Eur. J. Biochem. 182,267-275 (1989) I FEBS 1989 ((

The complex of actin and deoxyribonuclease I as a model system to study the interactions of nucleotides, cations and cytochalasin D with monomeric actin Bernhard POLZAR’, Ewa NOWAK ’, Roger S. GOODY and Hans Georg MANNHERZ’ lnstitut fur Anatomie und Zellbiologie, Abteilung Zellbiologie, Philipps-Universitat, Marburg Nencki Institute of Experimental Biology, Department of Muscle Biochemistry, Warszawa Abteilung Biophysik, Max-Planck-Institut fur medizinische Forschung, Heidelberg (Received October 21, 1988/February 13,1989) - EJB 88 1254

The stoichiometric actin - DNase-I complex was used to study the actin - nucleotide and actin - divalentcation interactions and its ATPase activity in the presence of MgClz and cytochalasin D. Treatment of actin DNase-I complex with 1 mM EDTA results in almost complete restoration of its otherwise inhibited DNase I activity, although the complex does not dissociate, as verified by size-exclusion chromatography. This effect is due to a loss of actin-bound nucleotide but is prevented by the presence of 0.1 -0.5 mM ATP, ADP and certain ATP analogues. In this case no increase in DNase I activity occurs, even in the presence of EDTA. At high salt concentrations and in the presence of Mg2+ (‘physiological conditions’) the association rate constants for ATP, ADP and EATP (1,N6-ethenoadenosine5’-triphosphate) and the dissociation rate constant for EATP were determined. Both the on and off rates were found to be reduced by a factor of about 10 when compared to uncomplexed actin. Thus the binding constant of cATP to actin is almost unaltered after complexing to DNase I (2.16 x 10’ M-’). Titrating the increase in DNase 1 activity of the actin-DNase I complex against nucleotide concentration in the presence of EDTA, the association constant of ATP to the cation-free form of actin - DNase I complex was found to be 5 x lo3 M - l , which is many orders of magnitude lower than in the presence of divalent metal ions. The binding constant of Ca2+ to the high-affinity metal-binding site of actin was found not to be altered when complexed to DNase 1, although the rate of Ca2+ release decreases by a factor of 8 after actin binding to DNase I. The rate of denaturation of nucleotide-free and metal-ion-free actin - DNase I complex was found to be reduced by a factor of about 15. The ATPase activity of the complex is stimulated by addition of Mg2+ and even more effectively by cytochalasin D, proving that this drug is able to interact with monomeric actin.

Complex formation leads to an inhibition of the DNase I activity and the ability of actin to polymerize [S, 61. Although the biological significance of the actin- DNase I interaction is still unclear, one can speculate that similar kinds of interactions are operative in the interactions of other proteins with actin, since competition of DNase I with tropomyosin and myosin subfragment 1 for binding to actin has been reported [3, 41. In addition to its F-actin-depolymerizing capacity, a capping activity has been shown to be exhibited by DNase I [7] (and our own unpublished results). Thus, the interaction of actin and DNase I may be similar to the mode of interaction of other actin-binding proteins with actin. Furthermore, since DNase 1 fixes actin in a quasi-monomeric state, even at high ionic strength the influence of cations on the actin -nucleotide interaction can be analyzed more directly at concentrations which would normally lead to actin polymerization. Correspondence to H. G. Mannherz, Institut fur Anatomie und Zellbiologie, Abteilung Zcllbiologic, Philipps-Univcrsitat Marburg, We have therefore investigated the effect of DNase I on Robcrt-Koch-StraDe 6 , D-3550 Marburg, F R G the interaction of actin with nucleotide and cations and the Abbreviations. EATP, 1,N6-ethenoadenosine 5’-triphosphate; interaction of nucleotide-free actin with DNase I. The stabiliADP[/lS], adenosine 5’-[fl-thio]diphosphate; ATP[yS], adenosine 5’- zation of actin in a quasi-monomeric state by DNase 1further[r-thioltriphosphate, ATP[B,yNH], adenosine 5’-[fl,y-imino]triphosphate; ATP[P,yCH,], adenosinc 5’-[/?,y-rnethylcne]triphosphdte; more allowed the measurement of the rate of ATP hydrolysis Ap,A, adenosine(5’)pentaphospho(5’)adenosine; Quin 2, {2-[2- by the actin - DNase-I complex and its modulation by cations and cytochalasin D. The results obtained specify further the bis(carboxymethy1)amino-5-methylphenoxy]methyl} -6-methoxy-8bis(carboxymethy1)aminoquinoline. mechanism of the actin - nucleotide interaction.

Actin is one of the most abundant proteins in nature, occurring in many eukaryotic cells. In striated muscle cells, actin is the main constituent of thin filaments. In non-muscle cells, actin can transiently form microfilaments or bundles thereof. In these cells the state of aggregation of actin is regulated by a number of proteins able to interact specifically with particular states of actin aggregation (for a review see [l]). In order to understand the mechanisms by which these proteins modify the polymerization behaviour of actin it is necessary also to investigate their effects on interactions of actin with nucleotides and cations since these interactions play a crucial role in polymerization and depolymerization processes of actin [I]. Deoxyribonuclease I (DNase I) has been shown to interact preferentially with monomeric actin (G-actin), thus forming a stoichiometric 1 : 1 complex [2 - 41.

