TRP Genes: Candidates for Nonselective Cation Channels and Store-Operated Channels in Insulin-Secreting Cells

A mouse βTC3 cell cDNA library was generously provided by R. Stein (Vanderbilt University). Mouse multiple tissue Northern blots and pEYFP-N1 were purchased from Clontech. Human pCDNA-Trp3-myc was a kind gift from Dr. Craig Montell (Johns Hopkins University).

The βTC3 cell cDNA library was screened with a cDNA fragment of mouse STrpC4 homolog, obtained by RT-PCR amplification from βTC3 insulinoma cell poly(A)+ RNA. PCR was performed with the oligonucleotide primer pair: 5′-CTGCAGATATCTCTGGGAAGG-3′ and 5′-GCTTTGTTCGAGCAAATTTCC-3′, based on the published partial sequences of TRP homolog cDNA isolated from mouse brain (X90697) (21). The PCR conditions were initial denaturation at 94°C for 3 min, followed by 30 cycles of denaturation at 94°C for 30 s, annealing at 58°C for 30 s, and extension at 72°C for 30 s with a final extension at 72°C for 8 min. The 417-bp fragment was subcloned into pCRII vector (Invitrogen) and sequenced. Screening of the βTC3 cell cDNA library (λZiplox; BRL Life Technologies) was performed as described (22) at a final stringency of 0.1× standard saline citrate (SSC), 0.1% SDS at 60°C. Both strands of the cDNA inserts were sequenced on an automated DNA sequencer (Perkin-Elmer). The complete clone of the alternative spliced form was constructed by ligation of the PstI-NotI fragment of the short splicing form with the SalI-PstI fragment of the unspliced form. Both clones were inserted between the SalI and NotI sites of the pBluescript II SK(+) phagemid.

As the nomenclature for the TRPC channels has recently been revised, the channel cDNA originally termed mTRP4 in the GenBank entry is now termed mSTRPC4, where “m” indicates mouse, “S” indicates the “short” TRPC family, “TRP” is the designation from the Drosophila transient receptor potential gene, “C” indicates channel, “4” indicates the family member (currently numbered 1–7), and “L” and “S” indicate the particular splice variant (23).

Mouse βTC3 insulinoma cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with high glucose, 5% fetal bovine serum (FBS), 15% horse serum (GIBCO/BRL, Life Technologies). HEK293 cells were cultured in DMEM with high glucose, 10% FBS (GIBCO/BRL, Life Technologies), and transfected (2 × 10⁶ cells/100-mm dish) with 5 μg DNA by Superfect reagent (Qiagen).

Mouse multiple tissue Northern blots (Clontech) were prehybridized and hybridized and washed to a final stringency of 0.1× SSC, 0.1% SDS at 60°C. A 2-kb human cDNA for β-actin was used as the probe to hybrid with the same blot as the control.

The 5′ untranslated region of mTrp4 cDNA was excised and subcloned between the XhoI and NotI sites of a modified pBK-CMV vector (Stratagen). Then a three-primer PCR strategy was used to insert a human insulin C-peptide cDNA in-frame between the last amino acid and the stop codon (24). The BamHI and NotI fragment containing the entire insertion was cloned into pCMV5 expression vectors between the restriction sites BglII and NotI. The corresponding splicing isoforms were deposited in GenBank as mTrp4S (U50921) for the shorter form and mTrp4 l (U50922) for the longer form.

The peptide N-CDSSIDYDLSPTDTAAHEDYV-C was synthesized and used to raise antisera in rabbits (Zeneca) in collaboration with R. Shehee, GlaxoWelcome Research Institute. The antisera from two animals, 523 and 524, were purified by affinity chromotography and tested in enzyme-linked immunosorbent assay (ELISA).

Cell lysates containing 40 μg total protein, prepared 48 h after transfection, was fractionated on 7.5% (37.5:1) denaturing polyacrylamide gels. Proteins were transferred to polyvinylidine difluoride membranes (Amersham) and incubated with immunoglobulin G followed by detection with an enhanced chemiluminescence reagent (Amersham). For immunohistochemistry, cells were fixed in Dulbecco’s phosphate-buffered saline (D-PBS)-buffered 4% paraformaldehyde solution, pH 7.4, for 10 min at 4°C. The primary and secondary antibodies were incubated with the fixed cells according to standard protocols (3% BSA, 0.1% Tween-20 in PBS, pH 7.4).

