Temporal switching and cell‐to‐cell variability in Ca 2+ release activity in mammalian cellsстатья из журнала
Аннотация: Article17 March 2009Open Access Temporal switching and cell-to-cell variability in Ca2+ release activity in mammalian cells Naotoshi Nakamura Naotoshi Nakamura Department of Pharmacology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan Search for more papers by this author Toshiko Yamazawa Toshiko Yamazawa Department of Pharmacology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan Search for more papers by this author Yohei Okubo Yohei Okubo Department of Pharmacology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan Search for more papers by this author Masamitsu Iino Corresponding Author Masamitsu Iino Department of Pharmacology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan Search for more papers by this author Naotoshi Nakamura Naotoshi Nakamura Department of Pharmacology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan Search for more papers by this author Toshiko Yamazawa Toshiko Yamazawa Department of Pharmacology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan Search for more papers by this author Yohei Okubo Yohei Okubo Department of Pharmacology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan Search for more papers by this author Masamitsu Iino Corresponding Author Masamitsu Iino Department of Pharmacology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan Search for more papers by this author Author Information Naotoshi Nakamura1, Toshiko Yamazawa1, Yohei Okubo1 and Masamitsu Iino 1 1Department of Pharmacology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan *Corresponding author. Department of Pharmacology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan. Tel.: +81 3 5841 3414; Fax: +81 3 5841 3390; E-mail: [email protected] Molecular Systems Biology (2009)5:247https://doi.org/10.1038/msb.2009.6 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Genetically identical cells in a uniform external environment can exhibit different phenotypes, which are often masked by conventional measurements that average over cell populations. Although most studies on this topic have used microorganisms, differentiated mammalian cells have rarely been explored. Here, we report that only approximately 40% of clonal human embryonic kidney 293 cells respond with an intracellular Ca2+ increase when ryanodine receptor Ca2+ release channels in the endoplasmic reticulum are maximally activated by caffeine. On the other hand, the expression levels of ryanodine receptor showed a unimodal distribution. We showed that the difference in the caffeine sensitivity depends on a critical balance between Ca2+ release and Ca2+ uptake activities, which is amplified by the regenerative nature of the Ca2+ release mechanism. Furthermore, individual cells switched between the caffeine-sensitive and caffeine-insensitive states with an average transition time of approximately 65 h, suggestive of temporal fluctuation in endogenous protein expression levels associated with caffeine response. These results suggest the significance of regenerative mechanisms that amplify protein expression noise and induce cell-to-cell phenotypic variation in mammalian cells. Synopsis Biochemical processes in cells are inherently noisy. This is due in part to the stochastic nature of gene expression systems, which typically involve small numbers of molecules such as DNA, mRNA and proteins (Kaern et al, 2005; Kaufmann and van Oudenaarden, 2007; Pedraza and Paulsson, 2008). In addition to such 'intrinsic noise', the internal states of cells and the structure of the signalling pathway also contribute to the fluctuation in the concentration of molecules, collectively termed 'extrinsic noise' (Hooshangi et al, 2005; Pedraza and van Oudenaarden, 2005; Rosenfeld et al, 2005; Shahrezaei et al, 2008). Intracellular noise can be exploited to play roles such as in the amplification of signals, the divergence of cell fates and the diversification of phenotypes (Arkin et al, 1998). In support of this idea, several recent reports suggested that individual clonal cells in the same external environment can exhibit qualitatively different phenotypes (Rao et al, 2002; Raser and O'Shea, 2005; Acar et al, 2008). Whereas