Introduction

The actin is an essential unit of the eukaryotic cytoskeleton and muscle sarcomeres with highly conservative sequences [1]. The specialized intracellular functions driven to low variety of proteins in their phylogenetic development [2,3,4]. Actin monomers (G-actin) and filaments (F-actin) have dynamical rearrangement thus manage cellular motility, division and transport processes [5,6,7,8,9,10,11]. The remodeling of actin filaments highly regulated by divalent cations (Ca2+ or Mg2+) [12] and nucleotides (ATP, ADP) [13,14,15,16,17,18,19,20]. The stability of G-actin depends on the bound cations and nucleotides but the majority in cytoplasm is in the ATP-G-actin form [21, 22] then subsequently hydrolyses to ADP and Pi with their polymerization [13,14,15,16,17,18,19,20]. The gelsolin regulates the rearrangement of cytoskeleton [21]. After the activation of gelsolin (GSN) by calcium ions, it can sever actin filaments then capped at its barbed end. In the cytoplasm nearby of the plasma membrane, GSN manages the cellular motions and morphology by capping, severing and uncapping of F-actin [22,23,24]. The process of uncapping is a crucial step to help the actin polymerization with an indirect influence on the membrane remodeling. Phosphatidylinositol 4,5-bisphosphate (PIP2) is involved in signal transduction and the regulation of the actin cytoskeleton by regulation of actin-binding proteins [25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42]. GSN can bind to PIP2 [43] and thus can be localized in the near of the plasma membrane when it is released from the end of F-actin [29, 44,45,46,47,48,49]. The PIP2-binding sites on gelsolin overlaps with the ATP binding site [47, 50,51,52]. The structural change of gelsolin with the mechanism of uncapping is not well known but in the absence of calcium the PIP2 competes with actin for binding to GSN [46, 49, 52,53,54]. However, we interpreted earlier that calcium activated GSN cannot bind PIP2 [49].

As previously shown the free calcium concentration in resting cells changing on the nanomolar level [55, 56], and in non-resting state can reach the micromolar levels [57, 58]. Cytoplasmic free magnesium levels are regulated in the nanomolar range but it does not have effect on GSN activity [49, 59, 60]. Interestingly, the reduction of calcium levels cannot cause directly the GSN release from actin, but a single calcium ion is still trapped in the complex [61]. The main concern with GSN inactivation is that it still depolymerizing actin filaments in the presence of millimolar EGTA provided nanomolar free calcium level [62, 63]. Possibly, only the PIP2 binding can fully inhibit any activity of GSN [49]. Structural rearrangement of F-actin and its binding molecules can be studied by DSC [64,65,66]. We probe here the isothermal enthalpy change within the interplay between GSN and F-actin under nano-, micro- and millimolar calcium concentrations and express the importance of PIP2 binding for the total inactivation of GSN. We can postulate that GSN shows remarkable thermal response to actin, addition under nanomolar calcium concentration which can be reduced by PIP2 binding.

Results

Figure 1 shows isothermal heatflow kinetics as well as the denaturation of mixing GSN with F-actin under different Ca2+ conditions, or in the presence of PIP2. All isotherm assays were taken 16 h but we focused on the first relevant 180 min after the relapse of the heatflow subsequently after the mixing effect, between the minimum (ΔQinitial) and the maximum points (ΔQtop) (Fig. 1A). The heatflow was dropped down (2.4 µW) (Table 1) if 12.8 µM GSN (1 mg mL−1) was added to nanomolar Ca2+ containing buffer then increased slowly (1.7 ± 0.3 × 10–4 µW s−1) in the first 3 h (Fig. 1B). In case of 46 µM F-actin (2 mg mL−1) bound GSN in a ratio of 1: 3.6; under nanomolar Ca2+ concentration, there were two remarkable exothermic peaks. The first peak was increased quickly 9.1 ± 2.5 W in the first 15 min by the rate of 82 ± 1.2 × 10–4 µW s−1, after a minor reduction it was followed by a second peak in 63 min. If GSN was formed complexes with 1.6 µM PIP2 [49] in the presence of nanomolar Ca2+ and F-actin, the heatflow was increased quickly with multiple rates (totally 4.6 ± 0.7 µW with an average rate of 17 ± 2.5 × 10–4 µW s−1) in 65 min. Just after the isothermal processes we performed denaturation with heating rate of 0.3 K min–1 (Fig. 1C). GSN in the presence of nanomolar Ca2+ shows a main transition at 39.5 °C (Table 2). GSN and F-actin in the presence of nanomolar Ca2+ shows four transitions at 33, 40, 54 and 60 °C and the GSN-PIP2 complex with F-actin show three transitions at 36, 56 and 60.5 °C. PIP2 binding shifted the main peak from 33 to 36 °C, however in case of free GSN the biggest transition was observable at 39.5 °C. Figure 1D shows that the isothermal heatflow of F-actin was dropped down (endothermic 3.6 µW)(Table 1) by mixing with millimolar Ca2+ containing buffer then it was increased by 3.23 µW first in a quick step 100 ± 0.8 × 10–4 µW s–1 then slowly 6 ± 1.5 × 10–4 µWs−1 in the first hour. The previously described two exothermic peaks of GSN and F-actin under nanomolar Ca2+ were turned to be reduced in the presence of micromolar Ca2+. Interestingly, we observed a minor drop (endothermic 1.36 µW) of the heatflow then it was followed by an exothermic change 3.84 ± 1.2 µW with the rate of 37 ± 0.8 × 10–4 µW s−1. If GSN was added to F-actin in the presence of millimolar Ca2+, the heatflow showed a big drop 9.7 µW at the beginning of the process then it was increased by 7.7 ± 2.8 µW with a relatively quick rate 17 ± 1.2 × 10–4 µW s−1 in the first two hours. Figure 1E shows that the shape of thermal denaturation curves in each cases from panel C seems to be Ca2+ dependent. F-actin in the presence of milimolar Ca2+ shows a main transition at 60 °C (Table 2). As we described above, the GSN and F-actin in the presence of nanomolar Ca2+ shows four transitions. However, in the presence of micromolar Ca2+, there were multiple peaks at 45.5; 49; 53.5 and 60 °C. Under millimolar Ca2+, there was only one main transition at 58 °C in case of GSN and F-actin.

