Acanthamoeba Profilin Affects the Mechanical Properties ... - CiteSeerX
THE JOURNAL
OF
BIOLOGICAL CHEMISTRY
Vol. 261, No. 23, Issue of August 15, pp. 10701-10706,1986 Printed in U.S. A.
0 1986 by The American Society of Biological Chemists, Inc.
Acanthamoeba Profilin Affects the MechanicalProperties of Nonfilamentous Actin* (Received for publication, March 13, 1986)
Masahiko Sato, William H. Schwarzz, and ThomasD. Pollard From the Department of Cell Biology and Anatomy and the $Department of Mechanical Engineering, J o h m Hopkins University School of Medicine, Baltimore, Maryland 21205
Refs. 1-3). Acanthamoeba has threeisoforms of profilin (profilin IA, IB (ll),and profilin I1 (12)) that are present throughout thecytoplasm (10) a t a concentration of over 100 p M (10). Large amounts of these profilins can be obtainedin high purity (9, 12). Both profilin I and I1 bind to actin monomers with a KO of 5-10 p~ and can inhibit the extentof polymerization. Electron microscope growth experiments show that both isoforms inhibittheelongation of filamentsmore strongly at the pointed end than at the barbed end (13). A new model for profilin interaction has been proposed which includes the possibility that profilin interacts with the barbed ends of actin filaments (13). We have measured several mechanical (rheological)parameters of profilin I, profilin 11, and their mixture asa function of time, concentration, salt, frequency, and shear rate. Becauseactin (4) and profilin were found to be viscoelastic materials, asingle parameter is inadequate to characterize interactions between components. Therefore, dynamic elasticity (G’) and viscosity ( a ’ ) were measured as a function of frequency, shear viscosity ( a ) was measured as a function of shear rate, andyield stress and recovery time were measured as functionsof strain and shearrate. The primary instrument used in this investigation was acone and plate rheometer which ensured that a constant shear rate was exerted on the Actin and actin-binding proteins are thoughtbetoessential sample at every point. The mechanical propertiesof profilin are of interest fortwo in maintaining the structural integrity of cytoplasm (1-3). To test this hypothesis, we (4) and others (5-8) have begun to reasons. First, profilin itself is a major cytoplasmic protein quantitate the mechanical properties of actin andsome actin- and may determine some mechanical properties of cytoplasm. actin polymerization and affects binding proteinsby using rheological techniques developed to Second,profilinregulates characterize complex materialswith solid- and liquid-like the filamentous structure of cytoplasm by binding to monomers and perhaps filaments as well. We were surprised to properties (called viscoelastic materials).Thisapproach should enable us to understand better how actin mechanical find that equilibrated samples of profilin alone (25-200 p ~ ) hadsignificantelasticityand viscositycompared to other properties are modulatedby actin-binding proteins. proteins. Furthermore, high concentrations of profilin also In a previous paper (4),we showed that actin filaments alone form a shear-sensitive viscoelastic solid and, remarka- changed the phaseof‘nonfilamentous actinfrom aviscoelastic bly, that nonfilamentous actin also formsa viscoelastic solid. solid to a liquid. Therefore, profilin not only limits the extent These results indicated that actin filaments and nonfilamen-of actin polymerization but also changes themechanical contousactin species weakly interactwitheachotherunder tributions of the nonfilamentous actin bound to the profilin. physiological conditions and that both filaments and nonG’ and 7’ are parameters measured in oscillatory experifilamentous species contribute to the viscoelastic properties ments in which minutedeformationsare used(14, 15). A of cytoplasm. Newtonian liquid, such as water, subjected to oscillatory deThe present study considers how Acanthamoeba profilin formation(strain) will exhibit the maximum stress at the affects the mechanical properties of actin. Profilins are small greatest shear rate. Therefore, the stress of a Newtonian liquid (12,000 daltons) soluble proteins that bind to actin monomerswill oscillate -90” out of phase with the input sinusoidal and inhibit actin polymerization into fiIaments (reviewed in strain. On the other hand, the stress of a Hookean solid, such as steel, will oscillate in phase with the input strain. The * This work was supported by National Institutes of Health Re- stress of a viscoelastic