most studies of intracellular noise so far have focused on unicellular organisms such as Escherichia coli or Saccharomyces cerevisiae, the importance of stochastic processes in multicellular organisms is now widely recognised (Laslo et al, 2006; Wernet et al, 2006; Chang et al, 2008). However, there are very few studies of differentiated cells of multicellular organisms (Ravasi et al, 2002; Feinerman et al, 2008), and the possibility and significance of such cells exhibiting different phenotypes in identical environments have rarely been discussed (Sigal et al, 2006; Cohen et al, 2008). The calcium ion (Ca2+) is a ubiquitous intracellular messenger that regulates a diverse array of cellular functions, such as muscle contraction, secretion, fertilisation, immune responses, gene expression and synaptic plasticity (Berridge et al, 2003). The endoplasmic reticulum (ER) is the major intracellular Ca2+ store, from which Ca2+ is released via two families of Ca2+ release channels: ryanodine receptors (RyRs) and inositol 1,4,5-trisphosphate receptors (IP3Rs). RyRs are activated by Ca2+ released by themselves (Endo, 1977)—a mechanism known as Ca2+-induced Ca2+ release (CICR). IP3Rs are also activated by Ca2+ in the presence of IP3 (Iino, 1990; Bezprozvanny et al, 1991; Finch et al, 1991). As Ca2+ response via these channels involves such a positive feedback, individual cells of the same type may show different Ca2+ responses with the amplification of intracellular noise. Here, we report that only approximately 40% of clonal human embryonic kidney 293 cells respond with an intracellular Ca2+ increase when RyRs are maximally activated by caffeine. The expression levels of RyRs and sarco/endoplasmic reticulum Ca2+ ATPase (SERCA) showed a unimodal distribution, which could not explain the existence of two populations in terms of the RyR-mediated Ca2+ response. Using a mathematical model of Ca2+ release via RyRs, we investigated a mechanism that generates two distinct phenotypes but no significant difference in the amount of molecular components. Model simulations predicted that the difference in the RyR-mediated Ca2+ response depends on a critical balance between Ca2+ release and Ca2+ uptake activities, which is amplified by the regenerative nature of the CICR (Figure 3). This prediction was confirmed by the modulation of RyR-mediated Ca2+ response by changing the activities of RyRs and SERCAs. Fluctuation of protein concentrations in individual cells over time (Rosenfeld et al, 2005; Cai et al, 2006; Sigal et al, 2006; Yu et al, 2006) points to the possibility that individual HEK293 cells switch between the caffeine-sensitive and caffeine-insensitive states. Long-term Ca2+ imaging with GCaMP2, a genetically encoded fluorescent Ca2+ indicator (Tallini et al, 2006) revealed that individual cells switch between the caffeine-sensitive and caffeine-insensitive states in a flip-flop manner with an average transition time of ∼65 h, suggestive of temporal fluctuation in endogenous protein expression levels associated with caffeine response (Figure 6). These results suggest the significance of regenerative mechanisms that amplify protein expression noise and induce cell-to-cell phenotypic variation in mammalian cells. Introduction Biochemical processes in cells inevitably fluctuate owing in part to the stochastic nature of gene expression systems, which typically involve small numbers of molecules such as DNA, mRNA and proteins (Kaern et al, 2005; Kaufmann and van Oudenaarden, 2007; Pedraza and Paulsson, 2008). In addition to such 'intrinsic noise', the internal states of cells and the structure of the signalling pathway also contribute to the fluctuation in the concentration of molecules, collectively termed 'extrinsic noise' (Hooshangi et al, 2005; Pedraza and van Oudenaarden, 2005; Rosenfeld et al, 2005; Shahrezaei et al, 2008). In some cases, intracellular noise in individual cells is filtered so that the system as a whole is precisely regulated, as is observed in the segmentation of a Drosophila melanogaster embryo (Houchmandzadeh et al, 2002; Gregor et al, 2007) and in circadian rhythms (Forger and Peskin, 2005; Gonze and Goldbeter, 2006). Yet in other cases, intracellular noise can be exploited to play roles such as in the amplification