Interaction of formin FH2 with skeletal muscle actin. EPR and DSC studies

Formins are highly conserved proteins that are essential in the formation and regulation of the actin cytoskeleton. The formin homology 2 (FH2) domain is responsible for actin binding and acts as an important nucleating factor in eukaryotic cells. In this work EPR and DSC were used to investigate the properties of the mDia1-FH2 formin fragment and its interaction with actin. MDia1-FH2 was labeled with a maleimide spin probe (MSL). EPR results suggested that the MSL was attached to a single SH group in the FH2. In DSC and temperature-dependent EPR experiments we observed that mDia1-FH2 has a flexible structure and observed a major temperature-induced conformational change at 41 °C. The results also confirmed the previous observation obtained by fluorescence methods that formin binding can destabilize the structure of actin filaments. In the EPR experiments the intermolecular connection between the monomers of formin dimers proved to be flexible. Considering the complex molecular mechanisms underlying the cellular roles of formins this internal flexibility of the dimers is probably important for manifestation of their biological functions. Electronic supplementary material The online version of this article (doi:10.1007/s00249-013-0922-0) contains supplementary material, which is available to authorized users.


SUPPLEMENT Simulations and integrals
To understand the origin of the appearance of the two subpopulations better we analyzed their contributions to the spectra. Changes in the relative contributions of the two subpopulations can be described by calculating the ratio of I +1 / I +1m . Here I +1m is the peak height of the low-field maximum in the spectrum of the MSL-formin, and I +1 is the peak-to-peak height of the first component of the spectrum due to the labels reflecting rotational motion of shorter correlation time.
In order to check the validity of estimating the relative contributions of the two subpopulations by calculating the I +1 / I +1m ratios spectral simulation was performed at five representative temperatures: 0, 15, 30, 40 and 60 o C (Table S1). In each case two components were taken to fit the experimental spectra by non-linear least square method developed by the Freed's group with the implementation as given in (Kupi et al. 2009). In line with the experimental observations the I +1 /I +1m ratio increases with temperature due to the increasing relative contribution of the fast component. Because the smallest and the highest ratios were found at 0 o and 60 o C, respectively, we calculated at these two temperatures the rate of deviations between the values determined either using complete spectra or the separated components.
Taking the values of Table S1 we can state that the relative error of the determination is smaller than 10 %. Our conclusion was that evaluation of the I +1 /I +1m ratios from the experimental spectra does not result in erroneous suppositions about changes in the relative contributions of the two subpopulations. On the bases of the spectral simulation we also concluded that the rotational asymmetry (R z compared to R x and R y ) of the slow population increases with the temperature mainly due to alteration of the R x and R y components; while the variation of the R z parameter was not necessary to arrive to the best fit from 0 o C to 60 o C.
In analyzing the I +1 / I +1m values, smaller ones indicate smaller relative contribution of the weakly immobilized subpopulation (in terminology of the above sentences this is the subpopulation which possesses shorter correlation time). Figure 7 in the paper shows the temperature dependence of the ratio of I +1 / I +1m determined from the experimental spectra in absence or presence of actin. The function of I +1 / I +1m against reciprocal absolute temperature 2 can be approximated with a near exponential function in both cases. Breakpoint did not appear in the plots. Another important finding from these measurements was that in the presence of actin the decay constant of the exponential function was significantly smaller showing that the binding of formin to actin affected the ratio of the two populations. Thus, the probe microenvironment was sensitive to the interaction of actin with formin.
The contribution of the weakly immobilized probe was smaller at all temperatures in the presence of actin than in the absence of it. The difference increased with increasing temperature. This behavior emphasizes the stabilizing effect of actin on formin and shows that this effect was more pronounced at higher temperatures. The actin binding shifted the equilibrium between the strongly and weakly immobilized states of the formin towards the former one. The component denoted as strongly immobilized corresponds to the subpopulation having longer correlation time (~ 20 -~ 26 ns) in the context of Table S1, i.e. the 'slow' component. Lack of any breakpoint on the temperature dependence of I +1 / I +1m proves that the breakpoint temperatures are the same for the strongly and weakly immobilized populations. We found in previous experiments that the observed breakpoint temperatures were not influenced by the presence of actin ( Fig. 3 and 4 in the paper), which suggests that the two populations of the formin dimers bind to identical binding sites in actin filaments.
The spectra of MSL-formin and its complex with actin showed two distinct spectral components (Fig. 1). The relative fractions of the components were temperature dependent ( Fig. 5). It is possible to approximate the ratio of the double integrals of the two fractions at the m = +1 EPR transition as a function of temperature by computer manipulation. A im represents the double integral of the spectral component characterizing the amount of the strongly immobilized labels at the low-field maximum, whereas A m is the double integral of the first component of the spectrum characterizing the weakly immobilized labels. The double integral of the whole spectrum was temperature independent within the limits of experimental error. The ratio of A im / A m showed an exponential dependence as a function of reciprocal absolute temperature. The populations associated with the two fractions were used to calculate an equilibrium constant K as the ratio of A im / A m . The value of R lnK shows a linear function against T -1 (Fig. 7).