FASCIN and alpha-actinin can regulate the conformation of actin filaments

Biochimica et Biophysica Acta 1850 (2015) 1855–1861

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbagen

FASCIN and alpha-actinin can regulate the conformation of actin filaments Katalin Türmer a,b, József Orbán a,b,c, Pál Gróf d, Miklós Nyitrai a,b,e,⁎ a

Department of Biophysics, Medical School, University of Pécs, Szigeti u. 12, Pécs H-7624, Hungary János Szentágothai Research Center, Pécs H-7624, Hungary MTA-PTE High Intensity Terahertz Research Group, Hungary d Department of Biophysics and Radiation Biology, Semmelweis University of Medicine, IX. Tűzoltó u. 37-47, Budapest H-1095, Hungary e MTA-PTE Nuclear-Mitochondrial Interactions Research Group, Hungary b c

a r t i c l e

i n f o

Article history: Received 5 April 2015 Accepted 21 May 2015 Available online 27 May 2015 Keywords: Actin Bundle Fascin Alpha-actinin EPR spectroscopy

a b s t r a c t Background: Actin filament bundling proteins mediate numerous processes in cells such as the formation of cell membrane protrusions or cell adhesions and stress fiber based locomotion. Among them alpha-actinin and fascin are the most abundant ones. This work characterizes differences in molecular motions in actin filaments due to the binding of these two actin bundling proteins. Methods: We investigated how alpha-actinin and fascin binding modify the conformation of actin filaments by using conventional and saturation transfer EPR methods. Results: The result characteristic for motions on the microsecond time scale showed that both actin bundling proteins made the bending and torsional twisting of the actin filaments slower. When nanosecond time scale molecular motions were described the two proteins were found to induce opposite changes in the actin filaments. The binding of one molecule of alpha-actinin or fascin modified the conformation of numerous actin protomers. Conclusion: As fascin and alpha-actinin participates in different cellular processes their binding can serve the proper tuning of the structure of actin by establishing the right conformation for the interactions with other actin binding proteins. Our observations are in correlation with the model where actin filaments fulfill their biological functions under the regulation by actin-binding proteins. General significance: Supporting the general model for the cellular regulation of the actin cytoskeleton we showed that two abundant actin bundling proteins, fascin and alpha-actinin, alter the conformation of actin filaments through long range allosteric interactions in two different ways providing the structural framework for the adaptation to specific biological functions. © 2015 Elsevier B.V. All rights reserved.

1. Introduction There are several actin filament bundling proteins that mediate diverse processes in cells such as the formation of cell membrane protrusions or cell adhesions and stress fiber based locomotion. Among the various cross-linkers, alpha-actinin and fascin are the most abundant ones. Their interaction with actin is different (see scheme in Fig. 1) and considered to be adapted to the related specific biological functions [1]. The globular fascin (55–58 kDa, ~5 nm diameter) has an evolutionarily highly conserved amino acid sequence folded up into four betatrefoil domains. There are three isoforms of fascin in mammals; fascin1, 2 and 3 and they are expressed in different tissues [2–5]. While in many cases phosphorylation activates proteins and signaling pathways, ⁎ Corresponding author at: University of Pécs, Medical School, Department of Biophysics, Szigeti str. 12, Pécs H-7624, Hungary. E-mail address: [email protected] (M. Nyitrai).

http://dx.doi.org/10.1016/j.bbagen.2015.05.018 0304-4165/© 2015 Elsevier B.V. All rights reserved.

fascin is one of the actin binding proteins that is negatively regulated by phosphorylation [6]. Fascin has two major F-actin binding sites, one in the N terminal (β trefoil-1) and another in the C terminal (β trefoil-3) position [7,8] and these binding sites assist the formation of stable and relatively rigid actin bundles. These actin bundles serve as highways of vesicle transport from the cell body to the constantly developing leading edge of the cell protrusion. Fascin localizes to filopodia, microspikes and actin-based protrusions underneath the plasma membrane and is involved in the assembly and maintaining of the parallel bundled F-actin structure at the filopodia tips in motile cells [4]. Its expression is important for the stability of actin bundles in invadopodia [9] and is upregulated in many human carcinomas; its intracellular concentration correlates with the clinical aggressiveness of tumors. Depletion or inhibition of fascin results in reduced migration of filopodia driven motile cells through diminished protrusion of filopodia [7]. Recently it has been shown that fascin can modulate the conformation and dynamic properties of actin filaments. Mansson and colleagues found that the flexibility of fascin bound filaments decreases on fascin bundling

1856

K. Türmer et al. / Biochimica et Biophysica Acta 1850 (2015) 1855–1861

alpha-actinin and actin

fascin and actin

bind to the bundles the same way as to individual actin filaments and the in vitro motility assays have shown no difference in actin filament and bundle sliding speed on myosin covered plates at relatively high, 1:2 fascin:actin concentration ratios [10]. On the other hand, during the sliding motion, disassembly of the bundles took place that may regulate the actin turnover in filopodia. Nagy and colleagues [19] found preference of myosin-X binding to fascin bundled actin when compared to single actin filaments. Because of this specified binding at the filopodial core part of bundles myosin-X can traffic along and accumulate at the distal tips of filopodia. Ricca and Rock [20] demonstrated processive stepping of myosin X along these fascin-actin bundles. All these observation suggested that the relationship between the shape and size of the actin bundles facilitated by fascin or alphaactinin is not the only feature that defines biological function and activity. It seems plausible to assume that the dynamic properties of actin filaments in these bundles play a central role, and the relative motions of the protein components must also be important for proper spatial configurations. Here we addressed the question of how the binding fascin or alpha-actinin changes the conformation of actin filaments by applying electron paramagnetic resonance (EPR) spectroscopy methods. We found that slower bending and torsional motions were restricted after the binding of these bundling proteins and the core structural elements of the actin filaments became more rigid. On a faster, nanosecond time scale molecular motions the two proteins were found to induce opposite changes in the actin filaments. Concentration dependent investigations showed that the binding of one molecule of alpha-actinin or fascin modified the conformation of 200 or 10 actin protomers, respectively. The observations are interpreted within the framework of the model in which actin-binding proteins are responsible for the regulation and adaptation of the conformational state of the actin filaments. 2. Materials and methods