material will oscillate somewhere besearch Grants GM 26338 and GM 26132 (to T. D. p.) and a postdoc- tween 0” and -90” depending on how solid- or liquid-like it toral fellowship from the Muscular Dystrophy Association (to M. S.). is. Therefore, the dynamic elasticity is the in-phase compoThe costs of publication of this article were defrayed in part by the response. The dynamicviscospayment of page charges. This article must therefore be hereby nent of the sinusoidal material response -90” out of phase with the applied marked “advertisement” in accordance with 18 U.S.C. Section 1734 ity is the material solely to indicate this fact. strain. By definition, Newtonian liquids do not have a G’ and
10701
Downloaded from www.jbc.org at Yale University on July 23, 2009
We investigated the mechanical properties of two abundant, cytoplasmic proteins from Acantharnoeba, profilin and actin, and found that while both profilin and nonfilamentous actin alone behaved as solids, mixtures of the two proteins were viscoelastic liquids. When allowed to equilibrate, profilin formed a viscoelastic solid with mechanical properties similar to filamentous and nonfilamentous actin. Consequently, profilin itself may contribute significantly to the elasticity and viscosity of cytoplasm. The addition of profilin to nonfilamentous actin caused a phase transition from gel (viscoelastic solid) to sol (viscoelastic liquid) when the concentration of free actinbecame too low to form a gel. In contrast, profilin had little effect on the rigidity and viscosity of actin filaments. We speculate that nonfilamentous actin and profilin, both of which form shear-sensitive structures, can bemodeled as flocculant materials. We conclude that profilin may regulate the rigidity (elasticity) of the cytoplasm not only by inhibiting polymerization of actin, but also by modulating the mechanical properties of nonfilamentous actin.
10702
Actin and Profilin Mechanical Properties
the leading portion of the profilin I peak and the trailing portion of the profilin I1 peak. The final actin purification step was gel filtration on Sephadex G150 in 2 mM imidazole, pH 7.0, 0.2 mM ATP, 0.5 mM dithiothreitol, 0.2 mM CaCIZ, 0.5 mM azide. Separate peak fractions of profilin and actin were dialyzed at 5 "C into Buffer P (2 mM Pipes, pH 7.0, 0.1 mM MgCl2, 1mM EGTA, 0.2 mM ATP, 1 mM dithiothreitol, 0.5 mM azide) changed daily. Profilin and actin samples were mixed just before loading onto stationary rheometer platens. For filamentous ~ in Buffer P was added with the actin actin-profilin samples, 1 0 salt and profilin to make Buffer P-2 (10 mM Pipes, pH 7.0, 1mM MgClz, 1 mM EGTA, 100 mM KCl, 0.2 mM ATP, 1 mM dithiothreitol, 0.5 mM azide). Cytochrome c (horse heart type VI, Sigma) was gelfiltered on Sephadex G-75 in the manner of profilin. Rheometry-A R18 Weissenberg rheogoniometer (Sangamo Controls, Ltd., Bognor Regis, Sussex, England) was used in both continuous shear and small amplitude fo.rced oscillation modes. Sat0 et al. (4) describe the cone and plate geometry and the equations used to derive the mechanical parameters. Samples were equilibrated on stationary platens overnight a t 25 'C before experimentation (4). Assays-The profilin concentration was determined by absorbance a t 280 nm ( E = 1.2 cm2/mg, 1.4 X 10' M" cm") in 10 mM imidazole, pH 7.5, 1.5 mM azide (10). The actin concentration was determined by absorbance a t 290 nm ( E = 0.63 cm2/mg, 2.66 X lo' M" cm") or with the Bradford assay (18) using actin as the standard. Cytochrome c concentration was determined by absorbance at 550 nm ( E = 29.5 X lo3 M" cm"). Polyacrylamide gel electrophoresis was performed by the method of Laemmli (19). Rheometer samples which had been incubated for about 30 h at 25 f 0.01 "C were prepared for electron microscopy by spreading onto carbon-coated, glow-discharged grids and staining for 5 s in 1% uranyl acetate. Actin grids were also pretreated with 0.02% cytochrome c in 0.1% isoamyl alcohol (20).
12
9
I
I
I
I
I
a
I
I
*1
,A
$+*itt t t i-
"MH*tt 0
I
I
I
I
I
'
115
230
345
460
575
Time (rnin)
FIG. 1. Dynamic elasticity and dynamic viscosity of profilin as a function of time. Dynamic elasticity (G' = 0) and dynamic viscosity (7' = +) of 85 GM profilin in Buffer P ( A ) and Buffer P-2 The abbreviations used are: Pipes, 1,4-piperazinediethanesulfonic ( B ) at 0.6 Hz, 25 & 0.01 "C. The extreme right values of G' and q' correspond to equilibrium values reached after 1000 min. acid; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraaceticacid.
