Dampening of cytosolic Ca2+ oscillations on propagation to nucleus.

Ca(2+) signals may regulate gene expression. The increase of the cytosolic Ca(2+) concentration ([Ca(2+)](c)) promotes activation and/or nuclear import of some transcription factors, but others require the increase of the nuclear Ca(2+) concentration ([Ca(2+)](N)) for activation. Whether the nuclear envelope may act as a diffusion barrier for propagation of [Ca(2+)](c) signals remains controversial. We have studied the spreading of Ca(2+) from the cytosol to the nucleus by comparing the cytosolic and the nuclear Ca(2+) signals reported by targeted aequorins in adrenal chromaffin, PC12, and GH(3) pituitary cells. Strong stimulation of either Ca(2+) entry (by depolarization with high K(+) or acethylcholine) or Ca(2+) release from the intracellular Ca(2+) stores (by stimulation with caffeine, UTP, bradykinin, or thyrotropin-releasing hormone (TRH)) produced similar Ca(2+) signals in cytosol and nucleus. In contrast, both spontaneous and TRH-stimulated oscillations of cytosolic Ca(2+) in single GH(3) cells were considerably dampened during propagation to the nucleus. These results are consistent with the existence of a kinetic barrier that filters high frequency physiological [Ca(2+)](c) oscillations without disturbing sustained [Ca(2+)](c) increases. Thus, encoding of the Ca(2+) signal may allow differential control of Ca(2+)-dependent mechanisms located at either the cytosol or the nucleus.

([Ca 2ϩ ] N ) are required for regulation of other transcriptional mechanisms, such as the ones mediated by CREB (1,4) or DREAM-DRE (5).
The mechanisms involved in controlling [Ca 2ϩ ] N are poorly known and controversial (6). The nucleus is surrounded by a double membrane, the nuclear envelope (NE), which can accumulate Ca 2ϩ inside its lumen. It has been proposed that Ca 2ϩ can be released directly from NE cisterna into the nucleoplasm through ion channels associated to either inositol trisphosphate or ryanodine receptors of the inner NE membrane (7)(8)(9). On the other hand, cytosol and nucleoplasm communicate through 10-nm diameter nuclear pores with large ion conductance (10). However, conductance can be drastically reduced by either ATP-dependent accumulation of Ca 2ϩ inside NE (gating) or macromolecular transport through nuclear pore complexes (plugging) (11). The permeability of the NE to cytosolic Ca 2ϩ is a matter of controversy. While several authors report that [Ca 2ϩ ] N follows passively changes of [Ca 2ϩ ] c (12)(13)(14) others identify a significant diffusion barrier at the NE (8,(15)(16)(17). This question is not trivial, as it may determine whether or not cytosolic Ca 2ϩ signals shall be transduced into changes of gene expression.
In the present work we have compared the cytosolic and the nuclear Ca 2ϩ signals reported by targeted aequorins in several resting and stimulated cell systems (adrenal chromaffin cells, PC12 cells, GH 3 pituitary cells). We find that strong stimulation of either Ca 2ϩ entry or Ca 2ϩ release from the intracellular Ca 2ϩ stores produced similar Ca 2ϩ signals in cytosol and nucleus. However, both spontaneous and thyrotropin-releasing hormone (TRH)-stimulated [Ca 2ϩ ] oscillations in GH 3 cells were considerably dampened in nucleus as compared with cytosol, suggesting that nucleus can filter high frequency physiological [Ca 2ϩ ] c signals.

EXPERIMENTAL PROCEDURES
Cell Culture and Expression of Aequorins-Bovine adrenal chromaffin cells, kindly provided by Professor Antonio García, Instituto Teófilo Hernando, Madrid, Spain, were handled as described previously (18,19). HEK-293 cells were maintained in DMEM medium (Invitrogen) supplemented with 10% fetal calf serum, 2 mM glutamine, and antibiotics. Culture and handling of GH 3 pituitary cells (20,21) and PC12 cells (22) were as described previously. Cytosolic and nuclear (nucleoplasmin chimera) aequorin cDNAs were obtained from Molecular Probes and cloned in the pHSVpUC plasmid (23). Packaging and titration of the pHSVnucAEQ (nuclear) and pHSVcytAEQ (cytosolic) viruses were performed as reported (22). The multiplicity of infection ranged between 0.01 and 0.1 for batch luminescence measurements and between 0.3 and 1 for bioluminescence imaging (19,23,24). Cells were cultured for 12-24 h before measurements.
