Electrostatic changes in phosphorylase kinase induced by its obligatory allosteric activator Ca²⁺

Abstract

Skeletal muscle phosphorylase kinase (PhK) is a 1.3-MDa hexadecameric complex that catalyzes the phosphorylation and activation of glycogen phosphorylase b. PhK has an absolute requirement for Ca²⁺ ions, which couples the cascade activation of glycogenolysis with muscle contraction. Ca²⁺ activates PhK by binding to its nondissociable calmodulin subunits; however, specific changes in the structure of the PhK complex associated with its activation by Ca²⁺ have been poorly understood. We present herein the first comparative investigation of the physical characteristics of highly purified hexadecameric PhK in the absence and presence of Ca²⁺ ions using a battery of biophysical probes as a function of temperature. Ca²⁺-induced differences in the tertiary and secondary structure of PhK measured by fluorescence, UV absorption, FTIR, and CD spectroscopies as low resolution probes of PhK's structure were subtle. In contrast, the surface electrostatic properties of solvent accessible charged and polar groups were altered upon the binding of Ca²⁺ ions to PhK, which substantially affected both its diffusion rate and electrophoretic mobility, as measured by dynamic light scattering and zeta potential analyses, respectively. Overall, the observed physicochemical effects of Ca²⁺ binding to PhK were numerous, including a decrease in its electrostatic surface charge that reduced particle mobility without inducing a large alteration in secondary structure content or hydrophobic tertiary interactions. Without exception, for all analyses in which the temperature was varied, the presence of Ca²⁺ rendered the enzyme increasingly labile to thermal perturbation.

In the five decades since the discovery of skeletal muscle phosphorylase kinase (PhK), the first protein kinase to be purified and characterized (Krebs and Fischer 1956), its activity (the phosphorylation and activation of glycogen phosphorylase b) and regulation thereof have been well characterized. The kinase activity of PhK has an absolute requirement for Ca²⁺ ions, which couples skeletal muscle contraction with the cascade activation of glycogenolysis to supply the short-term power demands of muscle (Shulman 2005). In contrast to the large amount of published data on the control of PhK's activity, structural information on the PhK holoenzyme complex has remained relatively limited, including those structural changes associated with its activation by Ca²⁺ ions.

PhK is a 1.3-MDa complex comprising four copies of four distinct subunits arranged as a hexadecameric (αβγδ)₄ pseudo-tetrahedron, a bridged, bilobate structure of apposing (αβγδ)₂ octameric lobes (Norcum et al. 1994; Nadeau et al. 2002, 2005; Venien-Bryan et al. 2002). The protein kinase activity of the PhK complex rests with its γ subunits, which in the absence of Ca²⁺ ions are constrained by its regulatory α, β, and δ subunits. The δ subunits, which are integral, nondissociable molecules of calmodulin (CaM), account for PhK's Ca²⁺-binding capacity. The catalytic γ subunits are structurally similar to other Ser/Thr protein kinases, especially cAMP-dependent protein kinase (PKA) (Owen et al. 1995); but unlike PKA, they contain a C-terminal extension, a regulatory domain that strongly binds CaM (Dasgupta et al. 1989), the δ subunit, and which also interacts with the α (Nadeau et al. 1999; Rice et al. 2002) and perhaps β (Nadeau et al. 2007) subunits within the PhK complex. It has been suggested that the binding of Ca²⁺ to δ directly or indirectly alters its interactions with the regulatory domain of γ, leading to activating (“deinhibiting”) conformational rearrangements involving multiple subunits of the PhK complex (Rice et al. 2002).

A variety of physical and chemical approaches have been used over the years in our own and other laboratories in attempts to gain structural information relating to the allosteric activation of the PhK complex by Ca²⁺ ions: differential ultraviolet (UV) absorbance (Dimitrov 1978), circular dichroism (CD) (Chan and Graves 1982), extrinsic 1-analinonaphthalene-8-sulfonate (ANS) fluorescence (Steiner and Sternberg 1982), differential proteolysis (Trempe and Carlson 1987), photoaffinity labeling (King et al. 1982), and chemical cross-linking (Nadeau et al. 1997). In these earlier studies, the PhK preparations that were used contained not just PhK hexadecamers, but also varying amounts of larger aggregates of soluble multimers (Cohen 1973), which could present an interfering complication for spectrophotometric approaches and possibly even for chemical ones. To eliminate these aggregates, a size exclusion–high performance liquid chromatography (SE-HPLC) method was developed (Traxler et al. 2001), allowing structural studies to be performed on monodisperse PhK (αβγδ)₄ hexadecamers. Utilization of such a separation step immediately prior to analyses is especially important when studying the effects of Ca²⁺, because Henderson et al. (1992) found that both Ca²⁺ ions and a small initial amount of soluble aggregates promoted further aggregation. Two recent structural studies have taken advantage of the SE-HPLC separation step prior to analyses to examine the effects of Ca²⁺ on the overall structure of the PhK complex. In the first of these, three-dimensional models were reconstructed from electron microscopy images of negatively stained PhK ± Ca²⁺ ions (Nadeau et al. 2002). Here, Ca²⁺ was observed to induce widely separated structural changes in both the lobes and bridges of the PhK structure, prompting those conformational changes to be referred to as global. The second and more recent study used small-angle X-ray scattering (SAXS) to study the effects of Ca²⁺ ions on PhK's overall dimensions and distribution of scattering densities in solution (Priddy et al. 2005). Modeling of those results suggested a modest Ca²⁺-dependent global redistribution of mass. To learn what changes in the intrinsic properties of PhK might be associated with those structural changes, we have now used a wide battery of biophysical techniques to characterize the physicochemical properties of hexadecameric PhK in the absence and presence of Ca²⁺ ions and as a function of temperature when possible. Our current findings indicate that, despite their being transmitted over a long distance in the PhK complex (Nadeau et al. 2002), the observable structural changes caused by Ca²⁺ are modest, but do give rise to a large change in the overall electrostatic surface charge of the complex.