Downloaded from www.jbc.org at Yale University on July 23, 2009
have a frequency-independent q' (14). Elastic solids have a RESULTS frequency-independent G' called a n equilibrium elastic modDynamicElasticity and Viscosity of Profilin-Similar to ulus (G,) and an infiniteviscosity 7'. actin (4), profilin (85 gM) required incubation times greater 7 is defined asthefrictionalresistance of a sampleto than 600 min before the G' and 7' stabilized in Buffer P (Fig. constant flow (shear strain rate) (15, 16). The shearviscosity or P-2 (Fig. 1B) when oscillated at 0.6 Hz with a maxi1A) of a Newtonian liquid is independent of the shear rate (16). ) actin For biological materials, this parameter is a complex function mum strainof 0.083. Mixtures of profilin (5-85 p ~ and (24 PM) in either Buffer P or P-2 required similar times for of shear rate,which is determinedby the rheometergeometry equilibration. The final values of G' and 7 ' were independent and the resultant velocity gradient. Shear viscosity is independent of the extent of displacement (strain) which is not of the duration of oscillation at 0.6 Hz as measured for an additional 1500 min and were reproducible to &50% for five thecase for 7' for biological materials(4).Thisandthe different batchesof profilin. following techniques are disruptive analyses which should not The G' and 7' of the equilibrated samples of profilin were be confused with material conditions used in oscillatory analhighly dependentonthe oscillationfrequencywhichwas ysis (4). varied from 0.0002 to 1.896 Hz (see Fig. 2). For profilin (85 Yield stress and recovery time are empirical parameters usedtoconfirmtheexistence of structurein a complex WM) in Buffer P or P-2, log 7' was an inverse curvilinear material. Yield stress is defined as the maximum stress a solid function of log frequency; log G' increased slightly with log can sustain without disruption by continued deformation at frequency(Fig.2A). The dynamic viscosity was five times a constant shear rate (16).Recovery time is a measure of the higher for profilin in Buffer P-2 than in Buffer P, but the time required for self-healing of the structure to 90% of the dynamic elasticitywas unaffected by the higher salt condition presheared dynamic elasticity. Recovery time is loosely re- (Fig. 2A). At a concentration of 200 PM profilin, log 7' showed a slight lated to the material relaxation time (14,16). Both parameters are dependent on the rate and extent of deformation (strain). upwardcurvature,and log G' leveled off to afrequencyNewtonian liquids have neither a yield stress nor recovery independent value of 5.5 dyne/cm2 below 0.0012 Hz (Fig. 2B). This value corresponds to the equilibrium elastic modulus time. and is characteristic of aviscoelastic solid. Atfrequencies MATERIALS ANDMETHODS greater than 0.01 Hz, the values of G' and 7' were approxiProtein Purification-Profilin and actin were prepared from a mately the same for profilin concentrations between 25 and sucrose extract of Acanthamoeba (Neff strain) following published 200 PM in Buffer P. procedures (9, 12, 17). The final purification step for conventional The mechanical properties of conventional profilin, profilin profilin (9) was gel filtration on Sephadex G-75 equilibrated with 10 I, and profilin I1 were approximately the same. The magnimM imidazole, pH 7.5, 1.5 mM azide. We will use "profilin" to refer to this mixture of isoforms in this paper. Profilins I and I1 were tudes of G' and 7' for profilin I and I1 were up to two times separated by isocratic cation exchange chromatography on a 1 X 51- that for conventional profilin, and at low frequencies had G, cm column of carboxymethyl-agarose (Bio-Rad CM Bio-Gel A, 100- values = 10 dyne/cm2 in Buffer P (Fig. 2C). The data for 200 mesh) in 10 mM Pipes,' pH 6.5 (12). Fractions were pooled from profilin I1 were taken within 2 days of the final purification
Actin and Profilin Mechanical Properties
10703
O t
-0
20
46
uM
80
68
106
PROFILIN
FIG. 3. Dependence of the dynamic elasticity of nonfilamentous actin on the profilin concentration. Dynamic elasticity at 0.0012Hz for 24 p~ actin in Buffer P (G’ = 0) is plotted as a . function of increasing profilin concentration (0-85 p ~ ) Predicted values for the mixture (G’ = A) are the sums of the elasticities for the individual proteins. These values are expected fortwo ideally noninteracting materials. Each data point corresponds to the calculated mean from 2 to 7 experiments. 01
0
-1
J
3
3 2 1
0 -1
-
-4
-3
-2
Log Frequency
-1
0
1
1
(Hz)
FIG. 2. Dynamic elasticity and dynamic viscosity of profilin as a function of frequency. A , 85 p~ profilin in Buffer P (G’ = 0, 7’ = 0)and Buffer P-2 (G’ = W, q’ = 0 ) .B, 200 b~ profilin in Buffer P (G’ = 0, TJ’= 0). C, 85 &M profilin I (G’ = 0, q’ = 0)and profilin I1 (G’ = W, q’ = 0 ) in Buffer P. The horizontal line is the viscosity (q ’ = 4.03 poise) of the oil (Cannon Instruments, College Park, PA) used to calibrate the rheometer. In these experiments, a constant oscillation amplitude of 0.0046 radians for a maximum strain of 0.083 was used for the frequency range of 0.0002-2 Hz.
