Expression and function of ryanodine receptors in nonexcitable cells.

We have used reverse transcriptase-polymerase chain reaction to investigate the expression of ryanodine receptors in several excitable and nonexcitable cell types. Consistent with previous reports, we detected ryanodine receptor expression in brain, heart, and skeletal muscle. In addition, we detected ryanodine receptor expression in various other excitable cells including PC 12 and A7r5 cells. Several muscle cell lines (BC3H1, C2C12, L6, and Sol8) weakly expressed ryanodine receptor when undifferentiated but strongly expressed type 1 and type 3 ryanodine receptor isoforms when differentiated into a muscle phenotype. Only 2 (HeLa and LLC-PK1 cells) out of 11 nonexcitable cell types examined expressed ryanodine receptors. Expression of ryanodine receptors at the protein level in these cells was confirmed using [3H]ryanodine binding. We also investigated the function of ryanodine receptors in Ca2+ signaling in HeLa cells using single-cell Fura-2 imaging. Neither caffeine nor ryanodine caused a detectable elevation of cytoplasmic Ca2+ in single HeLa cells. However, ryanodine caused a significant decrease in the amplitude of Ca 2+ signals evoked by repetitive stimulation with ATP. These studies show that ryanodine receptors are expressed in some nonexcitable cell types and furthermore suggest that the ryanodine receptors may be involved in a subtle regulation of intracellular Ca2+ responses.

We have used reverse transcriptase-polymerase chain reaction to investigate the expression of ryanodine receptors in several excitable and nonexcitable cell types. Consistent with previous reports, we detected ryanodine receptor expression in brain, heart, and skeletal muscle. In addition, we detected ryanodine receptor expression in various other excitable cells including PC12 and A7r5 cells. Several muscle cell lines (BC 3 H1, C2C12, L6, and Sol8) weakly expressed ryanodine receptor when undifferentiated but strongly expressed type 1 and type 3 ryanodine receptor isoforms when differentiated into a muscle phenotype. Only 2 (HeLa and LLC-PK1 cells) out of 11 nonexcitable cell types examined expressed ryanodine receptors. Expression of ryanodine receptors at the protein level in these cells was confirmed using [ 3 H]ryanodine binding. We also investigated the function of ryanodine receptors in Ca 2؉ signaling in HeLa cells using single-cell Fura-2 imaging. Neither caffeine nor ryanodine caused a detectable elevation of cytoplasmic Ca 2؉ in single HeLa cells. However, ryanodine caused a significant decrease in the amplitude of Ca 2؉ signals evoked by repetitive stimulation with ATP. These studies show that ryanodine receptors are expressed in some nonexcitable cell types and furthermore suggest that the ryanodine receptors may be involved in a subtle regulation of intracellular Ca 2؉ responses.
Cells have two major mechanisms available for regulating the release of internal Ca 2ϩ . In one case, external signals acting on receptors at the cell periphery generate the second messenger inositol 1,4,5-trisphosphate, which diffuses into the cell and mobilizes Ca 2ϩ by engaging inositol 1,4,5-trisphosphate receptors on the endoplasmic reticulum (1). The other mechanism employs a related, but distinct, family of intracellular channels, the ryanodine receptors (RyRs), 1 so called because they strongly bind the plant alkaloid ryanodine (2). RyRs were first identified as the intracellular calcium channels responsible for releasing calcium from the sarcoplasmic reticu-lum of both skeletal and cardiac muscle. The two striated muscle RyR isoforms have been designated RyR1 (skeletal muscle) and RyR2 (cardiac muscle). A recently identified isoform, referred to as RyR3, appears to be a major isoform in brain and smooth muscle (3)(4)(5).
Evidence for the expression of RyRs in various excitable and nonexcitable tissues has mostly been obtained using either molecular or pharmacological methods. For many excitable cells, these approaches have demonstrated the tissue-specific expression and function of RyRs (Refs. 6 and 7; for a review see Ref. 8). However, the situation concerning the expression of RyRs in some nonexcitable tissues and their possible participation in agonist-stimulated calcium signals is far from clear (8). In hepatocytes, for example, effects of ryanodine on agonistinduced calcium signals in intact cells (9, 10) and specific ryanodine binding sites in hepatocyte vesicles (11,12) have been demonstrated. However, the pharmacology of the putative RyR in hepatocytes is very different from those expressed in muscle tissues (9,13), and furthermore, the expression of RyR mRNA in hepatocytes has not been detected using molecular techniques (7,11,14,15).
