Phenyl-sepharose CL-4B, EDTA and EGTA were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Quin-2 was purchased from Calbiochem (La Jolla, CA, USA). Bis-ANS and IAANS were purchased from Invitrogen (Carlsbad, CA, USA). All other chemicals were of analytical grade.

Ohio Peptide LLC (Powell, OH) synthesized the CaM binding domain of Arabidopsis thaliana NADK2 (IYVHSKEGVWRTSAMVSRWK) (Turner et al., 2004). CelTek Peptides (Franklin, TN, USA) synthesized the CaM binding domain of human myosin light chain kinase (MLCK) (ARRKWQKTGNAVRAIGRLSS) (Garcia et al., 1997).

We used three intact CaM isoforms: vertebrate CaM (CaM (although we used a bacterially codon optimized rat CaM gene, the protein sequence is identical in all vertebrates); soybean CaM isoform 1 (sCaM1); soybean CaM isoform 4 (sCaM4). The DNA constructs for CaM^T5C, CaM^F19W, sCaM1^F19W, and sCaM4^F19W were generated from the respective wild type (WT) constructs in the pET17b vector (Studier and Moffatt, 1986; Black et al., 2000). Primer-selected site-directed mutagenesis was performed using the QuikChange Lightning Multi-site kit from Agilent (Santa Clara, CA, USA) following the manufacturer’s PCR protocol. All primers were synthesized by Integrated DNA Technologies (Coralville, IA, USA). Following PCR, the identity of each mutant was confirmed by DNA sequence analysis (GeneWiz). All CaMs were expressed in BL21 DE3 Escherichia coli and purified via phenyl-sepharose chromatography as previously described (Black et al., 2000).

We used several different fluorescent techniques to follow Ca²⁺ and Mg²⁺ binding and exchange with the different CaM constructs. Intrinsic tyrosine (Tyr) fluorescence can be used to follow the structural changes that occur within the C-domain of the WT CaMs (Black et al., 2000; Tikunova et al., 2001). We utilized the F19W mutation to exclusively follow the N-domain structural change via tryptophan (Trp) fluorescence, which occurs upon cation binding (Black et al., 2000; Tikunova et al., 2001). In order to observe the N- and C-domain pocket opening upon Ca²⁺-binding we utilized the fluorescent hydrophobic dyes BIS-ANS and 2,6-ANS (Fink, 1995). In order to directly follow Ca²⁺ dissociation from the CaM constructs we utilized the fluorescent Ca²⁺ chelator quin-2 (Davis et al., 2002). In order to follow WT CaM binding to their target peptides, we utilized the intrinsic Trp fluorescence of the peptides. We also utilized extrinsic fluorescent labeling of the native Cys residue in the two plant CaMs and engineered a single Cys mutant (T5C) in the vertebrate CaM and labeled the constructs with the environmentally sensitive fluorophore IAANS. IAANS fluorescence was sensitive to both N- and C-domain Ca²⁺-binding to the CaM constructs in the presence of the CaM target peptides.

The steady-state fluorescence measurements were performed using a Perkin-Elmer LS55 spectrofluorimeter at 20°C. The titration buffer consisted of 200 mM MOPS, 150 mM KCl, 2 mM EGTA (pH 7.0) (Tikunova et al., 2002, 2010; Tikunova and Davis, 2004). In order to follow the Ca²⁺-dependent change in the N-terminal domain of CaM, Trp fluorescence was excited at 295 nm and monitored at 350 nm. In order to follow the Ca²⁺-dependent change in the C-terminal domain of CaM, Tyr fluorescence was excited at 275 nm and monitored at 305 nm. The free Ca²⁺ concentration was calculated using the EGCA02 program by Robertson and Potter (1984). The free Mg²⁺ concentration was calculated using the MaxChelator program (Bers et al., 2010). Each reported K_d represents an average of at least three successive titrations ± standard error and fit with the Hill equation as previously described (George et al., 1996).