step because precipitates were observed in the solution by 8 days after isolation. For comparison, a protein of the same size as profilin, cytochrome c (12,000 daltons), was analyzed in Buffer P and found to be a Newtonian liquid with a viscosity of 0.03 poise with no measurable G‘. Paraffin oil and Buffer P alone were also Newtonian liquids with viscosities of 1.5 and about 0.01 poise, respectively.
1 O t
-4
-3 -2 -1 0 i o g F ~ P C C ~ C Ii ’d ? )
1
FIG. 4. Dynamic elasticity and dynamic viscosity of actin and mixtures of actin and profilin in buffer P. A , 24 p~ actin in buffer P (G’ = 0, q’ = 0).The equilibrium elastic modulus, G,, at Mechanical Properties of Mixtures of Actin a n d Profilinlow frequency was 8 dyne/cm*. No actin filaments were observed in The rheological measurements for mixtures of nonfilamen- this sample by electron microscopy. B, 24 p M actin plus 25 p M profilin tous actin and profilin confirmed previous evidence that the in Buffer P (G’ = 0, q’ = 0).At low frequency, G. = 8 dyne/cm2. C, two proteins interact; the dynamic elasticity and dynamic 24 p M actin plus 56 pM profilin in Buffer P (G’ = 0, TJ’= 0).Log G’ viscosity of the mixturewere much less than the sum predictedis linear with log frequency as measured to 0.0003 Hz. 24 pM actin from the two materials alone (see Figs. 3 and 4). This was plus 85 p~ profilin in Buffer P (G’ = W, q’ = 0).G’ for this mixture true for all concentrations of profilin tested (25-85 p M , see curves downward toward zero with decreasing frequency. The observed tendency of G’ to decrease at low frequency with added profilin Fig. 3), and above 56 p~ profilin the dynamic elasticity tendedwas reproduced for two separate batches of profilin and actin. toward G‘ = 0 a t low frequencies (Fig. 4 0 , indicating that
the mixture was now a viscoelastic liquid rather than a solid like actin alone (Fig. 4A) (4). Thus, high concentrations of profilin (56-85 p M ) caused actin to undergo a phase transition from a gel to a sol. The absence of actin filaments in these samples was confirmed by electron microscopy of negatively stained specimens (20). Profilin had no significant effect on the mechanical properties of actin filaments. That is, mixtures of filamentous
actin (24 p M in Buffer P-2) with 0-85 p M profilin did not show a phase transition behavior from a gel to a sol in our frequency range. The variation in the profilin concentration randomly shifted these plots along the vertical axis by less than a factor of 2. All of these mixtures formed viscoelastic solids with G, that were within *loo% of actin alone (see Fig. 5) and contained numerous actin filaments as observedin
Downloaded from www.jbc.org at Yale University on July 23, 2009
4r
0
10704
Actin and Profilin Mechanical Properties
0
g 2
-4
-3
-2
-1
LogFreauency
0
B
1
(Hz!
FIG.5. Dynamic elasticity and dynamic viscosity of 24 I.LM actin plus 86 p M profilin in Buffer P-2 (G'= 0, q' = 0).G. = 11 dyne/cm2 for this case as compared with G . = 16 & 6 dyne/cm* for filamentous actin alone.
