The Calmodulin Regulator Protein, PEP-19, Sensitizes ATP-induced Ca2+ Release*

Background: PEP-19 modulates the kinetics of Ca2+ binding to CaM. Results: An acidic region in PEP-19 binds Ca2+ and is essential for both modulating Ca2+ binding to CaM and sensitizing cells to ATP-induced Ca2+ release. Conclusion: Simply binding to CaM is not sufficient to account for the biological activities of PEP-19. Significance: Regulating ligand-induced Ca2+ release gives PEP-19 the potential to broadly affect cell signaling. PEP-19 is a small, intrinsically disordered protein that binds to the C-domain of calmodulin (CaM) via an IQ motif and tunes its Ca2+ binding properties via an acidic sequence. We show here that the acidic sequence of PEP-19 has intrinsic Ca2+ binding activity, which may modulate Ca2+ binding to CaM by stabilizing an initial Ca2+-CaM complex or by electrostatically steering Ca2+ to and from CaM. Because PEP-19 is expressed in cells that exhibit highly active Ca2+ dynamics, we tested the hypothesis that it influences ligand-dependent Ca2+ release. We show that PEP-19 increases the sensitivity of HeLa cells to ATP-induced Ca2+ release to greatly increase the percentage of cells responding to sub-saturating doses of ATP and increases the frequency of Ca2+ oscillations. Mutations in the acidic sequence of PEP-19 that inhibit or prevent it from modulating Ca2+ binding to CaM greatly inhibit its effect on ATP-induced Ca2+ release. Thus, this cellular effect of PEP-19 does not depend simply on binding to CaM via the IQ motif but requires its acidic metal binding domain. Tuning the activities of Ca2+ mobilization pathways places PEP-19 at the top of CaM signaling cascades, with great potential to exert broad effects on downstream CaM targets, thus expanding the biological significance of this small regulator of CaM signaling.


PEP-19 is a small, intrinsically disordered protein that binds to the C-domain of calmodulin (CaM) via an IQ motif and tunes
its Ca 2؉ binding properties via an acidic sequence. We show here that the acidic sequence of PEP-19 has intrinsic Ca 2؉ binding activity, which may modulate Ca 2؉ binding to CaM by stabilizing an initial Ca 2؉ -CaM complex or by electrostatically steering Ca 2؉ to and from CaM. Because PEP-19 is expressed in cells that exhibit highly active Ca 2؉ dynamics, we tested the hypothesis that it influences ligand-dependent Ca 2؉ release. We show that PEP-19 increases the sensitivity of HeLa cells to ATPinduced Ca 2؉ release to greatly increase the percentage of cells responding to sub-saturating doses of ATP and increases the frequency of Ca 2؉ oscillations. Mutations in the acidic sequence of PEP-19 that inhibit or prevent it from modulating Ca 2؉ binding to CaM greatly inhibit its effect on ATP-induced Ca 2؉ release. Thus, this cellular effect of PEP-19 does not depend simply on binding to CaM via the IQ motif but requires its acidic metal binding domain. Tuning the activities of Ca 2؉ mobilization pathways places PEP-19 at the top of CaM signaling cascades, with great potential to exert broad effects on downstream CaM targets, thus expanding the biological significance of this small regulator of CaM signaling.
PEP-19 (Purkinje cell protein 4, Pcp4) is a small protein (62 amino acids) with no known intrinsic activity other than binding to CaM 2 in the presence or absence of Ca 2ϩ . Although it was originally identified in the central nervous system, PEP-19 mRNA is also found in human bladder, kidney, prostate, uterus, thyroid, and adrenal tissues (1). Changes in expression levels suggest biological roles for PEP-19 in both normal and pathological conditions. For example, PEP-19 mRNA levels are significantly reduced in a mouse model for Parkinson disease (2) and in the prefrontal cortex of alcoholics (3), but its levels are increased in anergic B cells (4) and in human uterine leiomyomas (5). Animal and cellular model systems have demonstrated effects of PEP-19 on diverse cellular processes. PEP-19 null mice show a dramatic reduction in long term plasticity at synapses between granule cell parallel fibers and Purkinje cells (6). Overexpression of PEP-19 in PC12 cells increases neurite outgrowth (7), and premature neuronal differentiation is seen in transgenic mice with three copies of the PEP-19 gene (Pcp4) (8). The latter suggests a role for PEP-19 in Down syndrome because the human PEP-19 gene (PCP4) is present on chromosome 21. In addition, PEP-19 has anti-apoptotic activity when expressed in PC12 and HEK293T cells (9,10), and it provides protection against Ca 2ϩ overload in cortical neurons (10). These experimental observations are consistent with a proposed neuroprotective role for PEP-19 based on expression patterns in neuronal tissues that are susceptible to Huntington and Alzheimer diseases (11).