268 MATERIALS AND METHODS Protein preparations

Determination of the rate of EATPreassociation to actin - DNase-I complex

The determinations were performed using the three-syringe Actin preparations from rabbit skeletal muscle were rapid-mixing apparatus described above. cATP - G-actin and obtained as described earlier [3, 51. Final ATP-G-actin solu- DNase 1 were diluted to 3.6 pM and 6 pM, respectively, in tions in 0.2 mM ATP, 0.2 mM CaCl,, and 2 mM Hepes buffer, 4 mM Hepes pH 8.3, in one of the syringes directly before the pH 7.6, were kept at 0°C for no longer than one week. To measurements. Dissociation of the nucleotide was initiated by obtain actin with the fluorescent 1,N6-etheno analog of ATP mixing the protein complex with an equal volume of 2 mM (cATP) bound to it, ATP was substituted by sATP in all EDTA solution in 4 mM Hepes buffer, pH 8.3, from the sepreparative steps after acetone powder extraction. After two cond syringe. The mixture (1.8 pM actin and 3 pM DNase I, cycles of polymerization and depolymerizdtion the EATP- 1 mM EDTA, 4 m M Hepes pH 8.3) was allowed to react in actin was dialyzed against the solution of 0.2 mM EATP, an incubation tube for 40 s. After this time it was mixed 0.2 mM CaC12, 2 mM Hepes pH 7.6. with an equal volume of a solution from the third syringe DNase I from bovine pancreas was a commercial product containing 2.6 mM MgCl,, 0.2 M KCl, 4 mM Hepes pH 7.2, obtained from Worthington and was further purified as de- and various concentrations of either eATP, ATP or ADP and scribed earlier [5]. the fluorescence signal was recorded. It was experimentally established that, with the combination of buffers used, the final pH during the reassociation reaction was 7.6. The meaFluorescence measurements surements were performed at room temperature (21 _+ l -C). The binding of EATP to actin was monitored using the In each experiment, the time course of EATPdissociation from large ratio of relative molar fluorescence intensities of EATP- the initial EATP- G-actin- DNase 1 complex upon mixing G-actin complex and free eATP [7]. The kinetics of displace- with EDTA was recorded using a dual-mixing system (the ment of the bound EATP by added ATP and dissociation third syringe disconnected). From the fluorescence intensity of cATP in the presence of EDTA in excess over CaC12 at traces the concentration of EATP released from the complex 'physiological ionic strength' (i.e. 0.8 mM MgCI,, 0.1 M KCI, at the time of mixing with the MgCl,/KCl solution in the pH 7.6) were examined in a fluorescence spectrofluorimeter reassociation experiment was calculated, assuming that the (SLM 8000) equipped with a thermostatted cell holder which total fluorescence change in EDTA corresponded to cATP was maintained at 20.0 f- 0.5"C. Excitation and emission equimolar to actin. The initial concentration of free EATP wavelengths were 350 nm and 410 nm, respectively. Neither at the time of mixing with a MgCl,/KCl solution was then spectra changed by adding DNase I to EATP- G-actin. obtained as the sum of free nucleotide introduced with actin, Stopped-flow measurements of FATP dissociation from the the nucleotide released from actin - DNase-I complex and actin-DNase-I complex in the presence of EDTA and its that added together with MgCl,/KCl. reassociation upon addition of CaCl, in excess over EDTA were carried out with a Durrum D-132 three-syringe rapidmixing apparatus equipped with a D-137 dual-detector unit. Determination of the Ca2'-binding constant The mixing dead-time was 0.2 ms. Fluorescence was excited The affinity of Ca2+ for the actin - DNase-I complex was at 340 nm and the emitted light was monitored at 90" using a determined in the presence of 200 pM ATP by competition filter with a transmission cut-off at 360 nm. The fluorescence with Quin 2 essentially as described for G-actin by Gershman signal was divided by the signal obtained from the straight- et al. [9]. through beam to improve signal/noise ratio. The resulting signal was registered on a Nicolet 4094 digital oscilloscope. In the reassociation experiments usually between three and Other procedures six traces obtained from consecutive runs of the same reaction Actin concentration was determined spectrophotometwere averaged. For evaluation, traces on chart paper were rically at 290 nm using an absorption coefficient of 0.63 mg obtained from the transient recorder using an X- Y recorder. ml- cm- [3]. Molar actin concentrations were calculated They were analyzed by fitting to a single exponential using a with the value of42000 for the M , of G-actin. DNase I activity North 500/100 computer. was measured by the hyperchromicity test as given [lo] and is expressed in Kunitz units (1 KU causes a change of absorbance at 260 nm of 0.001 min-' cm-I). ATP hydrolysis Determinution of the rate of' displacement o f t h e bound EATP of actin-bound t3H]ATP was monitored by thin-layer by ATP chromatography of acid-quenched and neutralized samples EATP- G-actin and DNase I solutions were diluted in the as detailed [S]. Polyacrylamide gel electrophoresis in the presfluorescence cell to a final concentration of 0.9 pM actin and ence of sodium dodecyl sulfate (SDS/PAGE) was performed 1.5 pM DNase I directly before the measurement. The reac- according to Laemmli [ll]. The concentration of eATP was tion was started by addition of ATP at a final concentration determined at 265 nm using an absorption coefficient of of 1.1mM and the decrease in the fluorescence intensity was 5700 M - l cm-' [12]. followed. All measurements were performed under conditions of ionic strength and divalent metal ion (100 mM KCl, Reagents 0.8 mM MgC12) chosen to be near to physiological. The pH EATP was prepared by the method of Secrist et al. [I21 was maintained at pH 7.6 using 4 mM Hepes buffer. The concentration of free EATP introduced with actin was in the and separated from unreacted ATP by chromatography on a range of 2 -4 pM, i.e. 500-250 times lower than that of the QEAE-Sephadex G-SO column. ATP[yS] [13], ATP[aS] [14], added ATP. Under these conditions the reassociation of EATP ADP[fiS] [13], ATP[P,yNH] [15] and Ap,A [16] were synthesized as described. ATP[P,yCH,J was a product of Miles with actin could be neglected.