Cells were plated on glass coverslips coated with poly-l-lysine (Sigma) and/or with Pronectin F plus (Protein Polymer Technologies) at least 1 day before transfection. Transiently transfected cells were identified by yellow-shifted green fluorescence of enhanced yellow fluorescence protein (EYFP) (Clontech) cotransfected with the desired cDNA at a ratio of 1:9. This allowed direct identification of transfected cells that did not overlap with the fura-2 emission. Control cells were transfected with pEYFP alone. There was no detectable difference in responses between control cells expressing EYFP from those remaining untransfected. Cells were loaded with 5 μmol/l fura-2/AM and 0.05% pluronic F-127 in extracellular solution in the dark at 37°C for 30 min, at 24–48 h after transfection. Fluorescent images of the cells were recorded and analyzed at room temperature using the Metafluor Imaging System (Universal Imaging) by alternating excitation at 340 nm and 380 nm at 5-s intervals and detecting emission at 510 nm. The extracellular solution contains HEPES (15 mmol/l), glucose (10 mmol/l), NaCl (140 mmol/l), KCl (1 mmol/l), and MgCl₂ (1 mmol/l), supplemented with 1.8 mmol/l CaCl₂, 1.8 mmol/l SrCl₂, or 100 μmol/l EGTA (nominal calcium-free solution). All reagents were diluted to their final concentration in the extracellular solution and applied to cells by perfusion. Data from 10 to 30 cells per coverslip were collected per experiment and repeated to give the total number of cells.

RNA was extracted from human islets (kindly provided by B. Hering, University of Minnesota) using Trizol reagent. The cDNA was obtained with random primers or oligodT. For PCR analysis of LTRPC2 expression, we used three forward primers: For1, -TGTGCAGAATTGTACAGTTGCG; For22 -AAACCATGTCGCTGGCAGCTGG; and For40 -CTGGTGCTGGCG-GTGGAGACTT; and one reverse primer Rev584 -GCTCTTGAAGAGGCTGTGTTG. The amplified fragment (584 bp) was purified and fully sequenced in both directions using For22 and Rev584.

A partial mammalian TRPC sequence, originally termed “mTrp4” and previously identified by RT-PCR amplification from mouse brain (21), was employed to obtain the identical cDNA fragment from βTC3 cells. It was then used to screen a mouse βTC3 cell cDNA library. We identified five positive clones that contained mTrp4-related cDNAs. The longest was a 3.9-kb cDNA fragment that overlapped with the other four clones as determined by the restriction map. The sequence of this clone revealed a continuous single open reading frame encoding a protein of 974 amino acids (Fig. 1). The methionine initiating this open reading frame was in a Kozak (25) sequence context compatible with efficient translation and was preceded by an in-frame termination codon. The full-length cDNA clone was originally termed “mTrp4 l” (GenBank U50922). A 2.0-kb clone revealed a 252-bp in-frame deletion corresponding to amino acid 780–864 of mTrp4 l, a potential splice variant. The full-length alternatively spliced form of mTrp4 was termed mTrp4S (U50921) and obtained by interchanging the PstI–NotI fragment with the mTrp4 l clone.

To determine which mouse tissues express mSTRPC4, multiple tissue Northern blots (Clontech) containing poly(A)⁺ RNA prepared from various mouse tissues was hybridized with the same probe used to screen the library. We found that mSTRPC4 was highly expressed in brain as a mRNA doublet species that migrated at ∼4 kb and 7.5 kb, respectively (Fig. 2). No signals were detected in spleen, liver, or kidney. A smear was detected in lung, testis (Fig. 2), and, only after a much longer exposure, skeletal muscle and heart (not shown). Reprobing with another cDNA fragment from a different region of mSTRPC4 resulted in the same hybridization pattern (not shown). Since the β-actin control probe gave sharp bands, it is possible that the smear observed could be due to low homology interactions or nonspecific binding, rather than RNA degradation.