For other nonexcitable cells the converse problem exists, in that molecular techniques have identified RyR mRNA expression, but the functional evidence has been confusing, since some RyR-activating agents fail to evoke responses in these cells. For example, in mink lung epithelial cells and Jurkat T-lymphocytes, which appear to express RyR3, effects of ryanodine but not caffeine have been observed (3,16). In a separate study Guse et al. (17) found that Jurkat cells were caffeine-responsive.
In the present study we have examined RyR expression in a variety of tissues using RT-PCR. In addition to finding RyRs expressed in several excitable tissues and cell lines, we detected RyR expression in two nonexcitable cell types. For one of these nonexcitable cell types (HeLa cells), we investigated the function of the expressed RyRs using single-cell Fura-2 imaging.

MATERIALS AND METHODS
Cells-A7r5 embryonic rat aorta cells, RBL-2H3 rat mucosal mast cells, GH 3 rat pituitary tumor cells, CH 3 H10T 1 / 2 and Swiss 3T3 embryonic mouse fibroblasts, L cells from mouse connective tissue, BC 3 H1 mouse embryonic myoblasts, Rin m5F rat insulinoma cells, and PC12 pheochromocytoma cells were obtained and cultured as described by De Smedt et al. (18), Jurkat T cell leukemia cells as described by Parys et al. (19), HeLa carcinoma cells as described by Bootman et al. (20), and LLC-PK1 renal epithelial cells as described by Parys et al. (21). L6 cells and C2C12 cells were obtained from the ATCC (CRL 1458 and CRL 1772, respectively) and cultured as described by Florini and Magri (22). C6 cells were obtained from the ATCC (CCL 107) and cultured using the recommended conditions. Sol8 cells were a gift from Dr. C. Pinset (Institut Pasteur, Centre National de la Recherche Scientifique, Paris, France), and HUVEC cells freshly isolated from human umbilical vein * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Funded by the BBSRC and The Isaac Newton Trust. To whom correspondence should be addressed.  1 The abbreviations used are: RyR, ryanodine receptor; PCR, polymerase chain reaction; RT-PCR, reverse transcription-PCR; bp, base pair(s).
were provided by Dr. D. Collen (Centre for Molecular and Vascular Biology, KU Leuven, Belgium). PC12 cells were differentiated as described previously (23). Sol8 and BC 3 H1 cells were differentiated by reducing the concentration of fetal calf serum in the growing medium to 0.5%. For differentiation of C2C12 cells, cells were grown in medium containing 1% horse serum and 0.5% insulin transferrin selenite. Differentiation of L6 cells was evoked by growing cells in 5% horse serum. For single-cell imaging of HeLa cells, cells were transferred from plastic culture dishes to glass coverslips (22 mm diameter, Chance Propper Ltd, Smethick, Warley, UK). Cells were allowed to attach to the coverslips for 48 h before use.
Reverse Transcription and PCR Amplification-Total RNA was prepared from different tissues and cell lines according to Chirgwin et al. (24). Random primed first strand cDNA was synthesized from 1 g of RNA using avian myeloblastosis virus reverse transcriptase (Boehringer Mannheim). RyR cDNAs were amplified using degenerate primers based on sequences conserved among the three RyR isoforms (rabbit RyR1, RyR2, RyR3; human RyR1, RyR3; pig RyR1) and corresponding to the putative membrane-spanning regions M3 and M4 (25). The sequences of the primers and their position within the rabbit RyR sequences are shown below. Reactions were carried out in a Biometra UNO-Thermoblock for 30 cycles using a denaturing step for 1 min at 94°C, annealing for 1 min at 50°C, and extension for 1 min at 72°C. This was followed by a final extension step at 72°C for 10 min.
Cloning and Sequencing of RT-PCR Products-PCR products were fractionated on a 1.5% agarose gel, and individual bands were purified using the Geneclean II DNA purification kit (Bio 101, Inc.). The purified DNA was cloned into the plasmid vector pGEM-T (Promega) according to the manufacturer's instructions, and clones containing inserts of the expected size were sequenced in both directions by dideoxy sequencing using the Sequenase version 2.0 kit (Amersham Corp.).