Measurements of Aequorin Bioluminescence-Cells expressing apoaequorins were incubated for 1-2 h at room temperature with 1 M coelenterazine. Coelenterazine h (from Molecular Probes), which increases the affinity of reconstituted aequorin for Ca 2ϩ (25), was used for imaging experiments. The standard incubation medium had the following composition (in mM): NaCl, 145; KCl, 5; CaCl 2 , 1; MgCl 2 , 1; glucose, 10; sodium-HEPES, 10, pH 7.4. Batch cell aequorin photoluminescence measurements were performed as described previously (19,24), and calibrations in [Ca 2ϩ ] were done using the values published for the constant (26). For bioluminescence imaging measurements (23,24), cells were placed into a perfusion chamber thermostatted to 37°C under a Zeiss Axiovert 100 TV microscope and perfused at 5-10 ml/min with the test solutions, prewarmed at 37°C. At the end of each experiment cells were permeabilized with 0.1 mM digitonin in 10 mM CaCl 2 to release all the residual aequorin counts. Images were taken with a Hamamatsu VIM photon counting camera handled with an Argus-20 * This work was supported by a grant from the Spanish Ministerio de Ciencia y Tecnología (MCyT; BFI2001-2073). 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.
‡ image processor and integrated for 10-s periods. Photons/cell in each image were quantified using the Hamamatsu Aquacosmos software. Total counts per cell ranged between 2 ϫ 10 3 and 2 ϫ 10 5 , and noise was (mean Ϯ S.D.) 1 Ϯ 1 cps typical cell area (about 2000 pixels). Data were quantified as rates of photoluminescence emission/total cps remaining at each time and divided by the integration period (L/L TOTAL in s Ϫ1 ). Emission values of less than 4 cps were not used for calculations. Calibrations for [Ca 2ϩ ] are shown in Fig. 3. For calculation of oscillation indexes the differences (in absolute value) between each L/L TOTAL value and the following one were added and divided by the total number of measurements during the integration period. This parameter is sensitive to both the amplitude and the frequency of oscillations (20).

RESULTS AND DISCUSSION
To assess whether the NE affects the progression of the Ca 2ϩ wave we elicited a [Ca 2ϩ ] c increase, either by activation of Ca 2ϩ entry or of Ca 2ϩ release from the endoplasmic reticulum, and compared the size of the Ca 2ϩ peaks measured in cytosol and nucleus. Fig. 1A summarizes results in bovine adrenal chromaffin cells. Ca 2ϩ entry was elicited either by depolarization with high K ϩ (70 mM) solution or by stimulation with acetylcholine. The [Ca 2ϩ ] peaks were very similar in the cytosol (open bars) and in the nucleus (filled bars). Ca 2ϩ release from the endoplasmic reticulum was induced by stimulation with caffeine (18). Again the [Ca 2ϩ ] peaks elicited by caffeine in the cytosol and the nucleus were very similar (Fig. 1A). In PC12 cells the Ca 2ϩ peaks elicited in cytosol and nucleus by either Ca 2ϩ entry (High K ϩ ) or Ca 2ϩ release from the intracellular Ca 2ϩ stores (UTP or bradykinin) were similar. In GH 3 pituitary cells depolarization with high K ϩ or release of Ca 2ϩ from the stores by stimulation with 0.1 M TRH (22) also produced similar Ca 2ϩ peaks in cytosol and nucleus (Fig. 1C). In all the cases the increase of [Ca 2ϩ ] c tended to be somewhat smaller in the nucleus but the differences were small, at most 25%, even though some of them were statistically significant (Fig. 1).
The apparent permeability of the NE was directly assessed by monitoring changes of [Ca 2ϩ ] N on exposure of permeabilized HEK-293 cells to bathing solutions with controlled Ca 2ϩ con- centrations. When external [Ca 2ϩ ] was increased from 0.1 to 2 M, [Ca 2ϩ ] N increased quickly to near thermodynamic equilibrium ( Fig. 2A). Addition of the Ca 2ϩ ionophore ionomycin at this stage did not modify [Ca 2ϩ ] N . On switching perfusion to the low [Ca 2ϩ ] solution, [Ca 2ϩ ] N returned quickly to the resting level. Half-equilibration times were about 10 s for both the increase and the decrease of [Ca 2ϩ ] N . Similar results were obtained with permeabilized GH 3 pituitary cells (Fig. 2B). Single-cell measurements produced similar results (Fig. 2C). Several putative regulatory mechanisms of nuclear permeability to Ca 2ϩ were explored pharmacologically. Pretreatment of the cells with activators of either protein-kinase A (forskolin, 5 M, 5 min) or protein kinase C (phorbol myristate acetate, 0.1 M, 5 min) did not modify results in experiments similar to those in Fig. 2A. Replacement of K ϩ by Na ϩ , addition of protonophores (carbonyl cyanide m-chlorophenylhydrazone, 10 M) or inhibitors of the endoplasmic reticulum Ca 2ϩ pump (tert-butylhydroquinone, 40 M) did not have any effect either.