of signals, the divergence of cell fates and the diversification of phenotypes, as seen in the lysis/lysogeny decision circuit of the bacteriophage lambda (Arkin et al, 1998; Skupin et al, 2008). In support of this idea, several recent reports suggested that individual clonal cells in the same external environment can exhibit qualitatively different phenotypes, which may confer a selective advantage in adapting to changing external environments (Rao et al, 2002; Raser and O'Shea, 2005; Acar et al, 2008). Such phenomena question the implicit assumption behind cell-population-wide experiments that genetically identical cells are phenotypically identical (Levsky and Singer, 2003) and highlight the need for measuring individual cells. Thus far, most studies of intracellular noise have focused on unicellular organisms such as Escherichia coli or Saccharomyces cerevisiae. Earlier studies explored the origin of intracellular noise using artificial gene circuits (Elowitz et al, 2002; Ozbudak et al, 2002; Blake et al, 2003; Raser and O'Shea, 2004), whereas recent studies showed various phenotypic diversities in naturally arising biological systems (Samadani et al, 2006; Di Talia et al, 2007; Maamar et al, 2007; Nachman et al, 2007; Schultz et al, 2007; Suel et al, 2007). In multicellular organisms, the importance of stochastic processes is widely recognised in several biological systems in development, including haematopoietic lineage differentiation and retinal colour-vision mosaic development (Laslo et al, 2006; Wernet et al, 2006; Chang et al, 2008). However, there are very few studies of differentiated cells of multicellular organisms (Ravasi et al, 2002; Feinerman et al, 2008), and the possibility and significance of such cells exhibiting different phenotypes in identical environments have rarely been discussed (Sigal et al, 2006; Cohen et al, 2008). The calcium ion (Ca2+) is a ubiquitous intracellular messenger that regulates a diverse array of cellular functions, such as muscle contraction, secretion, fertilisation, immune responses, gene expression and synaptic plasticity (Berridge et al, 2003). The endoplasmic reticulum (ER) is the major intracellular Ca2+ store, from which Ca2+ is released via two families of Ca2+ release channels: ryanodine receptors (RyRs) and inositol 1,4,5-trisphosphate receptors (IP3Rs). RyRs are activated by Ca2+ released by themselves (Endo, 1977)—;a mechanism known as Ca2+-induced Ca2+ release (CICR). IP3Rs are also activated by Ca2+ in the presence of IP3 (Iino, 1990; Bezprozvanny et al, 1991; Finch et al, 1991). CICR is a positive feedback mechanism that amplifies microscopic Ca2+ release and helps Ca2+ signals propagate throughout the cell. As Ca2+ response via these channels involves such a positive feedback, individual cells of the same type may show different Ca2+ responses with the amplification of intracellular noise. Here, we report that human embryonic kidney 293 (HEK293) cells can be an excellent model system for studying the phenotypic diversity of clonal cells, owing to their interesting feature that only approximately 40% of them respond to caffeine with Ca2+ release via RyRs, whereas all of them release Ca2+ via IP3Rs in response to a purinergic agonist, ATP. Our present results suggest that Ca2+ responses to caffeine have threshold characteristics, dictated by a critical balance between the Ca2+ release and uptake activities of the Ca2+ store as well as the regenerative mechanism of CICR, which amplifies the small cell-to-cell differences of protein expression. Furthermore, we observed that with time, individual cells switch between the caffeine-sensitive and caffeine-insensitive states in a flip-flop manner. These results suggest that not only fluctuation in protein expression but also its amplification mechanisms contribute to phenotypic diversity in mammalian cells. Results Purinergic and caffeine responses in HEK293 cells When stimulated with ATP (a purinergic agonist) at the maximal concentration (10 μM), almost all (>99%) of the HEK293 cells responded with a transient increase in intracellular Ca2+ concentration (Figure 1A–D; Supplementary Figure 1B and Supplementary Movie 1), demonstrating their Ca2+ release activity via IP3Rs. On the other hand, when HEK293 cells were stimulated