'

'

269 4

Table 1. Ef3ciency of adenosine and various adenine nucleotides in maintaining the DNase I inhibitory capacity ofactin in the presence o j I mM EDTA Monomeric (G), polymeric (F) actin and actin in complex with DNase I were employed; 100% corresponds to full inhibitory capacity which was determined in the absence of EDTA and the presence of 0.1 mM ATP; 0% inhibitory capacity corresponds to the DNAdegrading activity in the presence of 1 mM EDTA and 0.1 mM adenosine. Actin at 10 FM, freed from unbound nucleotides by gel filtration, was incubated for 1 h at room temperature with the nucleotides indicated. Thereafter the activity of actin DNase-I complex was determined directly or G- and F-actins were preincubated with equimolar DNase I for a further 20 min at room temperature. For the DNase I assay, a final actin-DNase-I concentration of 10 nM was used in 1 ml tcst solution ~

'

0 0

10

20 30 LO rnin

50 60

Analogue Fig. 1. The rate of loss of the inhibitory action of actin on DNase I after addition of I mM EDTA to actin- DNase-I complex. Actin- DNase I was freed of excess nucleotide by gel filtration prior to the experiment and used at 10 pM final concentration in 0.1 mM CaC12, 0.5 mM NaN3 and 5 mM Hepes/NaOH pH 7.4. At time zero, 1 mM EDTA and nucleotides were added to give a final concentration of 0.1 mM ATP (A),ADP (a), AMP ( 8 )and adenosine (0). At time intervals indicated aliquots of 10 ~l were analyzed for DNase I activity using the hyperchromicity test as described in Materials and Methods. ~ ~ x min-' x cm-' DNase I activity is expressed as L I A of~ 0.001 (Kunitz units, KU)

Corp., Frankfurt, FRG. ATP, ADP, AMP and adenosine were purchased from Pharma-Waldhof (Dusseldorf, FRG). Both AMP and adenosine were found to be free of ADP and ATP by HPLC. [2,8-3H]-ATP was a product of Amersham Buchler (Braunschweig, FRG). Cellulose thin-layer chromatography sheets impregnated with poly(ethy1eneimine) were obtained from Machery and Nagel (Duren, FRG). Cytochalasin D was from Serva (Heidelberg, FRG) and Quin 2 from Calbiochem. Hepes was a product of Carl Roth (Karlsruhe, FRG). RESULTS Effict of'nucleotide binding to actin on its inhibitory capacity on DNuse I

In a previous communication investigating the effect of 5'-nucleotidase and other ATP-degrading enzymes on the inhibitory capacity of actin on DNase I [17] it was shown that dephosphorylation of actin-bound nucleotide is accompanied by a loss of its ability to inhibit the DNA-degrading activity of DNase I. We reinvestigated this effect by addition of EDTA in order to achieve rapid exchange of either G-actin and actin - DNase-I-bound ATP or F-actin-bound ADP for the nucleotide species of interest. The time course of such an experiment using actin- DNase-I complex is shown in Fig. 1. Here actin - DNase-I complex was freed from unbound nucleotide (ATP) by gel filtration over a Sephadex G-100 column immediately prior to the experiment and then incubated with 1 mM EDTA and 0.1 mM adenosine or the adenine nucleotides as indicated. Aliquots at given time intervals were analyzed for DNase I activity. It can be seen that ATP and ADP at 0.1 mM concentration completely maintain the actin in a state able to inhibit DNase 1. In contrast, in the presence of 0.1 mM AMP or adenosine, a time-dependent increase in the DNase I activity was observed. Identical results were obtained when G- or F-actin were incubated with these adenosine nucleotides in the presence of 1 mM EDTA for 1 h

Capacity to inhibit DNase I F-actin

G-actin

actin- DNase-1

100 0 0 95 92.5 35 100 100 35 15

100 0 0 96 100 94 100 100 33 88

%

0.1 mM 0.1 mM 0.1 mM 0.1 mM 0.5 mM 0.5 mM 0.5 mM 0.5 mM 0.5 mM 0.5 mM

ATP adenosine AMP ADP ATP[aS] ADP[BS] ATP[yS] ATP[B,yNH] ATP[j?,yCH2] Ap,A

100 0 0 78.5 95 94.7 84.6 92 93.3 19.5

at room temperature and then analyzed for their ability to inhibit DNase I We next analyzed whether the reactivation of DNase I activity of actin - DNase-I complex by EDTA was due to a separation of the stoichiometric complex into free actin and DNase I. To this end actin - DNase-I complex was analyzed by gel filtration over Sephadex G-150 equilibrated with Hepes buffer in the presence or absence of 1 mM EDTA. Prior to loading the complex on the column it had been incubated with [3H]ATPovernight at 4°C. Fractions collected were then analyzed for DNase I activity and radioactivity. The results obtained are shown in Fig.2. Under both buffer conditions, the protein elutes as a single peak at identical positions. Analysis of the protein peak by gel electrophoresis in the presence of SDS indicated the coelution of actin and DNase I for both buffer conditions (Fig.2c, d). In the absence of EDTA, the specific DNase I activity of the protein peak was found to be low (exhibiting a sigmoidal progression curve typical for actin-inhibited DNase I) and the associated radioactivity was high. In the presence of 1 mM EDTA the specific DNase I activity of the protein peak had increased almost 8-fold whereas the radioactivity associated with the protein peak had decreased by a factor of 6.5. In a separate experiment using unlabeled ATP under otherwise identical conditions, it was found that the specific DNase I activity of the actin - DNase-I complex increased from 27.3 KU/pg complex to 240 KUjpg after gel filtration in the presence of 1 mM EDTA (i. e. 8-fold). Since the elution profile of the actin - DNase-I complex is unchanged after addition of EDTA, the removal of divalent metal ions does not lead to separation of the complex into actin and DNase I. The arrow in Fig.2b marks the position at which free DNase I would