I!
4L i
A
Downloaded from www.jbc.org at Yale University on July 23, 2009
electron micrographs of negatively stained material. Continuous Shear Experiments-Continuous shear experiments confirmed that profilin diminishes the rigidity of nonfilamentous actin but not of filamentous actin. When subjected to continuously increasing strain, mixtures of filamentous actin and profilin responded with a curvilinear increase in stress until the material was disrupted at the yield stress (in this case8.2 dyne/cm2) and thenbegan to flow as a liquid (Fig. 6A). This was seen as an asymptotic drop in stress toa constant value after material disruption. Filamentous actin ) similarly to the mixture (shownin Fig. alone (24p ~ behaved 6A) with a yield stress of 10.1 k 2 dyne/cm2. 200 PM profilin in Buffer P showed a small yield stress of2.2 dyne/cm2 in ) similar Buffer P (Fig. 6 B ) .Nonfilamentous actin (24 p ~ was to conventional profilin (Fig. 6B) but had higher yield stress a 0 ' of 6.5 dyne/cm2. In contrast, mixtures of nonfilamentous actin 0 5 10 15 20 and profilin at the phase transition concentration did not 5;%r S t r o l l - , showareproducible yield stress (Fig. 6C). Thestress for FIG.6. Shear stress as a function of strain. Equilibrated samNewtonian liquids, such as viscosity standard oil, responded ples were strained (deformed) from time 0 at a constant shearrate of as a step function of continuously increasing strain, as seen 0.546 s-' by rotating the bottom rheometer plate at 0.09 rpm. The material shear stresswas monitored by the freely suspended top cone. in Fig. 6D. After cessation of shearing, all samplesof actin andprofilin See Sat0 et al. (4) for details. A , 24 p M filamentous actin plus 85 WM recovered their equilibrium mechanical properties with time, profilin in Buffer P-2. The yield stress was 8.2 dyne/cm2. B , 200 p~ profilin in Buffer P. The yield stress was 2.2 dyne/cm2. C, 24 confirming that a shear-sensitive structure was formed re- nonfilamentous actin plus 55 p~ profilin in Buffer P. The complex versibly in all of these profilin and actin-profilin samples (Fig. stress response supports the conclusion from the frequency data that 7). The recovery time, however,was very sensitive to the this concentration of actin and profilin forms a transitional intershear strain rate and the total deformation (strain). 90% The mediate between solid (gel) and liquid (sol). D,Viscosity standard oil recovery times for profilin and actin-profilin samples were (Cannon Instruments).The stress of this andother Newtonian liquids greater than 40 min and very difficult to reproduce, showing responds as a step function of strain. that these sampleswere more shear-sensitive than actin alone (4). Shear Viscosity of Profilin and Actin-Like actin (4-6, 8), shear viscositiesforprofilin I,conventional profilin, and mixtures of actin and profilin were highly dependent on the shear rate (Fig. 8). Plots of log 7 versus log shear rate were curvilinear butlevelled off asymptotically to a constant value at shear rates >50 s-'. This behavior indicates that profilin and mixturesof actin andprofilin form shear-sensitive (shearthinning) structures which break apart into component subU units at shear rates greater than 50s" (Fig. 8, A and B). Shear viscosity values for 25, 85 and 200 WM profilin in shear Buffer P were nearly identical (Fig. 8 A ) , which parallels the 1 5 0 01 4 6 01 4 2 01 3 8001 3 4 0" ' similarity of dynamic parametersobserved in frequency plots TIME (rnin) for profilin from 25 to 200 pM. All three curves were asympFIG.7. Recovery of the mechanical properties of filamentotic to a shear rate-independent viscosity of 0.04 poise. By tous actin plus profilin after disruption by shearing. 24 p~ comparison, the shear viscosity for nonfilamentous actin (Fig. filamentous actin plus 85 WM profilin in Buffer P-2 (G' = 0).Dynamic elasticity (G') was monitored at 0.6 Hz before and after disruption at 8C) was asymptotic to 0.05 poise at high shear rates. Shear viscosities of nonfilamentous actin andprofilin were a shear rateof 0.546 s-' for 1min, as in Fig. 6A. The arrow indicates the point along the time course at which the sample was sheared. All half the values for profilin at low shear rates.However, above actin and/or profilin samples in Buffer P or P-2 self-healed with time shear ratesof 10 s-', the shearviscosity levelled off to a higher that was dependent on the extent of deformation (strain) and shear value of 0.07 poise, suggesting that larger components exist strain rate. In this case, 90% recovery occurred at about 40 min.
