Calmodulin Oxidation and Methionine to Glutamine Substitutions Reveal Methionine Residues Critical for Functional Interaction with Ryanodine Receptor-1 *

Calmodulin (CaM) binds to the skeletal muscle ryanodine receptor Ca release channel (RyR1) with high affinity, and it may act as a Ca -sensing subunit of the channel. Apo-CaM increases RyR1 channel activity, but Ca -CaM is inhibitory. Here we examine the functional effects of CaM oxidation on RyR1 regulation by both apo-CaM and Ca -CaM, as assessed via determinations of [H]ryanodine and [S]CaM binding to skeletal muscle sarcoplasmic reticulum vesicles. Oxidation of all nine CaM Met residues abolished functional interactions of CaM with RyR1. Incomplete CaM oxidation, affecting 5–8 Met residues, increased the CaM concentration required to modulate RyR1, having a greater effect on the apo-CaM species. Mutating individual CaM Met residues to Gln demonstrated that Met-109 was required for apo-CaM activation of RyR1 but not for Ca -CaM inhibition of the channel. Furthermore, substitution of Gln for Met-124 increased the apoand Ca -CaM concentrations required to regulate RyR1. These results thus identify Met residues critical for the productive association of CaM with RyR1 channels and suggest that oxidation of CaM may contribute to altered regulation of sarcoplasmic reticulum Ca release during oxidative stress.

The 148-amino acid Ca 2ϩ -binding protein, CaM, is composed of N-terminal and C-terminal globular domains connected by a flexible, central tether. CaM has an unusually high Met content, indeed 9 of the 148 amino acids are Met residues, resulting in an ϳ6-fold higher Met content than the average protein (4). In vertebrate CaM these Met residues are clustered primarily in the N (residues 36, 51, 71, and 72) and the C (residues 109, 124, 144, and 145) termini. A 9th Met is located in the tether at residue 76. High affinity Ca 2ϩ binding to EF-hand motifs in each of the globular domains induces a structural rearrangement that reveals the Met-rich hydrophobic patches (5). These hydrophobic patches mediate Ca 2ϩ -CaM interaction with a large and diverse group of proteins that share little sequence homology (6). A less appreciated aspect of CaM regulation is the ability of Ca 2ϩ -free CaM (apo-CaM) to regulate certain targets (7). However, the role of Met residues in apo-CaM interaction with targets is not clear.
Although the high Met content of CaM contributes to effective target binding, the Met residues in the Ca 2ϩ -bound form of CaM are surface-exposed and susceptible to oxidation. Oxidation converts Met to Met sulfoxide, a physiologically relevant product (8). Indeed, Met sulfoxide-containing CaM has been isolated from the brains of aged animals (9).
The present study examines the functional effects of CaM oxidation on the regulation of RyR1. To identify which of the individual Met residues are important for the functional interaction of CaM with RyR1, we used site-directed mutagenesis to change each of the nine CaM Met residues to Gln, introducing an oxygen atom at the same position in the side chain as the sulfoxide. Our results define CaM Met residues that are critical for the functional interaction between CaM and RyR1 and suggest that CaM oxidation may contribute to altered regulation of SR Ca 2ϩ release during oxidative stress.

Materials
Pigs were obtained from the University of Minnesota Experimental Farm. Tran 35 S-labeled Met and Cys were obtained from ICN Radiochemicals (Costa Mesa, CA). [ 3 H]Ryanodine was purchased from PerkinElmer Life Sciences. Unlabeled ryanodine was obtained from Calbiochem. High performance liquid chromatography grade acetonitrile was purchased from Fisher Scientific. C4 ZipTips were from Millipore (Burlington, MA). RPMI 1640 medium was from ICN. Spectrophotometric grade trifluoroacetic acid, AMPPCP, and other reagents were from Sigma.