Fig. 1. Scheme for the arrangements of proteins established in the actin bundles created by alpha-actinin homodimers (A) or fascin (B). Actin protomers and monomers are indicated with orange circles and the corresponding bundling proteins are shown in blue.

resulting in ~ 15 times longer persistence length compared to unique actin filaments [10]. Alpha-actinin is an antiparallel homodimer (20–30 nm head to head length) with a single N terminal actin-binding domain followed by four spectrin-like triple helical motifs arranged in tandem structure [11]. The spectrin motifs are responsible for non-covalent dimer formation and the specific features of the dimerization domains define the head-tohead length. The orientation of the actin-binding heads is important in actin binding and consequent bundling [12,13]. Alpha-actinin can be found at many different sites in cells, such as Z-disks of skeletal muscle sarcomeres, and in non-muscle cell adhesion plaques or at the leading edge of motile cells [2,14–16]. The different isoforms of alpha-actinin may bundle actin filaments either to polar or bipolar actin bundles and consequently the polarity of the formed actin network depends on the proteins. Despite of their similar (crosslinking and bundling) basic functions alpha-actinin and fascin bound actin filaments display different mechanical properties [1,17]. There are differences in the dependence of elasticity and rate of bundle assembly on linker protein concentration. The most evident difference is that according to the size of these molecules and the distance between two actin-binding domains the formed actin network is different (Fig. 1.). Fascin forms tight bundles in contrast with alpha-actinin. However, the correlation between the characteristic geometrical properties and their functional behavior is not necessarily simple. Both fascin associated tight actin bundles in microspikes and stress fibers formed by alpha-actinin bundling are permissive for motility of myosin motors [18]. In fascin mediated bundles myosin heads can

2.1. Chemicals Tris(HCl), MOPS, KCl, MgCl2, CaCl2, and EGTA were purchased from Sigma-Aldrich (Budapest, Hungary). ATP, ADP and dithiothreitol (DTT) were obtained from Merck (Darmstadt, Germany). NaN3 was purchased from Fluka (St. Gallen, Switzerland). Maleimide spin label (MSL) for EPR measurements was purchased from Sigma-Aldrich (Budapest, Hungary). All chemicals were of analytical grade, solved in double distilled water on request. 2.2. Protein purification and modification Actin was purified from the domestic white rabbit skeletal back and leg muscles (Feuer, Spudich) and was stored in G-actin buffer (0.2 mM ATP, 0.1 mM CaCl2, 4 mM TRIS or MOPS, pH 8.0, 0.1 mM DTT and 0.005% NaN3) at 4 °C. The actin monomer (G-actin) concentration was determined from absorbance measured at 290 nm using a molar extinction coefficient of 0.63 mg−1 ml cm−1 using a Shimadzu UV2100 spectrophotometer. Actin filaments (F-actin) were polymerized by the addition of 2 mM MgCl2 and 100 mM KCl (final concentrations) to G-actin buffer for 2 h at room temperature and then kept on ice until use. Fascin and α-actinin were extracted as recombinant proteins from cells of Escherichia coli expression system. Cells were thawed at 4 °C in PBS (proportional to the amount of wet cell mass), protease inhibitor mix, and PMSF, and lysozyme was added to the cells. Then the cells were disrupted by handheld homogenizer and sonicator (5 × 1 minute pulses of 80% amplitude). To the cell suspension DNase I enzyme was added and then stirred for 1 h at 0 °C. After the incubation the excess cell debris was separated from the sample by ultracentrifugation (100,000 g, 4 °C, 30 min). The proteins (GST-fusion tagged recombinant alpha-actinin and fascin) were purified with affinity chromatography (Pharmacia FPLC), in which the GST-protein was bound to the beads of the GSH

K. Türmer et al. / Biochimica et Biophysica Acta 1850 (2015) 1855–1861

(glutathione–sepharose 4B) column. To remove nonspecifically bound proteins we used 20× column volume of washing buffer (50 mM Tris/ HCl (pH 7.5), 300 mM NaCl, 3 mM DTT, 1 mM EDTA). After cleavage with Factor Xa (0 °C, 16 h) the protein was eluted in buffer (10 mM Tris/HCl (pH 7.5), 150 mM NaCl, 3 mM DTT, 1 mM EDTA). The sample was concentrated in Amicon Ultra-15 MWCO 10 kDa centrifuge tubes using Janetzki K26 centrifuge (4000 g, 4 °C). The remaining contaminating proteins were removed by gel filtration (Superdex G-75 column). Samples were taken from the fractions corresponding to each absorbance peak and then purity was checked on SDS-acrylamide gel. The fractions containing the protein of our interest were collected and concentrated as described earlier. For storage our protein was dialyzed against F-actin buffer, and frozen in liquid nitrogen then stored at −80 °C. The bundling activity of the proteins was tested with fluorescently labeled (5–10%) actin using fluorescent confocal microscopy. The obtained images confirmed the formation of actin filaments and actin filament bundles in the presence of fascin or alpha-actinin (data are not shown), indicating that the recombinant proteins were biologically active. The previously polymerized actin was mixed with the bundling protein at around 100:1 actin:bundling protein concentration ratio then the appearance of bundled filaments was verified.