Actin and Profilin Mechanical Properties
10705
3r/
TABLEI Calculation of the concentrations of profilin and actin species The total actin was constant at 24 p ~ and , the concentrations of the other species were calculated assuming a 1:1 complex of profilin with actin having a KDof 5 p ~ .
2
Total profilin
8.7 1.8
Free actin
Free profilin
FM
/IM
PM
PM
0 24 56 85
24.0
0 8.7 35.0 62.8
0
15.3 21.0 3.0 22.2
p * ~ ~ ~ ~ f n
= 16 +. 6 dyne/cm') (4).Second, shear experiments showed
Downloaded from www.jbc.org at Yale University on July 23, 2009
that profilin at equilibrium has a yield stress that compares with actin (4)and with living cytoplasm of Physarum plasmodium (21). Third, shearviscosity experiments showed that profilin is very shear-sensitive when mechanically disturbed but will self-heal with time. Effects of Profilin on the Mechanical Properties of ActinThe observed phase transition of nonfilamentous actin from a gel to a sol can be explained in terms of the low concentration of free actin at highprofilin concentration (Table I). About 10 p~ nonfilamentous actin was required to form an actin gel.' With 56 or 85 p~ profilin, the calculated concen. tration of nonfilamentous actinwas less than 5 p ~Similarly, 0 1 2 3 - 2 - 1 the calculated concentrations of free profilin in these samples Log Shear Rate (5.') ) less than concentrationsof profilin alone (35 and 63p ~ were . these reasons, the mixture FIG. 8. Shear viscosity (t~) as a function of shear rate. A , 85 that failed to gel (e.g. 85 p ~ ) For p M profilin I in BufferP (I)= 0 ) .85 p M conventional profilin (profilin liquified at the concentrationsof profilin that we tested. We I and 11) in Buffer P (I) = 0).B , 24 p~ nonfilamentous actin plus 85 expect that these mixtures may form a viscoelastic solid at p M profilin in Buffer P (I) = 0). C, 24 p M nonfilamentous actin in higher concentrations of profilin, since free profilinalone (200 Buffer P (g = A). 24 P M cytochrome c in Buffer P (7 = A,0.02 It 0.01 ) a gel. Additional datawill be necessary to establish poise). The shear rate independence of the cytochrome c viscosity p ~ forms the rheological properties of the complex of nonfilamentous shows that it is a Newtonian liquid that does not form higher order actin with profilin. The available data are inconclusive but structure. consistent with the complex having weak mechanical propin the mixture than in actin or profilin alone (Fig. 8B). For erties compared to free actin and free profilin. Profilin had no significant effect on the mechanical propcomparison, the shear viscosity of gel-filtered cytochrome c erties of filamentous actin because there was no concentrawas plotted asa function of shear rate in Fig. 8C. Cytochrome c (24 p ~ could ) not be distinguished from Newtonian liquids tion-dependent changein G', q', or q observed for the mixture . think this of filamentous actin and profilin (25-85 p ~ ) We of 0.01 and 0.033 poise on our instrument. interpretation is correct because these concentrations of profilin would reduce the actin filament contentby less than 8% DISCUSSION because the critical concentrationfor actin polymerization is We characterized the effect of profilin on the mechanical PM in Buffer P-2. about 0.1 properties of actin to search for possible roles for profilin in Modeling-The molecular basis of the surprising mechanithecytoplasm.Towardthis goal, equilibratedsamples of amoeba actin and profilin were oscillated and sheared over a cal properties of profilin and nonfilamentous actin is not clear.Onone hand,electron micrographs of theseparate wide range (5 orders of magnitude) that included frequencies rheometer samplesin Buffer P were devoid of actin filaments and rates of deformation analogous to those expected inside and organized profilin structure. On the other hand, these a cell. We found that Acanthamoeba profilin alone is a comsamples behaved like viscoelastic solids (gels). plex viscoelastic material with significantrigidity (elasticity) Apossible