novel Ba 2؉ influx after activation of M 5 muscarinic acetylcholine receptors. By contrast, passive depletion of intracellular Ca 2؉ stores with thapsigargin did not induce Ba 2؉ influx in cells expressing rTRP6 isoforms. Ba 2؉ influx was also stimulated in rTRP6Aexpressing cells after exposure to the diacylglycerol analog, 1-oleoyl-2-acetyl-sn-glycerol (OAG), but rTRP6Bexpressing cells failed to show OAG-induced Ba 2؉ influx. Expression of a rTRP6 N-terminal fragment of rTRP6B or rTRP6A antisense RNA blocked M 5 muscarinic acetylcholine receptor-dependent Ba 2؉ influx in COS cells that were transfected with rTRP6 cDNAs. Together these results suggest that rTRP6 participates in the formation of Ca 2؉ channels that are regulated by a G-protein-coupled receptor, but not by intracellular Ca 2؉ stores. In contrast to the results we obtained with rTRP6A and rTRP6B, cells expressing rTRP6C showed no increased Ca 2؉ or Ba 2؉ influxes after stimulation with carbachol and also did not show OAG-induced Ba 2؉ influx. Glycosylation analysis indicated that rTRP6A and rTRP6B are glycosylated in COS cells, but that rTRP6C is mostly not glycosylated. Together these results suggest that the N terminus (3-56 amino acids) is crucial for the activation of rTRP6A by diacylglycerol and that the 735-802 amino acid segment located just downstream from the 6th transmembrane segment may be required for processing of the rTRP6 protein.
Cytosolic Ca 2ϩ signals play important roles in the development, function, and death of cells (1,2). Increases in intracel-lular Ca 2ϩ typically result from the release of Ca 2ϩ from intracellular stores and/or from an increased influx of extracellular Ca 2ϩ (3). In many types of cells, stimulation of seven-transmembrane receptors that couple to the production of inositol 1,4,5-triphosphate is followed by a biphasic increase in intracellular Ca 2ϩ . An initial transient increase is due to the opening of inositol 1,4,5-triphosphate-activated Ca 2ϩ channels that release the Ca 2ϩ from the endoplasmic reticulum. This is followed by a sustained increase in Ca 2ϩ due to the influx of Ca 2ϩ across the plasma membrane (4 -6). Calcium influx that is triggered by the emptying of intracellular stores was first termed capacitative calcium entry by Putney (7). The molecular identification of Ca 2ϩ channels mediating capacitative calcium entry is currently a topic of wide interest.
The molecular cloning of cDNAs encoding Ca 2ϩ channels that function in the visual system of Drosophila (d), the transient receptor potential (dTRP) 1 (8) and dTRP-like (dTRPL) (9) channels, provided the first two candidates for capacitative calcium entry channels. dTRP was shown to form Ca 2ϩ -permeable channels that are activated by store depletion, whereas dTRPL forms nonselective cation channels that are constitutively active and insensitive to store depletion when expressed in Sf9 cells (10,11). Coexpression of dTRP and dTRPL was also shown to give rise to a store-operated current that is distinct from that produced by either dTRP or dTRPL alone (12). Furthermore, dTRP and dTRPL interact directly, suggesting that they are likely to form heteromeric channels (12,13).
A search for mammalian homologues of the dTRP and dTRPL genes has yielded seven TRP homologues, TRP1-7, which are differentially expressed in various tissues. The functional characteristics of channels encoded by these genes have been studied primarily in heterologous cell systems (14 -32). Based on these studies, the seven mammalian TRP homologues can be divided into two groups: TRP1/2/4/5, which are thought to form channels activated by Ca 2ϩ store depletion, and TRP3/ 6/7, which are thought to be activated by G q -coupled receptors independently of Ca 2ϩ stores (14,20,33,34). There are still controversies, however, concerning the mechanisms of regulation of individual TRP channels (23-27, 29, 34, 35). For example, TRP6 has been reported to be both a G-protein receptorcoupled, Ca 2ϩ store-independent channel (25,26) and a Ca 2ϩ store-operated channel (27). Recently, human (h) TRP6, hTRP3 (26), and mouse (m) TRP7 (28) have been reported to be activated by diacylglycerol (DAG), but the activation pathway and domain within the TRP proteins required for activation are still unclear.