270

t

0

10

20

30

LO

fraction number

Fig.2. Gelfiltration qf (3H]ATP-uctin- DNase I complex in the presence and absence of EDTA. Actin-DNase I complex was passed over Sephadex G-75 equilibrated with Hepes pH 7.4 (a) or Hepes supplemented with 1 mM EDTA (b). Prior to the experiment G-actin was freed from unbound nucleotide by gel filtration over Sephadex G-100. Then DNase I was added at equimolar concentration and 5 mg of the thusformed actin - DNase-I complex in 2 ml was preincubated with ['HIATP (60 pCi) for 14 h at 4°C for each run. Fractions of 2 ml were collcctcd and analyzed for radioactivity by scintillation counting (0)and for DNase 1 activity ( 0 )which is expressed in Kunitz units (KU, see Fig. 1). The fractions of the first (protein) peak from the experiments of a and b were analyzed by electrophoresis on 10% polyacrylamide gels in the presence of sodium dodecyl sulfate (c. d). M = molecular mass markers, from top to bottom: bovine serum albumin (67 kDa), actin (42 kDa) and DNasc I(31 kDa)

have been eluted. Thus, under these conditions (i.e. in the absence of bound metal and nucleotides) an actin - DNase-I complex is formed which exhibits practically uninhibited DNase activity. We termed this behaviour 'desensitization of actin' [17]. For the observed desensitization of actin in the presence of AMP or adenosine two explanations are possible: (a) AMP or adenosine do not bind to actin thus allowing its denaturation in the presence of EDTA or (b) AMP or adenosine bind to actin - DNase-I complex, but a conformational change of actin renders it unable to inhibit DNase I. The two possibilities were distinguished experimentally as follows. Actin DNase-I complex previously freed of unbound nucleotide was treated with 0.1 mM ATP or AMP in the presence of 1 mM EDTA, supplemented with [3H]ATP or [3H]AMP, respectively. After a 30-min incubation at room temperature, the CaC12 concentration was raised to 1.2 mM and both samples were subjected to gel filtration over Sephadex G-75 equilibrated with Hepes buffer. The fractions collected were analyzed for protein and radioactivity. In both cases actin - DNase-I eluted as a single peak containing both actin and DNase I as verified by SDS/PAGE. Radioactivity, however, was only found to be associated with protein in the sample incubated with ATP (data not shown). Protective ejfect of adenosine nucleotide analogues

The experiments described so far indicate that the protective effect of ATP and ADP on actin when complexed with DNase 1is due to the maintenance of a nucleotide-actin complex even in the presence of EDTA. Several ATP analogues were also tested for their ability to protect actin against de-

sensitization in the presence of EDTA. From such experiments the ability of these analogues to interact with actin either in the G or F form or in complex with DNase I can be deduced. The results obtained are compiled in the Table 1. The degree of protection of the analogues at 0.1 mM is compared to ATP (complete protection) and adenosine (no protection). Most of the analogues used exhibited almost identical protective effects on all states of aggregation or complexation of actin. The reasons why ADP[PS], ATP[P,yCH,] and Ap,A show distinct protective effects is unclear at present. Kinetics of nucleotide interaction of actin - DNase-I complex Nucleotide dissociation. In an earlier communication it has been shown that the fluorescent ATP analogue EATP is released from actin considerably more slowly when in complex with DNase I [5]. From this result a higher affinity of ATP to actin in its complex with DNase I than for free actin was deduced [ S ] . The rate of FATP dissociation from the actin DNase-I complex was redetermined at 'physiological ionic strength' in two ways: (a) after addition of large excess of ATP (1.1 mM) to actin-DNase-I containing bound rATP or (b) by rapidly adding EDTA in excess over CaC12. The rate of EATP release after addition of 1.1 mM ATP was measured at pH 7.6. As expected a strong dependence of this rate on Ca2+ concentration is observed, as shown in Fig. 3 , ranging from 1 x s-' to 6 x s-' at high and low C a 2 + , respectively. The rate of dissociation by rapidly adding 1 mM EDTA was measured using the stopped-flow apparatus (see Methods). In view of the strong pH dependence of the nucleotide dissociation rate constant by EDTA [lS], the dissociation was measured at pH 8.3. As described for G-actin

271

I

I

0.1

0.2

I

0.3

I

0.L

I

0.5

ICaCI,] (mM1 Fig. 3. Dependence of' the rate of' EATP release ,from actin - DNase complex on total CaC12 concentration. Actin containing cATP as bound nucleotide was preincubated with DNase I for 1 h at room temperature. Release of EATPbound to actin was initiated by adding 1.1 mM ATP as displacing agent. Final conditions: 0.9 pM EATP, 0.1 M KCl, 0.8 mM MgC12, 4 mM Hepes pH 7.6, 20°C. Rate of EATP release ( k e x ) was determined at the CaC12 concentrations given by continously monitoring the rate of ATP fluorescence decrease using an SLM 8000 spectrofluorometer (see Materials and Methods). Different symbols represent three different actin - DNase-I complex preparations

