Báo cáo khoa học: Specific cleavage of the DNase-I binding loop dramatically decreases the thermal stability of actin pot

Similar to intact F-actin, both cleaved F-actins were significantly stabilized by phalloidin and aluminum fluoride; however, in all cases, the thermal stability of the cleaved F-actins was

decreases the thermal stability of actin

Anastasia V Pivovarova1, Sofia Yu Khaitlina2and Dmitrii I Levitsky1,3

1 A N Bach Institute of Biochemistry, Russian Academy of Sciences, Moscow, Russia

2 Institute of Cytology, Russian Academy of Sciences, St Petersburg, Russia

3 A N Belozersky Institute of Physico-Chemical Biology, Moscow State University, Moscow, Russia

Introduction

Actin is one of the most abundant and highly conserved

cell proteins It is involved in many different cellular

processes that are essential for growth, differentiation

and motility Actin is found in two main states: as

monomers (G-actin) and as a helical polymer (F-actin)

Polymerization of G-actin into F-actin is accompanied

by hydrolysis of tightly bound ATP followed by a

slower release of Pi; as a result, protomers of F-actin

contain tightly bound ATP, ADP or ADP-Pi

The atomic resolution structures of G-actin revealed that it is divided into two easily distinguishable domains by a deep cleft containing the tightly bound nucleotide and cation [1] The nucleotide-binding cleft was suggested to exist in two main states, closed and open [2,3], and solution studies on nucleotide exchange and susceptibility of the cleft to limited proteolysis appear to be consistent with the opening of the cleft upon the transition from the ATP- to ADP-G-actin

Keywords

actin; differential scanning calorimetry;

DNase-I binding loop; proteolytic cleavage;

thermal unfolding

Correspondence

D I Levitsky, A N Bach Institute of

Biochemistry, Russian Academy of

Sciences, Leninsky Prospect 33, 119071

Moscow, Russia

Fax: +7 495 954 2732;

Tel: +7 495 952 1384

E-mail: levitsky@inbi.ras.ru.

(Received 10 June 2010, revised 14 July

2010, accepted 16 July 2010)

doi:10.1111/j.1742-4658.2010.07782.x

Differential scanning calorimetry was used to investigate the thermal unfolding of actin specifically cleaved within the DNaseI-binding loop between residues Met47-Gly48 or Gly42-Val43 by two bacterial proteases, subtilisin or ECP32⁄ grimelysin (ECP), respectively The results obtained show that both cleavages strongly decreased the thermal stability of mono-meric actin with either ATP or ADP as a bound nucleotide An even more pronounced difference in the thermal stability between the cleaved and intact actin was observed when both actins were polymerized into fila-ments Similar to intact F-actin, both cleaved F-actins were significantly stabilized by phalloidin and aluminum fluoride; however, in all cases, the thermal stability of the cleaved F-actins was much lower than that of intact F-actin, and the stability of ECP-cleaved F-actin was lower than that of subtilisin-cleaved F-actin These results confirm that the DNaseI-binding loop is involved in the stabilization of the actin structure, both in mono-mers and in the filament subunits, and suggest that the thermal stability of actin depends, at least partially, on the conformation of the nucleotide-binding cleft Moreover, an additional destabilization of the unstable cleaved actin upon ATP⁄ ADP replacement provides experimental evidence for the highly dynamic actin structure that cannot be simply open or closed, but rather should be considered as being able to adopt multiple conformations

Structured digital abstract

l MINT-7980274 : Actin (uniprotkb: P68135 ) and Actin (uniprotkb: P68135 ) bind ( MI:0407 ) by biophysical ( MI:0013 )

Abbreviations

D-loop, DNase I-binding loop; DSC, differential scanning calorimetry; DHcal,calorimetric enthalpy; Tm,thermal transition temperature.

[4–6] By contrast, in the numerous crystal structures

published to date, including those of ADP-G-actin

[7,8], the open nucleotide-binding cleft has been

observed only in profilin-bound actin crystals [9] Only

the closed conformation was also revealed by

molecu-lar dynamics simulations of the crystal actin structures

[10–12], whereas metadynamics simulations have

dem-onstrated that the closed conformation of the

nucleo-tide-binding cleft is the most stable state only when

ATP is bound, and the ADP-bound state favors a

more open conformation of the cleft [13] It is possible

therefore that crystallization favors a closed state for

G-actin even though the state of the cleft in solution

may vary [14,15]

The other nucleotide-dependent conformational

transitions in actin crystals involve so-called DNase

I-binding loop (D-loop) in subdomain 2 of the actin

molecule [7] In tetramethylrhodamine-modified

ADP-G-actin, the D-loop was found to be in an a-helical

conformation, whereas this region was disordered in

the ATP-bound actin, suggesting that the D-loop folds

on ATP hydrolysis [7,16] This transition within the

D-loop was supported by the results of metadynamic

simulations demonstrating a distinct allosteric

relation-ship between the conformation of the D-loop (ordered

or disordered) and the state of the nucleotide-binding

cleft (open or closed) [13] These data are consistent

with the results of the biochemical observations

indi-cating that proteolytic modifications of the D-loop

affect the state of the interdomain cleft [5,17]