explanation is that nonfilamentous rctin and suggesting that it contributes substantially to the mechanical profilin flocculate into branched,aggregate structures similar properties of cytoplasm.Remarkably, we found that even to other nonfilamentous materials including kaolin, polystythough both profilin and nonfilamentous actin alone formed rene beads, boehmite, proteins (casein), and DNA fragments viscoelastic solids, the mixture of the two underwent a phase (23-27). Thesematerials alsoform viscoelastic solids and transition into a viscoelastic liquid. have rheological properties similar to those presented here for Evidence for Profilin Structure-Several lines of evidence profilin and actin. Specifically, emulsions of particles 1-1000 show that Acanthamoebaprofilin formed a structure with nm in diameterform shear-thinning, amplitude-sensitivegels substantialmechanicalpropertiescomparedwithanother (flocs) capable of undergoing phase transitions (28-30). Casmallprotein,cytochrome c. First, at about physiological sein (26) and polystyrene dispersions (24) also exhibit me, conventionalprofilin formed a chanical independence to concentration around the minimum concentrations (200 p ~ (IO)), viscoelastic solid with a G, of 5.5 dyne/cm'. Purified profilin concentration necessary to form a gel. This typeof modelling I and I1 a t 85 I . ~ Malso formed similar viscoelastic solids with may prove useful in understanding the unit interactions inG, values = 10 dyne/cm'. This behavior and the values of these mechanical propertieswere similar to both nonfilamen- ' M. Sato, W. H. Schwarz, and T. D. Pollard, unpublished obsertous actin ( G , = 8 2 3 dyne/cm') and filamentous actin (G, vations.
10706
Mechanical Properties Profilinand Actin
Acknowledgments-We
gratefully thank G. Leimbach and D. Chan
for help in collecting some of the data.
REFERENCES 1. Korn, E. D. (1982) Physiol. Reu. 62,672-737 2. Pollard, T. D., and Cooper, J. A. (1986) Annu. Reu. Biochem. 5 5 , 987-1035 3. Stossel, T. P., Chaponnier, C., Ezzell, R., Hartwig, J. H., Janmey, P., Kuiatkowski, D., Lind, S., Smith, D., Southwick, F. S., Yin, H. L., and Zaner, K. S. (1985) A n n u Reu. Cell Biol. 1, 353-402 4. Sato, M., Leimbach, G., Schwarz, W. H., and Pollard, T. D.
(1985) J . Biol. Chem. 260,8585-8592 5. Jen, C. J., McIntire, L. V., and Bryan, J. (1982) Arch. Biochem. B i o p h y ~216,126-132 . 6. Zaner, K. S., and Stossel, T. P. (1983) J. Biol. Chem. 258,1100411009 7. Schamus, E., Booth, S., Hallaway, B., and Rosenberg, A. (1985) J. Biol. Chem. 260,3724-3730 8. Opperman, W., and Jaberg, B. (1985) Rheol. Acta 2 4 , 525-529 9. Reichstein, E., and Korn, E. D. (1979) J. Biol. Chem. 254,61746179 10. Tseng, P. C., Runge, M. S., Cooper, J. A,, Williams, R. C., Jr., and Pollard, T. D. (1984) J. Cell Biol. 9 8 , 214-221 11. Ampe, C., Vandekerkove, J., Brenner, S. L., Tobacman, L., and Korn, E. D. (1985) J . Biol. Chem. 260,834-840 12. Kaiser, D. P., Sato, M., Ebert, R., and Pollard, T. D. (1986) J . Cell Biol. 1 0 2 , 221-226 13. Pollard, T. D.,and Cooper, J. A. (1984) Biochemistry 2 3 , 66316641 14. Ferry, J. D. (1970) ViscoelasticProperties of Polymers, pp. 11118, John Wiley & Sons, New York 15. Walters, K. (1968) Basic Concepts and Formulae for the Rheogo-
niometer, Sangamo Controls Ltd., Sussex, England 16. Van Wazer, J. R., Lyons, J. W., Kim, K. Y., and Colwell, R. E. (1963) Viscosity and Flow Measurement: A Laboratory Handbook of Rheology, Interscience Publishers, New York 17. Gordon, D. J., Eisenberg, E., and Korn, E. D. (1976) J. Biol. Chem. 251,4718-4786 18. Bradford, M. M. (1976) Anal. Eiochem. 7 2 , 248-254 19. Laemmli, U. K. (1970) Nature 227, 680-685 20. Jockusch, B. M., and Isenberg, G. (1981) Proc. Natl. Acad. Sci. U. S. A . 