6

.-

4

Ln I

0

so 2

[19], the time course of EATPdissociation at pH 8.3 showed the presence of more than one process. An apparent firstorder rate constant was calculated from the initial time course of fluorescence decrease and found to be 1.8 x l o p 2 s p l for actin-DNase-I complex. It was thus 8-fold lower than for uncomplexed G-actin (k = 1.37x10-'xs-') in the same conditions. Kinetics of nucleotide reussociation. After complexing divalent cations by EDTA, G-actin rapidly denatures and thus looses its ability to rebind nucleotide [8]. Nucleotide reassociation can therefore only be measured after rapidly adding (within the first 10 s) nucleotide and CaClz in excess over EDTA [8]. The multi-mixing set up (see Methods) was employed, by which, after the initial addition of 1 mM EDTA at pH 8.3, reassociation of EATPwas induced by addition of excess CaClz (and nucleotide) by a third syringe. Reassociation was determined at pH 7.6. In a first series of experiments

the rate of association of EATP was determined. Fig.4 illustrates the results obtained; here increasing EATP concentrations were added in the third syringe together with CaCl, after a 40-s incubation with 1 mM EDTA. Pseudo-first-order conditions were established by carrying out the experiments at high molar excess of EATPover actin - DNase-I (range of 2.5 - 25 pM). In Fig. 4 the observed pseudo-first-order rate constants are plotted as a function of free EATPconcentration. A straight line up to 25 pM EATP is obtained, whose slope gives the second-order rate constant for EATP association (k = 2.1 x lo5 M - ' s-'). Using identical conditions, competition experiments of ATP and ADP with EATP were performed to determine their rate of association. The simultaneous addition of ATP or ADP with EATP in the second mixing step resulted in an increased rate constant of cATP association to actin - DNase-I complex although with a decreased amplitude. The observed rate constants of EATPbind-

272

2.80

2.LO 2.00 ,.

0 1.60 x

-

1.20

i 0.80

1 3 6 9 0.LO

OO

IATPI or [ADPI(pM)

Fig. 5. Plots of the kinetics of’cATP rebinding to uctin- DNase-1 complex against A T P or A D P concentration. The measurements were performed as described in Fig.4 in the presence of either ADP ( A ) or ATP (0)at the final concentrations indicated

ing were plotted as a function of ATP or ADP (Fig.5). The relationships were found to be linear up to 10 pM nucleotide and the second-order rate constants for their association were calculated from the slopes, while their intercepts with the y axis give the rate constant of EATP association to actinDNase-I complex (for a detailed analysis see [19]). Values of 2.17 x 10’ M - l s - l and 2.3 x 10’ M - l were obtained for the second-order rate constants of association of ATP and ADP, respectively. They are thus about an order of magnitude lower than for G-actin under the ‘physiological conditions’ used in this analysis [20]. Furthermore, from the rate constants of association and dissociation of EATPits binding constant ( K d ) to actin - DNase-I complex is given by the ratio of kass/kdiss = 2.16 x lo8 M p l .

0 0

0.40

0.80

1.20

1.60 2.00 Time (s)

2.LO

2.80

10’

Fig. 6. Exrent of E A T P rebinding to actin - DNase-I complex after varying periods of preincubation with EDTA. EATP- G-actin at 0.88 pM and DNase I at 1.35 pM in 0.02 mM CaC1, and 4 mM Hepes pH 8.3, was preincubated with 1 mM EDTA for the time periods indicated. Then by the use of a third syringe (see Materials and Methods) MgC12,KCI, and CaCI, in 4 mM Hepes pH 7.6, was rapidly added to give a final buffer composition of 3 mM MgCI2, 0.1 M KCI, 0.5 mM EDTA, 1.16 mM CaC1, and Hepes pH 7.6. Total [EATP] during rebinding was 2.49 pM. The measurement of the amplitude of EATP fluorescence increase at 410 nm was triggered by the third syringe. The ordinate gives then amplitude of fluorescence increase in arbitrary units ( I ) and the abscissa the time of preincubation of EATP-actin-DNase-I with 1 mM EDTA in seconds. The curve shows the computer fit to the data assuming sequential nucleotide release and denaturation with the half-lives mentioned in the text

Rate of denaturution of cation-free actin when in complex wirh DNuse I Using the multimixing set up (see Methods), the amplitude of cATP rebinding to actin - DNase-I complex was analyzed after varying periods of preincubation with EDTA. The results are shown in Fig.6, which demonstrates a biphasic dependence of the amplitude of rebinding on the duration of preincubation with EDTA. Up to about 60 s a continuous increase of the fluorescence amplitude, i. e. EATPrebinding, is observed which thereafter continously decreases. Most probably the first phase is due to the retarded rate of EATP release from actin-DNase-I complex ( t l j 2 = 38 s) whereas the second phase is due to a denaturation process rendering the actin incapable of rebinding nucleotide. This process has a halftime of approximately 200 s. Thus the process of ‘denaturation’, i. e. loss of nucleotide-binding capacity, of cation-free actin is retarded by a factor of about 15 when complexed with DNase I ( t l j z for free actin = 13 s at pH 8.2) [S]. Although only a limited number of data points were collected, it appears that the denaturation process does not follow first-order kinetics. As will be shown later, the ‘denaturation’ of cationfree actin in complex with DNase I, i.e. the loss of its ability to inhibit DNase I, is dependent on nucleotide concentration (Fig. 7). Therefore the rate of this process is expected to depend o n the established equilibrium between nucleotide and cation-free actin.