Results

Tertiary structure assessment

Second derivative UV absorption spectroscopy

Derivative UV absorption spectroscopy was used to simultaneously monitor the spectroscopic properties of discrete populations of PhK's aromatic side chain chromophores (Mach and Middaugh 1994) in the absence or presence of its obligatory activator, Ca²⁺. The wavelength positions of PhK's six second derivative minima are plotted in Figure 1 as a function of temperature for the nonactivated and Ca²⁺-activated conformers. The environments of the three photoreactive aromatic chromophores vary little between the two conformers of PhK at the initial 30°C temperature; however, a differential response to increasing temperature is manifested as divergent transition initiation temperatures and thermal transition midpoints, with the differences ranging from 6.4°C for Trp (A ₀ ≈ 296 nm) to 10.2°C for Phe (A ₀ ≈ 251 nm). For both forms of PhK, increased temperature causes Phe (Fig. 1A,B) and Trp chromophores (Fig. 1F) to become on the average exposed to a more polar, presumably aqueous, environment, as evidenced by the short wavelength, blue shifted transitions (Demchenko 1986). With respect to the Tyr bands of Figure 1C,D, the transitions are red shifted as the temperature increases, indicative of chromophore burial, i.e., exposure to a less polar environment (Demchenko 1986). Figure 1E also displays the relative Tyr/Trp exposure ratio (Ragone et al. 1984) and their overlapped Tyr/Trp absorbances, which are also red shifted in response to increased temperature. The thermal transitions of Figure 1A–F strongly suggest that the similar initial tertiary structure arrangement of the PhK complex, as assessed by UV absorbance, is maintained to higher temperatures in the Ca²⁺-free conformer than in the Ca²⁺-activated form.

Figure 1.

Second derivative UV absorption spectroscopy of PhK. Each data point is a second derivative minimum in the wavelength range characteristic of protein aromatic chromophores, (A) Phe, (B) Phe, (C) Tyr, (D) Tyr, (E) Tyr/Trp combination, and (F) Trp, of nonactivated (○) and Ca²⁺-activated (▵) PhK as a function of temperature. The transition midpoint of each chromophore shift, ±Ca²⁺, is listed within the individual panels A–F. The relative Tyr/Trp exposure ratios determined by the method of Ragone et al. (1984) for the Tyr284 versus Trp296 second derivative minima are also displayed in panel E for nonactivated (—) and Ca²⁺-activated (···) PhK as a function of temperature and are indicated on the right ordinate.

Intrinsic Trp fluorescence

In another approach to analyzing tertiary structure changes, both the relative intrinsic Trp fluorescence emission intensity (I_E) and wavelength maximum (λ_max) were monitored as a function of temperature for PhK, ± Ca²⁺ (Fig. 2). The initial equivalent values for λ_max and I_E for the two conformers indicate that the overall polarity of the tertiary structure environments of their Trp fluorophores are similar at 30°C. Both conformers, however, undergo a slight blue shift in λ_max as a function of increasing temperature, with the Ca²⁺-activated form exhibiting a transition of greater magnitude (Fig. 2). In general, a λ_max blue shift is likely to be accompanied by a relative increase in I_E, or at least a sustained I_E, as the fluorophore upon being buried is protected from fluorescence quenching by the polar solvent, proximal charged or uncharged polar side chains, and/or polar peptide backbone groups (Demchenko 1986). Thermally induced conformational rearrangements of the Ca²⁺-activated PhK (▴) are indicated by a λ_max blue shift that is initiated near 50°C. In comparison, the shift of the λ_max for the nonactivated conformer (•) is gradual, without an obvious inflection point, and is minimal in magnitude throughout the entire temperature gradient. These results suggest that the overall fluorophore environment of the Ca²⁺-free PhK remains relatively fixed, while that of the Ca²⁺-bound conformer is more labile to thermal perturbation. For Ca²⁺-free PhK, the relatively large decrease in I_E(T) (○) with little change in λ_max is characteristic of globular proteins in aqueous solution, wherein the rate of decrease of I_E(T) is a function of temperature-dependent internal conversion and a decrease in effective concentration. For the more thermally labile Ca²⁺-activated conformer (▵), the I_E is sustained at temperatures beyond the 50°C λ_max transition, probably as a result of increased burial of the indole side chains.