The D-loop of actin can be specifically cleaved with

two bacterial proteases One of them is subtilisin,

which cleaves the D-loop between Met47 and Gly48

[18] The other protease, which specifically cleaves

actin at the only site between Gly42 and Val43

(Fig 1), was initially isolated and characterized as a

minor protein of lactose-negative Escherichia coli A2 strain and referred to as protease ECP32 [19,20] More recently, it was found that the A2 strain producing protease ECP32 is identical to Serratia grimesii, and therefore this enzyme was named grimelysin [21] Although ECP32 and grimelysin were suggested to be identical enzymes [21], both names are used in the lit-erature In accordance with previous studies [5,17], we refer to this protease as ECP in the present study The nucleotide-binding cleft in both subtilisin-cleaved and ECP-cleaved G-actin is clearly in a more open confor-mation compared to intact actin, as demonstrated by the increased nucleotide exchange rate in solution [17,22] and their higher susceptibility to limited prote-olysis [5,17] By contrast, the crystal structure of ECP-cleaved G-actin showed the nucleotide-binding cleft to

be in a typical closed conformation, probably as a result of crystallization preferentially trapping actin in only one of its possible conformations [14]

Taking into account the ambiguity of the nucleo-tide-binding cleft conformation and its relationship with the D-loop, the present study aimed to determine whether the specific cleavage of the D-loop affects the structural properties of the entire actin molecule and,

in particular, conformational transitions of the nucleo-tide-binding cleft For this purpose, we studied the effects of the D-loop cleavage on the thermal ing of G- and F-actin Previously, the thermal unfold-ing of G-actin containunfold-ing different nucleotides was indirectly studied with the DNase-I inhibition assay [23] and by monitoring the change in absorbance of tetramethylrhodamine-actin [24] The results obtained showed that replacement of the tightly bound ATP by ADP led to a significant decrease in the thermal stabil-ity of G-actin [23,24] In the present study, we applied differential scanning calorimetry (DSC), which is the most direct and effective method for studying the ther-mal unfolding of proteins Previous studies have shown that DSC can be successfully used to reveal the changes in the thermal unfolding of actin induced by interaction of G-actin with actin-binding proteins [25– 27], G–F transformation of actin, and stabilization of F-actin by phalloidin and Pi analogs [28] Moreover, the effects of nucleotides on the thermal unfolding of F-actin have been studied by this method and a ‘disso-ciative’ mechanism for the thermal denaturation of F-actin has been proposed [28,29]

In the present study, we used DSC to characterize the thermal unfolding of actin specifically cleaved within the D-loop by ECP or subtilisin The results obtained show for the first time that the cleavage strongly decreases the thermal stability of G-actin and especially that of F-actin, both in the absence and the

Fig 1 3D atomic structure of G-actin The four subdomains are

indicated by the encircled numbers Arrows show the cleavage

sites in the D-loop, between Gly42 and Val43 by ECP32⁄ grimelysin,

and between Met47 and Gly48 by subtilisin.

presence of phalloidin and Pi analogs These results

are discussed with regard to a more open

conforma-tion of the interdomain cleft in the cleaved actin

com-pared to intact actin, both in the monomers and in the

filament subunits

Results

Effects of the D-loop cleavage on the thermal

unfolding of G-actin

The excess heat capacity curves obtained for intact,

ECP-cleaved, and subtilisin-cleaved ATP-Ca-G-actins

are presented inFig 2 It is seen that the G-actin

spe-cies cleaved within the D-loop are clearly less

thermo-stable than noncleaved G-actin The thermal

transitions of both ECP-cleaved and subtilisin-cleaved

G-actin are shifted to a lower temperature, by 4–5C,

compared to that of intact G-actin, and the values

of calorimetric enthalpy, DHcal, determined for the

cleaved G-actins are  57–63% of those for

nonc-leaved G-actin (Table 1) Thus, both cleavages within

the D-loop strongly decrease the thermal stability of

ATP-Ca-G-actin, with no significant difference

between the effects of ECP and subtilisin It is

important to note that heating cleaved G-actins in the

calorimeter cell did not lead to any further proteolysis

of the proteins (Fig 2, inset)

We also compared the thermal unfolding of intact and ECP-cleaved G-actins in the different states, with the tightly bound Ca2+ replaced by Mg2+ and with the tightly bound ATP replaced by ADP (Fig 3) The replacement of Ca2+by Mg2+in ATP-G-actin had no appreciable effect on the thermal unfolding of either intact or ECP-cleaved G-actin: in both cases, it only slightly decreased the maximum thermal transition temperature (Tm), by 1–2C, with no effect on the

DHcalvalue (Table 1)

By contrast, the replacement of bound ATP by ADP caused a dramatic decrease in the thermal stabil-ity of G-actin Intact ADP-Mg-G-actin demonstrated the thermal transition with Tm of 48.8C (Fig 3A) (i.e 11C less than that of ATP-Mg-G-actin) and its calorimetric enthalpy (340 kJÆmol)1) was much less than that of ATP-Mg-G-actin (570 kJÆmol)1) (Table 1) Figure 3A shows that the sample contains only ADP-actin because no peak at 60C (correspond-ing to the thermal transition of ATP-Mg-actin) was seen on the thermogram A similar effect was observed

on ECP-cleaved Mg-G-actin with ATP replaced by ADP (Fig 3B) In this case, the nucleotide replace-ment decreased the Tmby 9C and led to a more than two-fold decrease in the DHcalvalue (Table 1)

Thermal unfolding of F-actin with the cleaved D-loop

Previous studies have shown that ECP-cleaved actin is unable to polymerize unless its tightly bound Ca2+ is replaced with Mg2+, and that the Mg2+-bound form has higher critical concentration and polymerizes more slowly than Mg-G-actin cleaved with subtilisin [17,20,30] In agreement with these data, in the present study, ECP-cleaved Mg-G-actin polymerized more