To study the mechanisms of regulation of TRP6, we used a RT-PCR strategy to clone the rat (r) TRP6 cDNA. This approach yielded cDNA encoding three distinct isoforms of TRP6. When expressed in COS cells, rTRP6A and rTRP6B form Ca 2ϩ channels that couple to M 5 muscarinic acetylcholine receptors (mAChR) and are regulated independently of intracellular Ca 2ϩ stores. The longest isoform, rTRP6A, is activated by the DAG analog, 1-oleoyl-2-acetyl-sn-glycerol (OAG), but rTRP6B is not, suggesting that amino acids present in the N terminus of rTRP6A are crucial for activation by DAG but are not required for activation by carbachol. Finally, the apparent inabil-ity of the shortest TRP6 isoform, rTRP6C, to undergo glycosylation and form functional channels suggests that a domain located on the cytoplasmic side of the sixth transmembrane segment may be required for processing of the rTRP6 protein.

EXPERIMENTAL PROCEDURES
Molecular Cloning-rTRP6A cDNA was cloned from rat lung total RNA, and rTRP6B and C cDNAs were cloned from PC12D cell (36) total RNA using a RT-PCR approach. The forward primer for rTRP6A was 5Ј-CCAGGCACTTGCCATGAGCCAGAG-3Ј; the forward primer for rTRP6B and C was 5Ј-ATGAGCCGGGGTAATGAAAACAGAC-3Ј, and the common reverse primer was 5Ј-CCAATCGATCTATCTGCGGCTT-TCC-3Ј. The RT-PCR products were isolated from 1% agarose gels and cloned in pGEM-T Easy (Promega). TRP6 cDNAs were excised from this FIG. 1. Primary structures of rat TRP6 isoforms. A, amino acid sequences of three rat TRP6 isoforms were aligned using the ClustalW program (43). Underlines indicate the peptides used for production of anti-TRP6 antibodies. B, schematic comparison of TRP6A, -B, and -C. MS, methionine-serine. The black regions indicate sequences deleted in TRP6B and/or TRP6C. C, Kyte and Doolittle hydropathy (DNA Strider 1.2) analysis of TRP6A using a window size of 9 amino acids (aa). The striped bars indicate the ankyrin repeats, the open bars indicate the putative transmembrane segments (S1-S6), and the black bar indicates the putative pore region (P) between S5 and S6 (SMART prediction tool (44)).
vector by digesting with NotI and SpeI and subcloned between the NotI and SpeI sites of pEF-BOS-SK, a derivative of the mammalian expression vector pEF-BOS (37) that contains the multiple cloning site of pBluescript-SK (Stratagene). The resulting plasmids were designated pBOS-TRP6A, -B, and -C. An antisense expression vector, pBOS-TRP6A-antisense, was constructed by subcloning a NotI/SpeI DNA fragment encoding rTRP6A in pEF-BOS-KS. An expression vector encoding the TRP6B N-terminal fragment fused to green fluorescent protein (GFP), pEGFP-TRP6B-N, was constructed by cloning a DNA fragment encoding amino acid residues 1-301 of rTRP6B between the EcoRI and XmaI sites of pEGFP-N2 (CLONTECH). pBOS-TRP6A⌬ was constructed by digesting pBOS-TRP6A and pBOS-TRP6C with XhoI and SpeI and ligating the small XhoI/SpeI fragment from pBOS-TRP6C to the large fragment of pBOS-TRP6A. In the same way, pBOS-TRP6C ϩ (which has the same primary structure as pBOS-TRP6B) was constructed by ligating the small XhoI/SpeI fragment from pBOS-TRP6A to the large fragment of pBOS-TRP6C.