To determine the affinity of each CaM for Ca²⁺ or Mg²⁺, fluorescence emission intensity was recorded as microliter amounts of CaCl₂ or MgCl₂ were added to 2 mL of each CaM (1 μM) in titration buffer with constant stirring (Tikunova et al., 2002; Zhang et al., 2011). The Ca²⁺ titrations in the presence of Mg²⁺ were performed as described above with the addition of 3 mM Mg²⁺ to the titration buffer. To determine the affinity of each CaM isoform for NADK, each CaM was titrated into 2 mL of titration buffer containing 1 μM NADK and 100 μM Ca²⁺ (pCa = 4) (Johnson et al., 1996). The Ca²⁺ titrations in the presence of NADK or MLCK peptide were performed with 4 μM peptide in the 1 μM CaM solution.

The kinetic data were collected using an Applied Photophysics Ltd. (Leatherhead, UK) model SX.18 MV stopped-flow instrument with a dead time of 1.4 ms at 20°C. The samples were excited using a 150W xenon arc source (Tikunova et al., 2002; Davis et al., 2004, 2007). The stopped-flow buffer for all experiments consisted of 10 mM MOPS, 150 mM KCl (pH 7.0). Tyr fluorescence was excited at 275 nm and emission was monitored using a UV-transmitting black glass (UG1) filter from Oriel (Stratford, CT) (VanScyoc et al., 2002). Trp fluorescence was excited at 295 nm and also monitored using a UG1 filter (Davis et al., 2004). Bis-ANS fluorescence was excited at 390 nm and emission monitored using a 500 nm long pass interference filter from Newport (Irvine, CA, USA) (Hawe et al., 2008). 2,6-ANS or IAANS fluorescence were excited at 320 or 330 nm, respectively, and emission for both probes was monitored using a 420–470 nm band pass interference filter from Oriel (Stratford, CT, USA) (Tikunova et al., 2010). Quin-2 fluorescence was excited at 330 nm and emission was monitored using a 510 nm broad band pass interference filter from Oriel (Stratford, CT, USA) (Tikunova et al., 2002).

In order to measure rates of Ca²⁺ dissociation via structural change, stopped-flow buffer with 200 μM Ca²⁺ and either 1 μM CaM^F19W (Trp) or 3 μM CaM (Tyr) was rapidly mixed with 10 mM EGTA in stopped-flow buffer. To measure the Ca²⁺-dependent rate of hydrophobic pocket closure, stopped-flow buffer with 200 μM Ca²⁺, 1 μM CaM and 0.5 μM bis-ANS was rapidly mixed with 10 mM EGTA in stopped-flow buffer. For the actual Ca²⁺ dissociation rate measurements, stopped-flow buffer with 30 μM Ca²⁺ and 6 μM CaM was rapidly mixed with 150 μM quin-2 in stopped-flow buffer. To directly determine the rate of Ca²⁺ dissociation in the presence of peptide, stopped-flow buffer with 30 μM Ca²⁺, 6 μM CaM and 18 μM peptide was rapidly mixed with 150 μM quin-2 in stopped-flow buffer. To determine the rate of Ca²⁺ dissociation via structural change in the presence of peptide, stopped-flow buffer with 1 μM IAANS-labeled CaM, 200 μM Ca²⁺, and 3 μM NADK or MLCK peptide was rapidly mixed with 10 mM EGTA in stopped-flow buffer. Labeling of CaM^T5C, WT sCaM1, and WT sCaM4 was performed as previously described, taking advantage of the native C26 in the soybean CaMs (Tikunova et al., 2010). To determine rate of Ca²⁺ dissociation in the presence of peptide via the intrinsic Trp in each peptide, 1 μM CaM, 200 μM Ca²⁺, and 5 μM NADK or MLCK peptide in stopped-flow buffer was rapidly mixed with 10 mM EGTA in stopped-flow buffer.