The above studies emphasize the need to understand the mechanism of action of PEP-19. Two models for PEP-19 have been proposed based on studies using peptides and the homologous proteins neurogranin (Ng) and neuromodulin (12)(13)(14). The first, or camstatin model, proposes that PEP-19 competitively inhibits activation of CaM target proteins. The second, or calpacitin model, proposes that PEP-19 binds with high affinity to apo-CaM to retard its release from PEP-19, thereby affecting the temporal profile of available CaM during a Ca 2ϩ pulse. We proposed an alternative or additional mechanism for PEP-19 based on its ability to modulate the Ca 2ϩ binding properties of CaM. Specifically, PEP-19 increases both the Ca 2ϩ k on and k off rates at the C-domain of CaM up to 40-fold with little effect on the K Ca (15). We also showed that an acidic sequence located adjacent to the IQ motif is required to modulate Ca 2ϩ binding to the C-domain of CaM, even though it has no apparent intrinsic affinity for CaM (16). Thus, the acidic/IQ motif of PEP-19 has the potential to modulate the rate-limiting kinetics of Ca 2ϩ binding to CaM.
This study investigates the molecular mechanism by which PEP-19 modulates Ca 2ϩ binding to CaM, and it tests the hypothesis that the biological activities of PEP-19 rely on synergy between the biochemical properties of its acidic and IQ sequences. Our results show that the acidic sequence in PEP-19 has intrinsic metal binding properties that play a role in increasing the rates of Ca 2ϩ binding to CaM, at least in part, by electrostatically steering Ca 2ϩ to and from Ca 2ϩ binding sites III and/or IV. We also show that PEP-19 sensitizes HeLa cells to ATP-dependent Ca 2ϩ release and that this effect is greatly reduced or eliminated by mutations in PEP-19 that inhibit or eliminate its ability to modulate Ca 2ϩ binding to CaM. Tuning the activities of Ca 2ϩ mobilization pathways by PEP-19 greatly expands the biological significance of this small regulator of CaM signaling.

EXPERIMENTAL PROCEDURES
Mutagenesis and Protein Purification-QuikChange II XL site-directed mutagenesis kit (Stratagene) was used to generate a panel of PEP-19 mutants. CaM and C-CaM (isolated C domain of calmodulin) were decalcified by addition of 5 mM EDTA and 0.1 mM BAPTA as a UV marker and then desalting on a Bio-Gel P2 column (Bio-Rad) in 10 mM NH 4 HCO 3 that had been decalcified using a Ca 2ϩ sponge column (Molecular Probes). Decalcified proteins were then lyophilized and resuspended in desired buffers. Protein concentrations were estimated using an extinction coefficient of ⑀ 276 nm ϭ 0.18 ml Ϫ1 mg Ϫ1 for C-CaM and ⑀ 215 nm ϭ 0.59 ml Ϫ1 mg Ϫ1 for PEP-19.
Ca 2ϩ Binding Measurements-The rate of Ca 2ϩ dissociation (k off ) from CaM or C-CaM in the presence or absence of PEP-19 derivatives was determined using stopped-flow fluorescence and the Ca 2ϩ -sensitive dye Quin-2 as described previously (15). Typically, solutions of 2-5 M CaM or C-CaM in 20 mM MOPS, pH 7.5, 100 mM KCl, 30 M CaCl 2 were rapidly mixed with 20 mM MOPS, pH 7.5, 300 M Quin-2. Excess free Ca 2ϩ and Ca 2ϩ that is rapidly released from the N-domain of CaM bind to Quin-2 in the 1.7-ms dead time of the stopped-flow instrument. The subsequent increase in Quin-2 fluorescence is due to binding Ca 2ϩ released slowly from the C-domain. Experiments were performed at 23°C using an Applied Photophysics Ltd. (Leatherhead, UK) model SX20 MV sequential stopped-flow spectrofluorimeter with a 150 watt Xe/Hg lamp.