Cell Culture and Transfection-COS-7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Confluent cultures were replated in 12-well plates at a density of 5 ϫ 10 4 cells/well. After 24 h, the medium was changed, and 100 l of transfection mixture was added to each well. Transfection mixtures contained 500 ng of the M 5 muscarinic receptor expression plasmid pBOS-M 5 (38) Generation of Polyclonal Anti-TRP6 Antibodies-A 17-amino acid peptide ( 914 KLGERLSLESKQEESRR 930 ) and a 12-amino acid peptide ( 788 QGHKKGFQEDAE 799 ) were used as the antigen peptides for the production of antiserum in rabbits (Biologica Co., Nagoya, Japan) and were designated as anti-TRP6-C terminus and anti-TRP6 (788 -799), respectively. These TRP6 antibodies were purified by affinity chromatography (SulfoLink kit, Pierce) using the antigen peptide as described in Harlow and Lane (39).
The samples were then mixed with an equal volume of 2ϫ SDS sample buffer and resolved by SDS-polyacrylamide gel electrophoresis and Western blots as described above.

Primary Structures of Rat TRP6
Isoforms-Using a RT-PCR strategy, we cloned cDNAs encoding three rat TRP6 isoforms: rTRP6A from rat lung and rTRP6B and C from PC12D cells. rTRP6A is composed of 930 amino acids, and rTRP6B and C contain 876 and 808 amino acids, respectively. Compared with rTRP6A, rTRP6B is missing 54 amino acids at the N terminus, and rTRP6C is missing an additional 68 amino acids near C terminus ( Fig. 1A and B). Hydropathy analysis suggests that rTRP6 has six hydrophobic transmembrane domains and one pore-forming region, located between the fifth and sixth transmembrane segments. Both N-and C-terminal regions contain a large percentage of hydrophilic amino acid residues, and the N terminus of rTRP6A-C is predicted to contain three ankyrinlike repeats (Fig. 1C).
Expression and Distribution of rTRP6 -When expressed in COS cells, rTRP6A, -B, and -C proteins of apparent molecular masses 107, 98, and 87 kDa, respectively, were detected using an antibody that recognizes the rTRP6 C terminus. By contrast, COS cells transfected with the empty expression vector, pBOS-SK, produced no detectable TRP6 protein ( Fig. 2A, left). The specificity of the rTRP6-C terminus antibody is demonstrated by the fact that preincubation with the TRP6-C-terminal oligopeptide completely blocked the appearance of TRP6 bands in the Western blots ( Fig. 2A, right). Endogenous rTRP6 proteins were found to be widely expressed in rat brain (Fig.  2B, left). The rTRP6 protein expressed in brain roughly comigrated with rTRP6A expressed in COS cells (Fig. 2B, right).
Functional Characterization of rTRP6A, -B, and -C-To study the roles of the three rTRP6 isoforms in calcium signaling, rTRP6A, -B, or -C cDNAs or pBOS-SK were transiently expressed in COS cells along with expression vectors for M 5 muscarinic acetylcholine receptor and GFP. An expression vector encoding M 5 mAChR, which couples to G q (38), was used to study the role of rTRP6 in carbachol-stimulated Ca 2ϩ entry. Cells were loaded with fura-2 in KRH (containing 2 mM CaCl 2 ), and extracellular Ca 2ϩ was chelated by adding 4 mM EGTA just before making the measurements. The addition of 500 M carbachol to these cells induced a rapid rise in cytosolic Ca 2ϩ , which decreased to basal levels within 2-3 min (Fig. 3, A-D). This transient increase in Ca 2ϩ corresponds to the rapid release of Ca 2ϩ from internal stores and subsequent expulsion of Ca 2ϩ from the cell. The addition of 4 mM CaCl 2 to the extracellular solution induced a second increase in cytosolic Ca 2ϩ due to the influx of extracellular Ca 2ϩ (Fig. 3, A-D). Compared with the control cells, the Ca 2ϩ increases due to influx were much larger in cells expressing rTRP6A or -B, but cells expressing rTRP6C showed no significant differences in Ca 2ϩ influx compared with the control (Fig. 3E).