Fig 2 Temperature dependences of the excess heat capacity (C p )

of intact (curve 1), ECP-cleaved (curve 2) and subtilisin-cleaved

(curve 3) ATP-Ca-G-actins The actin concentration was 24 l M

Other conditions: 2 m M Hepes (pH 7.6), 0.2 m M CaCl2and 0.2 m M

ATP The inset shows representative SDS ⁄ PAGE patterns of intact

(lanes 1 and 1¢), ECP-cleaved (lanes 2 and 2¢) and subtilisin-cleaved

(lanes 3 and 3¢) G-actin before (lanes 1, 2 and 3) and after heating

in the calorimetric cell up to 80 C (lanes 1¢, 2¢ and 3¢) Note that

the positions of actin (lanes 1 and 1¢) and its C-terminal fragments

produced by ECP (36 kDa) (lanes 2, 2¢) or by subtilisin (35 kDa)

(lanes 3 and 3¢) remain unchanged after the heating–cooling

procedure.

Table 1 Calorimetric parameters obtained from the DSC data for intact, ECP-cleaved and subtilisin-cleaved G-actins The parameters were extracted from Figs 2 and 3 The error of the given values of

Tmdid not exceed ±0.2 C The relative error of the given values of

DH cal did not exceed ±10%.

G-actin Nucleotide Cation T m (C) DH cal (kJÆmol)1)

slowly than subtilisin-cleaved Mg-G-actin, which, in

turn, demonstrated slower polymerization than intact,

noncleaved Mg-G-actin Nevertheless, light-scattering

measurements showed complete polymerization of all

the Mg-G-actin species to Mg-F-actin after 1.5 h of

incubation with 100 mm KCl and 1 mm MgCl2 in the

presence of 1 mm ATP (Fig 4, inset)

Figure 4 shows that Mg-F-actin obtained from the

cleaved Mg-G-actin is much less thermostable than

noncleaved Mg-F-actin, and a decrease in the thermal

stability is even more pronounced than in the case of

G-actin The thermal transitions of ECP-cleaved and

subtilisin-cleaved F-actin are shifted to a lower

temper-ature, by 11.3 and 8.8C, respectively, compared to

that of intact F-actin (Table 2) Importantly, a

pro-nounced difference is observed between the thermal

transitions of ECP-cleaved and subtilisin-cleaved

F-actin (Fig 4) ECP-cleaved F-actin unfolds not only

at lower temperature (58.6 versus 61.1C), but also with a much lower cooperativity The width at the half-height of the thermal transition, which can serve

as a relative measure for cooperativity of the transi-tion, was equal to 8.5 C for ECP-cleaved Mg-F-actin and 4.3C for subtilisin-cleaved Mg-F-actin Thus,

Fig 4 DSC curves of Mg-F-actin assembled from intact (curve 1), ECP-cleaved (curve 2) and subtilisin-cleaved (curve 3) ATP-Mg-G-actin The actin concentration was 24 l M Other conditions: 20 m M

Hepes (pH 7.3), 0.1 M KCl, 1 m M MgCl 2 and 0.7 m M ADP The inset shows time courses of polymerization of intact (curve 1), ECP-cleaved (curve 2) and subtilisin-cleaved (curve 3) actins Poly-merization was monitored by recording light-scattering intensity at

350 nm upon the addition of 0.1 M KCl and 1 m M MgCl 2 to ATP-Mg-G-actins.

Fig 3 DSC curves of intact G-actin (A) and ECP-cleaved G-actin

(B) with different tightly bound nucleotide and cation:

ATP-Ca-G-actin, ATP-Mg-G-actin and ADP-Mg-G-actin The actin concentration

was 24 l M Other conditions: 2 m M Hepes (pH 7.6), 0.2 m M CaCl2

or MgCl 2 , and 0.2 m M ATP or ADP.

Table 2 Calorimetric parameters obtained from the DSC data for Mg-F-actin assembled from intact, ECP-cleaved and subtilisin-cleaved Mg-G-actin The parameters were extracted from Figs 4 and 5 The error of the given values of Tmdid not exceed ± 0.2 C The relative error of the given values of DH cal did not exceed

±10%.

Mg-F-actin Stabilizer Tm(C) DH cal (kJÆmol)1)

ECP-cleaved Phalloidin + AlF 4 81.7 690

Subtilisin-cleaved Phalloidin + AlF 4 84.4 780

although both cleaved G-actins unfold similarly

(Fig 2), a pronounced difference in the thermal

unfolding between ECP-cleaved and subtilisin-cleaved

actin is revealed when these actins are polymerized into

filaments

Stabilization of the cleaved F-actin by phalloidin

and aluminum fluoride

It is well known that cyclic heptapeptide phalloidin

binds to F-actin with very high affinity at the interface

of three adjacent actin protomers [31,32] and stabilizes

actin filaments (i.e it significantly increases the thermal

stability of F-actin) [25,28,33–35] A very similar

stabi-lizing effect was observed in the presence of Pianalogs,

aluminum fluoride (AlF4) or beryllium fluoride (BeFx)

[25,28,34,36], which form complexes with F-actin

subunits that mimic their ADP-Pistate The stabilizing

effects of phalloidin and AlF4 (or BeFx) were similar

but independent of each other because simultaneous

addition of both stabilizers caused an additional

increase in the thermal stability of F-actin [28,34]