1857

2.5. Evaluation of the EPR spectra In order to characterize the changes in conventional EPR spectra we determined the distances between the two outer extremes (2Amax) that is characteristic of the rotational motion of the spin labeled molecules. In the case of molecules, which are nearly spherical, one can also use the formula introduced by the Freed's group for spherical tops, if the rotational correlation times are in the range not longer than about 10−7 s [21,22]. Although this condition is fulfilled by the G-actin, it is not correct for F-actin molecules, especially if they are bundled by the fascin or α-actinin. We have tried to make EPR spectral simulations too, however, the results pointed out that neither these spectral simulations allow a correct determination of the rotational correlation time. The rotational motion of the bundled F-actin molecules is beyond the time range where the slow motional simulation program, used earlier also by us for other systems [23–25], would allow a correct determination of the rotational correlation times. Evaluating the ST-EPR spectra the usual L″/L and the C″/C parameters [24,26–29], as shown in Fig. 3, were determined from the spectra. Correlation times in the ST-EPR time domain have been calculated based on the digitized values from the tables given for MSL-labeled hemoglobin according to Hyde and Dalton [29]. 3. Results and discussion

2.3. Spin labeling of actin Actin was spin-labeled with N-(1-oxyl-2,2,6,6,-tetramethyl-4piperidinyl)-maleimide spin label (MSL) in filamentous form in excess of MSL (a molar ratio of 1:1.2) for 12 h at 4 °C. Unreacted spin labels were removed by pelleting the actin by ultracentrifugation, and then pellet was re-suspended, homogenized and dialyzed in G-actin buffer. Labeling ratio has been determined by comparing the double integral of the EPR spectra measured at given actin concentrations and a known concentration of MSL prepared in G-actin buffer. EPR spectra for both cases were measured under identic spectrometer settings, unless receiver gains were taken into account during the calculations. The range of the labeling ratio (labeled protein to whole protein concentration), thus determined for different samples has been found to be between 0.8 and 0.9.

2.4. EPR spectroscopy Conventional EPR spectra of actin were recorded with an ESP 300E X-band spectrometer (Bruker Biospin, Rheinstetten, Germany). First harmonic in-phase, absorption spectra were obtained using 20 mW microwave power and 100 kHz magnetic field modulation with amplitude of 0.20 mT; receiver gain, conversion time, time constant and scan numbers of 8.93 · 104, 81.9 ms, 10.24 ms and 25 scans were used, respectively. Actin concentration varied between 30 and 120 μM. In ST-EPR measurement the second harmonic out-of-phase spectra were recorded (50 kHz magnetic field modulation, 0.5 mT modulation amplitude and 63 mW microwave power, 32 or 64 scans with receiver gain, conversion time, time constant and scan numbers of 8.93°104, 81.9 ms, 10.24 ms). For ST-EPR measurements only samples with the highest molar ratios of bundling proteins to actin (1: 20) were used. The temperature was set to 20 °C with a temperature controller. The protein samples were placed in two capillary tubes (Mettler ME-18552 melting point tubes), each of them contained 10 μl solution. The sample tubes were positioned parallel in the center region of the TM 110 cylindrical cavity. A small thermocouple was inserted in one of the capillary tubes, and the temperature was regulated with a diTC2007 type temperature controller. The EPR spectra were evaluated with the WINEPR program from Bruker and with a computer program developed in our laboratory.

The interaction of actin with alpha-actinin and fascin was described by using EPR methods using spin labeled actin. The conformational flexibility is the inherent property of both actin monomers and filaments [30–32]. While the motions of atoms and residues during their rearrangement occur within femto- or picoseconds, due to its size the rotation of the free actin monomer is characterized by a few tens of nanosecond rotational correlation time. Intramolecular domain motions within the monomers are faster falling on the few nanoseconds scale. The restricted motion of the protomers after polymerization is more complex but usually still fast enough to be on the few tens of nanosecond time scale. Larger segmental units of filaments composed by several actin protomers can be characterized by hundreds of nanosecond. The filaments as a whole can perform twisting or bending motions depending on its actual conformation and environment. These motions are much slower than those described above and their characteristic time is usually on the micro- or even millisecond time scale. Due to the applied physical principles, various spectroscopic methods are sensitive to molecular motions on different time scales [33–36]. Femto- and picosecond motions can be described by using structural (X-ray diffraction or electron microscopy) methods or ultrafast kinetic approaches. Fluorescence emission polarization and conventional EPR techniques can report regarding motions in a nanosecond or few tens of nanosecond time range, while more specific applications such as transient phosphorescence experiments or saturation transfer EPR method provides information about the molecular motions that occur in milli- or microseconds. Here we used both conventional and saturation transfer EPR techniques. The conventional EPR spectra allowed us to draw conclusions regarding the molecular transitions influencing fast, nanosecond scale motions in actin filaments, while the saturation transfer EPR experiments provided information regarding the effects of fascin or alpha-actinin induced changes in the actin filaments on a much longer, microsecond time scale. By the parallel application of the two techniques we could distinguish between the effect of bundling proteins on the conformation of the actin protomers and the influence of the formation of interfilamental connections on the slower motion of the entire actin filaments. 3.1. Conventional EPR measurements Conventional EPR spectra were recorded in the absence or presence of actin bundling proteins. Fig. 2 shows the results obtained for actin