After transfection with pCMV-mSTRPC4 l, or pCMV-mSTRPC4 l-cp fused with the COOH-terminal tag, homogenates displayed strong and specific signals on Western blots, at ∼97 kd, and appropriately higher in the case of the fusion proteins (Fig. 3).

We attempted to demonstrate functional expression of mSTRPC4 l and mSTRPC4S by measuring Ca²⁺ or Sr²⁺ influx. Sr²⁺ has been shown to be preferentially carried by endogenous SOC channels, but not by inositol 1,4,5-trisphosphate (IP₃)–activated channels (26). HEK293 cells express an endogenous muscarinic acetylcholine receptor coupled to G_q protein and phospholipase C. Activation by IP₃ release was achieved by treatment with carbachol. In the absence of extracellular Ca²⁺, application of 100 μmol/l carbachol to EYFP transfected control cells induced a transient rise in [Ca²⁺]_i, due to activation of IP₃ receptors on the endoplasmic reticulum (Fig. 4A1). When cells were treated with 1 μmol/l ionomycin in the absence of extracellular Ca²⁺, a transient rise in [Ca²⁺]_i was also observed as a result of the mobilization of the intracellular Ca²⁺ stores. Ca²⁺ influx induced by the addition of extracellular Ca²⁺ indicated activation of the endogenous SOC channels (e.g., Fig. 4A1), which was substantially decreased if Sr²⁺ was used instead of Ca²⁺ (not shown). A significant increase in carbachol-activated Sr²⁺ influx was detected in the HEK293 cells expressing hSTRPC3 protein (Fig. 4A2), confirming the previous reports of hSTRPC3 functioning as an IP₃-activated channel. Under the same experimental conditions, the level of Sr²⁺ influx after carbachol stimulation was similar in both mSTRPC4 transfected cells and mock-transfected HEK293 cells (Fig. 4A3 and 4A4). Contrary to a previous report (27), mSTRPC4 transfected cells did not display an increased level of Sr²⁺ influx after intracellular stores were depleted by ionomycin or when they were stimulated by carbachol (Fig. 4B and C). A direct comparison of receptor-activated calcium entry (RACE) and CCE in hSTRPC3, mSTRPC4 l, and mSTRPC4S in transfected HEK293 cells was also made. This showed that only the RACE pathway stimulated by carbachol in the hSTRPC3 cells was significantly increased over the controls or other transfectants (Fig. 4C).

As described above, the sequence originally termed TRP7 by Okada et al. (28) has recently been found to be a channel that expresses ADPR pyrophosphatase activity, and now is termed LTRPC2 by virtue of its homology to the long (L) TRP-related family (20). We found that human islets have readily detectable levels of expression of the LTRPC2 mRNA. Three of three NH₂-terminal oligonucleotide primers gave the expected bands at 584, 562, and 544 bp (Fig. 5, lanes 2–4). The largest amplicon was purified and sequenced using internal primers, and the sequence was found to be identical to LTRPC2.

We have cloned and characterized two splice variants of the mouse STRPC4 homolog, originally termed mTrp4 l and mTrp4S. The sequence similarity among STRPC genes is highest in the regions encoding the putative transmembrane domains, whereas the lowest degree of conservation is in the 3′ end. The mTRPC4 protein described here was 41, 42, and 48% identical and 62, 62, and 68% similar to the deduced sequences of Drosophila TRP, TRPL, and human STRPC1, respectively.

In contrast to some of the other TRPCs that are clearly capable of expressing channel activity, expression of STRPC4 has been problematic. Bovine Trp4 was originally described to be an ICRAC-like store-operated channel (29). However, both mouse STRPC4 and STRPC5 have also been reported to function as receptor-activated channels, possibly via phosphorylation signaling rather than via IP₃ (30). Our own studies, using exactly the same sequence and similar expression systems, did not show receptor activation. The mSTRPC4 protein studied here was recognized by NH₂-terminal antipeptide antisera, in both Western blot and immunohistochemical studies on the expressed protein, confirming the sequence of both cDNAs and the deduced mass of the expressed protein. However, recombinant expression of the mSTRPC4 cDNAs in HEK293 cells did not increase the calcium influx induced by either carbachol stimulation or intracellular calcium store depletion. Under the same conditions, the hSTRPC3 channel increased calcium influx evoked by carbachol stimulation and had no effect on endogenous store operated Ca²⁺ influx. The reasons for these conflicting results for STRPC4 function are not obvious, but several explanations are plausible. For example, HEK293 cells have endogenous TRPC protein expression, including human STRPC4 (31). Since it is known that HEK293 cells can vary in their expression of ion channels, it is possible that the particular subline of HEK293 cells expresses a critical subunit that is necessary for STRPC4 expression. This possibility is supported by the detection of nonselective cation channels when STRPC4 was coexpressed with STRPC1 (32).