The above results suggest that the NE does not act as a significant permeability barrier for Ca 2ϩ diffusion, and hence [Ca 2ϩ ] N mimics [Ca 2ϩ ] c , at least for events that last for several seconds, such as the ones tested here. However, we were surprised to find that the measured resting [Ca 2ϩ ] N in GH 3 cells was consistently much smaller than [Ca 2ϩ ] c . The values found were (mean Ϯ S.E.) 56 Ϯ 14 (n ϭ 11) and 177 Ϯ 39 nM (n ϭ 12), respectively. In the other cell types we found no such differences between the resting values of [Ca 2ϩ ] c and [Ca 2ϩ ] N . A relevant difference among these cell types is that GH 3 cells display spontaneous electric activity, Ca 2ϩ action potentials, and [Ca 2ϩ ] c oscillations (27,28). We reasoned that a kinetic barrier could filter propagation of high frequency [Ca 2ϩ ] c oscillations to the nucleus. This would result in a lower value of the average resting [Ca 2ϩ ] c in GH 3 cells, but not in the other cell types tested, which do not display spontaneous Ca 2ϩ oscillations. On the other hand, this kinetic barrier would not disturb significantly the propagation to nucleus of sustained [Ca 2ϩ ] c increases, such as those shown in Fig. 1. To test this working hypothesis we compared single-cell spontaneous [Ca 2ϩ ] c and [Ca 2ϩ ] N oscillations in GH 3 cells. Hypothalamic releasing factors regulate hormone secretion in several anterior pituitary cell types by increasing (TRH, growth hormone-releasing hormone) or decreasing (dopamine, somatostatin) the spontaneous activity (29). In GH 3 cells, TRH increases the rate of action potential firing, [Ca 2ϩ ] c oscillations, and prolactin secretion (27)(28)(29)(30). ( Fig. 1C). Fig 3C compares the [Ca 2ϩ ] c and the [Ca 2ϩ ] N oscillations, quantified as oscillation indexes. These indexes are computed by adding the differences (in absolute value) between every [Ca 2ϩ ] value and the subsequent one along the whole integration period (20). The average oscillation index was three to four times smaller in the nucleus than in the cytosol during the resting periods (Rest or Wash). During stimulation with TRH the oscillation index increased both in cytosol and in nucleus, but still it was two to three times smaller in the nucleus (Fig. 3C). In trace 3 of Fig. 3A we have modeled the effects of a nuclear permeability barrier with T1 ⁄2 ϭ 10 s on the transmission of the [Ca 2ϩ ] c oscillation (trace 1) to the nucleus (trace 3). The outcome was very similar to the experimental observation (compare traces 2 and 3). The oscillation index computed from trace 1 (cytosol) was 3.9 times larger than the one computed from trace 3 (nucleus), a difference very similar to the one measured experimentally (Fig. 3C). The kinetic barrier proposed here could fit equally well for either slowed diffusion at the NE or increased buffering of Ca 2ϩ inside nucleus (31). In both cases the appearance of free Ca 2ϩ inside the nucleus would be slower than expected.
In conclusion, our results indicate that nucleus dampens Ca 2ϩ oscillations of cytosolic origin in GH 3 cells, both the spontaneous and the TRH-induced. This nuclear barrier is, however, much less efficient to filter sustained increases of [Ca 2ϩ ] c such as those elicited here by depolarization with high K ϩ solutions for several seconds or maximal stimulation with inositol 1,4,5-trisphosphate-generating agonists or caffeine. In practical terms, this means that propagation of the [Ca 2ϩ ] c signal to the nucleus is very much dependent on signal encoding, the higher frequency signals being much less efficient than the more sustained ones. This may be a clue for understanding the effects of different stimulation patterns on nuclear events. For example, [Ca 2ϩ ] c oscillations increase the efficiency of gene expression via NFAT or NFB (3). Another example is the differential activation of serum response element-regulated genes and cyclic AMP response element-regulated genes by either cytosolic or nuclear calcium rises (32). According to our results, expression of genes controlled directly by [Ca 2ϩ ] N should preferentially respond to sustained [Ca 2ϩ ] c changes.