1858

K. Türmer et al. / Biochimica et Biophysica Acta 1850 (2015) 1855–1861

hyperfine splitting constant in the case of actin–fascin complexes indicates the opposite, the more flexible protein matrix around the spin probe in actin or a less polar environment of the monitoring spin label. To characterize further the effects of bundling proteins on actin filaments the conventional EPR experiments were repeated at various fascin and alpha-actinin concentrations. The hyperfine splitting constants obtained from these experiments are presented in Fig. 3. To emphasize differences in the 2Amax, the value determined in the absence of bundling proteins was subtracted from the hyperfine splitting constants obtained in the presence of bundling proteins. When alpha-actinin was added to actin filaments the increase of the hyperfine splitting constant saturated at relatively low alpha-actinin concentration, at around the 200:1 actin:alpha-actinin ratio. In the experiments with fascin the concentration dependence of the hyperfine splitting constant was less steep and did not reach saturation at the highest investigated fascin concentration (5 μM), where the actin:fascin ratio was 20:1. 3.2. Long-range interactions were induced in actin filaments

monomers, actin filaments, and filaments with fascin or alpha-actinin in a 20:1 actin:actin bundling protein ratio. In the absence of bundling proteins the hyperfine splitting constant (2Amax ± SEM) was determined to be 6.376 ± 0.0058 mT (n = 12) for the actin monomers and 6.796 ± 0.0265 mT (n = 15) for the filaments. Both of these values are in good agreement with previous results with monomeric and polymerized actin [24,37,38] and indicated that the spin labeled actin monomers incorporated into the filaments during polymerization. When alpha-actinin was added to the actin filaments in a 20:1 actin:alphaactinin concentration ratio the hyperfine splitting constant increased by ~0.05 mT (0.5 G). In the case of fascin (20:1) the value of this parameter decreased by ~ 0.06 mT (0.6 G). The opposite effects from alphaactinin and fascin indicated that these bundling proteins change the conformation of the actin in two opposite ways. The hyperfine splitting constant is sensitive to the change of the protein flexibility, polarity of the spin label environment and also to the change of the effective radius of the protein domains that can be taken as a unit rotating together. According to earlier observations [34, 39–41] the MSL spin label is rigidly attached to the Cys-374 site of actin. Thus, in our cases changes in the 2Amax values can be attributed to differences in the changes of effective radius or flexibilities due to the binding of the bundling proteins. Although a change in the effective rotating radius could be a consequence of an increase in effective molecular mass, we believe that the effect of this increase in the mass cannot explain the observed changes in the hyperfine splitting. The bundling proteins were present in the samples at a relatively low concentration (20:1) and thus the increase of the mass of the individual actin filaments after their binding was too little to explain the observations. On the other hand, the hyperfine splitting constant decreased in the case of fascin binding that is the opposite of what would be expected from the increase of the total mass. Our interpretation of the results is that the actin bundling proteins changed the conformation of actin protomers in the filaments. This altered conformation could result in either flexibility or polarity changes in the filaments. Note, that we are not aware of any methodology at present, which could clearly distinguish between changes in the polarity and rotational motions for the system studied in the present work. The larger hyperfine splitting constant for the actinalpha-actinin complex indicates that the intramolecular motions in actin became slower reflecting the more rigid structure of actin or increase in the polarity of the closer vicinity of the spin labeling molecule. The smaller

Previous studies showed that the affinity of fascin and alpha-actinin for actin filaments is in the 0.5–5 μM range. The affinity of mouse fascin2 – also used in the present work – for alpha-skeletal actin was 4.95 μM [42]. For chicken smooth muscle alpha-actinin the value of 0.6 μM, for Acantamoeba alpha-actinin 4.7 μM was found for this parameter [43]. Although we used human alpha-actinin, it can be assumed that its affinity for actin was similar to these values. This assumption is corroborated by the alpha-actinin concentration dependence of the EPR data, where saturation was reached at low alpha-actinin concentrations. In the EPR experiments we applied actin in large excess to either fascin or alphaactinin, and the actin concentration exceeded at least ten times the corresponding affinities. Accordingly, the ratio of the bound actin-binding protein concentration to the free actin binding protein concentration should have been 10 or more, indicating that most of the bundling proteins (at least 90%) were in complex with actin under these conditions. The effect of alpha-actinin on the conventional EPR spectra saturated at 200:1 actin:alpha-actinin concentration ratio (Fig. 3.), i.e., further addition of alpha-actinin did not cause any change in the spectra. Considering that most of the added bundling proteins bound to actin, the binding affinity played little role in the definition of the saturation concentration. Sub-stoichiometric saturation concentration showed that one alpha-actinin could modify the conformation of more than one actin protomer. If the number of protomers that were altered by one fascin molecule is defined as a cooperative unit, for alpha-actinin the

0.6

(2Amax) (Gauss)

Fig. 2. The EPR spectra of MSL-labeled actin and its complexes. G-actin and F-actin spectra were taken at actin concentration of 50 μM, while in the case of experiments with alphaactinin and fascin the final actin concentration was 40 μM. Protein molar ratios are indicated.

alpha-actinin

0.4 0.2 0.0 -0.2 -0.4

fascin

-0.6 0.00

0.01

0.02

0.03

0.04

0.05

[bundling protein] : [actin] ratio Fig. 3. The bundling protein concentration dependence of the hyperfine splitting constant. Differences of the 2Amax were calculated as the measured 2Amax minus the initial 2Amax,o values for the given F-actin samples. The x-axis shows the ratio of the bundling protein (fascin (empty circles) or alpha-actinin (filled squares)) concentration to the concentration of the actin monomers. In the case of alpha-actinin samples the actin concentration was set to 50 μM, while for fascin-samples it was 40 μM.