For the Mg²⁺ dissociation experiments, stopped-flow buffer with 3 mM Mg²⁺, 1 μM CaM^F19W and 15 μM EGTA (to chelate contaminating Ca²⁺) was rapidly mixed with 30 mM EDTA in stopped-flow buffer. The kinetics of Ca²⁺/Mg²⁺ competition were determined using F19W fluorescence. Stopped-flow buffer with 1 μM CaM^F19W, 3 mM Mg²⁺, and 15 μM EGTA was rapidly mixed with solution containing 2 mM Ca²⁺ in stopped-flow buffer. The competition experiments with 2,6-ANS rapidly mixed solution containing stopped-flow buffer, 2 μM CaM, 3 mM Mg²⁺, 15 μM EGTA and 1 μM 2,6-ANS with solution containing 2 mM Ca²⁺ in stopped-flow buffer.

Rapid Ca²⁺ transients were generated by mixing a solution containing stopped-flow buffer, 2 μM of each sCaM^F19W mutant and 2 mM EGTA with solution containing stopped-flow buffer, 0, 100 μM, 250 μM or 2 mM Ca²⁺ in both the presence and absence of 3 mM Mg²⁺. We followed the binding and subsequent dissociation of Ca²⁺ at the N-domain of CaM or its isoforms using fluorescence from the engineered Trp residue. The visible percent occupancy was calculated using 0 mM Ca²⁺ as the baseline and 2 mM Ca²⁺ as the maximum.

Slow Ca²⁺ transients were generated by mixing solution containing stopped-flow buffer, 3 μM of each sCaM isoform, 9 μM NADK, and 500 μM EDTA with solution containing stopped-flow buffer, 0, 10, 30, 60, 120, and 250 μM, or 2 mM Ca²⁺ in the presence of 1 mM Mg²⁺. As for the rapid transients, the visible percent occupancy was calculated using 0 mM Ca²⁺ as the baseline and 2 mM Ca²⁺ as the maximum. We followed the binding and subsequent dissociation of Ca²⁺ using Trp from the NADK peptide.

Bis-ANS fluorescence was used as described above to follow the structural change as Ca²⁺ dissociates from the N-domain of sCaM1 or sCaM4. The amplitudes from the Ca²⁺ dissociation without peptide were set at 100% and the amplitudes in the presence of increasing amounts of NADK (3, 6, and 15 μM) were then expressed as a percentage of this peptide-free amplitude.

All kinetic data were fit with a program (by P. J. King, Applied Photophysics Ltd.) that uses the non-linear Levenberg–Marquardt algorithm (Tikunova et al., 2010). Each k_off value represents an average of at least three separate experiments, each averaging at least five traces fit with either a single exponential equation (Trp, Tyr) or a double exponential equation (bis-ANS, quin-2) (Tikunova et al., 2002).

All values are expressed as mean ± SEM. Statistical significance (p < 0.05) was determined by one-way ANOVA followed by post hoc analysis with Tukey’s test using the statistical analysis software MiniTab16 (State College, PA, USA). Statistical significance (p < 0.05) for the rapid and slow transient occupancy experiments was determined by two-way ANOVA.

Figure 1 shows the steady-state Ca²⁺ binding curves from the C-domain of WT CaM isoforms (open squares; Tyr fluorescence) and from the N-domain of F19W CaM (Black et al., 2000; Tikunova et al., 2001) isoforms (solid squares; Trp fluorescence). Consistent with recent NMR data (Gifford et al., 2013), the C-domains for each soybean CaM had approximately fourfold higher Ca²⁺ affinity than the respective N-domains, which was also true for vertebrate CaM. In addition, both the N- and C-domains of sCaM4 had approximately two–threefold higher Ca²⁺ affinity than that of sCaM1 (Gifford et al., 2013). Considering steady-state behavior can be modulated by tuning the rate of Ca²⁺ exchange (Martin et al., 1992; Tikunova et al., 2010), it is important to understand the kinetic parameters of the CaM isoforms. Furthermore in a cell, Ca²⁺ levels are rarely in steady-state and rise and fall transiently.