Equilibrium Ca 2ϩ binding constants for CaM in the presence or absence of PEP-19 derivatives were determined using tyrosine fluorescence at 23°C as described previously (17). Data were collected with a QuantaMaster fluorimeter (Photo Technology International). Intrinsic Tyr emission spectra were recorded from 290 to 320 nm with the excitation wavelength of 276 nm. Solutions contained 20 mM MOPS, pH 7.5, 0 or 100 mM KCl, 1 mM EGTA, 1 mM HEDTA, 1 mM nitrilo-2,2Ј,2Љ-triacetic acid, 5 M CaM or C-CaM with or without PEP-19 or its derivatives. Calcium was added from a concentrated stock prepared in the same buffer with CaM, PEP-19, and chelators, so that only the concentration of Ca 2ϩ changes during the titration even though the volume increases. The concentration of total Ca 2ϩ needed to achieve a desired free Ca 2ϩ concentration was determined using the on-line calculator MaxChelator. Control titrations were performed using Br 2 BAPTA as an indicator instead of CaM or C-CaM to confirm that the calculated free Ca 2ϩ was accurate at high and low ionic strength. The K Ca for Br 2 BAPTA is 1.59 M at 100 mM KCl and 0.15 M at 10 mM KCl (18).
Tyrosine fluorescence intensity was plotted against the free Ca 2ϩ concentration and fit to the following form of the Hill Equation 1, where [Ca 2ϩ ] is the free Ca 2ϩ concentration; F is the fluorescence intensity at a given free Ca 2ϩ concentration; F min is the initial fluorescence intensity in the absence of added Ca 2ϩ ; F max is the fluorescence at maximal Ca 2ϩ ; K Ca is the concentration of Ca 2ϩ at which the change in fluorescence is half-maximal, and n is the Hill coefficient.
NMR Methodology-NMR experiments were performed on a Bruker DRX 600 MHz spectrometer equipped with a 5-mm triple resonance cryoprobe at 298 K. Protein samples were dissolved in buffer containing 10 mM imidazole, 5% D 2 O (v/v), pH 6.3, 100 mM KCl. 1 H, 15 N HSQC spectra were used to determine residues in PEP-19 that are affected by binding to C-CaM. Briefly, 1 H, 15 N HSQC spectra were collected during titration of 15 N-labeled PEP-19 with C-CaM in the presence or absence of Ca 2ϩ . Characteristics of fast exchange were seen at saturating Ca 2ϩ , so backbone amides could be assigned by following crosspeaks during the titration. Slow exchange was seen in the apostate, so assignments in the bound state were made using HNCO, HNCA, HN(CO)CA, HNCACB, CBCA(CO)NH, 15 N HSQC-TOCSY, and 15 N-edited NOESY-HSQC experiments. All NMR spectra were processed and analyzed using Topspin 2.0 (Bruker) and FELIX 2004 (MSI, San Diego). 1 H chemical shifts were referenced to 2,2-dimethyl-2-silapentane-5-sulfonate, and 15 N/ 13 C chemical shifts were referenced indirectly using their respective gyromagnetic ratios. The average amide chemical shift change was calculated using Equation 2, where ⌬␦H is the change in 1 H chemical shift and ⌬␦N is the change in 15 N chemical shift.
Calcium Imaging-Calcium imaging was performed exactly as described previously (19). HeLa cells were transfected with yellow fluorescent protein (YFP) only (control) or co-transfected with YFP and PEP-19 constructs at a DNA ratio of 1:4 using Lipofectamine 2000. Twenty four hours after transfection, single cell calcium responses evoked by NaATP were recorded from all YFP-positive cells in a given field. All experiments were repeated at least three times, and the data were pooled for statistical analysis. The actual number of single cell records averaged for each condition is indicated above the bars in Fig. 6c.