K. Türmer et al. / Biochimica et Biophysica Acta 1850 (2015) 1855–1861

1859

cooperative unit was 200. For fascin the concentration dependence of the hyperfine splitting did not reach saturation (Fig. 3). By hyperbola fitting we estimated that the saturation for the effect of fascin on the EPR spectra would occur at appr. 10:1 actin:fascin concentration ratio. In this case our conclusion is that the binding of one fascin molecule influenced the conformation of 10 actin protomers. Although the exact molecular mechanism by which these conformational changes propagated along the actin filaments cannot be deduced from our data, the results clearly showed that much less than stoichiometric binding of either fascin or alpha-actinin was enough to modify the entire actin filaments. This property of the bundling proteins could facilitate their regulatory roles in the definition of actin conformation by providing an economic way of regulation for the cells.

3.3. Saturation transfer EPR measurements The conventional EPR experiments showed that both alpha-actinin and fascin altered the faster molecular motions on the nanosecond time scale, and their effect was opposite. On the bases of our earlier work with formin–actin complexes we were able to detect that local conformational changes can be decoupled from global changes of the actin filament [24]. To gain further information and to check if there would be similar effects regarding the bundling protein induced changes in actin filaments we have carried out saturation transfer EPR experiments. Compared to the conventional EPR method, ST-EPR can reflect molecular motions on much longer (microsecond) time-scales, and resolved rotational times characterizing the torsional twisting or bending motions in larger entities of the actin filaments [44–46]. Evaluation of the ST-EPR spectra is based on rotational motions of spherical rigid rotators, in which case theoretical calculation supports determination of nearly exact correlation times [26,29,47]. For asymmetric rotators – such as the actin filaments – characterization of the molecular motions by correlation times can only be taken as approximations. Experimental observation and theoretical consideration on highly anisotropic rotational diffusion of gel-phase lipid systems have shown that the diagnostic parameters L″/L and C″/C reflect different sensitivity against changes in rotational diffusion on microsecond time-scale [48,49]. It has been found that the diagnostic parameter C″/C is a much more sensitive indicator for changes of the rotational motion in the above systems. Moreover, in gel-phase the torsional oscillations in the chain contribute also the observed ST-EPR signals. Therefore, in agreement with our and others' earlier works we used the diagnostic spectral parameters L″/L and C″/C to draw conclusions regarding the effects of binding of bundling proteins on the torsional motions of the actin filaments (Fig. 4.). Higher values of these parameters reflect slower correlation times. The parameters obtained from these experiments are summarized in Table 1. In the case of actin filaments in the absence of bundling proteins we observed similar values for L″/L and C″/C (1.214 ± 0.013 and 0.120 ± 0.024, respectively) as in previous studies [24,34]. When fascin bound to the actin filaments the values for L″/L and C″/C were 1.582 ± 0.015 and 0.606 ± 0.017, respectively, reflected decreased actin filament flexibility. In the case of alpha-actinin the C″/C parameter showed similar decrease (C″/C = 0.419 ± 0.013) in rotational motion. Despite of many experiments, for alpha-actinin bound actin filaments the L″/L changes could not reliably show any alterations in the molecular motions. The reason for this was the too noisy spectra at the required field strengths. Comparing the effective correlation times that could be calculated either from the L″/L or C″/C parameters it is clear that they show great differences, as it is expected, for a highly asymmetric molecular complex of the actin filament. We observed similar patterns in correlation times studying formin–actin interactions [24,37]. Correlation times calculated from the C″/C show the asymmetric torsional motions of the filaments with an increase for fascin–actin complex, and also alpha-actinin may

Fig. 4. ST-EPR spectra measured on spin labeled F-actin samples in the presence or absence of bundling proteins. The ratios L″/L and C″/C were determined corresponding to the indicated positions of the spectra and summarized in Table 1.

have the same characteristics revealed by the correlation time obtained from C″/C diagnostic parameter (Table 1.). The values obtained for L″/L or for C″/C increased after the binding of fascin or alpha-actinin, indicating that on the microsecond time scale the molecular motions were slower in the actin filaments after their binding. The slower filament motion may appear because the actin filaments became stiffer or can be the consequence of the formation of the interfilamental connections in the bundles. This observation is in correlation with the much longer persistence length measured for actin filament bundles after the binding of fascin [10]. In order to correlate the effects observed with conventional and saturation transfer EPR methods we suggest two coupled and parallel molecular events that follow the binding of bundling proteins to actin. One of them – detected by conventional EPR technique – modifies the conformation of a set of actin protomers in the filaments. The other one – detected on ST-EPR spectra – takes place at level of entire filaments or even at the supramolecular level. This second effect results Table 1 Parameters obtained from the ST-EPR spectra in the case of actin filaments in the absence or presence of alpha-actinin or fascin. ST parameter

Actin filament

Alpha-actinin–actin (1:20)

Fascin–actin (1:20)

L″/L τL″/L [μs]= C″/C τC″/C [μs]= 2Amax (mT)

1.214 ± 0.013 255 ± 15 0.120 ± 0.024 8±1 6.796 ± 0.0265

0.846a ± 0.045 77a ±15 0.419 ± 0.013 24 ± 2 +0.05

1.582 ± 0.015 920 ± 80 0.606 ± 0.017 59 ± 5 −0.06

a ST-EPR spectra at this spectral range were too noisy to allow further quantitative evaluation.