Figure 2.

Intrinsic Trp fluorescence emission for PhK as a function of temperature (λ_ex = 295 nm). The fluorescence intensity maxima relative to the initial measurements at 30°C are traced for nonactivated (○) and Ca²⁺-activated (▵) PhK. The λ_max shifts are plotted by monitoring fluorescence emission maxima for nonactivated (•) and Ca²⁺-activated (▴) PhK. Where so designated, the symmetrical error bars extend downward for nonactivated PhK and upward for Ca²⁺-activated PhK.

ANS fluorescence

The fluorescence emission of ANS is strongly quenched by water. Its emission intensity, however, is increased up to 100-fold in apolar environments (Rosen and Weber 1969). Thus, ANS was used as a probe for the Ca²⁺-induced appearance of dye (apolar) binding sites (Stryer 1965) by monitoring I_E and λ_max as a function of temperature (Fig. 3). Light scattering was simultaneously measured (Fig. 3A) as a monitor of potential protein aggregation throughout the temperature gradient. ANS I_E at 30°C in the presence of Ca²⁺-free PhK is 85% that of ANS in the presence of Ca²⁺-bound PhK, which is within the limit of error of the measurement (Fig. 3B), and the λ_max is also not significantly different for the two conformers (< 1 nm, Fig. 3B). Scattering, λ_max and I_E are relatively similar from 30°C to 40°C, but diverge at 42.5°C, indicating a greater exposure of ANS to a less polar environment in the presence of the Ca²⁺-activated form of PhK. The blue shifted λ_max that occurs concomitant with an increase in I_E (Fig. 3B) reaches its maximum at 55°C in the presence of Ca²⁺-activated PhK and is followed by a decrease in I_E, which is probably the result of protein aggregation and precipitation (Fig. 3A). Thermally induced transitions in the presence of the Ca²⁺-free conformer begin 6°C–8°C higher for the same parameters, reaching a maximum I_E at 60°C (Fig. 3B), while the λ_max shift (Fig. 3B) and maximum scattering intensity occur at 62.5°C (Fig. 3A). Thus, measures of the tertiary structure of PhK by second derivative UV absorbance, as well as intrinsic Trp and ANS fluorescence, indicate that the structures of both conformers are quite similar at 30°C, which is the temperature that has conventionally been used to study the activity of PhK and its regulation. At higher temperatures, their tertiary structures appear to diverge.

Figure 3.

Extrinsic ANS fluorescence as a function of temperature (λ_ex = 385 nm). (A) Protein scattering intensity measured at 390 nm is traced for nonactivated (•) and Ca²⁺-activated (▴) PhK. (B) The fluorescence intensity of ANS in the presence of nonactivated (•) and Ca²⁺-activated (▴) PhK, as well as the λ_max. shifts of the fluorophore in the presence of nonactivated (○) and Ca²⁺-activated (▵) PhK. Where so designated, the symmetrical error bars extend downward for nonactivated PhK and upward for Ca²⁺-activated PhK.

Secondary structure characteristics

Far UV CD

To further characterize the physicochemical effects caused by the binding of Ca²⁺, CD was used to estimate the secondary structure content of the two PhK conformers as a function of temperature. Their spectra at the initial temperature of 30°C are very similar (Fig. 4A). To estimate the secondary structure content, the 30°C spectra of Figure 4A were analyzed by two different secondary structure prediction programs, CDPro (Sreerama and Woody 2000) and DICHROWEB (Lobley et al. 2002). Both programs predict very similar secondary structure content for the two conformers, with each estimated to have ∼70% helix, 20% sheet, and 10% unordered regions. CD ellipticity at UV wavelengths associated with particular secondary structures is displayed as a function of temperature in Figure 4B (helix₂₀₈), 4C (sheet₂₁₅), and 4D (helix₂₂₂). In each case, Ca²⁺-free PhK maintains the negative ellipticity characteristic of these structural elements to higher temperatures than does its Ca²⁺-activated counterpart. The transition midpoint temperatures are similar to those observed in second derivative UV spectroscopy: in the mid to upper forties for Ca²⁺-activated PhK and the mid to upper fifties for the Ca²⁺-free conformer.

Figure 4.

Far UV CD analysis of PhK. (A) CD spectra of nonactivated (—) and Ca²⁺-activated (···) PhK at 30°C. Ellipticity of nonactivated (○) and Ca²⁺-activated (▵) PhK as a function of temperature for the secondary structural elements of proteins at (B) 208 nm (characteristic of α-helix), (C) 215 nm (characteristic of antiparallel β-sheet), and (D) 222 nm (characteristic of α-helix). The transition midpoint of each chromophore shift, ±Ca²⁺, is listed within the individual panels B–D.