[ 3 H]Ryanodine Binding to Skeletal Muscle Heavy Sarcoplasmic Reticulum
Isolation of SR Vesicles-Skeletal muscle SR vesicles were prepared from porcine longissimus dorsi muscle (10). Muscle was homogenized in * This work was supported by National Institutes of Health Grant GM 31382 and by the American Heart Association, Northland Affiliate. 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.
[ 3 H]Ryanodine Binding-Ryanodine selectively binds to the open RyR and therefore provides a useful indicator of channel activity (11). SR vesicles (0.2 mg/ml) were incubated at 37°C in medium containing 120 mM potassium propionate, 10 mM PIPES, pH 7.0, 3 mM AMPPCP, 100 nM [ 3 H]ryanodine, and a Ca-EGTA buffer set to give the desired free Ca 2ϩ concentration (12

Oxidation of Calmodulin
Because the thioether group of Met is not protonated at low pH, it can be oxidized selectively under acidic conditions (8). 60 M calmodulin was incubated in 50 mM Homopipes, pH 5.0, 0.1 M KCl, 2.0 mM MgCl 2 , 50 mM H 2 O 2 at room temperature for 0.5-24 h. The reaction was stopped by overnight dialysis (molecular mass cutoff ϭ 3,500) at 4°C in distilled water (5 ϫ 1 liter) buffered with 10 mM ammonium bicarbonate, pH 7.7.

Calmodulin Site-directed Mutagenesis, Expression, and Purification
Recombinant rat CaM was expressed in Escherichia coli using the pET-7 vector (16), purified via phenyl-Sepharose chromatography (17), and dialyzed overnight at 4°C against 2 mM HEPES, pH 7.0. CaM concentration was determined with the Micro BCA assay (Pierce) using wild-type CaM as a standard. The concentration of the CaM standard was determined using the published molar extinction coefficient, ⑀ 277-320 nm ϭ 3,029 M Ϫ1 cm Ϫ1 (18). Glutamine was substituted for each of the Met residues (residues 36, 51, 71, 72, 76, 109, 124, 144, and 145) using QuikChange mutagenesis kits (Stratagene, La Jolla, CA). DNA sequence analysis confirmed the correct generation of each mutant.

[ 35 S]Methionine Incorporation
The biosynthesis of 35 S-labeled CaM was carried out essentially as described previously for [ 35 S]FKBP12 (19). Briefly, bacterial growth was initiated in M9 medium containing ampicillin. When the A 600 value of the bacteria reached 0.6 the pelleted bacteria were resuspended in RPMI 1640 medium containing ampicillin, 1/40th of the Met and Cys concentration compared with the regular RPMI 1640 medium, and isopropy-D-thiogalactopyranoside was added to a final concentration of 1 mM. A 1.4-mCi aliquot of [ 35 S]methionine and [ 35 S]cysteine was then added to the medium and the bacteria cultured for 5-7 h at the same conditions.

Matrix-assisted Laser Desorption/Ionization-Time-of-Flight (MALDI-TOF) Mass Spectrometry
MALDI-TOF mass spectrometry was performed at the University of Minnesota Mass Spectrometry Consortium for the Life Sciences using a Bruker Biflex III mass spectrometer (Bruker, Boston, MA) equipped with a N 2 laser (337 nm, 3-ns pulse length) and a microchannel plate detector. Data were collected in the linear mode, positive polarity, with an accelerating potential of 19 kV. Each spectrum was the accumulation of ϳ200 laser shots. External calibration was performed using horse heart cytochrome c and horse skeletal muscle myoglobin. The matrix used for samples and standards was a saturated solution of 3,5-dimethoxy-4-hydroxycinnamic acid in 50:50, acetonitrile:nanopure water, 0.1% trifluoroacetic acid. Prior to MALDI-TOF analysis samples were desalted using Millipore's C4 ZipTips according to the manufacturer's protocol.

Steady-state Fluorescence
The Ca 2ϩ -induced increase in tyrosine fluorescence intensity is thought to reflect a reduced quenching in Ca 2ϩ -bound CaM and has been used to monitor Ca 2ϩ binding to the C terminus of CaM (20,21). Spectra were collected at 25°C using an ISS K2 fluorometer in ratio mode. The 3 M CaM samples were excited at 275 nm using a xenon lamp, and corrected emission spectra were acquired from 280 to 400 nm in 1-nm increments. Excitation and emission bandwidths were 8 nm. Ca 2ϩ titrations were performed by the addition of small aliquots of concentrated CaCl 2 to the sample in the apo buffer (120 mM KCl, 20 mM PIPES, 1.0 mM EGTA, pH 7.0). A matching buffer scan was subtracted from each spectrum. The fluorescence readings, at 305 nm, for each titration were normalized to the high and low end points before nonlinear least squares analysis.