with 25 mM caffeine (an activator of RyRs), approximately 40% of the cells responded with Ca2+ release, whereas the rest did not (Figure 1A–C; Supplementary Figure 1C and Supplementary Movie 2). We observed no distinct spatial heterogeneity in Ca2+ response within individual cells at the present spatial and temporal resolution (Supplementary Figure 2). The proportion of cells responding to caffeine increased up to a maximum of approximately 40% as caffeine concentration was increased from 1 to 30 mM (Figure 1E). To confirm that the caffeine-induced Ca2+ release is mediated by RyRs, we applied a plant alkaloid, ryanodine, along with caffeine to deplete the intracellular Ca2+ store via RyRs (Iino et al, 1988; Oyamada et al, 1993). A measure of 30 μM ryanodine plus 25 mM caffeine application indeed abolished the caffeine response (Figure 1F). To exclude the possibility that the cells are a mixture of multiple clones, we cloned the cells by limiting dilution, which yielded subclones with a similar cell-to-cell heterogeneity (Figure 1G). Thus, it was shown that HEK293 cells contain two types of cell that are genetically identical but phenotypically different: 'caf-positive' cells (with Ca2+ response via RyRs) and 'caf-negative' cells (without Ca2+ response via RyRs). We found a similar cell-to-cell variability in smooth muscle cells in the portal vein of guinea pigs (Supplementary Figure 3). Figure 1.Ca2+ responses in individual HEK293 cells. Measurement of [Ca2+]i after agonist stimulation. (A, B) Time courses of [Ca2+]i in two different cells ((A) caf-positive and (B) caf-negative cells). (C) Scatter plot of amplitude of ATP response (Δ[Ca2+]i) versus amplitude of caffeine response (Δ[Ca2+]i). Each point corresponds to each cell in the same imaging field (n=49 cells). (D, E) The proportion of cells responding to ATP increased up to a maximum of approximately 100% as ATP concentration was increased from 0.1 to 10 μM; the proportion of cells responding to caffeine increased up to a maximum of approximately 40% as caffeine concentration was increased from 1 to 30 mM. Data are means±s.e.m. (n=4 imaging fields including 230–280 cells). (F) Application of ryanodine plus caffeine abolished the caffeine response. Data are representative of n=370 cells. (G) Maintenance of intercellular heterogeneity in caffeine response. Proportions of caf-positive cells in HEK293 cells and their two subclones. Data are means±s.e.m. (n=6 dishes for each). A total of 379 HEK293 cells, 330 subclone 1 cells and 367 subclone 2 cells were analysed. There were no significant differences in the proportion of caf-positive cells among the three groups (p=0.28, single-factor ANOVA). Download figure Download PowerPoint Cell-to-cell variation does not depend on cell cycle or cell morphology We searched for the mechanism that allows clonal cells to exhibit two different phenotypes. We first examined whether the proportion of caf-positive cells depends on the cell cycle. However, cells synchronised at either G0/G1, G1/S or G2/M phase had the same proportion of caf-positive cells (Supplementary Figure 4A). Furthermore, the proportion of caf-positive cells did not differ significantly with the temperature (room temperature versus 37°C) at which caffeine response was measured (Supplementary Figure 4B). We also analysed the dependence of the magnitude of caffeine response on the extent of cell–cell contact. However, no significant correlation was observed (Supplementary Figure 4C). Neither did we observe any significant dependence on cell morphology (size, perimeter and ellipticity) (Supplementary Figure 4D–F). Immunocytochemistry of RyRs and SERCAs Several recent reports indicated that, even among genetically identical cells in the same environment, the level of gene expression can quantitatively differ depending on intracellular noise (Blake et al, 2003; Raser and O'Shea, 2004). Therefore, we examined the possibility that the protein concentrations associated with Ca2+ response are different, hence leading to distinct phenotypes. We carried out immunocytochemistry using antibodies that recognise different subtypes of RyR. The major subtypes of RyR expressed in HEK293 cells are types 1 and 2 (Querfurth et al, 1998); thus, we used an antibody that recognises RyR1 as