Restriction Enzyme Analysis of RT-PCR Products-RT-PCR products were subjected to restriction enzyme analysis to determine which isoforms were present. For mouse RT-PCR products, three restriction enzymes, SacI, FokI, and AvaII, were chosen. SacI was predicted to cut RyR1 only, to produce fragments of approximately 395 and 135 base pairs (bp). FokI was predicted to cut RyR1 into 500-and 30-bp fragments and cut RyR2 into 200-, 180-, and 150-bp fragments. AvaII was expected to cut RyR1 into 270-and 260-bp fragments and RyR3 into 400-and 130-bp fragments. For rat RT-PCR products, the restriction enzymes SacI, BglII, and EaeI were chosen. SacI should cut RyR1 only into 395-and 135-bp fragments, BglII should cut RyR2 only into 295and 235-bp fragments, and EaeI should cut RyR3 only into 350-and 180-bp fragments. For human and pig RT-PCR products, the restriction enzymes SacI, BglII, and BstEII were chosen. SacI was expected to cut RyR1 only into 300-, 100-, and 130-bp fragments, BglII was expected to cut RyR2 only into 295-and 235-bp fragments, and BstEII was expected to cut RyR3 only into 280-and 250-bp fragments.

Microsomal Preparation and [ 3 H]Ryanodine
Binding-Total microsomes were isolated from the various cell lines and from whole rabbit brain as described previously (19). The binding assay was based on the optimized conditions described for ryanodine binding to the rabbit brain ryanodine receptor (26). Binding to microsomes (1.25 mg/ml) was measured in 10 mM Hepes/NaOH (pH 7. Single-cell Fura-2 Imaging-The culture medium was replaced with an extracellular medium containing 121 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl 2 , 1.8 mM CaCl 2 , 6 mM NaHCO 3 , 5.5 mM glucose, 25 mM Hepes (pH 7.3). Cells were loaded with 2 M Fura-2 acetoxymethylester (Molecular Probes Inc.) by incubation for 35-40 min at room temperature (20°C) followed by an extracellular medium wash and a further 20-min incubation to allow de-esterification of the loaded dye. Coverslips were mounted at room temperature on the stage of a Nikon diaphot inverted epifluorescence microscope. Fluorescent images were obtained at video rate (40 ms each wavelength) by alternate excitation at 340 or 380 nm using an image-processing system (Imagine, Synoptics Ltd, Cambridge, UK) interfaced to a DEC MicroVAX II microcomputer and then filtered with a 200-ms time constant. The emission signal at 510 nm was collected by a charge-coupled device intensifying camera (Photonic Science, Robertsbridge, Kent, UK), and the digitized signals were stored and processed as described previously (28). The responses shown in the present study were obtained from four independent trials for each experiment.

RT-PCR Amplification of RyR Isoforms-To search for
RyRs in a range of excitable and nonexcitable cell types from different species we used a RT-PCR assay. In this technique, RNA is used as a template for an initial reverse transcription step to produce cDNA, followed by amplification of specific sequences using PCR. By aligning the previously published RyR sequences (see "Materials and Methods"), we designed degenerate PCR primers to amplify an approximately 530-bp product from the 3Ј region of each RyR isoform. This region is highly conserved between isoforms and species but, importantly, also contains sufficient differences to allow the three RyR isoforms to be distinguished. The intron-exon structure of this region has been determined for two RyR isoforms. In the case of pig RyR1 genomic DNA, the primers would be expected to amplify a fragment that spans 4 introns (approximately 1700 bp), 2 and within human RyR3 genomic DNA, the primers should amplify a fragment that spans at least 1 intron. 3 Interestingly, the intron identified within the human RyR3 gene is in the same position as an intron in the pig RyR1 gene, suggesting that some intron splice sites may be conserved between RyR isoforms and animal species. Therefore, although the intron-exon structure of all three RyR isoforms from each species tested has not been determined, it seems unlikely that PCR amplification of any contaminating genomic DNA would produce the correctly sized PCR product (530 bp).
As the majority of cell types tested were of mouse, rat, or human origin, and all three RyR isoforms have been found in various regions of the brain (7, 31), we tested the primers by carrying out PCR amplification of cDNA from mouse, rat, and human brain. A product of the predicted size (530 bp) was amplified from all three brain cDNAs (Fig. 1A). Control reactions containing H 2 O or RNA were negative in the PCR (Fig.  1A). Cloning and sequencing the brain PCR products revealed that the primers were able to specifically amplify all three RyR isoforms. The primers were also able to specifically amplify RyRs from mouse heart (RyR2 and RyR3) and skeletal muscle (RyR1 and RyR3) (data not shown).