0 10-6

10-5

10-3

10-4

(MI Fig. 7. Titration of the loss of DNase I inhibitory capacity of actin in uctin- DNase-I complex by different nucieotides. ActinDNase-I complex freed from nucleotide by gel filtration over Sephadex G-100 prior to the experiment was added at 1 0 p M final concentration to Hepes pH 7.4, supplemented with 1 mM EDTA and the nucleotides indicated at the concentrations given: ( A ) ATP; (0) ADP and ( 0 )ADP[flS]. After a 30-min incubation at room temperature, the DNase I activity of 25-pI aliquots was tested in 1 ml test solution using the hyperchromicity test

AjJinity of Ca2+ to uctin in complex with DNase I

Using ATP as displacing agent, a strong Ca2+dependence of ATP release from actin - DNase-I complex was shown [5]

213 '10 ADP 1 OB

L

0,s + C D 2.5 1.0 0.5

50

0

I

I

500

1000

1500

Time (5)

Fig. 8. Kinetics of Ca2+ release from its complexes with ATP-Guctin and ATP-actin- DNuse-I monitored by the increase in Quin 2 f2uorescence. Conditions: 6 mM Hepes pH 7.2,200 pM ATP, 200 pM MgC12, 148 pM Quin 2. The reaction was started by addition of 28.4 pM G-actin (0)or 14.7 pM actin-DNase-I ( 0 ) .Thesolid lines show tits to single exponential functions

(and this communication). In the previous communication [5] it was inferred that the affinity of C a 2 + to actin does not change after complexing with DNase I. Recently, the affinity of Ca2+ to G-actin has been reinvestigated using a direct approach and much higher values than previously assumed have been obtained [9]. We have also attempted to determine the affinity of Ca2+ to actin when in complex with DNase I. As described by Gershman et al. [9], actin - DNase-I complex was rapidly mixed with Quin2 and an excess of MgClz and the rate of Ca2+ release was determined by following the fluorescence increase of the Ca2'-chelator Quin 2. Fig. 8 gives a comparative experiment using free G-actin and actin - DNase-I complex. First-order rate constants of 1.57 x s-' and 2.07 x s - l can be computed for Ca2 release from free actin and actin - DNase-I complex, respectively. From titration curves of Quin 2 fluorescence of actin - DNase-I complex by MgCI2 a binding constant for Ca2+ of 3 x lo8 M - ' to actin-DNase-I complex could be computed. Aflinity of'nucleotide to cation-free actin - DNase-I complex

In a previous section it was demonstrated that cationfree actin when complexed with DNase I does not loose its DNase I inhibitory capacity, i. e. does not denature, provided the nucleotide concentration is sufficiently high. It appears, however, impossible to determine directly the rate constants of association of nucleotide in the presence of EDTA, since a continuous presence of high concentration of nucleotide (ATP) was found necessary to maintain actin in its native state. Therefore an equilibrium assay was employed to estimate the affinity of nucleotide to cation-free actin in complex with DNase 1. Fig.7 gives the dependence of the maintenance of the DNase 1 inhibitory ability of complexed actin on ATP, ADP and ADP[jS] concentration. In the presence of 1 mM EDTA the DNase I inhibitory capacity is titrable by these nucleotides and the 50% values of DNase I inhibition obtained are at 0.2 mM, 0.7 mM and 4 mM for ATP, ADP and ADP[jS], respectively. It thus appears that nucleotide binding to actin is reduced by nearly 5 orders of magnitude (ATP) when tightly bound cations are removed.

- --

a

0 0

1 4

8

12

16

20

24 h

Fig.9. Rate of A T P hydrolysis by actin - DNase-I complex at increasing MgC12 concentrations. Actin- DNase-1 complex at 20 pM was freed from unbound nucleotide by gel filtration and equilibrated with 40 pCi [3H]ATP (specific activity: 49 Ci/mmol) for 2 h at room temperature in 0.1 mM CaCI2, 0.5 mM NaN, and 5 mM Hepes pH 7.4. The reaction was started by addition of MgClz to the concentrations indicated on the right. Cytochalasin D (CD) concentration was 20 pM together with 0.5 mM MgCI2. At the time intervals given, aliquots were analyzed for the hydrolysis of actin-bound ATP to ADP as described in Materials and Methods

ATP hydrolysis of actin - DNase-I complex

In view of the discrepancies conccrning the rolc and the relative location of the binding sites of divalent cation and nucleotide to actin, we have tested the influence of cations on the hydrolysis of actin-bouind ATP of actin - DNase-I complex at concentrations higher than the critical concentration of actin polymerization. Recent reports had indicated that monomeric actin slowly hydrolyzes its bound ATP on incerasing the concentration of MgCI2 [21, 221. These results could also be corroborated for the actin- DNase-I complex. It was found that MgCI2 and KCl induce the hydrolysis of actin-bound ATP of actiii- DNase-I complex, whereas CaC12 had no effect. The rates of ATP hydrolysis at 0.5 mM MgCI2 and 0.1 M KC1 were found to be 0.8 and 0.13 mol ATP hydrolyzed (mol actin - DNase-I complex)- h- respectively. The rate of ATP hydrolysis was found to increase linearly with MgC12 concentration (see Fig. 9, which gives the time dependence of ATP hydrolysis of actin -DNase-I complex at different MgClz concentrations up to 2.5 mM). Addition of 20 pM cytochalasin D to the actin - DNase-I complex in 0.5 mM MgCI2 increased the rate of ATP hydrolysis by a factor of 20 to 1.5 mol ATP hydrolyzed (mol actin-DNaseI complex)-' h-' (Fig. 9). Fig. 10 shows the dependence of the rate of ATP hydrolysis on the Mg2+ concentration up to 5 mM which can be fitted by a hyperbolic function, suggesting saturation of one or more sites by metal ion. The dissociation constant obtained using a least-squares fit is 8.1 mM. Stimulation of the actin - DNase-I ATPase therefore appears to be regulated by the occupancy of low-affinity divalent-cationbinding sites.