To further characterize the rTRP6 channels, we examined the ability of carbachol to stimulate Ba 2ϩ influx in COS cells expressing individual rTRP6 isoforms. Ba 2ϩ is frequently used as a substitute for Ca 2ϩ since its entry into the cell can also be monitored by measuring increases in fura-2 fluorescence. Previously, we found that endogenously expressed store-operated Ca 2ϩ channels in neuronal PC12D cells are relatively impermeable to Ba 2ϩ ions. 2 pBOS-SK-transfected COS cells in nominally Ca 2ϩ -free KRH buffer did not show any Ba 2ϩ influx after stimulation with carbachol and subsequent addition of 200 M BaCl 2 (Fig. 4A). By contrast, significant carbachol-stimulated Ba 2ϩ influx was observed in TRP6A-or -B-transfected cells (Fig. 4, B and C). COS cells expressing rTRP6C, the shortest isoform, showed no Ba 2ϩ influx (Fig. 4D).
To determine whether the regulation of rTRP6 is dependent on the depletion of intracellular Ca 2ϩ stores, the effects of thapsigargin were examined. Thapsigargin indirectly causes the release of Ca 2ϩ stored in the endoplasmic reticulum by irreversibly inhibiting the sarco/endoplasmic reticulum calcium pump, which functions to fill these stores (40). In nominally Ca 2ϩ -free medium, the addition of 1 M thapsigargin depleted the intracellular Ca 2ϩ stores, resulting in a transient increase in cytosolic Ca 2ϩ . Ba 2ϩ influx was not observed in control cells or in rTRP6A-, -B-, or -C-transfected cells after the addition of 200 M BaCl 2 to the medium (Fig. 5, A-D). Similar results were obtained using 100 nM thapsigargin to deplete intracellular Ca 2ϩ stores. 3 Inhibition of Ba 2ϩ Influx by Expression of the rTRP6B Nterminal Fragment or rTRP6A Antisense RNA-Coexpression of rTRP6B-N (encoding the N-terminal 1-301 amino acid residues of rTRP6B) with rTRP6A totally blocked carbachol-stimulated Ba 2ϩ influx (Fig. 6, A and B). . Ca 2ϩ -dependent changes in fluorescence were measured in fura-2 loaded COS cells in KRH (containing 2 mM CaCl 2 ) using the Argus-50/Ca system. EGTA (4 mM) was added to the cells to chelate extracellular Ca 2ϩ before measurements. Cells were exposed to 500 M carbachol (Carb) and 4 mM CaCl 2 at the times indicated by the arrows. Expression of GFP was used as a marker for cells that had taken up plasmids. The traces shown are the averages of seven measurements performed on groups of GFP fluorescent cells (3-10 cells/group). E, mean amplitudes of carbacholinduced Ca 2ϩ influx in control and TRP6-expressing cells. The amplitudes were obtained by subtracting the base-line 340-nm/380-nm ratio from the 340-nm/380-nm ratio measured at the highest point of Ca 2ϩ influx. Asterisks (*) above the bars indicate the statistical significance of differences (p Ͻ 0.01) compared with the control (by Bonferroni's t test).
was cotransfected with the rTRP6B expression vector (Fig. 6, C  and D). In Fig. 6E, Western blots show that rTRP6B protein levels are significantly reduced in cells cotransfected with the rTRP6A antisense expression vector.
These data suggest that expression of rTRP6A or -B causes an increase in Ca 2ϩ influx compared with the control cells. To obtain additional evidence for this model, we examined the effects of rTRP6A antisense RNA expression on Ca 2ϩ influx. Fig. 7 shows that in nominally Ca 2ϩ -free medium, carbacholstimulated Ca 2ϩ influx was reduced to control levels when rTRP6A antisense RNA was coexpressed with rTRP6A.
