A structural basis for the pH-dependence of cofilin F-actin interactions Laurence Blondin 1 , Vasilia Sapountzi 2 , Sutherland K. Maciver 1 , Emeline Lagarrigue 2 , Yves Benyamin 1 and Claude Roustan 1 1 Laboratoire de motilite ´ cellulaire, Universite ´ de Montpellier, France; 2 Genes and Development Group, Department of Biomedical Sciences, University of Edinburgh, Scotland A marked pH-dependent interaction with F-actin is an important property of typical members of the actin depolymerizing factor (ADF)/cofilin family of abundant actin-binding proteins. ADF/cofilins tend to bind to F-actin with a ratio of 1 : 1 at pH values around 6.5, and to G-actin at pH 8.0. We have investigated the mechan- ism for the pH-sensitivity. We found no evidence for pH-dependent changes in the structure of cofilin itself, nor for the interaction of cofilin with G-actin. None of the actin-derived, cofilin-binding peptides that we had previ- ously identified [Renoult, C., Ternent, D., Maciver, S.K., Fattoum, A., Astier, C., Benyamin, Y. & Roustan, C. (1999) J. Biol. Chem. 274, 28893–28899] bound cofilin in a pH-sensitive manner. However, we have detected a conformational change in region 75–105 in the actin subdomain 1 by the use of a peptide-directed antibody. A pH-dependent conformational change has also been detected spectroscopically in a similar peptide (84–103) on binding to cofilin. These results are consistent with a model in which pH-dependent motion of subdomain 1 relative to subdomain 2 (through region 75–105) of actin reveals a second cofilin binding site on actin (centered around region 112–125) that allows ADF/cofilin associ- ation with the actin filament. This motion requires salt in addition to low pH. Keywords: cofilin; actin; pH dependency; synthetic peptide; actin antibodies. The ADF/cofilins are a family of actin-binding proteins that are pivotally involved in both the polymerization and depolymerization of actin filaments, most notably in the advancing lamellae of motile cells [1,2]. Cell motility, through the actin-based cytoskeleton, is tightly controlled by the interplay of a variety of signaling pathways. The importance of the contribution of ADF/cofilins to cell motility is reflected in their being regulated by many of these signals, including phosphorylation [3], polyphosphoinosi- tides [4–6], the presence of other actin-binding proteins [7– 10] and pH [11–13]. Evidence for the regulation of the ADF/ cofilins by pH has been present both in vitro [11–14] and in living cells [15]. Most members of the ADF/cofilin family show a complex pH-dependent behaviour with respect to F-actin binding; exceptions are depactin from sea urchin eggs [16] and actophorin [17] from the soil amoeba Acanthamoeba. ADF/cofilins in general tend to bind to F-actin around pH 6.5 and to G-actin around pH 8.0 [6,11,18], but actophorin binds rabbit skeletal muscle F-actin at both pH extremes [17]. Actin solutions can be reversibly transformed from the G to F state by changes in pH in the presence of cofilin [6,11,19]. The F-actin bound by cofilin at low pH has several properties distinct from that of F-actin alone. These cofilin–actin filaments are short [19], have an increased helical twist [20] and do not bind phalloidin [8,12], caldesmon [8] or tropomyosin [7,10,21]. The study of the pH sensitivity of the actin–cofilin interaction is complicated by the fact that actin itself is pH-sensitive across the same range. The spontaneous polymerization of actin is more rapid at pH 6.5 than at pH 8.0 [22] and there appears to be a difference in conformation of G-actin at the two pH extremes [23]. Transients in intracellular pH occur in a variety of situations such as chemotaxis [24], mitosis, depolarization [25] and ischemia [26]. The actin–cofilins are typically concentrated at the leading edge of cells [5,27,28] and the cell cortex, regions that are especially likely to experience local fluctuations in pH [25]. The lammelae of alkalized macro- phages ÔhyperruffleÕ, whereas ruffling ceased on intracellular acidification [29], as expected from the properties that the ADF/cofilins display in vitro. The position and geometry with which ADF/cofilins bind F-actin has been controversial. Image reconstructions have placed cofilin on the surface of the filament, between subdomain 1 of one actin monomer and subdomain 2 of the longitudinally associated monomer, immediately toward the barbed end of the filament [20,30,31]. Our previous studies [32,33] argue that cofilin is not on the surface of the filament but is instead buried between two longitudinally associated monomers within the filament, and that subdo- main 2 from one monomer and subdomain 1 from the other are pushed apart. This results in the increased twist of the Correspondence to C. Roustan, UMR 5539[CNRS] UM2 CC107, Universite ´ de Montpellier 2, Place E. Bataillon CC107, 34095 Montpellier Cedex 5, France. Fax: + 33 0467144927, E-mail: roustanc@crit.univ-montp2.fr Abbreviations: ADF, actin depolymerizing factor; FITC, fluorescein 5-isothiocyanate; RITC, rhodamine isothiocyanate; 1,5-I-AEDANS, N,-iodoacetyl-N¢-[sulfo-1-naphthyl]-ethylenediamine; G-actin, monomeric actin; F-actin, filamentous actin. Note: web pages are available at http://www.dbs.univ-montp2.fr/ umr5539/, http://www.ephe.univ-montp2.fr, http://www.bms.ed. ac.uk/research/smaciver/index.htm (Received 9 May 2002, accepted 27 June 2002) Eur. J. Biochem. 