Steady-state Ca²⁺ binding is defined by how quickly the domain can associate and dissociate Ca²⁺. Figure 2A shows the Ca²⁺ dissociation rates from the N-domain of the F19W CaM isoforms using Trp fluorescence. Consistent with the differences in Ca²⁺ affinity, the N-domain of sCaM4 had a ∼2.5-fold slower rate of Ca²⁺ dissociation compared to that of sCaM1 (Table 1). Figure 2B shows the Ca²⁺ dissociation rates from the C-domain of the WT CaM isoforms using Tyr fluorescence. Ca²⁺ dissociated from the C-domain of sCaM4 approximately fivefold more slowly than from sCaM1. Thus, Ca²⁺ dissociates from the N-domain faster than from the C-domain for both plant CaM isoforms, again in agreement with vertebrate CaM behavior and consistent with the higher Ca²⁺ affinity of the C-domain. Figure 2C shows the actual rate of Ca²⁺ dissociation from both domains of each F19W CaM using the high-affinity fluorescent Ca²⁺ chelator quin-2 (Martin et al., 1985; Tikunova et al., 2002). For all the CaMs, the actual Ca²⁺ dissociation rates were nearly identical to those reported by the local structural change sensed by the intrinsic fluorophores at each individual domain. Figure 2D shows the rate of hydrophobic pocket closure in the CaMs as Ca²⁺ dissociates, using the hydrophobic fluorescent dye bis-ANS. For all the CaMs, the rate of hydrophobic pocket closure is comparable to the rate of structural change reported by the intrinsic fluorophores at each domain, which occurs at the actual Ca²⁺ dissociation rate. Although it would appear from the steady-state data (Figure 1) that sCaM1 behaves more similarly to CaM, kinetically this appears to be true only for the N-terminal domain, whereas kinetically the C-terminal domain of sCaM4 appears more similar to CaM (Figure 2; Table 1). Overall, sCaM4 had a higher Ca²⁺ affinity and slower Ca²⁺ dissociation rates than sCaM1, suggesting that differences in Ca²⁺ binding might be one strategy the plants have used to tune CaM function (Gifford et al., 2013).

In addition to Ca²⁺, CaM is known to bind Mg²⁺ competitively, competing with up to 1 mM free Mg²⁺ in the cell (Ohki et al., 1997; Tikunova et al., 2001; Waters, 2011; Gifford et al., 2013). In agreement with previous findings, our data show the N-domain of all the CaMs had a physiologically relevant Mg²⁺ affinity, whereas the C-domain Mg²⁺ affinity falls well outside the physiological range (Figure 3A). The Mg²⁺ affinity of the N-domain of sCaM4^F19W was approximately threefold higher compared to sCaM1^F19W, but not significantly different at the C-domain. Based on these data, Mg²⁺should drastically decrease the apparent Ca²⁺ sensitivity of the N-domain while only slightly affecting the C-domain. Figure 3B shows the effect of 3 mM Mg²⁺ on Ca²⁺ binding to the vertebrate and soybean CaMs. As expected, the apparent Ca²⁺ sensitivity of the N-domain of all the CaMs decreased approximately sixfold in the presence of Mg²⁺. On the other hand, the apparent C-domain Ca²⁺ sensitivity for all the CaMs only marginally decreased by ∼1.5-fold. Thus, in a cell the high level of Mg²⁺ would be expected to exert a greater effect of competing for Ca²⁺ binding to the N-domain of CaM, with little effect on the C-terminal domain Ca²⁺ binding.

In order for Ca²⁺ to bind to the N-domain of CaM in the presence of Mg²⁺, Mg²⁺ must first dissociate (Tikunova et al., 2001), adding another layer of regulation for Ca²⁺ sensing (Davis et al., 2016). Figure 4A shows the rate of Mg²⁺ dissociation from the N-domain of the F19W CaM isoforms using Trp fluorescence. Despite having distinct N-terminal Mg²⁺ affinities, the soybean CaMs had nearly identical Mg²⁺ dissociation rates. Based on these data, Mg²⁺ should drastically slow the rate of N-domain pocket opening in response to a rise in Ca²⁺. Figure 4B shows the rate of N-domain pocket opening for the CaMs using 2,6-ANS. As expected, the rate of N-domain pocket opening is now limited by the rate of Mg²⁺ dissociation. These data suggest that Mg²⁺ would drastically affect the ability of the N-domain to respond to a rapid Ca²⁺ transient.