PEP-19
Proteins Generated for Study-We used amide chemical shift perturbation to identify residues in PEP-19 that experience significant structural transitions upon binding to the C-domain of CaM because these residues will likely play key roles in regulating Ca 2ϩ binding to CaM. C-CaM, which encodes residues 76 -148 of CaM, was used for these experiments because we showed previously that PEP-19 binds to C-CaM and had the same effects on its Ca 2ϩ -binding proteins as seen for full-length CaM (20). Fig. 1, a and b, shows that backbone amide chemical shifts for residues 1-30 in PEP-19 are unchanged upon binding to C-CaM in the absence or presence of Ca 2ϩ . Because free PEP-19 is intrinsically disordered (21), these data show that residues 1-30 remain disordered when bound to C-CaM. Amide chemical shift perturbations are restricted to residues in the acidic/IQ region of PEP-19 upon binding to either apo-or Ca 2ϩ -C-CaM.
Based on the above chemical shift perturbations, two sets of proteins were generated to test the biochemical and functional significance of the acidic sequence in PEP-19 (see Fig. 1c). The acidic sequence is deleted in PEP⌬Ac such that Val-26 effectively substitutes for Glu-40 of native PEP-19. We anticipated that a hydrophobic residue at this position would promote association of PEP⌬Ac with both the N-and C-domains of CaM because a Phe residue at the homologous position in the Ca V 1.2 channel anchors its IQ region to the N-domain (22). Residues in the acidic sequence of PEPscram are randomized to determine whether the native sequence is important for modulating Ca 2ϩ binding to CaM or whether a cluster of negative charges is sufficient.
The second set of proteins was designed to test the functional significance of sequence similarity between the acidic region of PEP-19 and the consensus EF-hand Ca 2ϩ -binding site where alternating residues provide oxygens to coordinate Ca 2ϩ at X, Y, Z and ϪY, ϪX, and ϪZ positions (see Fig. 1c). Thus, Ala was substituted individually for Glu-29, Asp-31, Asp-33, Asp-35, or Glu-40. In addition, Pro-37 was changed to Gly to test the hypothesis that backbone constraints imposed by the cyclized Pro side chain dictates the relative positions of adjacent acidic residues when PEP-19 is bound to CaM, thereby affecting Ca 2ϩ binding.
Deletion of the Acidic Sequence Prevents Modulation of Ca 2ϩ Binding to CaM-Calcium-dependent Tyr fluorescence was used to measure the K Ca of the C-domain of CaM in the presence or absence of native and mutated PEP-19. Table 1 shows that neither native PEP-19 nor its mutated derivatives have large effects on K Ca , although most decreased the cooperativity of Ca 2ϩ binding.
The relatively slow Ca 2ϩ k off rate of 10.4 s Ϫ1 for free CaM in Table 1 is due to dissociation of 2 Ca 2ϩ from the C-domain because dissociation of Ca 2ϩ from the N-domain is very rapid and occurs in the dead-time (1.7 ms) of the stopped-flow fluorimeter. PEP-19 greatly increases the rate of Ca 2ϩ dissociation to about 300 s Ϫ1 , but the stoichiometry remains 2 Ca 2ϩ released per CaM. Table 1 shows that deletion of the acidic sequence in PEP⌬Ac prevents the increase in Ca 2ϩ k off . Thus, the acidic sequence of PEP-19 is required for modulation of Ca 2ϩ binding to CaM.
Interestingly, the stoichiometry of Ca 2ϩ release in the presence of PEP⌬Ac is 4 Ca 2ϩ /mol of CaM instead of 2 seen the presence of all other PEP-19 proteins. This is consistent with the above prediction that PEP⌬Ac binds to both the N-and C-domains of CaM, thereby slowing the rate of release of Ca 2ϩ from the N-domain as is seen for other CaM-binding proteins and peptides (23). We confirmed this mode of binding using a donor-and acceptor-labeled CaM (CaM(DA)) (24), which gives a large decrease in fluorescence due to FRET when CaM adopts a compact structure upon binding both domains to one peptide. Fig. 2 shows that fluorescence from CaM(DA) is not greatly affected by native PEP-19 because it binds preferentially to the C-domain of CaM, but a large decrease in fluorescence is seen upon binding to either PEP⌬Ac or a CaM-binding peptide

Regulation of Ca 2؉ Signaling by PEP-19
2042 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 288 • NUMBER 3 • JANUARY 18, 2013 from CaM kinase II, CKII(293-312), which is known to bind to both domains of CaM (25). As a further test, we generated PEP(E40F), with Phe at the homologous position to the Phe that anchors the IQ motif of the Ca V 1.2 channel to the N-domain of CaM (22). Fig. 2 shows that PEP(E40F) also causes a large decrease in fluorescence upon binding to CaM(DA). These results show that the absence of an appropriately positioned hydrophobic group in the acidic region of PEP-19 allows preferential binding to the C-domain of CaM.