269, 4194–4201 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03101.x actin filament observed first by McGough and coworkers [20] and subsequently by others [31], and the thrusting forward of subdomain 2 with respect to the rest of the monomer. In this report we study the pH-dependence of the actin– cofilin interface and provide evidence for a pH-dependent movement of subdomain 1 that may be involved in the pH- dependence of the interaction of cofilin with actin. EXPERIMENTAL PROCEDURES Proteins and peptides Rabbit skeletal muscle actin was isolated from acetone powder [34]. Human cofilin was produced in E. coli [BL21(DE3)], transfected with a T7-based vector, pMW172, carrying a human nonmuscle cofilin encoding cDNA fragment and purified as described previously [13,35]. Cofilin labeling with fluorescein isothiocyanate (FITC) was carried out by incubating the reagent (dissolved in N,N-dimethyl- formamide) with the protein in a molar ratio of 1 : 4. The coupling reaction was carried out in 50 m M NaHCO 3 buffer, pH 8.5, for 3 h, and excess reagent removed by gel filtration (PD-10, Amersham Pharmacia Biotech.) and equilibrated with the same buffer. The stoichiometry of the labeling was determined to be 0.7 mol FITC per mol cofilin. The procedure for Rhodamine isothiocyanate (RITC) labeling of cofilin or actin was similar, except that thereactionwasperformedusinga3molarexcessof reagent. Antibodies directed towards cofilin or 75–105 peptide coupled to hemocyanin were elicited in rabbits [36]. They were labeled with Oregon green (Molecular Probes) by thesameproceduredescribedforFITCexceptthata20 molar excess of reagent was used. IgGs labeled with alkaline phosphatase were purchased from Sigma. Synthetic peptides derived from actin sequences were prepared on solid phase support using a 9050 Milligen PepSynthesizer (Millipore, U.K.) according to the Fmoc/tBu system. The crude peptides were deprotected and thoroughly purified by preparative reverse-phase HPLC. The purified peptides were shown to be homogenous by analytical HPLC. Electrospray mass spectra, carried out in the positive ion mode using a Trio 2000 VG Biotech Mass spectrometer (Altrincham, UK), were in line with the expected structures. Peptides were labeled at the cysteine residue with N-iodo- acetyl-N¢-[sulfo-1-naphthyl]-ethylenediamine (1.5-I-AE- DANS) or at amino groups by FITC [37,38]. Excess reagent was eliminated by sieving through a Biogel P2 col- umn equilibrated with 0.05 M NH 4 HCO 3 buffer, pH 8.0. Immunological techniques ELISA [39], previously described in detail [40], was used to monitor the interaction between coated peptides and cofilin. Peptides (5 lgÆmL )1 )in50m M NaHCO 3 /Na 2 CO 3 ,pH9.5, were immobilized on plastic microtiter wells. The plate was then saturated with 0.5% gelatin and 3% gelatin hydroly- sate, in 140 m M NaCl, 50 m M Tris buffer, pH 7.5. Binding was monitored at 405 nm using alkaline phosphatase- labeled IgGs (dilution 1 : 1000). Control assays were carried out in wells saturated with the mixture of gelatin and gelatin hydrolysate used alone. Each assay was conducted in triplicate and the mean value plotted after subtraction of nonspecific absorption. The binding parameters (apparent dissociation constant K d and the maximal binding A max ) were determined by non linear fitting A ¼ A max · [L]/ (K d + [L]) where A is the absorbance at 405 nm and L the ligand concentration, by using the CURVE FIT software developed by Kevin Raner software, Victoria, Australia. Additional details on the different experimental conditions are given in the figure legends. Fluorescence measurements Fluorescence experiments were conducted using a LS 50 Perkin-Elmer luminescence spectrometer. Spectra for FITC, Oregon green or RITC were obtained with the excitation wavelength set at 470, 480 and 540 nm, respectively. Fluorescence changes were deduced from the area of the emission spectra of FITC or Oregon green between 510 and 530 and 570–590 nm for RITC. The parameters K d (apparent dissociation constant) and A max (maximum effect) were calculated by nonlinear fitting of the experi- mental data points. Actin binding to immobilized cofilin Recombinant human nonmuscle cofilin was coupled to cyanogen-activated Sepharose 4B beads (Amersham Phar- macia Biotech.) according to the manufacturer’s recom- mendations. Excess reactive groups were quenched by washing with 0.1 M Tris buffer, pH 8.0. Prior to actin- binding experiments the beads were washed in Buffer G–ADP at either pH 6.5 (10 m M imidazole, 0.1 m M ADP, 0.2 m M CaCl 2 ,0.2m M dithiothreitol and 1 m M NaN 3 )or pH 8.0 (10 m M Tris, 0.1 m M ADP, 0.2 m M CaCl 2 ,0.2m M dithiothreitol and 1 m M NaN 3 ). ADP–G-actin was made from ATP–actin by incubation with hexokinase (Sigma) and glucose [41] before being added to the indicated total concentration. The beads were collected by centrifugation after incubation. The amount of actin bound to the beads and remaining in the supernatant was measured by scanning SDS/PAGE gels. Analytical methods Protein concentrations were determined by UV absorbency using a Varian MS 100 (Varian SA, Les Ulis, France), and a Pharmacia Ultraspec 2000 spectrophotometer. For cofilin the absorbance was measured at 280 nm, where one absorbance unit is equivalent to 74 l M . For actin solutions, the absorbance was measured at 290 nm where one absorbance unit is equivalent to 38 l M .SDS/PAGEwas carried out on 15% gels as described previously [42] and stained with Coomassie blue R-250. RESULTS pH and F-actin In a previous study [33] we have shown that at pH 6.5, FITC labeled cofilin binds to G- and F-actin, but a change in the fluorescence intensity of FITC occurs only when labeled cofilin interacts with F-actin. In the present study, similar experiments were performed at two pH values (6.5 and 8.0) for comparison. G-actin and FITC–cofilin were Ó FEBS 2002 pH-dependence of cofilin–actin interaction (Eur. J. Biochem. 