FTIR

To confirm the CD results, the protein Amide I band (1600–1700 cm⁻¹) of the infrared spectrum of PhK was analyzed (Krimm and Bandekar 1986). The second derivative spectrum in this region (Fig. 5) reveals well-resolved spectral minima that correspond to characteristic absorbance frequencies of the different possible secondary structures. The integrated peak areas of individual curves that are center-aligned to each second derivative minimum were used to estimate the relative composition of helix, sheet, turns, and unordered features as a percent of the total area of their corresponding zero order spectra (Table 1). Within experimental error, the total β-structure content of PhK at 25°C appears to be maintained at 44% upon the binding of Ca²⁺. A significant difference between the two conformers is apparent, however, at 1652 cm⁻¹ (Fig. 5), a frequency near the band usually assigned to α-helices (1653 ± 4 cm⁻¹; Susi and Byler 1986). The second derivative spectrum of the nonactivated kinase in the 1650 cm⁻¹ region suggests the presence of simple α-helices, while that of the Ca²⁺-activated conformer exhibits a shoulder emanating from this band at ∼1647 cm⁻¹ (Fig. 5). Second derivative minima near 1645 ± 4 cm⁻¹ are typically assigned to unordered structure (Susi and Byler 1986) or distorted helices (Trewhella et al. 1989). Based on these assignments, this suggests that the binding of Ca²⁺ may result in a decrease in the amount of highly structured α-helices (Fig. 5; Table 1).

Figure 5.

Second derivative infrared spectra of PhK. Second derivative intensity of the Amide I band (1600–1700 cm⁻¹) of the infrared spectrum for nonactivated (—) and Ca²⁺-activated (···) PhK at 25°C. Peaks were initially aligned to each minimum to quantify the percent secondary structure content listed in Table 1.

Table 1.

Percent secondary structure content of PhK from FTIR analysis at 25°C

Overall, the solution-based techniques employed to assess Ca²⁺-induced structural changes in PhK reveal only slight alterations in its secondary and tertiary structures at lower temperatures. In contrast, profound effects of the binding of Ca²⁺ ions are apparent at higher temperatures for all spectral features examined. To determine whether the spectral data obtained at higher temperatures are descriptive of native hexadecamers or aggregates thereof (Henderson et al. 1992) or possibly thermally disrupted smaller species, it was necessary to characterize the higher order structure of both PhK conformers throughout the temperature range examined.

Higher order structure

Dynamic light scattering (DLS)

The size (hydrodynamic diameter, d) of PhK in the absence and presence of Ca²⁺ was estimated by dynamic light scattering. At the initial experimental temperature, Ca²⁺ causes an increase in d from 39 ± 0.4 nm to 43 ± 1.1 nm (Fig. 6A). The size of the Ca²⁺-ligated enzyme then progressively grows as a function of increased temperature. In contrast, the nonactivated PhK maintains its initially smaller diameter until nearly 50°C (Fig. 6A). In a related experiment in which the temperature was held constant at 30°C for 90 min after the addition of Ca²⁺, the diameter did not continue to grow during this period (data not shown), ruling out time-dependent effects. Likewise, nonspecific effects of increased ionic strength were eliminated in separate analyses by introducing a series of both monovalent (Na⁺ and K⁺) and divalent (Mg²⁺ and Sr²⁺) cation-chloride salts to match the ionic strength of the CaCl₂ used in all other studies. No change in d was observed, indicating that the effects of Ca²⁺ are specific and unrelated to increased ionic strength.

Figure 6.

Dynamic light scattering of PhK. (A) Hydrodynamic diameter of the nonactivated (○) and Ca²⁺-activated (▵) kinase as a function of temperature compared with the simultaneous static scattering count rates of nonactivated (•) and Ca²⁺-activated (▴) PhK. (B) A measure of the heterogeneity of the scattering particles, the polydispersity index, as a function of temperature for nonactivated (—) and Ca²⁺-activated (···) PhK.

The polydispersity index (PDI) (Fig. 6B) is a measure of the heterogeneity of the scattering species within a given sample, with the lower value of PDI indicating a more uniform distribution of size (Berne and Pecora 1976). The PDI values from 10°C to 40°C are on the order of 0.2, which suggests the possibility of some heterogeneity in the particle size (Fig. 6B). The temperature-induced increase in size of Ca²⁺-activated PhK from 10°C to 40°C occurs concomitant with a PDI value decrease, implying an increase in size uniformity to 40°C (Fig. 6B), which is, coincidentally, the upper limit of the normal body temperature of the rabbit (39.6°C–40°C) (McClure 2005). In contrast, the PDI of nonactivated PhK remains constant, up to and well beyond 40°C (Fig. 6B), as would be expected of a population of molecules that does not exhibit a temperature-dependent increase in diameter below 50°C (Fig. 6A).