Figures 5A,B show the response of the N-domain of the plant CaMs to rapid (approximately 0.4 ms half-life) artificial Ca²⁺ transients (Davis et al., 1999). The percentage of sCaM4 able to bind Ca²⁺ during the rapid transients was significantly lower than that of sCaM1 (Table 2). Figures 5C,D show the response of the N-domain of the plant CaMs to rapid artificial transients in the presence of Mg²⁺. Consistent with their similar N-domain Mg²⁺ dissociation rates, in the presence of Mg²⁺ the two plant CaMs responded similarly to the rapid artificial Ca²⁺ transients. Thus, the addition of Mg²⁺ appears to minimize the differences in the Ca²⁺ sensing ability between the plant CaMs. Thus, in a cell the rate of Ca²⁺-binding to the N-domain of CaM will be limited by Mg²⁺ dissociation.

Another critical element to understanding the differential regulation of these plant CaMs is their ability to bind targets. We chose to study this phenomenon using two peptides from enzymes known to be competitively inhibited by sCaM4, myosin light chain kinase (MLCK) and NAD kinase (NADK) (Lee et al., 2000). Currently NADK is the only plant enzyme known to be differentially regulated by the soybean CaMs (Lee et al., 1997; Turner et al., 2004). Furthermore, vertebrate CaM is a poor activator of NADK, but a potent activator of MLCK (Roberts et al., 1984). Figure 6 shows the NADK peptide affinity for the Ca²⁺-saturated CaMs using the intrinsic Trp fluorescence of the peptide. Similar to MLCK (Van Lierop et al., 2002), both plant CaMs bound the NADK peptide with a nearly identical affinity, albeit with a twofold lower affinity than CaM. In agreement with previous work, NADK did not bind the CaMs in the absence of Ca²⁺ (data not shown) (Turner et al., 2004).

Figures 7A,B show the apparent Ca²⁺ sensitivity of the CaMs in the presence of MLCK and NADK peptides using the intrinsic Trp fluorescence of the peptides. In the presence of the MLCK peptide, there was no significant difference in apparent Ca²⁺ affinity between the CaMs. In the presence of the NADK peptide, the apparent Ca²⁺ affinity of sCaM4 was only ∼1.5-fold higher than that of sCaM1, but ∼1.5-fold lower than that of CaM. In general, the NADK peptide had a weaker affinity for the Ca²⁺-saturated CaMs and also had a substantially weaker effect at sensitizing the CaMs to Ca²⁺ compared to MLCK.

Figures 8A,B show the actual Ca²⁺ dissociation rates from both domains of the CaMs bound to either MLCK or NADK peptides using quin-2 fluorescence. Surprisingly, unlike the MLCK peptide, NADK only marginally slowed the Ca²⁺ dissociation from the C-terminal domain of sCaM4 and vertebrate CaM (Table 1). This effect was also observed as Ca²⁺ dissociated from the CaM-peptide complexes by either following the intrinsic Trp fluorescence of the peptides (Figures 8C,D) or the change in fluorescence from IAANS-labeled CaMs (Figures 8E,F). This data suggests that the C-domains of CaM and SCaM4 bind NADK differently than the C-domain of SCaM1. Although there was minimal influence on the C-domain Ca²⁺ dissociation rate of sCaM4 with NADK, the Ca²⁺ dissociation rates and structural changes observed were similar to those of sCaM1. Thus, the differences in Ca²⁺ binding observed in the absence of the peptide are minimized in the presence of a CaM-binding peptide, and cannot explain the differential regulation of the enzyme. Therefore, we expect that the soybean CaMs should respond similarly to a Ca²⁺ transient in the presence of Mg²⁺ and NADK.

Figures 9A,B show the response of the plant CaMs to artificial Ca²⁺ transients (up to 70 ms half-life) in the presence of physiological Mg²⁺ and NADK peptide following the change in intrinsic Trp fluorescence from NADK. As we predicted, there was no significant difference in the response of either plant CaM to the Ca²⁺ transients (Table 2). These data suggest that the plant CaMs may have evolved their differential enzyme regulation without changing their ability to respond to Ca²⁺ signals.