Effects of OAG and Phorbol 12-Myristate 13-Acetate (PMA) on Ba 2ϩ Influx-Recent studies show that hTRP3 and hTRP6 (26) and also mTRP7 (28) can be directly activated by the analogs of diacylglycerol. We therefore tested the effects of a membrane-permeable diacylglycerol analog, OAG, on Ba 2ϩ influx in COS cells expressing rTRP6. Cells in nominally Ca 2ϩfree medium were exposed to 100 M OAG and then 200 M BaCl 2 . No Ba 2ϩ influx was observed in pBOS-SK-transfected control cells (Fig. 8A), but significant stimulation of the influx of Ba 2ϩ by OAG was observed in rTRP6A-expressing cells (Fig.  8B). By contrast, there was no OAG-induced Ba 2ϩ influx in rTRP6B-or -C-transfected cells (Fig. 8, C and D). Pretreatment of cells with 500 nM PMA, a PKC activator, abolished OAGactivated, rTRP6A-mediated Ba 2ϩ influx (Fig. 8F), whereas pretreatment with PMA alone had no effect on Ba 2ϩ influx (Fig.  8E). PMA also failed to stimulate Ba 2ϩ influx when added just before the addition of BaCl 2 . 3 Similar results were obtained using 100 nM PMA. 3 Effects of PMA and GF 109203X on M 5 Receptor-mediated Ba 2ϩ Influx-To further investigate the involvement of PKC in M 5 receptor-mediated Ba 2ϩ influx, we examined the effects of the phorbol ester PMA and the PKC inhibitor GF 109203X (41,42). Pretreatment of COS cells expressing rTRP6A and M 5 receptor with PMA blocked carbachol-activated, rTRP6A-mediated Ba 2ϩ influx (Fig. 9, A and B). GF 109203X alone did not significantly affect carbachol-induced Ba 2ϩ influx (Fig. 9C), but pretreatment of the cells with GF 109203X before PMA abolished the inhibitory effect of phorbol ester (Fig. 9D). These results indicate that PMA inhibits carbachol-activated Ba 2ϩ influx by activating PKC.
Glycosylation Analysis of rTRP6 -The predicted amino acid sequence of rTRP6A contains nine consensus NX(S/T) motifs for N-glycosylation (Fig. 10A). Because only sites that are exposed to the extracellular side of the membrane are expected to be glycosylated, the topological model for rTRP6 depicted in Fig. 1 suggests that only Asn at position 711 is a candidate for glycosylation. To analyze whether rTRP6 is glycosylated in COS cells and to provide evidence that rTRP6 proteins are expressed on the surface of plasma membrane, we treated the cell lysates with two glycosidases, endoglycosidase H and peptide N-glycosidase F before performing the Western blot analysis. Treatment with endoglycosidase H, an enzyme that cleaves the N-glycosidic bond between the first and second GlcNAc residue of high mannose-containing, immature glycoproteins did not change the pattern of rTRP6A significantly, but eliminated two sharp bands located just above the rTRP6B band with the highest mobility. Treatment with peptide N-glycosidase F, an enzyme that cleaves the N-glycosidic bond between the sugar chain and asparagine of the mature and immature forms of glycoproteins, almost totally eliminated the hazy material between 83 and 175 kDa and increased the intensities of rTRP6A and -B high mobility bands, but did not change the appearance of the rTRP6C band (Fig. 10B).
Properties of the Constructs TRP6A⌬ and TRP6C ϩ -The results in Fig. 10 show that cells expressing rTRP6C have very low levels of the mature glycosylated protein. This suggested that the 735-802 segment missing in rTRP6C is crucial for processing of the rTRP6 protein. To test this hypothesis, we prepared the constructs TRP6A⌬ and TRP6C ϩ . As shown in Fig. 11A, TRP6A⌬  is equivalent to rTRP6A but is missing the 735-802 segment, and TRP6C ϩ is equivalent to rTRP6C with the addition of the 735-802 segment (TRP6C ϩ is structurally the same as rTRP6B). Fig. 11B shows the results of Western blot analysis using the anti-TRP6-C terminus Ab. As shown in the figure, cells expressing rTRP6A⌬ lost most of the hazy material in the region between 83 and 175 kDa compared with rTRP6A. By contrast, cells expressing TRP6C ϩ contained more of this hazy material compared with rTRP6C-expressing cells and showed the same pattern with rTRP6B. Fig. 11C shows the Western blot result obtained by probing with an antibody that recognizes a 12-amino acid peptide located within the 735-802 segment region. As expected, TRP6A⌬ was not recognized by anti-TRP6 (788 -799) Ab. By contrast, TRP6C ϩ was recognized by anti-TRP6 (788 -799) Ab and showed the same pattern as rTRP6B (Fig. 11C).