Surface charge/zeta potential

Zeta potential measurements obtained by analyzing the shear surface electric potential of colloid particles in solution provide an estimate of effective surface charge. Using phase analysis light scattering (McNeil-Watson et al. 1998), the zeta potential of nonactivated PhK at 25°C and pH 6.8 was measured to be −32.8 ± 0.6 mV. In contrast, the zeta potential of the Ca²⁺-activated conformer under the same conditions is −22.1 ± 0.7 mV, a considerably more positive and presumably less soluble form of the enzyme (Hunter 1981). This Ca²⁺-induced change is consistent with either the burial or neutralization of acidic surface side chains or increased exposure of basic moieties.

To determine whether the effect of Ca²⁺ might be a general solvent effect or specific for PhK, these experiments were repeated with two proteins not known to bind Ca²⁺, namely porcine gelatin and human IgG. The zeta potential of these proteins remained unaltered when measured in solution at 25°C and pH 6.8 in the absence or presence of Ca²⁺ ions equal to or 10 times greater in concentration than that used with PhK.

Discussion

Differences in the tertiary structure of PhK ± Ca²⁺, as assessed by second derivative UV absorbance, intrinsic Trp fluorescence, and ANS fluorescence, were found to be very subtle or nonexistent at low temperatures, but to become more pronounced at increased temperatures. It is not surprising that PhK's side chain chromophores are not particularly sensitive conformational probes, given the total of 1012 aromatic chromophores (428 Phe, 460 Tyr, and 124 Trp) within the 1.3-MDa complex. As such, the information obtained by second derivative UV absorbance and intrinsic Trp fluorescence reflects only the cumulative state of all the chromophores pertinent to each technique. With increasing temperature, the ratio of the relative exposure of the Tyr versus Trp side chains decreases and is relatively unaffected by Ca²⁺ (Fig. 1E). With a Tyr/Trp ratio of greater than 1.25 at 30°C, more Tyr are solvent exposed than at increased temperatures, where the ratio is decreased and the Tyr residues become relatively less exposed than the Trp, in agreement with the red and blue shifts of their individual chromophore thermal transitions (Fig. 1C,D,F). Overall, the Ca²⁺-induced changes that were observed in tertiary structure did not indicate measurable large-scale alterations, especially at low temperatures, which supports previous SAXS results that Ca²⁺ causes only a subtle redistribution of density involving regions in the PhK complex that are separated by short to midrange vector lengths (Priddy et al. 2005). The major effect of Ca²⁺ on tertiary structure is to yield a form of the enzyme that is less tolerant to thermal perturbation, with the Ca²⁺-free conformer maintaining the integrity of its tertiary arrangement from 5°C to 10°C higher temperatures than its Ca²⁺-bound counterpart. Because the loss of protein structure as a function of increasing temperature was irreversible for the techniques screened, the thermal transition midpoints given in this study are meant to be used only to compare the differential response (± Ca²⁺) of the type of protein structure being probed; they are not meant to imply anything about the relative stabilities of the two initial conformers.

Differences in secondary structure between the two PhK conformers are also minor when estimated by CD at low temperatures (e.g., 30°C in Fig. 4A). Furthermore, the ratio of negative ellipticity at 208 nm to 222 nm, which is characteristic of α/β proteins (Levitt and Chothia 1976), is essentially the same for both conformers at this temperature. This spectral feature, greater negative ellipticity at 208 nm than 222 nm, has been shown to be characteristic of proteins containing helices and sheets that are associated with each other, in contrast to α + β proteins, wherein helices and sheets do not interact to a large extent (Chothia et al. 1977). Given that proteins containing mixed secondary content display predominantly helical character by CD (Manavalan and Johnson 1983), the large fraction of helix calculated from the CD data by the CDPro (Sreerama and Woody 2000) and DICHROWEB (Lobley et al. 2002) algorithms is probably an overestimate of the amount of helix in these conformers. As was the case with the tertiary structure studies, CD reveals that in the absence of Ca²⁺ ions, PhK maintains its secondary structure integrity to higher temperatures than when Ca²⁺ is bound. Moreover, throughout the temperature gradient for both conformers, the negative ellipticity at 208 nm grew progressively less negative at a faster rate than did the ellipticities at 215 and 222 nm (Fig. 4B vs. Figs. 4C,D). These results suggest that α/β interactions are disrupted at relatively low temperatures, while the helix and sheet structures remain independently stable. This agrees with the commonly accepted process of thermal unfolding, wherein quaternary and tertiary arrangements are more labile to thermal perturbation than the relatively stable secondary structure elements. We attempted to directly assess the energetics of the thermal unfolding of the two PhK conformers by differential scanning calorimetry; however, the resultant thermograms were not reproducible, presumably because of the significantly higher concentrations of protein required to measure the thermal transitions.