269) 4195 mixed, the addition of salts (0.1 M KCl and 2 m M MgCl 2 ) then induces oligomerization and the fluorescence was measured at 520 nm as a function of time (Fig. 1). A significant fluorescent enhancement was observed only at pH 6.5, immediately after salt addition and before a significant amount of actin has polymerized [33], even if the very rapid kinetics of cofilin–actin polymerization at pH 6.5 are considered [19]. In a control experiment, no change was observed in the fluorescence intensity of FITC- labeled cofilin used alone after salt addition to the sample (data not shown). Effect of pH on cofilin conformation The regulation of the cofilin activity by pH occurs in a pH range suggesting the involvement of histidine residues. In fact, the single histidine in human and yeast cofilin is not located in the same position [18,43,44] and more generally its position is not conserved during evolution. However, we have looked for a possible structural change in cofilin induced by pH shift. Two kinds of fluorescence experi- ments were performed. We have measured the intrinsic fluorescence of cofilin via its unique tryptophan residue and the extrinsic fluorescence of RITC covalently linked to cofilin at various values of pH between 6.5 and 8.0. We observed no significant changes in fluorescence intensity (not shown), indicating that the environment of these two chromophores in cofilin is independent of pH at least in the range tested. The pH dependence of the cofilin–G-actin interaction We then tested for a pH dependence in the interaction of cofilin with G-actin by two independent methods. G-actin was labeled with RITC and increasing concentrations of cofilin were added. In Fig. 2A, we show a decrease in the fluorescence intensity that is higher at pH 6.5 than pH 8. Analysis of these data shows that the fluorescence decrease extrapolated to infinite cofilin concentrations is significantly different for the two pH (32% and 22% for pH 6.5 and pH 8.0, respectively, Figs 2A and 3). In contrast, an apparent K d of about 1 l M was estimated in both cases. In a control experiment we observed that the fluorescence of the RITC-labeled actin is not affected by pH changes within the same range (not shown). We have confirmed that there is no difference in affinity between G-actin and cofilin by measuring the G-actin binding to cofilin immobilizing on sepharose beads (Fig. 2B). We were able to demonstrate that this was the case for both ADP and ATP–actin (not Fig. 1. Cofilin–actin copolymerization. FITC-cofilin (0.5 l M )and G-actin (5 l M ) were mixed in 50 m M Mops, 0.1 m M ATP, buffer pH 6.5 or 8.0, then 0.1 M KCl, 2 m M MgCl 2 were added. FITC-cofilin fluorescence was monitored at 520 nm versus time at pH 6.5 (—) or pH 8.0 ( ). Fig.2. EffectofpHontheinteractionofactinwithcofilin.(A) Effect of pH on the interaction of RITC-actin with cofilin. Binding of RITC-labeled actin (1.5 l M ) to cofilin in 50 m M Mops, 0.05 m M CaCl 2 ,0.05m M ATP, buffer pH 6.5 or 7.8 was monitored by fluorescence. Changes in the intensity of the emission spectra were recorded at pH 6.5 (d)and7.8(s), in the presence of increasing cofilin concentrations (between 0 and 3.5 l M ). An apparent K d of about 1 l M was estimated in both cases. (B) Binding of ADP-actin to cofilin immobilized on beads in ADP buffer G at either pH 6.5 (d), or pH 8.0 (s). No significant difference in actin binding was found as pH was varied. 4196 L. Blondin et al.(Eur. J. Biochem. 269) Ó FEBS 2002 shown). In order to deduce a more precise location in the actin sequence for pH-dependent structural changes induced by cofilin in F-actin, a peptidic approach was then carried out. Actin sequence correlated with pH effect Two interfaces on the actin surface have been characterized previously as interacting with cofilin [32,33]. Site 1 includes the 18–28 sequence and the C-terminal part of the protein, including the 360–372 sequence. In contrast, site 2 includes sequences between residues 75–135. These two sites contain some histidine residues: three residues in the 84–103 fragment and one in the 360–372 fragment. Another histidine is located in the 38–52 fragment, but this sequence was previously excluded from the interfaces [33]. The possible effect of pH on the actin site 1 was first tested using the C-terminal peptides 356–375 or 360–372. The competition between cofilin towards actin and sequence 360–375 belonging to site 1 was studied at two pH values (6.6 and 7.5) by ELISA. In this experiment the peptide was coated to plastic and the binding of cofilin, fixed at 1.8 l M , was monitored in the presence of increasing actin concen- trations (between 0 and 4.8 l M ).AsshowninFig.4we observed a decrease in the cofilin binding in the presence of actin suggesting that the actin–cofilin complex impedes the interaction of cofilin with the actin peptide. The complex formation between the C-terminal sequence of actin with cofilin was then investigated. Cofilin labeled with RITC was incubated in the presence of increasing concentrations of 355–375 actin peptide and the fluores- cence monitored at pH 6.5 and 8.0. A decrease of fluores- cence was observed. The binding occurs with similar K d of about 2 l M (not shown) and similar fluorescence changes (34% effect) in both cases (Fig. 3). Then, we have checked the peptide 355–375 labeled with IAEDANS at its cysteine residue (at position 374), but the interaction with actin does not induce any fluorescent change. Therefore, we have labeled the peptide with IAEDANS corresponding to actin sequence 360–372 in which an extra cysteine residue was added at its N-terminal extremity. In the present case, we observed an increase of the fluorescence intensity upon cofilin binding. However, the observed variation was similar for the two pH values used (12% effect) (Fig. 3). The interaction of the peptide 84–103 corresponding to a part of site 2 was also checked. As previously reported for site 1, competition between actin and the peptide 84–103 belonging to site 2 was also investigated. As shown in Fig. 5 we observed a decrease in the binding of cofilin to peptide 84–103 in the presence of increasing actin concentration at the two pH values tested. The complex formation between 84 and 103 actin fragment and cofilin was then determined. To perform these experiments, either peptide 84–103 was labeled with Oregon green, or cofilin was labeled with RITC. As shown in Fig. 6, in both cases and for both pH, the peptide binds to cofilin with a K d of about 3 l M . The interaction of cofilin with Oregon green labeled peptide induces a fluorescence decrease that is pH-dependant. The maximum effect extrapolated at infinite cofilin concentration is of 7% at pH 6.5 and 25% at pH 8.0 (Fig. 3). Similarly, in the second experiment where fluorescence intensity change of RITC in labeled cofilin was monitored vs. peptide concentrations, a decrease of 25% is observed at pH 6.5 and only of 14% at pH8.0(Figs3and7). Fig. 3. Effect of pH on the fluorescence changes induced by the inter- action of cofilin with actin or actin derivative synthetic peptides. Results are expressed as the maximum fluorescence variation. Enhance- ment (%) ¼ (A max /F 0 ) · 100 where A max is the maximum fluores- cence change extrapolated at an infinite ligand concentration, and F 0 the initial fluorescence in the absence of ligand. The experiments were performedin50m M Mops buffer pH 6.5 or pH 8.0 with cofilin in the presence of different ligands. (A) Oregon green 84–103 pep- tide + cofilin, (B) rhodamine-labeled cofilin + 84–103 peptide, (C) rhodamine-labeled G-actin + cofilin, (D) rhodamine-labeled cofi- lin + 355–375 peptide, (E) Dansylated 360–372 peptide + cofilin. Fig. 4. Competition binding study between actin fragment 360–372 and G-actin monitored by ELISA. The binding of cofilin (1.8 l M )tocoated actin peptide of sequence 360–372 in 50 m M Mops buffer pH 7.5 (s) or 6.6 (d), supplemented with 3% gelatin hydrolysate, in the presence of increasing G-actin concentrations (0–4.8 l M ). Binding was detected by using anti-cofilin Ig and monitored at 405 nm. Ó FEBS 2002 pH-dependence of cofilin–actin interaction (Eur. J. Biochem. 269) 4197 Evidence for a change in the conformation of actin in the 75–103-actin region The significance of the modifications observed in the interface between cofilin and actin upon pH effect was checked by using a fluorescent probe specific for the site 2 in actin. We have labeled specific purified antibodies directed towards 75–105 actin sequence with Oregon green. The binding of actin to this antibody was monitored at pH 6.5 and 8.0. As shown in Fig. 8, the fluorescence enhancement is about 4 fold higher at pH 8.0 than pH 6.5 while the apparent affinities appear unchanged. In contrast, no change between the two pH, in the fluorescence of antibodies alone, was obtained. This last result showed that antibodies interact in a different local environment with the antigenic epitope located within cofilin site 2 in actin sequence. DISCUSSION The ADF/cofilins are so far unique amongst the many distinct types of actin-binding proteins in their ability to alter the twist of actin filaments [20]. This property possibly explains the extreme cooperativity of F-actin binding [12,13,17,20] and perhaps severing [45]. The manner in which cofilin achieves this feat remains contentious and two broad models have been proposed. Both propose a binding geometry where cofilin binds one actin monomer at subdomain 1, and a second, longitudinally associated monomer immediately toward the barbed end at subdo- main 2. The major difference between the models is in the Fig. 5. Competition binding study between actin fragment 84–103 and G-actin monitored by ELISA. The binding of cofilin (1.1 l M )tocoated actin peptide of sequence 360–372 in 50 m M Mops buffer pH 7.5 (s) or 6.6 (d), supplemented with 3% gelatin hydrolysate, in the presence of increasing G-actin concentrations (0–2.4 l M ). Binding was detected by using anti-cofilin Ig and