Native Sequence of the Acidic Region Is Necessary to Modulate Ca 2ϩ Binding to CaM- Table 1 shows that PEPscram has essentially no effect on K Ca , k off , k on , or the stoichiometry of Ca 2ϩ binding to CaM. This lack of effect was so striking that we used NMR to determine whether PEPscram binds to CaM with the same domain specificity and exchange properties as native PEP-19. We showed previously that native PEP-19 binds to apo-CaM and Ca 2ϩ -CaM with characteristics of slow and fast exchange, respectively, on the NMR time scale (21). Fig. 3 shows that PEPscram retains these properties. Specifically, Fig.  3a shows that PEPscram binds to apo-CaM with slow to intermediate exchange on the NMR time scale, causing severe broadening of backbone amide cross-peaks for residues in the C-domain, but it has little effect on amides in the N-domain (full spectra are supplied as supplemental material). Fig. 3b shows that PEPscram also selectively binds to the C-domain of Ca 2ϩ -CaM, but with characteristics of fast exchange. Thus, both PEPscram and PEP-19 bind to the C-domain of apo-or Ca 2ϩ -CaM, and with similar exchange characteristics, but PEPscram is incapable of modulating the Ca 2ϩ binding properties of CaM.
None of the PEP-19 point mutations had significant effects on K Ca of CaM, but Table 1 and Fig. 4a show that they have varying effects on k off and k on . Conversion of Glu-29 to Ala at the putative X coordination position had no effect. Mutation of Asp-31, Asp-33, or Glu-40 to Ala inhibited the ability of PEP-19 to increase k off but to different extents. The properties of PEP(D35A) are very similar to native PEP-19, even though Asp-35 at the putative ϪY coordination position is centered between residues 31, 33, and 40. This could be explained by the fact that the ϪY position is highly variable in canonical EF-hand Ca 2ϩ -binding loops because the backbone carbonyl oxygen, not the side chain, of this residue coordinates Ca 2ϩ . Fig. 4b shows that conversion of Pro-37 to Gly significantly decreased the ability of PEP-19 to modulate Ca 2ϩ binding to CaM, although not to the extent seen for PEPscram. This suggests that backbone constraints imposed by the imide side chain of Pro-37 positions acidic residues in PEP-19 such that they can properly modulate Ca 2ϩ binding to CaM. Therefore, mutation of Pro would be equivalent to mutating multiple acidic residues. This is consistent with the fact that PEPscram is incapable of modulating Ca 2ϩ binding to CaM because it effectively has multiple acidic mutations.
Acidic Region of PEP-19 Binds Ca 2ϩ -The distribution of acidic residues in PEP-19 led us to determine whether the acidic sequence has intrinsic Ca 2ϩ binding activity. Its similar ionic  radii and metal coordination geometries to Ca 2ϩ make paramagnetic Tb 3ϩ a sensitive probe for identifying Ca 2ϩ -binding sites (26). Fig. 5a shows that Tb 3ϩ broadens backbone amide chemical shifts for residues in the acidic sequence of PEP-19, especially residues 31-36, which are severely broadened at a Tb 3ϩ /PEP-19 ratio of 1:50. Amides for Asp-33 and Asp-35 are most affected and are broadened beyond detection at a Tb 3ϩ / PEP-19 ratio of 1:100. These spectral perturbations indicate that Tb 3ϩ binds to the acidic region in PEP-19. Although Ca 2ϩ is not paramagnetic, we reasoned that it might affect specific amide resonance intensities due to exchange broadening if Ca 2ϩ binds to PEP-19. Indeed, Fig. 5b shows that addition of Ca 2ϩ to PEP-19 causes exchange broadening of amide resonances in the acidic sequence relative to other regions in PEP-19. Similar to the effect of Tb 3ϩ , Asp-33 and Asp-35 are most affected by addition of Ca 2ϩ and show maximal broadening at a Ca 2ϩ :PEP-19 molar ratio between 1 and 2 as shown in Fig. 5c with Arg-4 in the N-domain of CaM as a control. These spectral perturbations indicate that Ca 2ϩ binds weakly to the acidic sequence of PEP-19.