Considering FTIR and CD as complementary techniques for analyzing protein secondary structure can be helpful, especially in the absence of higher resolution data. For PhK, the percentages of specific secondary structures estimated by the two techniques do not wholly agree, but this is not uncommon. What is consistent between the two techniques is that both suggest only minor alterations in secondary structure at low temperatures in response to the binding of Ca²⁺ ions. Where Ca²⁺-induced alterations are apparent by FTIR, at the 1645 ± 3 cm⁻¹ and 1652 ± 4 cm⁻¹ bands that are commonly assigned to the vibrational frequencies of disordered and alpha helix structures, respectively (Byler and Susi 1986; Susi and Byler 1986), the actual alteration could reflect helix distortion. It has been observed that proteins containing distorted helices absorb in the infrared region at 1645 cm⁻¹. Coincidentally, one such protein is CaM (Trewhella et al. 1989). In fact, the presence of a 1645-cm⁻¹ second derivative shoulder in our Ca²⁺-bound PhK spectrum parallels the observations of Trewhella et al. (1989) regarding the effects of the addition of Ca²⁺ to CaM. In that study, the Ca²⁺/CaM second derivative IR spectrum contained a similar poorly resolved shoulder at 1644 cm⁻¹, and was suggested to result from distortion of the linker-helix. When a band of this vibrational frequency is more clearly resolved, and not merely present as a shoulder emanating from the main helix band, it is usually assigned to disordered content (Byler and Susi 1986; Susi and Byler 1986). In the present case, it seems prudent to assign the area of the shoulder at 1645 cm⁻¹ of the Ca²⁺-bound PhK spectrum to distorted helix, as did Trewhella et al. (1989), and then to sum the areas of the two bands to estimate the total amount of helix. This approach suggests that the percent helix contents of the two PhK conformers could differ by ∼4% (Table 1). Such a difference in helical content for the two conformers was not observed in the CD analyses; however, because CD tends to overestimate helix (Manavalan and Johnson 1983), whereas FTIR tends to overestimate sheet content (Susi and Byler 1986), an exact match is rarely achieved.

Besides the small Ca²⁺-induced alterations in secondary structure that were observed in the FTIR spectra of PhK, Ca²⁺ also increased its hydrodynamic diameter from 39 ± 0.4 to 43 ± 1.1 nm, as measured by DLS. This change in diameter detected by DLS could theoretically be attributed to (1) a change in symmetry, (2) an actual increase in hydrodynamic diameter (diffusion constant), or (3) a change in the surface of shear interface. The first two of these possibilities seem improbable based on the results of previous nsEM (Nadeau et al. 2002) and SAXS (Priddy et al. 2005) studies. In fact, in the latter study, both conformers of PhK were found to have nearly identical maximum linear dimensions and radii of gyration. Thus, the third alternative, a change in the surface of shear, seems a more likely cause for the observed increase in the apparent diameter of PhK. This was further examined by measuring PhK's zeta potential (McNeil-Watson et al. 1998). The zeta potential reflects the colloid shear surface to solvent interface environment (the Stern Layer), a relatively ordered layer of solvent molecules that diffuses as part of the colloid particle. A change in electrostatic surface potential as a result of an alteration in the interactions of solvent-accessible charged residues will alter the Stern Layer. For PhK at pH 6.8, the Ca²⁺-free conformers have a −32.8 ± 0.6 mV zeta potential, characteristic of soluble proteins, compared to a considerably less negative zeta potential of −22.1 ± 0.7 mV for Ca²⁺-bound PhK. Considering the large differences that were observed by DLS and zeta potential upon binding Ca²⁺, together with the comparatively small rearrangements of tertiary and secondary structure observed by the other spectroscopic methods and the unchanged dimensions from SAXS, we conclude that the increase in apparent hydrodynamic size brought about by Ca²⁺ results from an altered Stern Layer, with an associated decrease in particle mobility, and does not result from an actual increase in the dimensions of the PhK complex.

That the binding of Ca²⁺ leads to an alteration in the state of solvent-accessible charged residues of the PhK complex is consistent with results from a variety of previous studies on the enzyme. For example, in the presence of neutral salts, the Ca²⁺-dependent catalytic activity of PhK is dramatically decreased (Carlson and Graves 1976), suggesting that these salts may disrupt critical ionic interactions. Conversely, the presence of polar molecules, such as ethanol or acetone, has been shown to greatly increase PhK's activity (Singh and Wang 1979), perhaps by enhancing favorable polar interactions. Furthermore, virtually all chemical cross-linkers of PhK, which are attacked by its polar or charged amino acid side chains, yield differential subunit cross-linking patterns ± Ca²⁺ (Nadeau et al. 1997, 1999; Rice et al. 2002). Similarly, the yields of chemically modified subunits that have been either carbamidomethylated by iodoacetamide (thiol selective) or reductively methylated by formaldehyde (amine selective) are increased in the presence of Ca²⁺ (Nadeau et al. 1997). For differential chemical modification of PhK to be achieved as a function of its degree of saturation by Ca²⁺ would require at least some structural rearrangements to reorder the solvent accessibility of the reactive residues, a fact that supports the DLS and zeta potential results.