LTRPC channels are the most recent members of the TRPC family to be described, and their functions are not fully understood. We have also now found that LTRPC2/TRPC7 is expressed in human islets. This channel was recently cloned by employing a Nudix motif homology strategy, a motif that functions as a specific ADPR pyrophosphatase (20). The LTRPC2 channel was shown both to express functional ADPR pyrophosphatases activity and to be a nonselective ion channel that is activated by low micromolar amounts of ADPR. A similar current was found to be present in RINm5f cells dialyzed with ADPR in the patch pipette (reference 33 and A. Flieg, personal communication). This suggests that ADPR, the breakdown product of cyclic ADPR (cADPR), could have a role as an intracellular signaling molecule in islet β-cells. This is the first suggested biological role for ADPR. cADPR has been proposed to be a signaling molecule in insulin secretion by stimulating Ca²⁺ release from ryanodine receptor–regulated intracellular stores (34). In this hypothesis, the increased ATP concentration after glucose stimulation of islets leads to the accumulation of cADPR via the inhibition of the cADPR hydrolase activity of CD38 (ADP-ribosyl cyclase/cyclic ADP-ribose hydrolase). The demonstration of ADPR activation of LTRPC7 suggests that as cADPR is hydrolyzed to ADPR, activation of LTRPC7 could lead to further depolarization of the β-cell. In this way, LTRPC7 could contribute to the depolarization-induced Ca²⁺ entry via l-type, voltage-dependent calcium channels. The ADPR hydrolase activity of LTRPC7 would then make this a self-limiting process. This leads to the prediction that if intracellular concentrations of ADPR are in the micromolar range, inhibition or targeted disruption of the LTRPC7 gene should have specific inhibitory effects on regenerative calcium oscillations and second-phase insulin secretion. Note that, at this meeting, G. Rutter et al. presented evidence that cADPR can release Ca²⁺ from insulin secretory granules via ryanodine receptors, giving credence to the idea that there are separate IP₃- and ryanodine-sensitive pools in insulin-secreting cells. A scheme combining some of these ideas is shown in Fig. 6.

In summary, we report the cloning and expression of mouse STRPC4 cDNA and its alternatively spliced form from a mouse insulinoma cell library. We had previously reported that mSTRPC4 was the most readily detectable STRPC message in βTC3 cells and was also detectable in mouse pancreatic islets (12). When mSTRPC4 was expressed in HEK293 cells, it was not activated by calcium-store depletion or IP₃, despite expression of a full-length protein. We suggest that although STRPC4 is a highly conserved member of the STRPC family, additional subunits or regulatory proteins, such as STRPC1, are necessary for its function. These might be variably expressed in tumor cell lines, such as CHO and HEK293, which have been employed by others to determine STRPC4 function. The finding that LTRPC2 is also expressed in islets suggests that additional mechanisms related to nonselective cation channels in β-cells may be involved in regulating excitability and insulin secretion.

This work was supported by National Institutes of Health Grant PO1544681, RO1DK48494 (L.H.P.), and the Blum-Kovler Foundation. F.Q. was the recipient of a predoctoral fellowship award from the Markey Foundation. The authors would like to gratefully acknowledge I.D. Dukes, M.W. Roe, and J. Worley for previous and ongoing collaborations; K.S. Polonsky, G.I. Bell, Dr. Roland Stein (Vanderbilt University), Dr. Mike Zhu (Ohio State University), and Dr. Craig Montell (Johns Hopkins University) for providing reagents; and Dr. A. Fleig for helpful discussions on LTRPC7.