Effect of Electrostatics on Ca 2ϩ Binding-We reasoned that the acidic sequence of PEP-19 with intrinsic Ca 2ϩ binding properties may increase the Ca 2ϩ k on if positioned near site III and/or IV of CaM by attracting or electrostatically steering Ca 2ϩ to these binding sites. Because the contribution of electrostatic interactions would be decreased by monovalent cations, we predicted that decreasing the KCl concentration would increase the k on for Ca 2ϩ binding to the C-domain of CaM in the presence or absence of PEP-19. Table 2 shows the K Ca , k off , and k on values for Ca 2ϩ binding to the C-domain of CaM with or without 100 mM KCl and with or without 30 M PEP-19. The Ca 2ϩ binding affinity is increased about 13-fold at low ionic strength due primarily to a large increase in k on . The effect of KCl on k on can be explained by electrostatic shielding of acidic side chains on CaM that coordinate or attract Ca 2ϩ . PEP-19 increases Ca 2ϩ k on by 27-and 45-fold at 100 and 0 mM KCl, respectively. This effect of PEP-19 can be attributed, at least in part, to electrostatic steering of Ca 2ϩ ions via weak Ca 2ϩ binding activity of the acidic sequence in PEP-19.
PEP-19 Sensitizes HeLa Cells to ATP-induced Ca 2ϩ Release-ATP-induced Ca 2ϩ release in HeLa cells was selected as an   Table 1). b shows the effect of increasing concentrations of PEP-19, PEP(P37G), and PEPscram on the rate of Ca 2ϩ dissociation.
initial model system to determine whether PEP-19 can impact a Ca 2ϩ release pathway because this pathway involves multiple potential points of regulation by CaM, including P2Y G-protein-coupled receptors, phospholipase C, and the IP 3 receptor. PEPscram, PEP(P37G), and PEP⌬Ac were selected to test the biological significance of the acidic sequence because they all bind to CaM but have little or no effect on its Ca 2ϩ binding properties. Expression plasmids for native and mutated PEP-19 were engineered with N-terminal Myc tags to readily determine relative expression levels in transfected cells. We anticipated that the Myc tag would not affect interactions between PEP-19 and CaM because residues 1-23 in PEP-19 are disordered when bound to CaM. As shown in Table 1, PEP-19 and mycPEP-19 have essentially identical effects on Ca 2ϩ binding to CaM.
HeLa cells were transfected with a control YFP plasmid or cotransfected with YFP and PEP-19 plasmids at a 1:4 ratio. YFPpositive cells from different coverslips were selected for analysis. The Western blot in Fig. 6a shows comparable levels of expression of PEP-19 proteins in transfected cells, and it confirms that the apparent molecular mass of PEP⌬Ac is smaller than the other proteins due to deletion of the acidic sequence. Fig. 6b shows intracellular Ca 2ϩ in response to stimulation with 0.1, 1, and 10 M ATP. The most striking observation is that only cells expressing PEP-19 showed a robust increase in intracellular Ca 2ϩ in response to 0.1 M ATP. As summarized in Fig.  6c, control cells were unresponsive to 0.1 M ATP, but 65% (35/54) of cells expressing PEP-19 responded with a significant increase in intracellular Ca 2ϩ levels. Fig. 6d shows that peak intracellular Ca 2ϩ release stimulated by 1 and 10 M ATP was also significantly higher in cells expressing PEP-19 relative to control cells. Finally, Fig. 6e shows that PEP-19 increases the frequency of Ca 2ϩ oscillations induced by 1 M ATP relative to control cells.