Recently an electrostatic switch has been reported to be important in the activation of PKA (Vigil et al. 2006), a structural homolog of PhK's catalytic γ subunit. Perhaps even more relevant to the PhK complex is the switch involving a group of critical residues that participate in the charge–charge interactions between the inhibitory region of cardiac troponin I (cTnI) and its cognate linker helix region of cardiac troponin C (cTnC) (Lindhout et al. 2005). The cTnI subunit shares a high degree of sequence similarity with the C-terminal regulatory domain of PhKγ (Paudel and Carlson 1990), which has been shown to interact with CaM (Dasgupta et al. 1989; Harris et al. 1990), PhK's intrinsic δ subunit. This subunit is, in turn, a homolog of cTnC, the Ca²⁺-binding subunit of the troponin complex. Because these regulatory elements of PhK's γ and δ subunits are structurally and functionally comparable to TnI and TnC, respectively, there is a basis for assuming that the Ca²⁺-dependent, variant electrostatic interactions that regulate TnI and TnC interactions (Lindhout et al. 2005) may also occur within the PhK complex to regulate its Ca²⁺-dependent activity.

In a reconstituted binary complex of just the γ and δ subunits of PhK, charge reversal mutations of residues E82, E83, and E84 of the linker helix of the δ subunit were shown to disrupt activation of the γ subunit (Farrar et al. 1993). Comparing the sequences of PhKδ and cTnC, E82 of PhKδ is homologous to E94 of cTnC, a residue that is critical in forming a salt bridge with K108 of cTnI (Lindhout et al. 2005), whose sequence in turn aligns with R343 of PhKγ in the CaM-binding regulatory region of this subunit. A separate study that sought to analyze the asymmetric charge distribution of CaM revealed that the same three Glu residues found by Farrar et al. (1993) to be important in the activation of PhKγ by CaM were also essential in maintaining the Ca²⁺-dependent interactions between CaM and myosin light chain kinase (Weber et al. 1989). In short, charge reversal mutations of EEE82–84KKK did not affect the Ca²⁺-binding ability, stability, secondary structure, or macroscopic structure of CaM itself, but did significantly suppress the binding to and activation of myosin light chain kinase by CaM (Weber et al. 1989). It will be of considerable interest to learn how many similarities are shared among the Ca²⁺-dependent, activating interactions between PhKδ and PhKγ, between TnC and TnI, and in general between CaM and other CaM-dependent enzymes. In summary, because we have determined that the rabbit fast-twitch skeletal muscle isoform of PhK does not undergo a great deal of tertiary or secondary structure rearrangement upon binding Ca²⁺ ions at low temperatures, we conclude that the structural differences that do exist between these two conformers are predominantly electrostatic in nature. Although speculative, it might be noted that the C-terminal regulatory region of PhKγ that binds PhKδ (CaM) has a pI of 10.2; thus, increased surface exposure of this region of γ upon the binding of Ca²⁺ by δ would lead to a more positive (i.e., less negative) zeta potential, as was observed.

Materials and methods

PhK preparation

Nonactivated PhK was purified from fast-twitch skeletal muscle of New Zealand White rabbits as previously described (King and Carlson 1981) and stored at −80°C in buffer containing 50 mM HEPES, 0.2 mM EDTA, and 10% (w/v) sucrose at pH 6.8. Freshly thawed PhK was further purified immediately before use in two steps: (1) centrifugation for 30 sec at 10,000g to remove insoluble precipitates, and (2) SE-HPLC (Traxler et al. 2001) to remove the soluble aggregates. Fractions containing only hexadecamers were retained for spectroscopic analyses after elution from the BioSep SEC-S4000 (Phenomenex) column developed with a mobile phase containing 6 mM HEPES, 200 mM NaCl, 0.2 mM EGTA (pH 6.8) at a flow rate of 0.6 mL/min. The enzyme solution was subsequently dialyzed twice in 16 h against a 6 mM HEPES, 0.2 mM EGTA (pH 6.8) buffer to reduce Cl⁻ from the HPLC mobile phase. The PhK concentration was determined for each sample by UV A₂₈₀ using an absorptivity coefficient of 12.4 for a 1% protein solution (Cohen 1973). For analyses of nonactivated PhK, the dialyzed samples were diluted with dialysis buffer to 100 μg/mL (unless otherwise noted) and, to achieve the same protein concentration for Ca²⁺-activated PhK, the solution was diluted with the same buffer containing CaCl₂ to achieve a final Ca²⁺ ion concentration of 0.5 mM, or 0.3 mM greater than the chelator. Reference spectra were subtracted from a minimum of three replicates of protein raw spectra, except for FTIR (n = 2), and analyzed using Origin Professional v. 7.5 Scientific Graphing and Analysis Software (OriginLab, formerly Microcal).