To test the functional importance of the 735-802 segment, we measured carbachol-and OAG-stimulated Ba 2ϩ influx in pBOS-TRP6A⌬-or pBOS-TRP6C ϩ -transfected COS cells. As shown in Fig. 12A, cells expressing TRP6A⌬ failed to show Ba 2ϩ influx response to carbachol stimulation. By contrast, cells expressing TRP6C ϩ showed significant Ba 2ϩ influx, whereas the original rTRP6C-expressing cells did not (Fig. 12B  and Fig. 4D). There was also no Ba 2ϩ influx observed after stimulation with OAG in either TRP6A⌬ or TRP6C ϩ -expressing cells (Fig. 12, C and D). DISCUSSION In the present study, we cloned three isoforms of rat TRP6, designated rTRP6A, -B, and -C. These isoforms may result from alternative RNA splicing as suggested for other TRP subtypes (15-20, 28, 45). rTRP6A closely resembles human TRP6 (26; GenBank TM accession number AF080394) and mouse TRP6 (25; GenBank TM accession number U49069), with nucleic acid identities of 88.38 and 94.41%, respectively. rTRP6B closely resembles a second mouse isoform of TRP6 (unpublished sequence; GenBank TM accession number AF057748) with a nucleic acid identity of 94.83%. rTRP6C is a novel isoform. Mizuno et al. (27) also report the cloning of TRP6 from rat brain. Their sequence is similar to our rTRP6A cDNA, except that it contains an insertion of a G residue at position of 1210 and a deletion of a C residue at position 1326, resulting in . EGTA (4 mM) was added to the cells to chelate extracellular Ca 2ϩ before measurements. Cells were exposed to 500 M carbachol and 4 mM CaCl 2 at the times indicated by the arrows. The traces shown are the averages of 7 measurements of GFP fluorescent cells (3-10 cells/measurements). D, mean amplitudes of carbachol-induced Ca 2ϩ influx in control and TRP6expressing cells. The amplitudes were obtained by subtracting the base-line 340-nm/380-nm ratio from 340-nm/380-nm ratio measured at the highest point of Ca 2ϩ influx. The asterisk (*) above the bar indicates the statistical significance of difference (p Ͻ 0.01) compared with the control (by Bonferroni's t test). a shift in the reading frame in this region compared with our sequences and also with respect to the mouse and human TRP6 sequences. Previous studies found that mouse TRP6 mRNA is expressed at low levels in brain (25). Using Western blot analysis, we found that rTRP6 protein is expressed throughout the brain, including the cortex, cerebellum, hippocampus, brain stem, and midbrain regions (Fig. 2B).
In COS cells, transient expression of rTRP6A and -B resulted in a 2.5-and a 1.5-fold increase in Ca 2ϩ influx compared with control cells when Ca 2ϩ was added back to cultures stimulated with carbachol in the absence of extracellular Ca 2ϩ . By contrast, rTRP6C-transfected cells showed no significant Ca 2ϩ increase compared with the control cells in the same assay (Fig. 3).
Hofmann et al. (26) first showed that human TRP6 and TRP3 are activated by DAG analogs, and Okada et al. (28) observed a similar activation for mouse TRP7. In our experiments, OAG induced significant Ba 2ϩ influx in rTRP6A-transfected cells (Fig. 8B), but not in control cells (Fig. 8A). Interestingly, cells expressing rTRP6B failed to show OAG-induced Ba 2ϩ influx (Fig. 8C). rTRP6B is missing a 54-amino acid (3-56) segment that is present at the N terminus of rTRP6A, suggesting that these amino acids are crucial for activation by DAG but are not required for rTRP6 activation by mAChR (Fig. 4B). Together these results imply that activation of rTRP6 after stimulation of mAChR is mediated by DAG-dependent and DAG-independent pathways. Further study will be needed to elucidate the role of the N-terminal 54-amino acid segment in the activation of rTRP6A by DAG.