1860

K. Türmer et al. / Biochimica et Biophysica Acta 1850 (2015) 1855–1861

Fig. 5. A scheme to summarize the observations.

in a change of the torsional twisting or bending of actin filaments [33]. In addition, the formation of the bundles of actin filaments may re-act by interfilamental interaction provoking changes in a detectably higher amount of protomers. By measuring persistence lengths for fascin–actin complexes the flexural rigidity of actin filament bundles was greater than that of single actin filaments [10]. This observation has been attributed to increased number of parallel filaments and increased coupling of these filaments at higher fascin–actin ratios. Torsional rigidity has been determined by direct visualization of torsional Brownian motion [50]. In this latter work flexural rigidity and torsional rigidity have been found to be independent from each other, at least in the case when these parameters have been used for the interpretation of cation dependent changes in actin. With the saturation transfer EPR method we observed slower molecular motions on the microsecond time scale after the binding of either fascin or alpha-actinin. The decrease of correlation times, expressed by the diagnostic parameters of the observed anisotropic rotational diffusion, can be in line with increase in either flexural or torsional rigidities. The model we proposed above assumes the existence of long cooperative units within which similar conformational transitions occur in several actin protomers simultaneously. It is to be decided in future experiments whether these conformational changes propagate along a single actin filament or transferred trough the interacting connections between the filaments. 4. Conclusions The EPR experiments reported here provided information regarding the fascin and alpha-actinin induced molecular changes in actin filaments on two distinctly different time scales, on the nanosecond and on the microsecond time scale. The faster motions are attributed in actin to the motions of actin subdomains, protomers or short segments of actin filaments. The longer time-scale motions are more characteristic for the correlated motion of the entire actin filaments, including bending and torsional twisting modes. Saturation transfer EPR experiments showed that the larger scale bending and torsional flexibilities along the actin filaments followed decreasing tendency after the binding of either fascin or alpha-actinin and we conclude that the twisting and

bending motion of the entire filaments became slower in these cases. The core structure of the actin filaments was stabilized providing a more rigid structural framework for biological functions. Considering that in this respect similar tendency was observed with the two crosslinking proteins the stabilization may simply be the effect of the formation of bundles, in which the motion of the actin filaments is restricted by their actin filament neighbors [51]. Actin filaments can probably fulfill their numerous biological functions through the regulation of their structure by actin-binding proteins [52–55], i.e., the adaptation of the filament structure to the actual function requires the modification and adjustment by the binding of actinbinding proteins. This general tendency is expected to work for crosslinking proteins too. Our conventional EPR results showed that despite of the similar effect of fascin and alpha-actinin on the microsecond time scale, their influence on the faster motional modes was opposite. The hyperfine splitting constant increased after the binding of fascin and decreased with alpha-actinin. The effect of both proteins propagated on a relatively long actin filament sessions; alpha-actinin and fascin modified the conformation of 200 and 10 protomers, respectively. Although from the data set we could not determine whether the flexibility of the protein matrix or the polarity change was responsible for these observations, it is clear that the two cross-linking proteins induced different molecular transitions and alterations in the structure of actin filaments. The EPR results obtained here suggested that the smaller structural elements in the periphery of the actin filaments were changed in opposite ways by fascin and alpha-actinin, while the motion of the entire actin filaments was slower after the binding of either of the two actin cross-linking protein. The simplest explanation for the latter observation is that the core structure of the actin filaments was not altered by the cross-linking proteins and the establishment of the interfilamental connection provided structural constrains and thus slowed down the filamental motions. These structural changes are depicted in a simple scheme in Fig. 5. Considering that fascin and alpha-actinin participates in various different cellular processes [4,9,15] these observations suggest that their binding can serve the proper tuning and regulation of the actin structure and thus to establish the right conformation for the interactions of the other actin binding proteins involved in the actual biological function.