The subsequent experiments were designed to

inves-tigate the effects of the two F-actin stabilizers,

phalloi-din and AlF4 , on the thermal unfolding of F-actin

specifically cleaved within the D-loop In agreement

with previous studies [25,28,29,34], the binding of

phalloidin or AlF4 significantly increased the thermal

stability of Mg-F-actin Both stabilizers shifted the

maximum of the F-actin thermal transition from

69.9C to 82–83 C (Table 2), and their simultaneous

addition increased the Tm up to  91 C (Fig 5A)

Similar to intact F-actin, both cleaved F-actin species

are significantly stabilized by phalloidin and AlF4

(Fig 5B,C) Each of these stabilizers increased the Tm

of the cleaved F-actin, by 10–12C for ECP-cleaved

F-actin and by  15 C for subtilisin-cleaved F-actin

(Table 2), and their simultaneous addition resulted in

an additive effect that is expressed in the further

increase of the Tm value by  12–13 C (Fig 5B) or

8C (Fig 5C) However, in all these stabilized states,

the Tm value for the cleaved F-actin was significantly

lower than that of intact F-actin, by 9–14C for

ECP-cleaved F-actin and by 6–7C for subtilisin-cleaved

F-actin (Table 2) This means that ECP-cleaved F-actin

is less thermostable than subtilisin-cleaved F-actin not

only in the absence of stabilizers (Fig 4), but also in

the presence of phalloidin and AlF4 (Fig 5B,C)

There are also other distinct differences between

ECP-cleaved F-actin and subtilisin-cleaved F-actin,

whose thermal denaturation is more similar to that of

intact F-actin First, along with the main transition at

68.5C, the DSC profile of the phalloidin-stabilized

ECP-cleaved F-actin demonstrated a pronounced shoulder at 60 C (Fig 5B) Second, in the presence

of AlF4 , this cleaved F-actin demonstrated, along with the main thermal transition at 70C, a clear peak at

 57 C corresponding to the thermal unfolding of this protein in the absence of AlF4 (Fig 5B) This suggests

a much lower affinity of ECP-cleaved F-actin for phalloidin and AlF4 than that in intact and subtilisin-cleaved F-actin To test this assumption, we investi-gated the thermal unfolding of ECP-cleaved F-actin in the presence of different concentrations of phalloidin and AlF4 (Fig 6)

At relatively low phalloidin⁄ actin molar ratio of

1 : 4, two peaks are observed on the DSC profile

Fig 5 DSC curves of intact actin (A), ECP-cleaved Mg-F-actin (B) and subtilisin-cleaved Mg-F-Mg-F-actin (C) stabilized by phalloidin

or AlF 4 , or simultaneously by both stabilizers Concentrations of stabilizers: 24 l M phalloidin and 1 m M AlF 4 (5 m M NaF and 1 m M

AlCl 3 ) Other conditions were as described in Fig 4.

(Fig 6A), and the large peak with Tmat 58.6C

corre-sponds to the nonstabilized ECP-cleaved F-actin (i.e it

reflects the thermal unfolding of those actin protomers,

which are not affected by phalloidin) This means that

effect of phalloidin on the thermal stability of

ECP-cleaved F-actin is much less cooperative than in the

case of intact F-actin, when one bound phalloidin was

shown to stabilize up to seven neighboring protomers

in the actin filament [37] The peak of nonstabilized

actin disappeared with an increase in the

phalloi-din⁄ actin molar ratio (Figs 5B and 6A) However, the

pronounced shoulder at  61–65 C was observed on

the DSC profile of ECP-cleaved F-actin even in the

presence of a three-fold molar excess of phalloidin

(Fig 6A), thus suggesting that protomers of

phalloi-din-stabilized F-actin exist in two structural states with

different thermal stability

At a low concentration of AlF4 (0.1 mm), we again

observed a pronounced peak at 58.6C corresponding

to the nonstabilized ECP-cleaved F-actin (Fig 6B)

Thus, a much higher concentration of AlF4 (more than 1 mm) is required to achieve complete thermal stabilization of ECP-cleaved F-actin compared to intact F-actin, for which full stabilization was observed even in the presence of 50 lm AlF4 [36] This reflects

at least an order of magnitude lower affinity of the cleaved F-actin to AlF4

Importantly, upon simultaneous addition of AlF4 and phalloidin, we observed neither the peak of nonstabilized actin protomers, nor the shoulder char-acteristic of phalloidin-stabilized ECP-cleaved F-actin (Fig 6B) These results suggest that the binding of AlF4 to ADP-F-actin substantially modifies the struc-tural state of cleaved actin subunits stabilized by phal-loidin or phalphal-loidin increases the affinity of the cleaved actin subunits to AlF4

Discussion

The data reported in the present study show that cleavage of actin between Gly42-Val43 or Met47-Gly48 within the D-loop strongly decreases the thermal stability both of monomers and polymers According to previous studies, these cleavages increased the rate of the nucleotide exchange on the cleaved G-actin and its susceptibility to limited proteolysis, probably as a result

of the transition of the nucleotide-binding cleft to a more open conformation [5,17,22] The relationship between the conformation of the D-loop and the nucleo-tide-binding cleft was recently demonstrated in metady-namic simulations experiments [13] We assume therefore that the decrease in the thermal stability observed by DSC on actin species cleaved within the D-loop is associated with opening of the cleft

Does the thermal stability of G-actin reflect the conformational state of the nucleotide-binding cleft?