In contrast to native PEP-19, Fig. 6c shows that only 5% of all cells expressing PEP(P37G), PEPscram, or PEP⌬Ac responded to 0.1 M ATP with an increase in Ca 2ϩ . Fig. 6d shows that peak Ca 2ϩ levels induced by 1 and 10 M ATP are also significantly lower in cells expressing mutant PEP(P37G), PEPscram, or PEP⌬Ac relative to PEP-19. Moreover, Fig. 6d shows the mutant PEP-19 proteins do not mimic the effect of native PEP-19 on Ca 2ϩ oscillation frequency at 1 mM ATP. These data demonstrate that simply binding CaM is not sufficient and that the native acidic sequence in PEP-19 is required to sensitize HeLa cells to ATP-dependent Ca 2ϩ release.

DISCUSSION
Cell signaling pathways must be regulated at multiple levels to control the amplitude and temporal characteristics of cellular responses and to prevent chaotic signaling that can lead to cell damage or death. Calmodulin is primarily regulated by intracellular Ca 2ϩ , which is in turn controlled by cell-specific arrays of Ca 2ϩ channels, pores, and pumps (27). A poorly understood regulatory mechanism involves the actions of dedicated regulators of CaM signaling, which have no known intrinsic activity other than binding to CaM. For example, the small neuronal phosphoprotein called ARPP-21, or regulator of calmodulin signaling, binds to Ca 2ϩ -CaM to competitively inhibit activation of calcineurin and block suppression of L-type Ca 2ϩ currents (28). PEP-19 is also a small protein with the potential to broadly affect CaM signaling by binding to apoor Ca 2ϩ -CaM via its IQ motif.
An obvious potential mechanism for PEP-19 is to competitively inhibit activation of CaM targets as proposed in the camstatin model (12). A caveat to this is that enzymes such as CaM kinase II bind CaM with 10,000-fold greater affinity than does PEP-19. Nevertheless, CaM binds to many proteins with low affinity, and PEP-19 would be particularly effective as an antag- onist of proteins that bind preferentially to the apo-or Ca 2ϩbound C-domain of CaM. Another mechanism, the calpacitin model (14), proposes that higher affinity binding of PEP-19 to apo-CaM relative to Ca 2ϩ -CaM retards its release during a Ca 2ϩ pulse thereby affecting the temporal profile of available CaM and decreasing the overall rate of association of CaM with Ca 2ϩ -dependent target proteins, especially at low Ca 2ϩ levels. This model stems from early studies showing that the homologous protein, neuromodulin, binds preferentially to apo-CaM (29). However, this selectivity is only observed at low salt, and neuromodulin binds with equal affinity to apo-and Ca 2ϩ -CaM in buffers containing 150 mM KCl (30). PEP-19 also has little selectivity for apoversus Ca 2ϩ -CaM at physiologically relevant concentrations of salt (12,16).
These caveats to proposed mechanisms for PEP-19 led us to explore alternatives. We first showed that PEP-19 increased k on and k off rates for Ca 2ϩ binding to the C-domain of CaM by 30 -40-fold without greatly affecting K Ca (15). Importantly, an acidic sequence located adjacent to the IQ motif is required for PEP-19 to modulate Ca 2ϩ binding to CaM (16). Thus, PEP-19 has the potential to modulate the rate-limiting kinetics of Ca 2ϩ binding to CaM and to provide a regulatory mechanism that is analogous to regulators of G-protein signaling, or RGS proteins, that modulates nucleotide hydrolysis (31).