' ',

274

-

s x ..->

c. 4-

u

0 H

aJ u1

0

Z

n Fig. 10. Dependence of the initial rates of ATP hydrolysis of actin DNuse-I c.omp1e.x on MgClz concentrations. Values were calculated from Fig.9, additionally including an experiment at 5 mM MgCI2 not shown in Fig.9

DISCUSSION Results presented in this communication define the reciprocal effects of actin and DNaseI on each other in the actin - DNase-I complex. The well known inhibitory effect of actin on the DNase I activity is only exhibited by the actinDNase-I complex if ATP or ADP (with or without divalent metal ion) is bound at the actin nucleotide-binding site. Removal of nucleotide results in almost complete restoration of the DNase I activity, but not in dissociation of the actinDNase-I complex. In the absence of bound nucleotide, the rate of denaturation of actin in this complex, as judged by its ability to rebind nucleotides, is less than that for actin alone, but still occurs with a half-life of several minutes at room temperature and pH 8.3. The results presented do not allow one to decide whether the loss of nucleotide alone or the subsequent denaturation of actin is responsible for the restoration of DNase I activity. It is of interest to note that actin which is denatured to the extent that nucleotide can no longer be bound can still bind DNase I. This suggests that nucleotidebinding ability is lost without complete denaturation of actin, an assumption which appears to be supported by the experiment depicted in Fig. 11. Here the DNase I inhibitory capacity of native G-actin, G-actin treated for 1 h with 1 mM EDTA, and heat-denatured (97°C) G-actin was tested. It can be seen that only heat-denatured actin has completely lost its DNase 1 inhibitory capacity, whereas it retains some inhibitory capacity after treatment with 1 mM EDTA. After EDTA treatment, higher concentration of actin (4-fold) are necessary to obtain 50% inhibition of DNase 1 activity. Thus the data in Fig. 11 indicate that the affinity of EDTA-treated G-actin to DNase 1 is reduced by a factor of about 4, i.e. from 5 x lo8 M - ' [5] to about 1.25 x 10' M - ' . This affinity is still high enough to secure tight association of both proteins during the gel filtration. Since actin has been shown to behave as an allosteric competitive inhibitor of substrate (DNA) binding to DNase I [5], the reduction in actin affinity to DNase I appears to result in a higher DNA-degrading activity of EDTA-treated actin- DNase-I complex. DNase I activity itself has been shown to depend on the presence of divalent cations, i. e. is inhibited by 1 mM EDTA. Its catalytic activity, however, is immediately and fully restored after readdition of divalent cations (Mg2+ and C a 2 + )1231. Thus we regard the observed loss of actin inhibitory capacity not to be due to a direct effect of EDTA on DNase I.

0

1D

0.5

2.0

1.5

25

[actin] (pM) Fig. 11 Inhibition of DNase I activity by G-uctin, G-actin treated with 1 mM EDTA for I h and G-actin heated at 97'C for 1 h. G-actin freed of unbound nucleotide by gel filtration was incubated at room temperature for 1 h in the absence ( 0 )or presence (m) of 1 mM EDTA or was heated at 97°C (A)for 1 h. The trcated actins were incubated in Hepes pH 7.4 at the concentrations given with 0.5 pM DNase I for I h at room temperature. DNase I activity was determined for 10 pl of the incubation mixtures using 1 ml test solution as described in Materials and Methods

Earlier work from our laboratories had led to the conclusion that the rate of nucleotide release from the actinDNase-I - nucleotide - metal-ion complex was about one order of magnitude slower than from the corresponding complex in the absence of DNase I [ 5 ] . In the meantime, it has been shown [19, 241 that it is the rate of metal ion release, not that of nucleotide, which is rate-limiting in nucleotide exchange reactions. By directly monitoring Ca2 release, we show here that this also holds for the actin-DNase-I complex, and that the rate of C a 2 +release is similar to the maximal rate of nucleotide exchange observed earlier for the ATP actin - DNase-I complex [S]. Thus the following reaction scheme describes nucleotide (N) dissociation from actin - DNase-I complex (AD): +

K,";' '

N - A D - c ~ ~I+_ N-AD N-AD AD

+ ca2

ZAD + N A*D

where K:&*+ is the dissociation constant of divalent cation from its complex with actin-DNase-I. For free G-actin it was argued that the high-affinity metal-binding site is probably not identical to the binding site directly associated with the nucleotide binding [19, 241. We assume this to be also true for actin - DNase-I complex. After nucleotide release the irreversible denaturation process leads to A*D. Under metal-free conditions, the value for the dissociation constant of ATP from actin - DNase-I complex (0.2 mM) is much higher than that obtained recently for ATP from its complex with actin alone (0.01 mM [24]). However, it is possible that the method used in the present work is relatively inaccurate, since the constant is not measured using a strict equilibrium system due to the probable partial denaturation of actin during the assay. Thus, it cannot be concluded that