A range of cell types was subjected to PCR amplification using these primers (Fig. 1B). Only two of the nonexcitable cell types tested, HeLa and LLC-PK1 cells, yielded detectable PCR products. The other nonexcitable cell types, including fibroblasts and Jurkat cells, appeared negative in the PCR. As a positive control for the quality of the cDNA, amplification reactions were also carried out using ␤-actin PCR primers. Products of the correct size and of equivalent intensity were ampli-fied from all cDNAs tested in this study (data not shown).
Using the RyR primeres, products of the predicted size were detected in most excitable cell types, including differentiated and undifferentiated PC12 cells and A7r5 cells. A very faint PCR product was detected in GH 3 and Rin m5F cells, indicating that these cell types probably also express RyRs, but very weakly. An interesting pattern of expression was detected in the BC 3 H1, Sol8, C2C12, and L6 muscle cell lines. By adjusting the culture conditions appropriately (see "Materials and Methods"), these cells can either be maintained in a rapidly growing non-muscle state or differentiated into a muscle phenotype. In the undifferentiated cells, we detected only very weak expression of RyR by PCR. However, differentiation of the cells correlated with the appearance of much more abundant levels of the 530-bp product (Fig. 1B).
Tissue-specific Distribution of RyR Isoforms-The amplification products obtained from most of the PCR-positive tissues were cloned and sequenced to determine which isoforms were expressed. Although we sequenced between 5 and 20 clones obtained from these PCR-positive cell types, in some only one or two RyR isoforms were detected. To confirm that these were the only isoforms amplified, we performed restriction digests on these PCR products, using enzymes specific for each isoform (see "Materials and Methods"). Fig. 2 shows the restriction digest analysis of RyR PCR products amplified from differentiated BC 3 H1 cells, differentiated L6 cells, and HeLa cells. The BC 3 H1 PCR product was partially cut with SacI (RyR1-specific) and completely cut with AvaII (digests RyR1 and -3) to produce fragments of the expected sizes, suggesting that BC 3 H1 cells express RyR1 and RyR3. The L6 PCR product was completely cut with SacI and not cut by either BglII (RyR2-specific) or EaeI (RyR3-specific). The HeLa cell PCR product was completely cut with BglII

FIG. 1. RT-PCR detection of RyRs in excitable and nonexcitable cells. RT-PCR was performed as described under "Materials and
Methods." Samples (one-tenth of each PCR reaction) were subjected to electrophoresis through a 1.5% agarose gel. A, RT-PCR detection of a 530-bp fragment of different RyR isoforms in mouse, rat, and human brain cDNA but not in control reactions containing H 2 O or brain RNA. B, RT-PCR detection of RyRs in various excitable and nonexcitable cell types. The results shown are representative of four different experiments performed with two different RNA preparations.

FIG. 2. Restriction enzyme digest of RyR RT-PCR products.
RT-PCR products were produced as described under "Materials and Methods." PCR products were digested with the restriction enzymes identified using the manufacturer's suggested protocols, and the fragments produced were subjected to electrophoresis through a 1.5% agarose gel. A, differentiated BC 3 H1 cells. B, differentiated L6 cells. C, HeLa cells.
(RyR2-specific) to produce fragments of the expected size. The combined results obtained from sequencing and restriction enzyme digest of the PCR products are presented in Table I.
[ 3 H]Ryanodine Binding-The finding that some nonexcit-able cells expressed RyR mRNA was surprising, since RyRs were previously thought to be restricted to excitable tissues. In addition, there is conflicting evidence on the expression of ryanodine receptors in A7r5 cells (32,33). We therefore sought to confirm the PCR results and to obtain evidence for RyR expression at the protein level, using a [ 3 H]ryanodine binding assay. Fig. 3 compares the [ 3 H]ryanodine binding for several cell types. The data presented in Fig. 3 are not strictly quantitative, since the different RyR isoforms may have variable affinities for ryanodine. However, the data presented qualitatively demonstrate the expression of RyRs and furthermore may reflect the relative abundance of the proteins if the binding affinities were similar. All the cell lines tested appeared to have much lower [ 3 H]ryanodine binding than in rabbit brain, suggesting that these cells lines have either lower affinity RyR or lower RyR density than brain. The undifferentiated PC12 cells, which display robust responses to RyR agonists such as caffeine (34), had 5.7 times less ryanodine binding activity than brain. A7r5, HeLa cells, and LLC-PK1 epithelial cells displayed a much lower (16 -26 times less than brain), but significant, binding activity. The specific [ 3 H]ryanodine binding to Jurkat cells was negligible (not shown), which is consistent with the PCR experiments in which we were unable to detect RyR ex-  LLC-PK1 kidney proximal tubule ϩ ϩ a Cell types that were positive in the PCR are indicated with a plus (ϩ), and those that were negative with a minus (Ϫ).
b The particular isoforms that were detected by sequencing/restriction digestion are indicated with a plus (ϩ).
c Specific isoforms expressed not determined.
pression in Jurkat cells.