monitored at 405 nm. Fig. 6. Binding of cofilin with 84–103 actin sequence evidenced by fluorescence. Changes in the emission spectrum intensities of 84–103 peptide (0.6 l M ) labeled with Oregon green were monitored in the presence of cofilin (0–7 l M ). The experiments were carried out in 50 m M MOPS buffer pH 6.5 (d)orpH8.0(s). Fig. 7. Interaction of RITC-labeled cofilin with 84–103 actin sequence evidenced by fluorescence. Changes in the emission spectrum intensities of RITC-cofilin (2 l M ) were measured in the presence of 84–103 peptide (0–20 l M ). The experiments were carried out in 50 m M Mops buffer pH 6.5 (d)orpH8.0(s). Fig. 8. Binding of purified antibodies directed to 75–105 sequence of actin to G-actin monitored by fluorescence measurements. Changes in the emission spectrum intensities of antibodies (0.25 l M )labeledwith Oregon green were monitored in the presence of G-actin (0–4.2 l M ). The experiments were carried out in 50 m M Mops buffer pH 6.5 (d)or pH 8.0 (s). 4198 L. Blondin et al.(Eur. J. Biochem. 269) Ó FEBS 2002 position of the cofilin with respect to the second associated monomer. We [32,33] propose that cofilin intercalates into the filament between the two associated monomers to bind the second through an interaction with the upper ÔrearÕ of subdomain 2. A number of other groups [20,31,46], suggest that cofilin binds the ÔforwardÕ facing surface of subdomain 2 (that is, in the standard orientation as first displayed by Kabsch and colleagues [47]). These models predict profound differences in the surfaces of cofilin that would be exposed at the surface of the cofilin:actin filament, and in the interfaces between the molecules. An additional complexity is that in one reconstruction, a second ADF/cofilin was proposed to bind the filament [31] so that the over all ratio in these filaments was two ADF/cofilins to every actin, this would perhaps explain why others have found bundling activity associated with ADF/cofilins [48]. However, both pheno- menon could be explained by oxidation of the many cysteine residues carried by these proteins. The interaction of typical ADF/cofilins is pH sensitive but the molecular mechanism has not yet been explained. The pH sensitivity could result from three nonexclusive possibilities: it may arise from titratable residues on either surface; alternatively, the tertiary structure of cofilin may undergo a pH-sensitive change, or finally, a conformational change could be displayed by actin, either by actin monomers or between monomers associated within the filament. Binding of cofilin to G-actin at site 1 is pH-insensitive The ADF/cofilin family bind G-actin through subdomain 1 [49,50]. We have shown here by two independent means that the interaction of ADF/cofilin with G-actin through site 1 is not sensitive to pH within normal physiological range (pH 6.5–8.0). We measured the affinity of actin and cofilin by changes in the fluorescence of RITC labelled actin. Although the fluorescence change between RITC-actin and cofilin was larger when measured at pH 6.5 than at pH 7.8 (Fig. 2A) the calculated affinities were similar (K d ¼ 1 l M ). This value is higher than that typically measured for actin- G–actin interaction at 0.1 l M , probably as a result of the label as it is known that modification of Cys374 by other agents inhibits the interaction with cofilin [19]. We measured the affinity of binding of cofilin to unmodified ADP-G-actin and ATP-G-actin as a function of pH by direct means in order to confirm that the lack of pH sensitivity was not an artifact of labelled actin. We found no difference in binding of either ADP or ATP-actin to cofilin immobilized on beads between pH 6.5 and pH 8.0 (Fig. 2B). The G-actin binding footprint of yeast cofilin has been determined at pH 8.0 by synchrotron protein footprinting [51]. The G-actin binding footprint in surprisingly large, encompassing roughly a third of the surface of cofilin. No evidence for pH dependent conformational changes in the structure of cofilin We could find no evidence for substantial pH-dependent conformational changes in cofilin that might explain the pH-dependent nature of the interaction of ADF/cofilins with F-actin. We measured the intrinsic fluorescence of cofilin via its unique tryptophan residue and the extrinsic fluorescence of RITC covalently linked to cofilin at various pHs between 6.5 and 8.0 and observed no significant changes in fluorescence intensity in agreement with other studies using circular dichroism and limited proteolysis [52]. No evidence for pH-sensitivity in the actin:cofilin surfaces directly None of the actin-derived peptides that we have previously showntobindactin,dosoinapH-sensitivemanner. However, actin itself is pH-sensitive, the spontaneous polymerization of actin is more rapid at pH 6.5 than at pH 8.0 [22,53], the intermonomeric flexibility of Mg 2+ -actin filaments is larger at pH 7.4 than at pH 6.5 [54], and actin filaments are stabilized at low pH [55]. PH-dependence may result from conformational changes in the actin monomer itself We have detected a pH-sensitive change in the structure of actin subdomain 1 that may explain the overt pH sensitivity of the cofilin-F–actin interaction. The interaction of Oregon green coupled antibodies directed to residues 75–105 of actin is strongly pH-sensitive, most probably because