215 metal-ion-free nucleotide binding is weaker in the actin DNase-I complex than for actin alone. ATP cleavage, which in G-actin or actin - DNase-I is very slow at low ionic strength and low divalent metal ion concentration, can be stimulated by addition of Mg2 to the actinDNase-I complex. It has recently been suggested that one of the so-called weakly bound magnesium ions is bound to ATP in the actin-ATP complex and that this leads to cleavage of ATP. The relationship between these binding sites and the low-affinity sites implicated in polymerization to F-actin is unclear. It should be mentioned that during polymerization, ATP cleavage appears to occur after incorporation of G-actin into the growing polymer [25]. In the case of actin - DNase I, it seems to be possible to separate these two consequences of Mg2+binding, which may allow a more detailed study of the cleavage reaction alone. It would be of interest to investigate whether the actin in the actin - DNase-I complex adopts Factin-like properties (e. g. conformation) after ATP hydrolysis, especially in view of the current three-dimensional structural analysis of the actin - DNase-1 complex [26]. The cytochalasins may directly transform G-actin - ATP into G-actin - ADP, which exhibits a higher critical concentration of polymerization [27]. In view of the reported high affinity of cytochalasins to polymeric forms of actin [28], this may be an additional mechanism that causes the decreased extent of actin polymerization in the presence of these drugs. In contrast to other models of cytochalasin action [29], this effect is directly exerted on monomeric actin. This could only unambigously be demonstrated by using the actin - DNase-I complex that fixes actin in a quasi-monomeric state even in the presence of high salt. However, locally restricted conformational changes of actin when complexed with DNase I cannot be excluded. In summary, complexing actin with DNase I appears only to decrease the rates of conformational transitions of actin, but not to freeze it in a particular configuration. Therefore the actin - DNase-I complex will be a valuable tool to study conformational transitions of actin independently of its polymerization. It is a pleasure to acknowledge the expert technical assistance of J. Koch and M. Isakov. Ewa Nowak acknowledges the support of the Alexander von Humboldt-Stijtung. The work was supported by the Deutsche Forschungsgemeinschafi and by the Polish Academy of Sciences as part of the project C.P.B.P. 04.01.

REFERENCES 1. Pollard, T. D. & Cooper, J. A. (1986) Annu. Rev. Biofhem. 55. 987 - 1035. 2. Lazarides, E. & Lindberg, U. (1974) Proc. Natl Acad. Sci. U S A 71,4742-4746. 3. Mannherz, H. G., Barrington Leigh, J., Leberman, R. & Pfrang, H. (1975) FEBS Lett. 60, 34-38. 4. Hitchcock, S. E., Carlsson, L. & Lindberg, U. (1976) Cell 7,531 543. 5. Mannhcrz, H. G., Goody, R. S., Konrad, M. &Nowak, E. (1980) Eur. J . Biochem. 104, 367 - 379. 6. Hitchcock, S. E. (1980) J . B i d . Chem. 255, 5668-5673. 7. Pinder, J. C. & Gratzer, W. B. (1982) Biochemistry 21, 48864890. 8. Waechtcr, F. & Engel, J. (1975) Eur. J . Biochem. 57, 453-459. 9. Gershman, L. C., Selden, L. A. & Estes, J. E. (1986) Biochem. Biophys. Rus. Commun. 135,607 - 634. 10. Kunitz, M. (1950) J . Gen. Pltysiol. 33, 349-362. 11. Laemmli, U. K. (1970) Nature (Lond.) 227, 680-685. 12. Secrist, J. A,, Barrio, J. R., Leonard, N. J. & Weber, G. (1972) Biochemistry 11, 3499-3506. 13. Goody, R. S. & Eckstein, F. (1970)J. Am. Chem. Soc. Y3.62526257. 14. Eckstein, F. &Goody, R. S. (1976) Biochemistry 15, 1685- 1691. 15. Yount, R. G., Babcock, D., Ballantyne, W. & Ojala, D. (1971) Biochemistry 10, 2484-2489. 16. Feldhaus, P., Frohlich, Th., Goody, R. S., Isakov. M. & Schirmer, H. (1976) Eur. J . Biuchem. 57, 197-204. 17. Dieckhoff, J . & Mannherz, H. G. (1985) Biochim. Biophys. Acta 829,209 - 220. 18. Wdechler, F. & Engel, J . (1977) Eur. J . Biochem. 74, 227-232. 19. Nowak, E., Strzelecka-Golaszewska, H. & Goody, R. S. (1988) Biochemistry 27, 1785- 1792. 20. Wanger, M. & Wegner, A. (1983) FEBS Lett. 162, 122-1 16. 21. Brenner, S. L. & Korn, E. D. (1980) J . Biol. Chem. 255. 841 844. 22. Brenner, S. L. & Korn, E. D. (1981) J . Biol. Chem. 256, 86638670. 23. Kreuder, V., Dieckhoff, J., Sittig, M. & Mannherz, H. G. (1984) Ew. J . Bioclwm. 139, 389 -400. 24. Nowak, E. & Goody, R. S. (1988) Biochemistry 27, 8613-8617. 25. Pantaloni, D., Carlier, M. F. & Korn, E. D. (1985) J . Biol. Chem. 260,6572-6578. 26. Kabsch, W., Mannherz, H. G. & Suck, D. (1985) EMBO J . 4 , 21 13 -21 18. 27. Wegner, A. (1976) J . Mol. Biol. 108, 139- 150. 28. Lin, D. C., Towbin, K. D., Grumet, M. & Lin, S. (1980) J. Ce// Biol. 84, 455-460. 29. Goddette, D. W. & Frieden, C. (1986) J . B i d . Chem. 216,1597415980.