An intact actin structure is maintained by the presence

of high-affinity cation and nucleotide tightly bound in the interdomain cleft; removal of the nucleotide or cat-ion results in actin denaturatcat-ion Therefore, the stabil-ity of actin depends on the affinstabil-ity of the tightly bound cation and nucleotide that involves both pro-tein–ligand interaction and conformation of the inter-domain cleft Upon heating, irreversible unfolding of G-actin is preceded by reversible loss of the nucleo-tide–cation complex [23] Obviously, the more tightly nucleotide and cation are bound in the interdomain cleft and the more ‘closed’ is the cleft, the higher the temperature needed to remove them from the cleft and

to induce thermal unfolding of G-actin The relative

Fig 6 DSC curves for ECP-cleaved Mg-F-actin (24 l M ) either in

the presence of phalloidin (Ph) at different concentrations (6, 12 or

72 l M ) (A), or in the presence of 0.1 m M AlF 4 in the absence or in

the presence of 24 l M Ph, and in the presence of 0.5 m M

AlF 4 + 24 l M Ph (B) Other conditions were as described in Fig 4.

affinity of G-actin for ATP is much higher than for

ADP [4], and the interdomain cleft is suggested to be

in a more open conformation in the ADP-bound state

than in the ATP-bound state [5,12,13] In agreement

with this and with previous studies [23,24], the results

obtained in the present study show that ADP-G-actin

is much less thermostable than ATP-G-actin (Fig 3A

and Table 1) ATP-G-actin is less thermostable with

bound Mg2+ than with Ca2+ [23] (Table 1), and this

reduction in stability may be explained by the lower

affinity of ATP-G-actin for Mg2+ than for Ca2+

[4,38] Thus, the ligand-dependent thermal stability of

actin monomer can be accounted for by the different

affinity of these ligands to actin However, the thermal

stability of the cleaved actins is lower than the

corre-sponding stability of non-modified actin both in the

ATP- and ADP-states This cannot be explained by

the different affinity but suggests that the thermal

sta-bility of G-actin may depend on the conformation of

the nucleotide-binding cleft This suggestion is

sup-ported by the studies on the effects of actin-binding

proteins on actin structure

Actin-binding proteins profilin and cofilin, when

bound to G-actin between subdomains 1 and 3, have

antagonistic effects on the conformation of the

nucleo-tide-binding cleft Profilin stabilizes the ‘open’

confor-mation of the cleft [7,39,40], whereas cofilin appears to

lock the cleft in its ‘closed’ conformation [39–42]

Pre-vious studies on the thermal unfolding of G-actin

showed that profilin binding decreased the actin

ther-mal stability [23], whereas significant stabilization of

G-actin was observed in its complexes with cofilin

[25,26] Stabilization of G-actin was also observed in

the complexes of G-actin with thymosin b4 [27] and

gelsolin segment 1 [24], which appear to induce

confor-mational transitions closing the nucleotide-binding

cleft [6,27,43,44]

Thus, the increased thermal stability of G-actin

appears to correspond to the closed conformation of

the nucleotide-binding cleft, whereas the decreased

thermal stability is a feature of the actin with the open

cleft conformation The cleavage within the D-loop

enhances the nucleotide exchange [17] and increases

accessibility of the cleft to limited proteolysis [5], which

characterizes the cleft opening It is therefore likely

that the decreased thermal stability of G-actin cleaved

within the D-loop also results from the opening of the

nucleotide-binding cleft in these actin species

It is noteworthy that the replacement of tightly

bound ATP by ADP in ECP-cleaved G-actin induces

an additional decrease in the thermal stability of this

actin species already destabilized by the cleavage

within the D-loop (Fig 3B and Table 1) This suggests

that the nucleotide-binding cleft is highly dynamic and cannot be simply open or closed but rather should be considered as being more open or more closed In these terms, by analogy with the ‘superclosed’ state recently revealed in ATP-G-actin by molecular dynam-ics simulations [12], the nucleotide-binding cleft of ADP-G-actin cleaved within the D-loop appears to adopt the extra open conformation

Comparison of the effects produced by the cleavage of the D-loop with ECP and subtilisin Although the cleavages of the D-loop between Gly42-Val43 and Met47-Gly48 decreased the thermal stability

of G-actin to a similar extent (Fig 2), the effects of the cleavages became quite different when the cleaved actins were polymerized into filaments The thermal stability of F-actin assembled from ECP-cleaved actin was noticeably less than that of subtilisin-cleaved F-actin (Fig 4) These results are consistent with the earlier observed effects of these cleavages on the sus-ceptibility of the nucleotide-binding cleft to limited proteolysis with trypsin [5,17] In the cleaved G-actins, susceptibility of trypsin cleavage sites at Arg62 and Lys68 in the nucleotide-binding cleft was increased similarly [17] After polymerization, these sites became almost inaccessible for trypsin in intact F-actin and only slightly accessible for trypsin in subtilisin-cleaved F-actin By contrast, F-actin assembled from ECP-cleaved G-actin was easily fragmented by trypsin These observations indicate that the open conforma-tion of ECP-cleaved actin was preserved upon poly-merization, whereas F-actin assembled from subtilisin-cleaved monomers more closely resembled intact F-actin than ECP-cleaved F-actin [17] Thus, the lower thermal stability of ECP-cleaved versus subtilisin-cleaved F-actin corresponds to a more open nucleo-tide-binding cleft