of a difference in conformation of G-actin at the two pH extremes. This finding was confirmed by fluorescence measurements of a similar peptide 83–103 labelled with Oregon green that again showed pH-dependent changes in the presence of cofilin. Evidence for a pH sensitive change in conformation in subdomain 1 of G-actin has come from studies with AEDANS labeled actin [23]. Residues 75–105 encompass part of cofilin binding site 2 [33] and is situated between subdomains 1 and 2. FRET analysis has shown that cofilin binding alters the orientation of subdomain 1 and 2 of actin [33]. We have proposed that cofilin binds a second site (site 2) on F-actin [32], consisting of a helical region 112–125 that lies on the Ôupper, rearÕ surface of subdomain 1 close to subdomains 2 when viewed in the standard actin orientation [47]. This second site of actin is proposed to be cryptic, pH sensitive movements of region 75–105, may make site 2 (in region 112–125) available for binding probably by the C-terminal helix of ADF/cofilin [46]. We have previously shown that FITC-cofilin binds to both G- and F-actin and that this induces an increase in fluorescence in conditions that allow actin oligomerization to occur [33]. This increase in fluorescence is very much more rapid than polymerization and probably reflects a conformational change. We now show that this conforma- tional change only occurs at pH 6.5 and not at pH 8.0 (Fig. 1), suggesting that site 2 is present only in F-actin and in G-actin bound through site 1 by cofilin at pH 6.5. Three actin-binding sites of cofilin? Present evidence suggests that cofilin binds actin at site 1 through the N-terminal region [49], and site 2 possibly through the C-terminal helix [46]. Many studies have indicated that the so called long helix is important to actin-binding and additionally mutations here dissociate severing from pointed end off rate increase. It has been suggested that cofilin binds a third site on actin by a region around K114 on cofilin’s long helix [56]. This site may be an as yet unrecognized distinct region on actin, or a region Ó FEBS 2002 pH-dependence of cofilin–actin interaction (Eur. J. Biochem. 269) 4199 contiguous with those surfaces already identified as sites 1 or 2. We favour the latter hypothesis, and since K114 appears at the surface of cofilin so close to the N-terminus, we further hypothesize that site 1 is contiguous with site 3, in agreement with data obtained by synchrotron protein footprinting [51]. This is also in agreement with the finding that a peptide including K114 can be crosslinked to Cys374 on actin [57]. Implications for other actin-binding proteins The ADF homology domain (ADF-H) is defined as a protein sequence motif shared between the AC family members and a number of other proteins distinct from the ACs [58]. These include twinfilin, which has tandem ADF- H domains [59], coactosin and Abp1p (see [60]). Surpris- ingly, the gelsolin repeat is similar to the ADF-H fold despite having little sequence homology [61]. However gelsolin domain 2 binds actin through an interface distinct from that of gelsolin domain1 and both through interfaces distinct from cofilin [32]. Thus, the group of proteins that possess ADF-H sequence motifs or that share homologous folds, tend to share some actin binding properties such as PIP 2 sensitivity, ADP-actin monomer preference and (in some cases) pH dependence, yet paradoxically bind actin through distinct interfaces [62]. Actophorin and depactin (from Acanthamoeba and star fish eggs, respectively) are members of the ADF/cofilin family that are not pH sensitive. Depactin is reported as being not pH sensitive [16] and actophorin binds to F-actin at both pH 6.5 and pH 8.0 [17]. Any explanation of pH sensitivity of the ADF/cofilins must also explain why these otherwise typical ADF/cofilins are not pH sensitive. It is possible that actophorin does not share typical pH dependence of F-actin binding because site 2 is not hidden from binding as it is in the case of cofilin. These experiments are in progress. ACKNOWLEDGEMENTS This research was supported by grants from AFM and Amoebics Ltd. Edinburgh. REFERENCES 1. Bamburg, J.R. (1999) Proteins of the ADF/cofiln family: Essential regulators of actin dynamics. Annu.Rev.CellDev.Biol.15,185– 230. 2. Pollard, T.D., Blanchoin, L. & Mullins, R.D. (2000) Molecular mechanisms controlling actin filament dynamics in nonmuscle cells. Annu. Rev. Biophys. Biomol. Sruct. 29, 545–576. 3. Morgan, T.E., Lockerbie, R.O., Minamide, L.S., Browning, M.D. & Bamburg, J.R. (1993) Isolation and characterization of a regulated form of actin depolymerizing factor. J. Cell Biol. 122, 623–633. 4. Yonezawa, N., Nishida, E., Iida, K., Yahara, I. & Sakai, H. (1990) Inhibition of the interactions of cofilin, destrin, and deoxy- ribonuclease-1 with actin by phosphoinositides. J. Biol. Chem. 265, 8382–8386. 5. Quirk, S. & Maciver, S.K. (1993) Primary structure of and studies on Acanthamoeba actophorin. Biochemistry 32, 8525– 8533. 6. Gungabissoon,R.A.,Jiang,C J.,Drøbak,B.K.,Maciver,S.K.& Hussey, P.J. (1998) Interaction of maize actin-depolymerising factor with actin and phosphoinositides and its inhibition of plant phospholipase C. Plant J. 16, 689–696. 7. Bernstein, B.W. & Bamburg, J.R. (1982) Tropomyosin binding to F-actin protects the F-actin from disassembly by brain actin depolymerizing factor (ADF). Cell Motility 2,1–8. 