Fig. 1
figure 1

Isothermal severing and capping of F-actin by GSN. A Schematic curve of isothermal heat flow change by the time with the mixing effect and the investigated parameters. B Time dependent change of isothermal heat flow in case of GSN with (light brown line /GF) or without (blue line /G) F-actin in the presence of nanomolar Ca2+. The response of GSN-PIP2 complex in the presence of F-actin (green line /GFP). C All cases from panel B were terminated with thermal denaturation DSC scans. D Isothermal heat flow kinetics of F-actin in the absence (cyan line /F) or in the presence of GSN under different Ca2+ concentrations, nanomolar (light brown line), micromolar (magenta line) or millimolar (dark brown line). E Heat denaturation curves of processes from panel D. The scans are the average of three measurements, and endotherm effect is deflected downwards which is valid with all panels

Full size image
Table 1 Isothermal kinetic data achieved by exponential fit to the heat flow change between the minimum and maximum value of the curves. The values are the average of three measurements
Full size table
Table 2 Heat denaturation parameters were collected after the end of the isothermal processes
Full size table

Discussion and conclusions

The aim of our present work is to show thermodynamical changes due to the activation and inactivation of gelsolin. Heat flow follows the structural entropy change as increases with structural stabilization.

Under nanomolar Ca2+ concentration, GSN responded with a minor isothermal change to the mixing of sample. GSN addition to F-actin resulted a strong exothermic change linked to the reduction of their structural entropy. Possibly, free actin monomers were depleted by GSN binding [62, 63] which mechanism can be described as the Type I. Ca2+ binding sites [67] are partially saturated, and thus initialized a structural change of GSN for a cooperative binding to actin monomer. Here we found that the actin binding leads to a lower thermodynamical stability of GSN. The PIP2 binding was stabilized the structure of GSN similar to the free form with slow and weak response to the actin addition (Fig. 2). In good agreement with our previous results, since the intrinsic tryptophan of GSN responds with increased emission to PIP2 addition, the response was a less flexible structural dynamics of GSN [49]. At the end of the isotherm assays, the majority of actin was denatured around 60 °C, it can be interpreted by the hydrolyzation of ATP due to the long term processes [68].

Fig. 2
figure 2

Schematic model of calcium induced GSN activation in the near of the plasma membrane

Full size image

Under micromolar Ca2+ concentration, the Ca2+ binding to Type II. binding sites [67] were initialized structural change of GSN and the mixing of sample can make fragments of long actin filaments and all newly exposed barbed end can be quickly capped by the partially active GSN. However, a slow structural transition in all capped actin filaments was observable.

Under millimolar Ca2+ concentration, the fully active GSN can do very fast severing within the mixing of the sample and induced a relatively big endotherm drop of the heat flow at the beginning of the process. Then subsequently, there was a relatively slow exotherm change which can be explained by the capping and within caused slow structural rearrangement of short filaments. However, if we mixed F-actin with buffer only it shows a very fast then a very slow isothermal process, possibly it was a recovery after certain fragmentations. As Nag et al. published in 2009 [69] the transition temperature of GSN domains was increased from 35 to 60 °C with the change of nanomolar to millimolar Ca2+ concentration. Here we found that the heat denaturation of actin-GSN shows two main components nearby 30 °C and 60 °C in the presence of nanomolar, multiple peaks between 30 and 60 °C in the presence of micromolar and one main component nearby 60 °C in the presence of millimolar Ca2+. What we can interpret as the GSN domains were turned to be more compact by the calcium binding which were connected together by flexible linkers seems like beads on string.