well as one that recognises both RyR1 and RyR2. The intensity of cytoplasmic immunofluorescence showed a unimodal distribution with a relatively small standard deviation (Figure 2A and B), contrary to our expectation of a bimodal distribution corresponding to two distinct phenotypes. The detection sensitivity of this immunostaining was verified with a heterologous expression of RyR1. Higher levels of RyR were observed in RyR1-overexpressing cells (Supplementary Figure 5), demonstrating that the antibody reflects the expression level of the receptor. We also used an antibody against RyR3, but again found a unimodal distribution of immunofluorescence intensity among cells (Figure 2C). Thus, the expression level of each RyR subtype seems to be almost uniform among individual cells, although small cell-to-cell variations may be present. Figure 2.Immunofluorescence detection of RyRs or SERCAs in HEK293 cells. (A–D) Upper panels: confocal images showing immunofluorescence localisations of RyRs and SERCAs. Scale bars=20 μm. Lower panels: intercellular histograms of fluorescence intensity in cytoplasmic region. (A) RyR1 was detected using an anti-RyR1 antibody. (B) RyR1 and RyR2 were detected. (C) RyR3 was detected. (D) All SERCA subtypes were detected. The total numbers of cells analysed in the bottom panels are as follows: 214 in (A), 190 in (B), 241 in (C) and 220 in (D). Download figure Download PowerPoint As Ca2+ release from the Ca2+ store is antagonised by Ca2+ uptake by sarco/endoplasmic reticulum Ca2+ ATPase (SERCA), the difference in SERCA expression level may result in the observed heterogeneous Ca2+ responses. We, therefore, analysed the expression level of SERCA by immunocytochemistry. However, the distribution of the cytoplasmic immunofluorescence intensity was again unimodal, and we were unable to find two populations of cells with different SERCA expression levels (Figure 2D). To further study possible cell-to-cell variability in the Ca2+ sequestration activity, we analysed the falling phase of Ca2+ response after stimulation with 10 μM ATP. The falling phase could be fitted by a single exponential (Supplementary Figure 6, inset), and its time constant showed a unimodal distribution (Supplementary Figure 6), suggesting that the Ca2+ sequestration system has a similar activity among cells. Mathematical model Provided that there are no qualitative intercellular differences in the concentration of channels and pumps involved in the caffeine response, it must be the process of Ca2+ release that generates two distinct phenotypes in caffeine response. We investigated such a mechanism using a mathematical model of Ca2+ release via RyRs. We modified the conventional model proposed by Keizer and Levine (1996), which incorporates Ca2+ release channels and Ca2+ pumps (Supplementary information and Supplementary Figure 7A). In this model, Ca2+ responses are assumed to be spatially homogeneous, which is supported experimentally in Supplementary Figure 2. Figure 3A shows the response of the model with various numbers of RyRs and a constant number of SERCAs. RyR agonists were applied at t=0. When the number of RyRs in the ER was small, there was no response (magenta and blue traces). As we gradually increased the number of RyRs, a Ca2+ response was suddenly generated (green trace). Further increase in the number of RyRs had little effect on the peak Ca2+ response except for the decrease in lag time (orange and red traces). Although variations in the relative expression levels of RyR subtypes, which have different Ca2+ release activities, would also generate the cell-to-cell variability, we used the simplest formulation in the present model. Conversely, the number of SERCAs was varied while the number of RyRs was kept constant (Figure 3B). When the SERCA number was sufficiently large, no Ca2+ response (magenta and blue traces) was observed upon agonist application. When the SERCA number was gradually decreased, a Ca2+ response was suddenly observed (green trace). With further decrease in the number of SERCAs, there was almost no change in peak height, but the delay time decreased (orange and red traces). Thus, the Ca2+ release via RyRs activated by caffeine has threshold characteristics (Keizer and Levine, 1996; Marchant et al, 1999; Falcke, 