8. Yonezawa, N., Nishida, E., Maekawa, S. & Sakai, H. (1988) Studies on the interaction between actin and cofilin purified by a new method. Biochem. J. 251, 121–127. 9. Okada, K., Obinata, T. & Abe, H. (1999) XAIP1: a Xenopus homoloque of yeast actin interacting protein 1 (AIP1), which induces disassembly of actin filaments cooperatively with ADF/ cofilin family proteins. J. Cell Sci. 112, 1553–1565. 10. Ono, S. & Ono, K. (2002) Tropomyosin inhibits ADF/cofilin- dependent actin filament dynamics. J. Cell Biol. 156, 1065–1076. 11. Yonezawa, N., Nishida, E. & Sakai, H. (1985) pH control of actin polymerization by cofilin. J. Biol. Chem. 260, 14410–14412. 12. Hayden, S.M., Miller, P.S., Brauweiler, A. & Bamburg, J.R. (1993) Analysis of the interactions of actin depolymerizing factor (ADF)withG-andF-actin.Biochemistry 32, 9994–10004. 13. Hawkins, M., Pope, B., Maciver, S.K. & Weeds, A.G. (1993) Human actin depolymerizing factor mediates a pH-sensitive destruction of actin filaments. Biochemistry 32, 9985–9993. 14.Yeoh,S.,Pope,B.,Mannherz,H.G.&Weeds,A.(2002) Determining the differences in actin binding by human ADF and cofilin. J. Mol. Biol. 315, 911–925. 15. Bernstein, B.W., Painter, W.B., Chen, H., Minamide, L.S., Abe, H. & Bamburg, J.R. (2000) Intracellular pH modulation of ADF/ Cofilin proteins. Cell Motility Cytoskeleton 47, 319–336. 16. Mabuchi, I. (1983) An actin-depolymerizing protein (depactin) from starfish oocytes: properties and interaction with actin. J. Cell Biol. 97, 1612–1621. 17. Maciver,S.K.,Pope,B.J.,Whytock,S.&Weeds,A.G.(1998)The effect of two ADF/cofilins on actin filament turnover: pH sensi- tivity of F-actin by human ADF, but not of Acanthamoeba actophorin. Biochemistry 256, 388–397. 18. Iida, K., Moriyama, K., Matsumoto, S., Kawasaki, H., Nishida, E. & Yahara, I. (1993) Isolation of a yeast essential gene, COF1, that encodes a homologue of mammalian cofilin, a low-M r actin- binding and depolymerizing protein. Gene 124, 115–120. 19. Bonet, C., Ternent, D., Maciver, S.K. & Mozo-Villarias, A. (2000) Rapid formation and high diffusibility of actin-cofilin cofilaments at low pH. Eur. J. Biochem. 267, 1–8. 20. McGough,A.,Pope,B.,Chiu,W.&Weeds,A.(1997)Cofilin changes the twist of F-actin: Implications for actin filament dynamics and cellular function. J. Cell Biol. 138, 771–781. 21. Maciver, S.K., Zot, H.G. & Pollard, T.D. (1991) Characterization of actin filament severing by actophorin from Acanthamoeba castellanii. J. Cell Biol. 115, 1611–1620. 22. Zimmerle, C.T. & Frieden, C. (1988) Effect of pH on the mechanism of actin polymerization. Biochemistry 27, 7766–7772. 23. Zimmerle, C.T. & Frieden, C. (1988) pH-induced changes in G-actin conformation and metal affinity. Biochemistry 27, 7759– 7765. 24. Simchowitz, L. & Cragoe, E.J. Jr (1986) Regulation of neutrophil chemotaxis by intracellular pH. J. Biol. Chem. 261, 6492–6500. 25. Schwiening, C.J. & Willoughby, D. (2002) Depolarization-induced pH microdomains and their relationship to calcium transients in isolated snail neurones. J. Physiol. 538, 371–382. 26. Lipton, P. (1999) Ischemic cell death in brain neurons. Physiol. Rev. 79, 1431–1568. 27. Bamburg, J.R. & Bray, D. (1987) Distribution and cellular loca- lization of Actin Depolymerizing Factor. J. Cell Biol. 105, 2817– 2825. 28. Jiang, C.J., Weeds, A.G. & Hussey, P.J. (1997) The maize actin- depolymerizing factor, ZmADF3, redistributes to the growing tip of elongating root hairs and can be induced to translocate into the nucleus with actin. Plant J. 12, 1035–1043. 4200 L. Blondin et al.(Eur. J. Biochem. 269) Ó FEBS 2002 29. Heuser, J. (1989) Changes in lysosome shape and distribution correlated with changes in cytoplasmic pH. J. Cell Biol. 108, 855– 864. 30. Pope, B.J., Gonsior, S.M., Yeoh, S., McGough, A. & Weeds, A.G. (2000) Uncoupling actin filament fragmentation by cofilin from increased subunit turnover. J. Mol. Biol. 298, 649–661. 31. Galkin, V.E., Orlova, A., Lukoyanova, N., Wriggers, W. & Egelman, E.H. (2001) Actin depolymerization factor stabilizes an existing state of F-actin and can change the tilt of F-actin subunits. J. Cell Biol. 153, 75–86. 32. Renoult, C., Ternent, D., Maciver, S.K., Fattoum, A., Astier, C., Benyamin, Y. & Roustan, C. (1999) The identification of a second cofilin binding site on actin suggests a novel, intercalated arrange- ment of F-actin binding. J. Biol. Chem. 274, 28893–28899. 33. Blondin, L., Sapountzi, V., Maciver, S.K., Renoult, C., Benyamin, Y. & Roustan, C. (2001) The second ADF/cofilin actin-binding site exists in F-actin, the cofilin: G-actin complex, but not in G-actin. Eur. J. Biochem. 268, 6426–6434. 34. Spudich, J.A. & Watt, S. (1971) The regulation of rabbit skeletal muscle contraction. Biochemical studies of the interaction of the tropomyosin-troponin complex with actin and the proteolytic fragments of myosin. J. Biol. Chem. 246, 4866–4871. 35. Maciver, S.K. & Harrington, C.R. (1995) Two actin-binding proteins, actin depolymerizing factor and cofilin, are associated with hirano bodies. Neuroreport 6, 1985–1988. 36. Benyamin, Y., Roustan, C. & Boyer, M. (1986) Anti-actin anti- bodies. Chemical modification allows the selective production of antibodies to the N-terminal region. J. Immunol. Meth. 86, 21–29. 