The first goal of this study was to investigate the molecular mechanism of action of PEP-19. Our results show the following: 1) the native sequence of the acidic region as well as backbone constraints imposed by Pro-37 are required for PEP-19 to modulate Ca 2ϩ binding to CaM; 2) the acidic sequence has weak Ca 2ϩ binding properties. Interestingly, mutations that compromise the ability of PEP-19 to modulate Ca 2ϩ binding to CaM have proportional effects on both k on and k off (see Table 1), which suggests that a similar mechanism is responsible, at least in part, for modulating both parameters. A role for acidic residues in tuning the Ca 2ϩ k on but not k off for binding to Ca 2ϩ EF-hand proteins was demonstrated by Martin et al. (32), who showed that neutralizing three acidic surface residues near EF loop I in calbindin D9k decreased the k on up to 50-fold. By analogy, the acidic sequence of PEP-19 may mimic an increase in negative surface charge near site III and/or IV of CaM, thereby increasing the Ca 2ϩ k on CaM by stabilizing a Ca 2ϩ -CaM initiation complex or by electrostatically steering Ca 2ϩ to sites III and/or IV. PEP-19 may increase the Ca 2ϩ k off of CaM by providing a low affinity transition Ca 2ϩ -binding site that shuttles Ca 2ϩ to the solvent rather than allowing it to rebind to the EF-hands of CaM. The inability of PEP(P37G) and PEPscram to modulate Ca 2ϩ k on and k off may be due to repositioning the acidic residues relative to the EF-hand Ca 2ϩ -binding loops in CaM and/or compromising Ca 2ϩ binding to PEP-19.
The second goal of this study was to determine whether PEP-19 modulates CaM-dependent signaling pathways that affect intercellular Ca 2ϩ homeostasis. We selected purinergic ATP-induced Ca 2ϩ release as a model system because this pathway involves multiple potential points of regulation by CaM. The data in Fig. 6 show that PEP-19 sensitizes HeLa cells to ATP-dependent Ca 2ϩ release and also alters the frequency of Ca 2ϩ oscillations. Importantly, these biological effects require an intact acidic sequence, not simply binding of PEP-19 to CaM. Additional studies will be necessary to identify the level at which PEP-19 impacts Ca 2ϩ release, but these effects reinforce the idea that both PEP-19 and Ng play roles in intercellular Ca 2ϩ homeostasis (33). Such a role would be consistent with expression of PEP-19 in neuroendocrine and neuronal cells such as Purkinje cells (34) that have highly active Ca 2ϩ signaling dynamics with robust and prolonged trains of action potentials (35). Ng knock-out mice show multiple effects on Ca 2ϩ dynamics, including increased base-line Ca 2ϩ levels and blunted Ca 2ϩ transients induced by synaptic activity or glutamate receptor agonists (36). We anticipate PEP-19 and Ng will influence distinct sets of Ca 2ϩ mobilization proteins and/or have different effects on the same proteins because PEP-19 increases both k on and k off of Ca 2ϩ binding to the C-domain (15), whereas Ng increases only Ca 2ϩ k off leading to decreased Ca 2ϩ binding affinity (37). Different cellular effects of PEP-19 and Ng are also suggested by different patterns of expression and because PEP-19 has anti-apoptotic effects (9, 10), whereas RC3 is reported to have pro-apoptotic activity (38,39).
Calmodulin regulates numerous proteins involved in Ca 2ϩ mobilization that could be tuned by PEP-19. With respect to ATP-dependent Ca 2ϩ release, CaM directly and indirectly impacts phospholipase C activity (40), and it also modulates the activity of the IP 3 receptor (41) and store-operated Ca 2ϩ entry channels (42) subsequent to IP 3 generation. Other CaM-dependent channels and extrusion proteins include the ryanodine receptor (43), plasma membrane Ca 2ϩ pumps (44), and the Na ϩ /Ca 2ϩ exchanger (45). Interestingly, the modes of interaction between CaM and several key Ca 2ϩ mobilization proteins may make them particularly susceptible to PEP-19 because it binds selectively to the C-domain of CaM. For example, voltage-operated Ca 2ϩ channels (46) and the IP 3 receptor (47) rely on selective, sequential, or stepwise interactions with the C-domain of CaM in its apo-or Ca 2ϩ -bound forms.
In summary, this study reveals new mechanisms of action for PEP-19 and demonstrates novel effects on ATP-dependent Ca 2ϩ release that do not depend solely on binding PEP-19 to CaM, but it also requires its ability to modulate Ca 2ϩ binding to CaM. Tuning the activities of Ca 2ϩ mobilization pathways would place PEP-19 at the top of CaM signaling cascades, with great potential to exert broad effects on downstream CaM targets, thus expanding the biological significance of this small regulator of CaM signaling.