2004) and depends on the balance between the numbers or activities of RyRs and SERCAs (Supplementary Figure 7B). We plotted the peak cytosolic Ca2+ concentration against RyR and SERCA activities in a three-dimensional graph (Figure 3C). A boundary is observed between the caf-positive region and the caf-negative region. The cells are caf-positive in the region on the rear left of the boundary line; they are caf-negative in the region on the front right of the boundary line. CICR plays an essential role in this threshold behaviour of Ca2+ response. Indeed, when RyR activity was made Ca2+-insensitive in the above model, no distinct boundary lines between the caf-positive and caf-negative regions were observed (Supplementary Figure 7C). Figure 3.Mathematical model of intracellular Ca2+ dynamics. (A, B) Response of model when RyR activity or SERCA activity changes. (A) RyR activity was increased in a stepwise manner (0.5 (magenta), 2.5 (blue), 4.5 (green), 6.5 (orange) and 8.5 (red) in s−1) at a constant SERCA activity (3 μM s−1). (B) SERCA activity was decreased in a stepwise manner (5 (magenta), 4 (blue), 3 (green), 2 (orange) and 1 (red) in μM s−1) at a constant RyR activity (4.5 s−1). (C) The balance between the numbers or activities of RyRs and SERCAs determines Ca2+ response via RyRs with threshold characteristics. Download figure Download PowerPoint To examine the possibility that the threshold behaviour of Ca2+ response is peculiar to the above model, we tested another mathematical model of Ca2+ release via RyRs originally proposed by Roux and Marhl (2004). We confirmed that also in this model, the Ca2+ response via RyRs activated by caffeine has threshold characteristics (Supplementary Figure 7D), and when RyR activity was made Ca2+-insensitive, no distinct boundary lines between the caf-positive and caf-negative regions were observed (Supplementary Figure 7E). The threshold behaviour of Ca2+ response demonstrated in the above two models suggests that a relatively small difference in RyR or SERCA expression level/activity can induce two distinct phenotypes, caf positivity and caf negativity. Perturbation experiments Results of the immunocytochemistry analysis suggest that RyRs are present in caf-negative cells, and the predictions of the mathematical model suggest that caf-negative cells release no Ca2+ upon caffeine application because the balance between Ca2+ release and uptake does not favour Ca2+ release. We tested this notion by artificially shifting the balance between RyR and SERCA activities. First, we partly inhibited SERCAs using its inhibitor. When we applied 10 μM cyclopiazonic acid (CPA) to inhibit SERCA activity, there was a gradual increase in the resting [Ca2+]i (Figure 4A, middle; see also Supplementary Figure 8). When caffeine was further applied, the caf-negative cells responded with a large increase in [Ca2+]i (Figure 4A, right). Milder inhibition of SERCA by 3 μM CPA, which corresponds to the half maximal inhibitory concentration (Makabe et al, 1996), also yielded essentially the same results (Supplementary Figure 9). These results indicate that the leftward shift in Figure 3C (i.e., a decrease in SERCA activity) brought the cells from the caf-negative region to the caf-positive region. We next altered caffeine concentration to change RyR activity (Lee et al, 2002). With increasing caffeine concentration applied to the cells, an increasing number of cells exhibited Ca2+ response in a threshold manner (Figure 4B), indicating that the front-to-rear shift in Figure 3C (i.e., an increase in RyR activity) brought the cell from the caf-negative region to the caf-positive region. To confirm that the caffeine response requires a threshold concentration of caffeine, we carried out delay time analysis (Skupin et al, 2008). Delay time from the application of caffeine to reach 50% peak [Ca2+]i was measured at different caffeine concentrations. Then, inverse delay time was plotted against caffeine concentration, and the linear fit to the plots has a positive intersection with the caffeine axis (Supplementary Figure 10). The result of this analysis is consistent with the presence of threshold caffeine concentration. We also examined caffeine response in HEK293 cells cotransfected with RyR1 and EGFP. Most (approximately 89%) of the cells with a 'high EGFP expression level' were caf-positive (Figure 4C), again indicating that the front-to-rear shift in Figure 3C brought the cells to the caf-positive region. Thus, results of these perturbation experiments confirmed that caffeine sensitivity is determined by the balance between Ca2+ release and uptake. Figure 4.Perturbation experiments for verifying the model. (A) Cyclopiazonic acid (CPA) turned caf-negative cells caf-positive. Time courses of [Ca2+]i in the same cell are shown. Data are representative of four experiments including 251 cells. (B) With increasing caffeine concentration, an increasing number of cells showed Ca2+ response in a threshold manner. HEK293 cells loaded with fura-2 were exposed to different caffeine concentrations (3, 10 and 30 mM). Time courses of [Ca2+]i in three different cells are shown. Data are representative of n=221 cells. (C) HEK293 cells overexpressing RyR1 and EGFP were mostly caf-positive. Forty-six cells with a high EGFP expression level and 195 cells with a low EGFP expression level (see Materials and methods for their definitions) were analysed. Asterisk denotes a significant difference between the two proportions (P<0.01; one-proportion Z-test). Download figure Download PowerPoint Time-lapse Ca2+ imaging of individual HEK293 cells Previous studies using fluorescent reporter proteins have demonstrated that protein concentrations in individual cells fluctuate over time (Rosenfeld et al, 2005; Cai et al, 2006; Sigal et al, 2006; Yu et al, 2006; Cohen et al, 2008). In light of these findings, the results of the cloning experiment (Figure 1F) point to the possibility that individual HEK293 cells switch between the caf-positive and caf-negative states, and that they dwell in the caf-positive state approximately 40% of the time on average. If this is indeed the case, one should be able to observe the conversion between the states by the time-lapse Ca2+ imaging of individual cells. When caffeine response was examined every 10 min, no significant change was observed within 0.5 h (Supplementary Figure 11). We next examined whether a caf-positive (or a caf-negative) cell maintains its phenotype in terms of caffeine response over longer time or upon cell division. For long-term Ca2+ imaging, we used HEK293 cells retrovirally transduced with GCaMP2, a genetically encoded fluorescent Ca2+ indicator (Tallini et al, 2006). The cells were cultured on the stage of a microscope at 37°C in 5% CO2, and simultaneously observed every 10 min to monitor cell movement and division (Figure 5A). Responses of the agonists (ATP and caffeine) were examined at 12-h intervals. Figure 5.Time-lapse Ca2+ imaging of individual HEK293 cells. (A) Snapshots of cells were recorded every 10 min to monitor movement and cell division. Note that cell 1 divided between 9 and 12 h. Scale bars=20 μm. (B, C) The results obtained at 12-h intervals were compared. We recorded 550 cells (including 172 caf-positive cells) in 12 sets of experiments, of which we were able to monitor 192 cells for 12 h. (B) Representative plots of phenotypic switching of caffeine response. A measure of 10 μM ATP and 25 mM caffeine were used in all time-lapse experiments. Upper panel: caf-negative to caf-positive. Lower panel: caf-positive to caf-negative. (C) The pattern of the caffeine response in these 192 cells is shown. '+' and '−' are abbreviations of 'caf-positive' and 'caf-negative', respectively; '+/+' (or '−/−') indicates that the cell divided into two cells within 12 h, and the resulting two cells were both caf-positive (or caf-negative). Download figure Download PowerPoint We analysed 550 cells from 12 sets of experiments, of which 172 ce
Год издания: 2009
Издательство: Springer Nature
Источник: Molecular Systems Biology
Ключевые слова: Neuroscience and Neuropharmacology Research, Receptor Mechanisms and Signaling, Ion channel regulation and function
Другие ссылки: Molecular Systems Biology (PDF)
Molecular Systems Biology (HTML)
Europe PMC (PubMed Central) (PDF)
Europe PMC (PubMed Central) (HTML)
PubMed Central (HTML)
PubMed (HTML)
Molecular Systems Biology (HTML)
Europe PMC (PubMed Central) (PDF)
Europe PMC (PubMed Central) (HTML)
PubMed Central (HTML)
PubMed (HTML)
Открытый доступ: gold
Том: 5
Выпуск: 1