Circular Dichroism (CD)
CD spectra were recorded from 250 to 200 nM with a JASCO J-710 spectrophotometer coupled with a data processor. Spectra were recorded digitally and fed through the data processor for signal averaging and base line subtraction. Spectra were recorded at 25°C with a CaM concentration of 150 M in a solution of 2 mM HEPES, pH 7.0, and either 500 M Na 2 EGTA or CaCl 2 using quartz cuvettes with a path length of 1.0 mm. Spectra were recorded with a scan speed of 20 nm/min, signal-averaged six times, and an equally signal-averaged solvent base line was subtracted.

PAGE
CaMs were analyzed under denaturing conditions using SDS-PAGE (22). Samples were incubated for 30 min in sample buffer containing either 5 mM CaCl 2 or 5 mM EGTA before loading onto 15% gel. No Ca 2ϩ or EGTA was added to the gel or running buffer.

Analysis
The CaM concentration dependence of SR vesicle [ 3 H]ryanodine binding and the inhibition of [ 35 S]CaM binding by unlabeled CaM were fit with the Hill equation. The Ca 2ϩ dependence of ryanodine binding was fit with Equation 1, which assumes a high affinity Ca 2ϩ binding site, which when bound will activate the RyR and a lower affinity Ca 2ϩ binding site which when bound will inhibit channel opening, where B is the ryanodine bound, B max is the maximal ryanodine binding, EC 50 and IC 50 are the half-activating and half-inhibiting Ca 2ϩ concentrations, respectively, and n a and n i are the Hill coefficients for activation and inhibition, respectively.

Statistics
Data are presented as the means Ϯ S.E. [ 3 H]Ryanodine and [ 3 S]CaM binding curves in the presence and absence of CaM and CaM mutants were studied using a one-way analysis of variance with Dunnett's multiple comparison as a post hoc test or by Student's paired and unpaired t tests as appropriate. The level of significance was Ͻ0.05.  (Fig. 1, A and  B). In addition, after incubation with 50 mM H 2 O 2 for 24 h CaM was no longer able to inhibit [ 35 S]CaM binding to SR vesicles in a medium containing either 100 nM or 700 M Ca 2ϩ (Fig. 1, C  and D). Thus, the inability of oxidized CaM to modulate SR vesicle ryanodine binding was caused by the loss of CaM binding to SR vesicles. By comparison, partial oxidation of CaM by incubation in 50 mM H 2 O 2 for 30 min did not fully abolish CaM modulation of ryanodine binding; however, both the half-activating (EC 50 Ͼ 1000 nM) and half-inhibiting (IC 50 ϭ 104 Ϯ 8, n H ϭ 1.3 Ϯ 0.2) CaM concentrations were increased (Fig. 1, A  and B), with a larger effect occurring at 100 nM Ca 2ϩ . These results suggest that the oxidation of critical CaM Met residues alters the productive association of CaM with the RyR1.

Effects of CaM Oxidation on Regulation of
Mass Spectrometry of Oxidatively Modified CaM-MALDI-TOF mass spectrometry was performed to determine the extent of oxidative modification of CaM incubated with 50 mM H 2 O 2 for either 30 min or 24 h. As shown in Fig. 2 Regulation of RyR1 by Met 3 Gln CaM Mutants-To determine the role of specific CaM Met residues in regulating RyR1, we used site-directed mutagenesis to change each CaM Met to Gln. This substitution introduced an oxygen atom at the same position in the side chain as the sulfoxide. The greater polarity of the Gln side chain relative to Met was expected to decrease the hydrophobic interactions that normally stabilize the association of CaM with its target. However, substituting Gln for Met is unlikely to significantly disturb the structure of CaM because both amino acids have a similar propensity to form ␣-helices (23). Fig. 3 and Tables I and II summarize  pendence of Ca 2ϩ -CaM inhibition by all the N-terminal Met 3 Gln mutants and M76Q were similar to wild-type CaM.
With the exception of the M109Q CaM mutant, in 100 nM Ca 2ϩ , all of the C-terminal Met 3 Gln mutants enhanced ryanodine binding to an extent similar to that of wild-type CaM. Substitution of Gln for Met at position 109 completely abolished apo-CaM activation of RyR1. Replacing Met-124 with Gln increased the apo-CaM EC 50 nearly 23-fold compared with wild-type CaM. In a medium containing 700 M Ca 2ϩ , all of the C-terminal Met 3 Gln mutants, including surprisingly M109Q, inhibited ryanodine binding to an extent similar to that of wild-type CaM. Similar to the effect in 100 nM Ca 2ϩ , the M124Q mutation increased the IC 50 for Ca 2ϩ -CaM inhibition of ryanodine binding. However, the 5-fold increase in the Ca 2ϩ -CaM IC 50 was much smaller than the 23-fold increase in the apo-CaM EC 50 caused by this mutation.
The Ca 2ϩ dependence of SR vesicle [ 3 H]ryanodine binding was examined in the absence and presence of wild-type CaM and selected Met 3 Gln CaM mutants in medium containing 3 mM AMPPCP and 3 mM MgCl 2 ( Fig. 4A (Fig. 4B and Table III). Under these conditions, wild-type, M71Q and M124Q CaM clearly enhanced the Ca 2ϩ sensitivity of channel activation while the M109Q CaM did not.