According to the recent model of actin filament [45], the N-terminal part of the D-loop is located at the inter-monomer interface, participating both in the intra-strand contacts between actin subunits along the filament and in the lateral contacts stabilizing the inter-strand interaction, whereas the C-terminal part of the loop is not involved in the inter-strand contacts Recently, this structural difference was supported in mutational cross-linking experiments showing that the N-terminal part of the D-loop (residues 41–45) is in close proximity to residue 265 of the actin subunit in the opposite strand and can be easily cross-linked to this residue, whereas the rate and extent of the cross-linking reaction strongly declined for the C-terminal residues of the D-loop [46] Therefore, the inter-strand

contacts of the N-terminal part of the D-loop appear

to play a crucial role in stabilization of the actin

fila-ment [17,30,47] The cleavage of the D-loop between

Gly42 and Val43 impairs these contacts [47], and this

may explain why the cleavage of the D-loop in its

N-terminal part with ECP more strongly destabilizes

F-actin than cleavage by subtilisin between Met47 and

Gly48 in the C-terminal part of the loop It is also

important that the cleavage with ECP did not affect

the filament length but more strongly enhanced the

turnover rate of polymer subunits than the cleavage

with subtilisin [17,30] Thus, the low thermal stability

of F-actin assembled from ECP-cleaved monomers

strongly correlates with the high dynamics of this actin

species [17], supporting the idea of the monomer

disso-ciation being the first step of thermal inactivation of

F-actin [28,29]

Although both cleaved actins are stabilized with

phalloidin and AlF4 , stabilization of ECP-cleaved

F-actin demonstrates specific features that are not

characteristic of subtilisin-cleaved or intact actin The

most interesting features are the extremely low affinity

of ECP-cleaved F-actin to AlF4 and the pronounced

shoulder observed on the DSC profile of this actin

spe-cies even in the presence of a three-fold molar excess

of phalloidin (Figs 5B and 6) This suggests that

pro-tomers of phalloidin-stabilized cleaved F-actin exist in

two different structural states Phalloidin binds to

F-actin at the interface of three adjacent actin

protom-ers [31] and appears to stabilize actin filament in two

inter-related ways: by stabilizing lateral interactions

between the two filament strands and by inducing

con-formational changes in actin subunits resulting in the

state of the nucleotide-binding cleft being similar to

that in ATP-actin filaments without phalloidin [32] It

is plausible that the shoulder on the DSC profile of

phalloidin-stabilized ECP-cleaved F-actin belongs to a

population of the protomers in which the

conforma-tional effect of phalloidin is not completed This

expla-nation, although requiring further examination with

independent approaches, is supported by the

disap-pearance of the shoulder after the addition of AlF4

(Fig 6B) This Pi analog (as well as another analog,

BeFx) is known to bind to Pi site in the

nucleotide-binding cleft and mimic ADP-Pi or ATP actin

fila-ments [48], thus stabilizing the filament by closing the

cleft in actin subunits [49,50] Hence, the increase in

the thermal stability of ECP-cleaved F-actin and the

disappearance of the shoulder on the DSC profile can

be accounted for by the combined effect of phalloidin

and AlF4 on the nucleotide-containing cleft

Accord-ingly, the phalloidin-induced effect may increase the

affinity of AlF4 to actin, whereas AlF4 -induced

clo-sure of the cleft diminishes the population of the su-bunits remaining nonstabilized by phalloidin via its effect on the cleft conformation This interpretation is consistent with recently published DSC data showing that cooperative effect of phalloidin on the thermal stability of F-actin becomes noncooperative in the presence of AlF4 [51]

Phalloidin can stabilize F-actin with a very high coo-perativity, with the half-maximal effect being observed

at a phalloidin⁄ actin molar ratio of 1 : 20 [52] In the DSC experiments on intact actin [37], only 10–15% of actin protomers remained unaffected by phalloidin at a phalloidin⁄ actin molar ratio of 1 : 4 By contrast, more than half of subunits of ECP-cleaved F-actin remained nonstabilized by phalloidin under the same conditions (Fig 6A), consistent with a reduced cooperativity in the effect of phalloidin on the steady-state ATPase activity of ECP-cleaved actin [17] Taken together with the evidence concerning the critical role of the lateral contacts for stabilization of filaments assembled from ECP-actin monomers [47], these data allow us to assume that only the effect of phalloidin on the con-formation of the nucleotide-binding cleft is coopera-tive; it is propagated along the filament by allosteric interactions between phalloidin-bound and free pro-tomers By contrast, the stabilizing effect of phalloidin

on the lateral inter-strand interactions is noncoopera-tive; it requires direct binding of phalloidin to actin protomers

According to this interpretation, an explanation for the appearance of the pronounced shoulder on the DSC profile of the ECP-cleaved F-actin stabilized by phalloidin (Fig 6A) can be proposed This shoulder appears to reflect the thermal unfolding of the actin protomers whose cleft remains open, and therefore they are stabilized only by lateral inter-strand interac-tions induced by the direct binding of phalloidin On the other hand, the main transition at 68.5C (Fig 6A) most likely corresponds to the thermal unfolding of actin subunits that are stabilized not only

by the inter-strand interactions, but also by phalloidin-induced closing of the nucleotide-binding cleft