Second derivative UV absorption spectroscopy

UV absorption spectra were collected with 10-sec integration times on protein samples in Teflon stoppered, 1-cm, quartz cuvettes. Spectra were collected every 2°C with 300 sec equilibration times from 30°C to 70°C, while monitoring turbidity at 350 nm employing an Agilent 8453 diode-array spectrophotometer. Sample temperature was regulated by an Agilent 8990A Peltier-type temperature control device. Second derivatives of individual spectra were calculated by applying a nine-point filter and fifth-degree Savitsky–Golay polynomial fitted to a cubic function with a 99-point interpolation per raw data point in the UV-Visible Chemstation software suite (Agilent Technologies). Second derivative peak minima were assigned by the peak picking tool within the OriginPro software program.

Intrinsic Trp fluorescence spectroscopy

Fluorescence emission spectra were collected from PhK samples in Teflon stoppered, 1-cm, quartz cuvettes scanned from 305 to 405 nm at 1-nm intervals with a scan rate of 20 nm/min on a Jasco Series 810 spectrofluorometer equipped with an internal Peltier-type temperature controller. The instrument was set to emit at a 295-nm excitation wavelength, with 4-nm excitation and 4-nm emission slit widths, with data collection every 2°C from 30°C to 70°C and a 300-sec equilibration time between data collection intervals.

ANS fluorescence

Subsequent to dialysis, PhK samples were mixed with dialysis buffer containing the fluorescent probe ANS and diluted to 40 μg/mL protein and 6 μM ANS. Fluorescence emission spectra were collected from these samples in Teflon stoppered, 1-cm, quartz cuvettes from 400 to 600 nm at 1-nm intervals at a scan rate of 1 nm/sec using a 385-nm excitation wavelength on a QuantaMaster (Photon Technology International) spectrofluorometer equipped with an internal Peltier-type temperature controller. A 4-nm excitation and 4-nm emission slit width was employed for both detectors aligned at 90° to the incident beam in a “T” configuration, with a 300-sec equilibration time between every 2.5°C data collection interval from 30°C to 70°C.

CD spectroscopy

CD spectra were collected from samples in Teflon stoppered, 0.1-cm, quartz cuvettes from 260 to 190 nm at 1-nm intervals with a 20 nm/min scan rate with a Jasco Series 710 spectropolarimeter equipped with an internal Peltier-type temperature controller under a constant nitrogen purge. The excitation slit width was set to 5 nm with data collection from 30°C to 70°C and a 300-sec equilibration time interspersed every 2°C temperature increment. Secondary structural assignments were calculated for the 30°C data sets using CDPro (Sreerama and Woody 2000) and DICHROWEB (Lobley et al. 2002) software.

FTIR spectroscopy

PhK solutions were concentrated after dialysis in centrifugal concentrators (Amicon) to a final concentration of ∼4 mg/mL. Spectra (128) were collected at 4-cm⁻¹ resolution at 25°C on a Nicolet Magna-IR 560 equipped with a mercury–cadmium–telluride detector under a constant dry air purge using a 45° attenuated total reflectance crystal. The water association band near 2300 cm⁻¹ of a reference solution was subtracted from individual sample spectra to zero baseline the data for optimal contrast in the Amide I region (1600–1700 cm⁻¹) using the Omnic (Nicolet) software package. The resultant spectra were transferred to GRAMS/AI (Thermo Galactic Software) for derivative spectral analysis and assignment of secondary structure elements (Susi and Byler 1986). Peak position and area were adjusted to convergence between the residual and the baseline to an RMSD < 3.0 and linearity of R ² = 1.000 ± 0.0009.

DLS

The hydrodynamic diameter of PhK was measured for solutions of 25 μg/mL protein as a function of temperature from 10°C to 70°C by photon correlation spectroscopy employing a BLS-9000 (Brookhaven Instruments Corp.) DLS instrument. Particle hydrodynamic diameter (d) was calculated from measurements of the translational diffusion coefficient (D_t) based on the Stokes–Einstein equation

where k_b is the Boltzmann constant, T is the Kelvin temperature, η is the pure solvent viscosity, and d is the hydrodynamic diameter. The BLS-9000 was fitted with a 50-mW HeNe diode laser (JDS Uniphase) operating at 532 nm and a BI-200SM Goniometer with an EMI 9863 photomultiplier tube at 90° to the incident beam, connected to a BI-9000AT digital autocorrelator. The 10°C per hour temperature gradient was linear and controlled by a circulating water bath, with five consecutive 30-sec data collection intervals every 2.5°C throughout. The diffusion coefficient and the polydispersity of the protein solution were obtained from the method of cumulants, and static light scattering was determined by raw counts per second × 1000 (kcps) reaching the detector.

Surface of shear electrostatic potential

The zeta potential at pH 6.8 was collected at 25°C on a Brookhaven Instrument Corp. ZetaPALS analyzer. Light scattering at 15° to the incident 676-nm, 25-mW, solid-state laser beam was measured to calculate the electrophoretic mobility of the dissolved protein applying the Smoluchowski approximation, which is appropriate for the typical colloid and salt concentrations that were used for these experiments.