SDS-PAGE, CD, and Intrinsic Tyrosine Fluorescence-When
CaM is denatured for SDS-PAGE in the presence of Ca 2ϩ , the mobility of the protein on SDS-PAGE is increased relative to that seen after denaturation in the presence of EGTA (26). Although the mechanisms underlying the mobility shift are not understood, it is thought to reflect some difference in the ability of SDS to bind and/or denature apo-and Ca 2ϩ -CaM (27). Thus, mutation-induced structural changes in the Met 3 Gln mutant CaMs might be reflected in altered mobility on SDS-PAGE. As can be seen from Fig. 6, none of the Met 3 Gln mutations altered the mobility of CaM in the presence of EGTA. Furthermore, none of the mutations prevented the mobility shift upon the addition of Ca 2ϩ .
To assess further the potential structural alterations induced by the M109Q and M124Q CaM mutations, the CD spectra arising from these two mutants were compared with the spectra of wild-type CaM. The far UV spectra of wild-type CaM and the M109Q and M124Q mutant CaMs were not significantly different, either in the presence of 500 M EGTA (Fig. 7A) or in the presence of 500 M Ca 2ϩ (Fig. 7B). Thus the Met 3 Gln mutations did not cause major changes in the secondary structure of CaM.
To determine whether the M109Q or M124Q mutation altered the CaM Ca 2ϩ affinity we determined the Ca 2ϩ dependence of the change in intrinsic tyrosine fluorescence (20,21). The apparent Ca 2ϩ affinities of wild-type CaM, M109Q CaM, and M124Q CaM did not significantly differ (K Ca : wild-type CaM, 2.