In addition to its ability to activate Ba 2ϩ influx through rTRP6A, OAG is also known to activate PKC (46). To determine whether PKC activation is correlated with Ba 2ϩ influx, we examined the effects of another PKC activator, PMA, on Ba 2ϩ influx. Pretreatment with PMA did not induce Ba 2ϩ influx (Fig.  8E), but instead blocked OAG-induced Ba 2ϩ influx in rTRP6Aexpressing cells (Fig. 8F). These data indicate that DAG-induced Ba 2ϩ influx does not involve the activation of protein kinase C but, rather, that activation of PKC blocks the activation of the rTRP6A channel. Pretreatment of cells with PMA also blocked carbachol-activated, rTRP6A-mediated Ba 2ϩ influx (Fig. 9, A and B). This inhibition is mediated by PKC since pretreatment of the cells with GF 109203X before PMA abolished the inhibitory effects of this phorbol ester (Fig. 9D).
If DAG can both activate and inhibit TRP6 channels, one might expect that pretreatment of the cells with GF 109203X alone would potentiate carbachol-induced Ba 2ϩ influx. Pretreatment with GF 109203X, however, did not significantly change the rate or extent of carbachol-stimulated Ba 2ϩ influx (Fig. 9C). This result suggests that DAG produced after activation of mAChR might stimulate Ba 2ϩ influx more efficiently (or perhaps more rapidly) than it activates PKC. Activation by OAG and inhibition by PMA has also been reported for mTRP7 (28). Clearly, further study is needed to elucidate the mechanisms underlying the regulation of TRP Ca 2ϩ channels by DAG.
Previous studies from Birnbaumer and co-workers showed that human TRP3 proteins expressed in HEK293 (30) and COS cells (31,32) and mouse TRP6 proteins in COS cells (25) are glycosylated. In our study, we also found that rat TRP6 proteins are glycosylated when expressed in COS cells. rTRP6A is present mostly in the mature glycosylated form, whereas rTRP6B contains both mature glycosylated and immature and unglycosylated forms. By contrast, most of the rTRP6C is not glycosylated (Fig. 10B), suggesting that rTRP6C may not be processed normally. This may explain why rTRP6C did not respond to stimulation with carbachol and OAG.
The glycosylation results led us to examine the importance of the 68 amino acids (735-802 segment) missing from the rTRP6C, which is located on the cytosolic side immediately adjacent to the sixth transmembrane domain. To determine the function of this segment, we constructed TRP6A⌬, which is rTRP6A deleted for this segment, and TRP6C ϩ , which is rTRP6C with this fragment added on. As expected, TRP6A⌬ did not show glycosylation characteristic of the mature forms and also failed to show Ba 2ϩ influx response to the carbachol stimulation ( Fig. 11B and Fig. 12A). By contrast, TRP6C ϩ did express mature glycosylated forms compared with rTRP6C, showing a pattern similar to rTRP6B (to which it is structurally identical). As for rTRP6B, significant Ba 2ϩ influx was observed in TRP6C ϩ -expressing cells after stimulation with carbachol ( Fig. 11B and Fig. 12B). The first 35 amino acids in the 735-802 segment ( 735 DDADVEWKFARAKLWFSYFEE-GRTLPVPFNLVPSP 769 ) together with the sixth transmembrane segment are highly conserved in TRPs from Drosophila, Caenorhabditis elegans, and mammals, suggesting the potential functional importance of this domain. By contrast, the last 33 amino acids of this fragment ( 770 KSLLYLLLKFKKWM-SELIQGHKKGFQEDAEMNK 802 ) are only conserved among TRP6s from different species. Expression of mRNA encoding rTRP6C can be detected at low levels in brain and PC12D cells by RT-PCR, 3 but the function of this isoform remains unknown.
In summary, we have cloned three distinct rTRP6 isoforms and demonstrated that rTRP6 contributes to the formation of G-protein-coupled receptor-regulated store-depletion-independent Ca 2ϩ channels when expressed in COS cells. We have also identified two domains that are important for the function of rTRP6 channels: the N terminus (3-56 amino acids), which is crucial for the activation of rTRP6A by DAG, and the 735-802 segment located just downstream from the sixth transmembrane segment, which may be required for processing of the rTRP6 protein.