K. Türmer et al. / Biochimica et Biophysica Acta 1850 (2015) 1855–1861

Acknowledgement This work was supported by grants from the Hungarian Scientific Research Fund (OTKA grants NN 107776 and K112794 to MN). We thank Kinga Futo for providing alpha-actinin and fascin proteins and Alf Mansson for his critical comments on the manuscript. References [1] Y. Tseng, E. Fedorov, J.M. McCaffery, S.C. Almo, D. Wirtz, Micromechanics and ultrastructure of actin filament networks crosslinked by human fascin: a comparison with alpha-actinin, J. Mol. Biol. 310 (2001) 351–366. [2] N. Kureishy, V. Sapountzi, S. Prag, N. Anilkumar, J.C. Adams, Fascins, and their roles in cell structure and function, BioEssays 24 (2002) 350–361. [3] Y. Saishin, S. Shimada, H. Morimura, K. Sato, I. Ishimoto, Y. Tano, M. Tohyama, Isolation of a cDNA encoding a photoreceptor cell-specific actin-bundling protein: retinal fascin, FEBS Lett. 414 (1997) 381–386. [4] R.A. Edwards, J. Bryan, Fascins, a family of actin bundling proteins, Cell Motil. Cytoskeleton 32 (1995) 1–9. [5] B. Tubb, D.J. Mulholland, W. Vogl, Z.-J. Lan, C. Niederberger, A. Cooney, J. Bryan, Testis fascin (FSCN3): a novel paralog of the actin-bundling protein fascin expressed specifically in the elongate spermatid head, Exp. Cell Res. 275 (2002) 92–109. [6] A. Li, J.C. Dawson, M. Forero-Vargas, H.J. Spence, X. Yu, I. Konig, K. Anderson, L.M. Machesky, The actin-bundling protein fascin stabilizes actin in invadopodia and potentiates protrusive invasion, Curr. Biol. 20 (2010) 339–345. [7] S. Jansen, A. Collins, C. Yang, G. Rebowski, T. Svitkina, R. Dominguez, Mechanism of actin filament bundling by fascin, J. Biol. Chem. 286 (2011) 30087–30096. [8] Y. Hashimoto, M. Skacel, J.C. Adams, Roles of fascin in human carcinoma motility and signaling: prospects for a novel biomarker? Int. J. Biochem. Cell Biol. 37 (2005) 1787–1804. [9] L.M. Machesky, A. Li, Fascin: invasive filopodia promoting metastasis, Commun. Integr. Biol. 3 (2010) 263–270. [10] H. Takatsuki, E. Bengtsson, A. Mansson, Persistence length of fascin-cross-linked actin filament bundles in solution and the in vitro motility assay, Biochim. Biophys. Acta 1840 (2014) 1933–1942. [11] A. Blanchard, V. Ohanian, D. Critchley, The structure and function of alpha-actinin, J. Muscle Res. Cell Motil. 10 (1989) 280–289. [12] M.G. Klein, W. Shi, U. Ramagopal, Y. Tseng, D. Wirtz, D.R. Kovar, C.J. Staiger, S.C. Almo, Structure of the actin crosslinking core of fimbrin, Structure 12 (2004) 999–1013 (London, England: 1993). [13] V. Ortiz, S.O. Nielsen, M.L. Klein, D.E. Discher, Unfolding a linker between helical repeats, J. Mol. Biol. 349 (2005) 638–647. [14] A. Ribeiro Ede Jr., N. Pinotsis, A. Ghisleni, A. Salmazo, P.V. Konarev, J. Kostan, B. Sjoblom, C. Schreiner, A.A. Polyansky, E.A. Gkougkoulia, M.R. Holt, F.L. Aachmann, B. Zagrovic, E. Bordignon, K.F. Pirker, D.I. Svergun, M. Gautel, K. Djinovic-Carugo, The structure and regulation of human muscle alpha-actinin, Cell 159 (2014) 1447–1460. [15] B. Sjoblom, A. Salmazo, K. Djinovic-Carugo, Alpha-actinin structure and regulation, Cell. Mol. Life Sci. 65 (2008) 2688–2701. [16] G. Franzot, B. Sjoblom, M. Gautel, K. Djinovic Carugo, The crystal structure of the actin binding domain from alpha-actinin in its closed conformation: structural insight into phospholipid regulation of alpha-actinin, J. Mol. Biol. 348 (2005) 151–165. [17] M.M. Claessens, M. Bathe, E. Frey, A.R. Bausch, Actin-binding proteins sensitively mediate F-actin bundle stiffness, Nat. Mater. 5 (2006) 748–753. [18] R. Ishikawa, T. Sakamoto, T. Ando, S. Higashi-Fujime, K. Kohama, Polarized actin bundles formed by human fascin-1: their sliding and disassembly on myosin II and myosin V in vitro, J. Neurochem. 87 (2003) 676–685. [19] S. Nagy, B.L. Ricca, M.F. Norstrom, D.S. Courson, C.M. Brawley, P.A. Smithback, R.S. Rock, A myosin motor that selects bundled actin for motility, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 9616–9620. [20] B.L. Ricca, R.S. Rock, The stepping pattern of myosin X is adapted for processive motility on bundled actin, Biophys. J. 99 (2010) 1818–1826. [21] D.E. Budil, S. Lee, S. Saxena, J.H. Freed, Nonlinear-least-squares analysis of slowmotion EPR spectra in one and two dimensions using a modified Levenberg– Marquardt algorithm, J. Magn. Reson. Ser. A 120 (1996) 155–189. [22] S.A. Goldman, G.V. Bruno, J.H. Freed, Estimating slow-motional rotational correlation times for nitroxides by electron spin resonance, J. Phys. Chem. 76 (1972) 1858–1860. [23] N. Kaszas, T. Bozo, M. Budai, P. Grof, Ciprofloxacin encapsulation into giant unilamellar vesicles: membrane binding and release, J. Pharm. Sci. 102 (2013) 694–705. [24] T. Kupi, P. Grof, M. Nyitrai, J. Belagyi, The uncoupling of the effects of formins on the local and global dynamics of actin filaments, Biophys. J. 96 (2009) 2901–2911. [25] Z. Szabo, M. Budai, K. Blasko, P. Grof, Molecular dynamics of the cyclic lipodepsipeptides' action on model membranes: effects of syringopeptin22A,

[26] [27]

[28] [29]

[30] [31] [32] [33] [34]

[35]

[36]

[37] [38] [39]

[40] [41] [42]

[43]

[44]

[45]

[46]

[47] [48]

[49] [50] [51] [52] [53] [54] [55]