37. Takashi, R. (1979) Fluorescence energy transfer between sub- fragment-1 and actin points in the rigor complex of actosubfrag- ment-1. Biochemistry 18, 5164–5169. 38. Miki, M., dos Remedios, C.G. & Barden, J.A. (1987) Spatial relationship between the nucleotide-binding site, Lys-61 and Cys- 374 in actin and a conformational change induced by myosin subfragment-1 binding. Eur. J. Biochem. 168, 339–345. 39. Engvall, E. (1980) Enzyme immunoassay ELISA and EMIT. Methods Enzymol. 70, 419–439. 40. Me ´ jean, C., Lebart, M.C., Poyer, M., Roustan, C. & Benyamin, Y. (1992) Localization and identification of actin structures involved in the filamin–actin interaction. Eur. J. Biochem. 209, 555–562. 41. Pollard, T.D. (1986) Rate constants for the reactions of ATP- and ADP-actinwiththeendsofactinfilaments.J. Cell Biol. 103, 2747– 2754. 42. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. 43. Ogawa, K., Tashima, M., Yumato, Y., Okuda, T., Sawada, H., Okuma, M. & Maruyama, Y. (1990) Coding sequence of human placenta cofilin cDNA. Nucleic Acids Res. 18, 7169. 44. Moon, A.L., Janmey, P.A., Louie, K.A. & Drubin, D.G. (1992) Cofilin is an essential component of the yeast cortical cytoskeleton. J.Cell Biol. 120, 421–435. 45. Maciver, S.K. (1998) How ADF/cofilin depolymerizes actin fila- ments. Curr. Op. Cell Biol. 10, 140–144. 46. Ono, S., McGough, A., Pope, B.J., Tolbert, V.T., Bui, A., Pohl, J., Benian, G.M., Gernert, K.M. & Weeds, A.G. (2001) The C-terminal tail of UNC-60B (actin depolymerizing factor/cofilin) is critical for maintaining its stable association with F-actin and is implicated in the second actin-binding site. J. Biol. Chem. 276, 5952–5958. 47. Kabsch, W., Mannherz, H.G., Suck, D., Pai, E.F. & Holmes, K.C. (1990) Atomic structure of the actin–DNase I complex. Nature 347, 37–44. 48. Pfannstiel,J.,Cyrklaff,M.,Habermann,A.,Stoeva,S.,Griffiths, G., Shoeman, R. & Faulstich, H. (2001) Human cofilin forms oligomers exhibiting actin bundling activity. J. Biol. Chem. 276, 49476–49484. 49. Sutoh, K. & Mabuchi, I. (1984) N-terminal and C-terminal segments of actin participate in binding depactin, an actin- depolymerizing protein from starfish oocytes. Biochemistry 23, 6757–6761. 50. Muneyuki, E., Nishida, E., Sutoh, K. & Sakai, H. (1985) Puri- fication of cofilin, a 21,000 molecular weight actin-binding protein, from porcine kidney and identification of the cofilin-binding site in the actin sequence. J. Biochem. 97, 563–568. 51. Guan, J.Q., Vorobiev, S., Almo, S.C. & Chance, M.R. (2002) Mapping the G-actin binding surface of cofilin using synchrotron protein footprinting. Biochemistry 41, 5765–5775. 52. Arima,K.,Imanaka,M.,Okuzono,S.,Kazuta,Y.&Kotani,S. (1998) Evidence for structural differences between the two highly homologous actin-regulatory proteins: destrin and cofilin. Biosci. Biotechn Biochem. 62, 215–220. 53. Wang, F., Sampogna, R.V. & Ware, B.R. (1989) pH dependence of actin self-assembly. Biophys. J. 55, 293–298. 54. Hild, G., Nyitrai, M. & Somogyi, B. (2002) Intermonomer flexibility of Ca- and Mg-actin filaments at different pH values. Eur. J. Biochem. 269, 842–849. 55. Oda,T.,Makino,K.,Yamashita,I.,Namba,K.&Maeda,Y. (2001) Distinct structural changes detected by X-ray diffraction in stabilization of F-actin by lowering pH and increasing ionic strength. Biophys. J. 80, 841–851. 56. Moriyama, K. & Yahara, I. (2002) The actin-severing activity of cofilin is exerted by the interplay of three distinct sites on cofilin and essential for cell viability. Biochem. J. 365, 147–155. 57. Yonezawa, N., Nishida, E., Iida, K., Kumagai, H.I., Yahara & Sakai, H. (1991) Inhibition of actin polymerization by a synthetic dodecapeptide patterned on the sequence around the actin-bind- ing site of cofilin. J. Biol. Chem. 266, 10485–10489. 58. Lappalainen, P., Kessels, M.M., Cope, M.J.T.V. & Drubin, D.G. (1998) The ADF homolog (ADF-H) domain: a highly exploited actin-binding module. Mol. Biol. Cell 9, 1951–1959. 59. Palmgren, S., Vartiainen, M. & Lappalainen, P. (2002) Twinfilin, a molecular mailman for actin monomers. J. Cell Sci. 115, 881–886. 60. Maciver, S.K. & Hussey, P.J. (2002) The ADF/cofilin family: actin-remodeling proteins. BMC Genome Biol. 3 (5), 3007.1– 3007.3007.12. 61. Hatanaka,H.,Ogura,K.,Moriyama,M.,Ichikawa,S.,Yahara,I. & Inagaki, F. (1996) Tertiay structure of destrin and structural similarity between two actin-regulating protein families. Cell 85, 1047–1055. 62. Renoult, C., Blondin, L., Fattoum, A., Ternent, D., Maciver, S.K., Raynaud, F., Benyamin, Y. & Roustan, C. (2001) Binding of gelsolin domain 2 to actin. An actin interface distinct from that of gelsolin domain 1 and from ADF/cofilin. Eur. J. Biochem. 268, 6165–6175. Ó FEBS 2002 pH-dependence of cofilin–actin interaction (Eur. J. Biochem. 269) 4201 . A structural basis for the pH-dependence of cofilin F-actin interactions Laurence Blondin 1 , Vasilia Sapountzi 2 , Sutherland K. Maciver 1 , Emeline Lagarrigue 2 , Yves Benyamin 1 and Claude. by human ADF, but not of Acanthamoeba actophorin. Biochemistry 256, 388–397. 18. Iida, K., Moriyama, K., Matsumoto, S., Kawasaki, H., Nishida, E. & Yahara, I. (1993) Isolation of a yeast essential. 100 (Varian SA, Les Ulis, France), and a Pharmacia Ultraspec 2000 spectrophotometer. For cofilin the absorbance was measured at 280 nm, where one absorbance unit is equivalent to 74 l M . For actin