Attenuation of HeLa Cell Ca 2ϩ Responses by Ryanodine-
The PCR and restriction digest analysis ( Fig. 2 and Table I) and [ 3 H]ryanodine binding data (Fig. 3) suggested that HeLa cells express RyR2. We investigated the potential contribution of these channels to intracellular calcium signals using singlecell imaging. Application of caffeine (2-40 mM) to the cells did not evoke any measurable [Ca 2ϩ ] i increase (Fig. 4A). In addition, treatment with ryanodine (10 M) failed to give detectable changes in [Ca 2ϩ ] i when applied alone or in combination with 10 mM caffeine (Fig. 4B).
The lack of caffeine effect suggested that the HeLa cells either expressed a caffeine-insensitive RyR isoform or had a low RyR density. We therefore altered the experimental protocol to maximize the potential effect of the limited number of RyRs. HeLa cells were repetitively stimulated with ATP (100 M), which mobilizes Ca 2ϩ from intracellular stores in HeLa cells (35), either in the absence or presence of ryanodine (10 M) (Fig. 5, A and B, respectively). The averaged [Ca 2ϩ ] i values for peaks 2-7 are shown in Fig. 5, C and D. In control cells, the magnitude of the response to repeated ATP applications decayed only slightly, probably due to desensitization of the Ca 2ϩsignaling pathway (35). The magnitude of ATP-evoked responses in the cells treated with ryanodine diminished significantly, so that the response decayed to about 50% after six ATP applications. The latency of the response to ATP slightly increased in the presence of ryanodine (Fig. 5E).

DISCUSSION
The aim of the present study was to investigate the possible expression and function of RyRs in nonexcitable cells. To confirm that our PCR technique could detect RyR expression, we also analyzed several tissues known to have functional RyRs. Consistent with other studies (7, 31), we could amplify and clone all three RyR isoforms from brain samples ( Fig. 1A and Table I). In addition, we found RyR expression in skeletal and cardiac muscle. RyR3 appeared to be expressed in both these muscle types, whereas RyR1 and RyR2 were expressed in skeletal and cardiac muscle, respectively (Table I).
Of particular interest were the myogenic cell lines BC 3 H1, Sol8, C2C12, and L6, where the intensity of the 530-bp PCR product correlated with the differentiation of the cells from a non-muscle to a muscle phenotype. Although the PCR method we used was not quantitative, for a limited number of PCR cycles the intensity of the 530-bp PCR product may give an indication of the level of RyR mRNA expression. Our observations are consistent with several other studies linking muscle differentiation with RyR expression (7,23,36). It appears that In C the magnitude of the response was calculated by normalizing the area of each ATP-evoked Ca 2ϩ response relative to the area of peak 2. In D the amplitude of each ATP-evoked Ca 2ϩ response was calculated by normalizing the maximum [Ca 2ϩ ] i increase during each response relative to the height of peak 2. The graph in E shows the latency of the response to ATP (time from ATP application to peak) in the presence and absence of ryanodine. The data points are mean Ϯ S.E. of 39 (f) and 63 (q) cells, respectively. *, statistically different with p Ͻ 0.002. the myogenic cells express RyR1 (L6) or RyR1 and RyR3 (BC 3 H1, Sol8, and C2C12) after differentiation, similar to the situation in skeletal muscle (7) (Table I). Both undifferentiated and nerve growth factor-differentiated PC12 cells were RyRpositive (Table I). However, there was an apparent change in isoform expression pattern from RyR1 and RyR2 in undifferentiated cells to only RyR1 in differentiated cells. The expression of RyR1 in undifferentiated PC12 cells may explain the observation that theophylline is equipotent with caffeine at stimulating Ca 2ϩ release from this clone (34), since RyR1 appears equally sensitive to these two methylxanthines (37).
Several previous studies have suggested that RyR expression is growth state-dependent; for example, aortic smooth muscle cells lose caffeine sensitivity during their logarithmic proliferation phase, which can be reversed or augmented by the removal or addition of growth factors, respectively (38). Similarly, the expression of RyR3 in mink lung epithelial cells can be induced by exposure to TGF-␤ (3). These data suggest that RyR expression can be somewhat labile and may be associated with the state of cell proliferation and differentiation.