1861

syringomycin E, and syringotoxin studied by EPR technique, Biochim. Biophys. Acta 1660 (2004) 118–130. D.D. Thomas, L.R. Dalton, J.S. Hyde, Rotational diffusion studied by passage saturation transfer electron-paramagnetic resonance, J. Chem. Phys. 65 (1976) 3006–3024. D.D. Thomas, J.C. Seidel, J. Gergely, J.S. Hyde, The quantitative measurement of rotational motion of the subfragment-1 region of myosin by saturation transfer EPR spectroscopy, J. Supramol. Struct. 3 (1975) 376–390. J.S. Hyde, L.R. Dalton, Saturation-transfer spectroscopy, in: L.J. Berliner (Ed.)Spin Labeling II Theory and Applications, vol. II, Academic Press, New York 1979, pp. 1–70. J.S. Hyde, L.R. Dalton, Saturation-transfer spectroscopy, in: L.J. Berliner (Ed.)Spin Labeling Theory and Applications, vol. II, Academic Press, N.Y., San Francisco, London 1979, pp. 1–70. K.C. Holmes, D. Popp, W. Gebhard, W. Kabsch, Atomic model of the actin filament, Nature 347 (1990) 44–49. W. Kabsch, H.G. Mannherz, D. Suck, E.F. Pai, K.C. Holmes, Atomic structure of the actin:DNase I complex, Nature 347 (1990) 37–44. T. Oda, M. Iwasa, T. Aihara, Y. Maeda, A. Narita, The nature of the globular- to fibrous-actin transition, Nature 457 (2009) 441–445. G. Hild, B. Bugyi, M. Nyitrai, Conformational dynamics of actin: effectors and implications for biological function, Cytoskeleton 67 (2010) 609–629. G. Hild, M. Nyitrai, J. Belagyi, B. Somogyi, The influence of divalent cations on the dynamic properties of actin filaments: a spectroscopic study, Biophys. J. 75 (1998) 3015–3022. E. Prochniewicz, T.F. Walseth, D.D. Thomas, Structural dynamics of actin during active interaction with myosin: different effects of weakly and strongly bound myosin heads, Biochemistry 43 (2004) 10642–10652. C.A. Rebello, R.D. Ludescher, Differential dynamic behavior of actin filaments containing tightly-bound Ca2+ or Mg2+ in the presence of myosin heads actively hydrolyzing ATP, Biochemistry 38 (1999) 13288–13295. T. Kupi, P. Grof, M. Nyitrai, J. Belagyi, Interaction of formin FH2 with skeletal muscle actin. EPR and DSC studies, Eur. Biophys. J. 42 (10) (2013) 757–765. J. Belagyi, P. Grof, Rotational motion of actin monomer at low and high salt concentration, Eur. J. Biochem. 130 (1983) 353–358. M. Mossakowska, J. Belagyi, H. Strzelecka-Golaszewska, An EPR study of the rotational dynamics of actins from striated and smooth muscle and their complexes with heavy meromyosin, Eur. J. Biochem. 175 (1988) 557–564. A. Orlova, E.H. Egelman, A conformational change in the actin subunit can change the flexibility of the actin filament, J. Mol. Biol. 232 (1993) 334–341. D.D. Thomas, J.C. Seidel, J. Gergely, Rotational dynamics of spin-labeled F-actin in the sub-millisecond time range, J. Mol. Biol. 132 (1979) 257–273. B.J. Perrin, D.M. Strandjord, P. Narayanan, D.M. Henderson, K.R. Johnson, J.M. Ervasti, Beta-actin and fascin-2 cooperate to maintain stereocilia length, J. Neurosci. 33 (2013) 8114–8121. D.H. Wachsstock, W.H. Schwartz, T.D. Pollard, Affinity of alpha-actinin for actin determines the structure and mechanical properties of actin filament gels, Biophys. J. 65 (1993) 205–214. E. Prochniewicz, Q. Zhang, E.C. Howard, D.D. Thomas, Microsecond rotational dynamics of actin: spectroscopic detection and theoretical simulation, J. Mol. Biol. 255 (1996) 446–457. E. Prochniewicz, Q. Zhang, P.A. Janmey, D.D. Thomas, Cooperativity in F-actin: binding of gelsolin at the barbed end affects structure and dynamics of the whole filament, J. Mol. Biol. 260 (1996) 756–766. H. Yoshimura, T. Nishio, K. Mihashi, K. Kinosita Jr., A. Ikegami, Torsional motion of eosin-labeled F-actin as detected in the time-resolved anisotropy decay of the probe in the sub-millisecond time range, J. Mol. Biol. 179 (1984) 453–467. J.S. Hyde, D.D. Thomas, New EPR methods for the study of very slow motion: application to spin-labeled hemoglobin, Ann. N. Y. Acad. Sci. 222 (1973) 680–692. P. Fajer, A. Watts, D. Marsh, Saturation transfer, continuous wave saturation, and saturation recovery electron spin resonance studies of chain-spin labeled phosphatidylcholines in the low temperature phases of dipalmitoyl phosphatidylcholine bilayers. Effects of rotational dynamics and spin–spin interactions, Biophys. J. 61 (1992) 879–891. D. Marsh, Molecular motion in phospholipid bilayers in the gel phase: long axis rotation, Biochemistry 19 (1980) 1632–1637. R. Yasuda, H. Miyata, K. Kinosita Jr., Direct measurement of the torsional rigidity of single actin filaments, J. Mol. Biol. 263 (1996) 227–236. M.M. Claessens, C. Semmrich, L. Ramos, A.R. Bausch, Helical twist controls the thickness of F-actin bundles, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 8819–8822. E.H. Egelman, A tale of two polymers: new insights into helical filaments, Nat. Rev. Mol. Cell Biol. 4 (2003) 621–630. E. Reisler, E.H. Egelman, Actin structure and function: what we still do not understand, J. Biol. Chem. 282 (2007) 36133–36137. D. Popp, R.C. Robinson, Many ways to build an actin filament, Mol. Microbiol. 80 (2011) 300–308. D. Popp, R.C. Robinson, Supramolecular cellular filament systems: how and why do they form? Cytoskeleton 69 (2012) 71–87.