In conclusion, the results obtained in the present study suggest that the thermal stability of actin, regard-less of whether it is modified by limited proteolysis or

by stabilizers, depends on the conformation of the interdomain nucleotide-binding cleft Accordingly, the lower thermal stability of subtilisin- or ECP-cleaved actin compared to intact actin supports the idea [5,13] and also provides additional experimental evidence for

a distinct allosteric relationship between conformation

of the D-loop and the state of the nucleotide-binding cleft

Experimental procedures

Reagents

Subtilisin (type VIII bacterial protease), ATP, ADP,

EGTA, Hepes, phenylmethylsulfonyl fluoride, KCl, CaCl2,

MgCl2, AlCl3, NaF and phalloidin were purchased from

Sigma Chemical Co (St Louis, MO, USA); hexokinase was

kindly provided by Dr N Yu Goncharova (Department of

Biochemistry, School of Biology, Moscow State University,

Russia)

Protein preparations

Rabbit skeletal muscle actin was prepared from

acetone-dried muscle powder according to the method of Spudich

and Watt [53] G-actin was stored in buffer containing

The actin molar concentration was determined by

monitor-ing A290 using an E1% of 6.3 cm)1 [54] and a molecular

mass of 42.3 kDa ECP-cleaved G-actin was obtained as

described previously [17,30] Ca-G-actin (3.0 mgÆmL)1) was

digested at an enzyme⁄ protein mass ratio of  1 : 100 for

cleaved with ECP between Gly42 and Val43 is fairly

resis-tant to further proteolysis by this protease, it was not

nec-essary to use any protease inhibitor The cleaved actin was

used within 8–10 h Subtilisin-cleaved actin was prepared

essentially as described by Schwyter et al [18] Ca-G-actin

(3 mgÆmL)1) was digested for 1 h at an enzyme⁄ protein

stopped with 2 mm phenylmethylsulfonyl fluoride The

[55] Usually, more than 85% of actin was cleaved It is

important that the main part of the noncleaved actin

appears to correspond to small aggregates of unfolded

(so-called ‘inactivated’) G-actin [56], in which the D-loop

becomes almost inaccessible to proteolytic cleavage [57]

ATP-Ca-G-actin was transformed into ATP-Mg-G-actin

actin-bound ATP was converted into ADP by incubation of

ATP-Mg-G-actin with 0.8 mm ADP, 1 mm glucose and

hexokinase (8 UÆmL)1) for 2 h at 4C [5] It is known that,

deter-mined in the actin samples after 1 h of incubation with

glu-cose and hexokinase [58]

Intact, ECP-cleaved and subtilisin-cleaved Mg-G-actins

(3 mgÆmL)1) were polymerized by the addition of 100 mm

Poly-merization was monitored by an increase in intensity of

light scattering at 90 measured at 350 nm on a Cary

Eclipse fluorescence spectrophotometer (Varian Australia

Pty Ltd, Mulgrave, Victoria, Australia)

Stabilization of F-actin (24 lm) by phalloidin or by alu-minum fluoride (AlF4 ) was performed as described previ-ously [25,28], by the addition of 6–72 lm phalloidin or 0.1–

ADP

DSC

DSC experiments were performed on a DASM-4M differen-tial scanning microcalorimeter (Institute for Biological Instrumentation, Pushchino, Russia) as described previously [25,28,29,36] All measurements were carried out at a scan-ning rate of 1 KÆmin)1 The experiments with G-actin were performed in 2 mm Hepes, pH 7.6, containing 0.2 mm CaCl2

ADP-Mg-G-actin), whereas the thermal unfolding of F-actin was studied in 20 mm Hepes (pH 7.3), 0.1 m KCl, 1 mm

was 24 lm The reversibility of the thermal transitions was assessed by reheating of the sample immediately after cool-ing from the previous scan The thermal denaturation of all actin samples was fully irreversible Calorimetric traces were corrected for instrumental background and possible aggre-gation artifacts by subtracting the scans obtained from the reheating of the samples The temperature dependence of the excess heat capacity was further analyzed and plotted using Origin software (MicroCal, Northampton, MA, USA) The thermal stability of actin was described by the

Tm, and DHcal was calculated as the area under the excess heat capacity function DSC experiments with different actin species were performed at least twice with very good repro-ducibility, and the representative curves are shown

Acknowledgements

We are grateful to Dr Alevtina Morozova for provid-ing us with protease ECP32⁄ grimelysin This work was supported by the Russian Foundation for Basic Research (grants 09-04-00266 to D.I.L and

08-04-00408 to S.Yu.Kh), the Program ‘Molecular and Cell Biology’ of the Russian Academy of Sciences, and by the grant from the President of Russian Federation (grant MK 2965.2009.4 to A.V.P.)