DISCUSSION
The unusually high CaM Met content (9 of 148 residues) is thought to allow CaM to associate with and regulate a large number of structurally diverse proteins. Here we defined the role of specific CaM Met residues in the regulation of RyR1.
Oxidation of all 9 CaM Met residues to their corresponding sulfoxide abolished both apo-and Ca 2ϩ -CaM binding and modulation of RyR1. In comparison, incomplete oxidation, i.e. oxi-dation of 5-8 CaM Met residues, did not alter the extent of either apo-CaM activation or Ca 2ϩ -CaM inhibition, but rather increased the CaM concentration required for these effects. Thus it appears that the presence of 4 unoxidized Met residues is sufficient for the full extent of CaM regulation of RyR1 but not to provide CaM with the normal high affinity for RyR1. It is not clear, however, whether the ability of the incompletely oxidized CaM to regulate RyR1 fully was the result of the preservation of specific, vital Met residues or a critical Met surface area.
Apo-CaM function appeared to be more sensitive to Met modification than Ca 2ϩ -CaM function. Incomplete oxidation caused a larger increase in apo-CaM EC 50 than in Ca 2ϩ -CaM IC 50 . The differential effect of oxidation on the function of apoand Ca 2ϩ -CaM was also reflected in the differing effects of some of the Met 3 Gln mutants. Thus, the M124Q mutation increased the apo-CaM EC 50 by more than 20-fold but increased the Ca 2ϩ -CaM IC 50 by only 5-fold. Even more dramatic were the effects of the M109Q mutant on apo-CaM versus Ca 2ϩ -CaM function. This single Met 3 Gln substitution completely abolished activation of RyR1 but did not alter Ca 2ϩ -CaM inhibition of the channel.
Met-109 and Met-124 were necessary for the high affinity interaction of CaM with RyR1. Yuan et al. (28) found that in solution, the N-terminal Met residues in apo-CaM are largely buried, whereas the C-terminal Met residues are more exposed. They suggest that these C-terminal Met residues may play a role in the binding of apo-CaM to targets. Thus, in apo-CaM, Met-109 and Met-124 may be available to interact with RyR1. Consequently, a mutation in either of these residues significantly affects the functional interaction of apo-CaM with RyR1. Upon Ca 2ϩ binding to CaM, there is an increased Met exposure (5). Therefore, replacement of an individual Met is less deleterious to Ca 2ϩ -CaM than to apo-CaM function.
The M109Q mutation could potentially abolish apo-CaM activation of RyR1 via three mechanisms. The mutation could disrupt the structure of CaM to such an extent that the mutant apo-CaM could not functionally interact with the channel. In agreement with Chin and Means (29), the Ca 2ϩ -induced mobility shifts on an SDS gel by M109Q CaM and wild-type CaM were similar. In addition the mutant could be purified via phenyl-Sepharose chromatography. Thus M109Q CaM preserved sufficient hydrophobic surface to be retained by the phenyl-Sepharose column and underwent a Ca 2ϩ -induced structural rearrangement similar to wild-type CaM. Finally, the CD spectra of the M109Q and M124Q CaM mutants were indistinguishable from wild-type CaM in the absence of Ca 2ϩ and, also in agreement with Chin and Means (29), in the presence of Ca 2ϩ . Thus we were unable to detect any substantial structural modification in these mutants.
Alternatively, the M109Q mutation could increase the Ca 2ϩ affinity of the CaM such that a substantial fraction of the mutant CaM would exist as inhibitory Ca 2ϩ -CaM in medium containing 100 nM Ca 2ϩ . However, the Ca 2ϩ dependence of the change in intrinsic fluorescence did not differ between wildtype CaM and either M109Q or M124Q CaM.
Finally, a Met residue might be required in position 109 to make specific interactions with RyR1. All of the Met 3 Gln mutants interacted, although with varying affinity, with RyR1. The initial nonspecific association of CaM with targets is thought to be followed by more precise interactions between specific residues (30). In low Ca 2ϩ , M109Q CaM associated with RyR1, albeit with a low affinity, but did not activate the channel. Therefore, it is likely apo-CaM activation of RyR1 requires a specific interaction between Met-109 and the channel.
Met 3 Gln substitutions have been used previously to define the Met residues required for Ca 2ϩ -CaM activation of a number of CaM-dependent protein kinases (29,31) and the plasma membrane Ca 2ϩ pump (32). Although there was variability in the Met residues required for normal enzyme regulation, substitution of Gln for Met-124 decreased the affinity of CaM for all of these targets. The M124Q mutation also decreased the maximal CaM-dependent kinase activation but not CaM activation of the Ca 2ϩ pump. Whereas apo-and Ca 2ϩ -bound M124Q CaM fully regulated RyR1, the substitution decreased the affinity of CaM for the channel. Thus, Met-124 appears to be an important determinant of the CaM affinity for all of these targets; however, its importance in determining CaM regulatory efficacy is target-dependent.
CaM is functionally bifurcated (29,33), thus the N and C termini of CaM may serve different roles in both apo-and Ca 2ϩ -CaM regulation of RyR1. The effects of Met 3 Gln mutations clearly demonstrate the importance of the C-terminal lobe Met-109 and Met-124 in apo-CaM activation of RyR1. In addition, M124Q was the only substitution that significantly altered the interaction of Ca 2ϩ -CaM with RyR1. Rodney et al. (34) proposed a model of RyR1 regulation by CaM in which Ca 2ϩ binding to the CaM C-terminal pair of EF-hands mediates the conversion of CaM from an activator of RyR1 to an inhibitor. The requirement of Met-109 for apo-CaM activation of RyR1 but not for Ca 2ϩ -CaM inhibition of the channel sug-gests that a critical component of the Ca 2ϩ -induced structural change converting CaM from an activator to an inhibitor entails altering the interaction between CaM Met-109 and RyR1.
In summary, oxidation of all 9 CaM Met residues abolished the functional interaction between CaM and RyR1. Incomplete oxidation decreased CaM affinity for RyR1 but not the extent of channel regulation. Site-specific substitution of Met with Gln at residue 109 abolished apo-CaM activation of RyR1 without altering Ca 2ϩ -CaM inhibition of the channel. Substitution of Met-124 with Gln decreased the affinity of both apo-and Ca 2ϩ -CaM for RyR1. Thus these results identify Met residues critical for the productive interaction of CaM for RyR1 and suggest that oxidation of CaM may contribute to RyR1 dysfunction during oxidative stress.