After confirming that we could detect RyR expression in these excitable tissues, we investigated the potential expression of these receptors in 11 nonexcitable cell types, many of which are commonly used for studies of Ca 2ϩ signaling (Table  I). Of these cell types, only two appeared to express RyR mRNA. Increasing the number of PCR cycles or performing sequential PCR reactions, using the first reaction to prime the second, did not yield detectable products for the cell types that were RyR-negative. The group of cell types that were found not to express RyRs surprisingly included Jurkat T-lymphocytes (Table I). This result is consistent with the negligible [ 3 H]ryanodine binding displayed by Jurkat cells (data not shown). These data contrast with other studies, which suggested that RyRs were expressed in these cells (16,17). The explanation for our contrasting data is unclear but may reflect variable RyR expression between different Jurkat cell clones.
The two RyR-positive nonexcitable cell types, HeLa and LLC-PK1 cells, expressed RyR2 and RyR3, respectively. RyR expression in HeLa cells was also recently reported by Giannini et al. (7), although they identified RyR2 and RyR3 in their control HeLa cells. The expression of only RyR2 in our HeLa cells was confirmed by sequencing and restriction digestion of PCR products and may again point to a clonal variation. To extend the molecular characterization of RyR expression in these tissues, we sought to obtain evidence for RyR at the protein level. HeLa and LLC-PK1 cells were found to display significant levels of [ 3 H]ryanodine binding, although it was about 17-fold less than in rabbit brain and approximately onethird of that found in undifferentiated PC12 cells (Fig. 3).
Despite the molecular evidence and [ 3 H]ryanodine binding data, which clearly indicate that HeLa cells express RyRs, we were unable to directly demonstrate a RyR-dependent [Ca 2ϩ ] i increase in the cells (Fig. 4). Neither caffeine nor ryanodine evoked a measurable increase in [Ca 2ϩ ] i when applied either on their own or in combination. A similar lack of caffeine responsiveness in HeLa cells was previously shown (39,40). However, ryanodine caused a progressive decrease in the magnitude of the Ca 2ϩ signal evoked by repetitive ATP applications and a slight increase in the latency before the [Ca 2ϩ ] i rise (Fig. 5). The decreased magnitude of the Ca 2ϩ signals is consistent with the previously described use-dependent block of RyR function by ryanodine (34,41,42), whereby the ryanodine-bound receptors remain in a constitutively open low conductance state. The finding that ryanodine inhibited ATP-induced Ca 2ϩ signals in a use-dependent manner suggests that hormonal stimulation of HeLa cells brings about activation of RyRs to amplify the normal inositol 1,4,5-trisphosphate-dependent elevation of Ca 2ϩ . Just how this channel opening is achieved is unknown, but possibilities include activation via Ca 2ϩ -induced Ca 2ϩ release (8) or production of a RyR-sensitizing ligand such as cyclic adenosine diphosphate-ribose (43).
The apparent paradox between the effects of ryanodine and caffeine suggests that HeLa cells either express a caffeineinsensitive RyR isoform or express a very low level of RyR, such that an acute opening of the RyRs does little to influence [Ca 2ϩ ] i , but a prolonged RyR opening can gradually deplete the stores. The latter explanation seems to be the more likely, since we detected a low density of RyR in the [ 3 H]ryanodine binding studies (Fig. 3), and all RyR isoforms have been shown to be caffeine-sensitive (44). Furthermore, similar results i.e. apparent caffeine-insensitivity and a ryanodine-induced Ca 2ϩ store depletion, have been reported by Giannini et al. (3) for mink lung epithelial cells expressing RyR3. It seems likely that a low density of RyR may explain the lack of effect of caffeine on [Ca 2ϩ ] i in several cell types shown by molecular techniques to express RyRs (7).
The function of the RyRs expressed in nonexcitable cells is not fully established. In a few cell types, such as sea urchin eggs (29) and pancreatic acinar cells (30), RyRs may contribute to the initiation of Ca 2ϩ signals. In hepatocytes, the recruitment of RyRs has been reported to be agonist-specific (9). Data from the present study suggest that RyRs in nonexcitable HeLa cells may provide a subtle regulation of the magnitude and kinetics (Fig. 5) of hormone-evoked [Ca 2ϩ ] i responses. These data indicate that RyRs make an important contribution to intracellular Ca 2ϩ signals in a variety of nonexcitable cell types.