References

1 Kabsch W & Holmes KC (1995) The actin fold FASEB

J 9, 167–174

2 Tirion MM & ben-Avraham D (1993) Normal mode analysis of G-actin J Mol Biol 230, 186–195

3 Page R, Lindberg U & Schtt CE (1998) Domain motions in actin J Mol Biol 280, 463–474

4 Kinosian HJ, Selden LA, Estes JE & Gershman LC (1993) Nucleotide binding to actin Cation dependence

of nucleotide dissociation and exchange rates J Biol

Chem 268, 8683–8691

5 Strzelecka-Golaszewska H, Moraczewska J, Khaitlina

SYu & Mossakowska M (1993) Localization of the

tightly bound divalent-cation-dependent and

nucleotide-dependent conformational changes in G-actin using

lim-ited proteolytic digestion Eur J Biochem 211, 731–742

6 Kudryashov DS & Reisler E (2003) Solution properties

of tetramethylrhodamine-modified G-actin Biophys J

85, 2466–2475

7 Otterbein LR, Graceffa P & Dominguez R (2001) The

crystal structure of uncomplexed actin in the ADP state

Science 293, 708–711

8 Rould MA, Wan Q, Joel PB, Lowey S & Trybus KM

(2006) Crystal structures of expressed

non-polymeriz-able monomeric actin in the ADP and ATP states

J Biol Chem 281, 31909–31919

9 Chik JK, Lindberg U & Schutt CE (1996) The structure

of an open state of b-actin at 2.65 A˚ resolution J Mol

Biol 263, 607–623

10 Zheng X, Diraviyam K & Sept D (2007) Nucleotide

effects on the structure and dynamics of actin Biophys

J 93, 1277–1283

11 Dalhaimer P, Pollard TD & Nolen BJ (2008)

Nucleo-tide-mediated conformational changes of monomeric

actin and Arp3 studied by molecular dynamics

simula-tions J Mol Biol 376, 166–183

12 Splettstoesser T, Noe´ F, Oda T & Smith JC (2009)

Nucleotide-dependence of G-actin conformation from

multiple molecular dynamics simulations and

observa-tion of a putatively polymerizaobserva-tion-competent

super-closed state Proteins 76, 353–364

13 Pfaendtner J, Branduardi D, Parrinello M, Pollard TD

& Voth GA (2009) Nucleotide-dependent

conforma-tional states of actin Proc Natl Acad Sci USA 106,

12723–12728

14 Klenchin VA, Khaitlina SYu & Rayment I (2006)

Crys-tal structure of polymerization-competent actin J Mol

Biol 362, 140–150

15 Reisler E & Egelman EH (2007) Actin structure and

function: what we still do not understand J Biol Chem

282, 36133–36137

16 Graceffa P & Dominguez R (2003) Crystal structure of

monomeric actin in the ATP state Structural basis of

nucleotide-dependent actin dynamics J Biol Chem 278,

34172–34180

17 Khaitlina SYu & Strzelecka-Goaszewska H (2002) Role

of the DNase-I-binding loop in dynamic properties of

actin filament Biophys J 82, 321–334

18 Schwyter D, Phillips M & Reisler E (1989)

Subtilisin-cleaved actin: polymerization and interaction with

myo-sin subfragment 1 Biochemistry 28, 5889–5895

19 Khaitlina SYu, Smirnova TD & Usmanova AM (1988)

Limited proteolysis of actin by a specific bacterial

prote-ase FEBS Lett 228, 172–174

20 Khaitlina SYu, Collins JH, Kuznetsova IM, Pershina

VP, Synakevich IG, Turoverov KK & Usmanova AM (1991) Physicochemical properties of actin cleaved with bacterial protease from E coli A2 strain FEBS Lett

279, 49–51

21 Bozhokina E, Khaitlina S & Adam T (2008) Grimely-sin, a novel metalloprotease from Serratia grimesii, is similar to ECP32 Biochem Biophys Res Commun 367, 888–892

22 Ooi A & Mihashi K (1996) Effects of subtilisin cleavage

of monomeric actin on its nucleotide binding J Bio-chem 120, 1104–1110

23 Schuler H, Lindberg U, Schutt CE & Karlsson R (2000) Thermal unfolding of G-actin monitored with the DNase I-inhibition assay stabilities of actin isoforms Eur J Biochem 267, 476–486

24 Perieteanu AA & Dawson JF (2008) The real-time mon-itoring of the thermal unfolding of tetramethylrhod-amine-labeled actin Biochemistry 47, 9688–9696

25 Dedova IV, Nikolaeva OP, Mikhailova VV, dos Remedios CG & Levitsky DI (2004) Two opposite effects of cofilin on the thermal unfolding of F-actin: a differential scanning calorimetric study Biophys Chem

110, 119–128

26 Bobkov AA, Muhlrad A, Pavlov DA, Kokabi K, Yilmaz A & Reisler E (2006) Cooperative effects of cofilin (ADF) on actin structure suggest allosteric mech-anism of cofilin function J Mol Biol 356, 325–334

27 Dedova IV, Nikolaeva OP, Safer D, De La Cruz EM & dos Remedios CG (2006) Thymosin b4 induces a con-formational change in actin monomers Biophys J 90, 985–992

28 Levitsky DI, Pivovarova AV, Mikhailova VV & Nikolaeva OP (2008) Thermal unfolding and aggre-gation of actin Stabilization and destabilization of actin filaments FEBS J 275, 4280–4295

29 Mikhailova VV, Kurganov BI, Pivovarova AV & Levitsky DI (2006) Dissociative mechanism of F-actin thermal denaturation Biochemistry (Mosc) 71, 1261–1269

30 Khaitlina SYu, Moraczewska J & Strzelecka-Goas-zewska H (1993) The actin⁄ actin interactions involving the N-terminus of the DNase-I binding loop are crucial for stabilization of the actin filament Eur J Biochem

218, 911–920

31 Oda T, Namba K & Maeda Y (2005) Position and ori-entation of phalloidin in F-actin determined by X-ray fiber diffraction analysis Biophys J 88, 2727–2736

32 Pfaendtner J, Lyman E, Pollard TD & Voth GA (2010) Structure and dynamics of the actin filament J Mol Biol 396, 252–263

33 Le Bihan T & Gicquaud C (1991) Stabilization of actin

by phalloidin: a differential scanning calorimetric study Biochem Biophys Res Commun 181, 542–547