Here we have demonstrated that PIP2 binding stabilized the whole structure of gelsolin and possibly reduced its actin monomer binding activity under nanomolar calcium as the typical cytoplasmic calcium concentration of resting cells. If gelsolin-capped filaments point at the plasma membrane, the local concentration of gelsolin increases by forming PIP2 clusters [70]. All the conditions which help the binding between gelsolin and PIP2, and hence, filament uncapping in case of resting cells. The total inactivation of gelsolin is required PIP2.

We can presume that the low free calcium concentration keeps on the structure of gelsolin which is able to bind actin within the cooperativity of actin bound calcium (Fig. 2). Gelsolin can help to manage monomer pool far from the membrane and it can be linked to a basic sensory mechanism which influences the direction of filament growth in the near of the membrane.

Materials and methods

Proteins

His-tagged human wild-type gelsolin was expressed in E. coli strain Rosetta2 (DE3) pLysS cells from a pSY5 plasmid [71]. The protein was subjected to Ni-NTA affinity chromatography, HRV 3C protease cleavage, followed by gel filtration (Superdex 200, GE Healthcare) in 10 mM MOPS, 150 mM NaCl, pH 8. Traces of calcium were removed by dialysis (4 mM MOPS, 1 mM EGTA, pH 7.4, overnight). Rabbit skeletal muscle actin was prepared from acetone powder by a modified protocol from Spudich and Watt [72]. Actin was stored in buffer A (4 mM MOPS, 0.2 mM ATP (ATP disodium trihydrate, Sigma-Aldrich), 0.1 mM CaCl2, pH 7.4). We applied 2 mM EGTA then 2 mM MgCl2. Actin polymerization process was initialized by addition of 100 mM KCl following the same protocol as in our previous study [64]. 12.8 μM gelsolin was incubated under physiological salt conditions (100 μM CaCl2, 100 mM KCl, 1 mM MgCl2, 0.2 mM ATP, 4 mM MOPS, pH 7.4) supplemented with EGTA or CaCl2 to vary the free calcium levels (calculated with Maxchelator Stanford); for nM: 2 mM EGTA; for µM: 100 μM EGTA; for mM: 1 mM CaCl2. We used 1.6 µM PIP2 (PtdIns-(3,5)-P2(1,2-Dihexanoyl), Cayman Chemicals) which is below the CMC of PIP2 [73,74,75], incubated with F-actin in the presence of 2 mM EGTA.

DSC measurements

The actin samples were freshly prepared before all measurements. The isotherm kinetics analysis was made by a SETARAM Micro-DSCII calorimeter on room temperature, 22 °C for 16 h. After the third hour of measurements, we observed the continuous dropping of heat flow which can be described by the decay of ATP and actin filaments. We used Hastelloy designed pair of “mixing batch” vessels for isotherm measurements (Vlower = 500 µL, Vupper = 200 µL) in the reference vessel the GSN was mixed with buffer only and in the sample one with F-actin in the presence of nanomolar, micromolar or millimolar free Ca2+ (calculated by MaxChelator, https://somapp.ucdmc.ucdavis.edu) and after it subsequently undergone for denaturation measurements. Kinetics measurements were carried out with an isotherm program then finally denatured by heating up them with 0.3 K/min up to 80 °C. MOPS buffer was used as a reference. The reference and sample vessels were equilibrated with a precision of ± 0.1 mg; this way we did not need to do any correction between vessels’ heat capacity. The first point of our analysis was the minimum of the heat flow change (ΔQinitial) after a single sharp peak which shows the mixing effect, and thus the curve was increased slowly to the maximum point (ΔQtop) (Fig. 1A). The exponential fitting resulted the time rate (k) of the peaks, the measurements were started in 5 min after reaching the thermal equilibrium between the vessels, but we used for the analysis the part of the curve after the mixing effect. Actin and gelsolin based molecular events were happened only in the first three hours then all curves decreased slowly possibly by the decay of ATP thus the most reasonable time duration of kinetics was analyzed in the first 10,800 s. All kinetic processes were terminated with heat denaturation assays. With the help of a two-point SETARAM peak integration setting, the denaturation calorimetric enthalpy was calculated from the area under the heat absorption curve, and then, the results [denaturation or melting temperature (Tm) and calorimetric enthalpy (ΔHcal) data of samples] were compared. This method is identical with the protocol as we applied in our previous study [64].