78,3005-3009 21. Sato, M., Wong, T. Z., and Allen, R. D. (1983) J . Cell Biol. 9 7 , 1089-1097 22. Porter, K. R., Beckerle, M., and McNiven, M. (1983) Mod. Cell Biol. 2,259-302 23. Michaels, A. S., and Bolger, J. C. (1962) I and EC Fundamentals 1,153-162 24. Tadros, T. F. (1984) ACS (Am. Chem. SOC.)Symp. Ser. 240, 411-431 25. Lih, M. (1961) J. Colloid Sci. 1 6 , 297-310 26. Tokita, M., Niki, R., and Hikichi, K. (1984) J. Physiol. SOC.Jpn. 53,480-482 27. Bloomfield, V. A. (1986) ACS (Am. Chem. Soc.) Polymer Preprints 2 7 , 249-250 28. Hirtzel, C. S., and Rajagopalan (1983) ResearchandResearch Needs in Colloidal Interactions, Office of Interdisciplinary Research, National Science Foundation, Wash., D. C. 29. Van de Ven. T. G. M.. and Hunter, R. J. (1979) J. Colloid Interface sci. 6 8 , i35-143 30. Parsegian, V. A. (1982) Adu. Colloid Interface Sci. 1 6 , 49-56
Downloaded from www.jbc.org at Yale University on July 23, 2009
volved in the actin and profilin system and may explain the similar mechanical independence to concentration observed for profilin from 25 to 85 pM in Buffer P. Aggregates (flocs) are generally composed of unit structures called floccules which may themselves beaggregates. Electron micrographs of negatively stained material suggest that actin floccules and profilin floccules may be larger than monomers (4). Structural detailsof the aggregate species are not clear at this time. A floccule is generally able to interact with another particle anywhere on its surface, but this may not be thecase with nonfilamentous actin and profilin. We think that mixtures of nonfilamentous actin and profilin have less than the expected mechanical properties because sites required for selfassociation between actin species and between profilin species were blocked in the complex. Details of these complex interactions are not clearat this time. Interpretation-Actin and profilin are two of the most (10). The prevalentproteinsin Acantharnoebacytoplasm rheological properties measured here should be relevant to living cytoplasm because we included deformations,shear rates, frequencies, and protein concentrations expectedfor the cell. Previously, it was known that profilin modulates the actin pool available to polymerize into filaments (1-3), but our results demonstrate two additional functions for profilin. First, profilin itself can forma viscoelastic solidat physiological conditions and may contribute to the structural integrity of the cytoplasm. Second, high concentrations of profilin can alter the phase state of nonfilamentous actin and so may regulate the rigidity of actin-rich cytoplasm ( 2 2 ) . That is, profilin may be used to make regions of the cytoplasm more liquid (such as the endoplasm) and thereforemore amenable to motility.
OF
BIOLOGICAL CHEMISTRY
Vol. 261, No. 23, Issue of August 15, pp. 10701-10706,1986 Printed in U.S. A.
0 1986 by The American Society of Biological Chemists, Inc.
Acanthamoeba Profilin Affects the MechanicalProperties of Nonfilamentous Actin* (Received for publication, March 13, 1986)
Masahiko Sato, William H. Schwarzz, and ThomasD. Pollard From the Department of Cell Biology and Anatomy and the $Department of Mechanical Engineering, J o h m Hopkins University School of Medicine, Baltimore, Maryland 21205
Refs. 1-3). Acanthamoeba has threeisoforms of profilin (profilin IA, IB (ll),and profilin I1 (12)) that are present throughout thecytoplasm (10) a t a concentration of over 100 p M (10). Large amounts of these profilins can be obtainedin high purity (9, 12). Both profilin I and I1 bind to actin monomers with a KO of 5-10 p~ and can inhibit the extentof polymerization. Electron microscope growth experiments show that both isoforms inhibittheelongation of filamentsmore strongly at the pointed end than at the barbed end (13). A new model for profilin interaction has been proposed which includes the possibility that profilin interacts with the barbed ends of actin filaments (13). We have measured several mechanical (rheological)parameters of profilin I, profilin 11, and their mixture asa function of time, concentration, salt, frequency, and shear rate. Becauseactin (4) and profilin were found to be viscoelastic materials, asingle parameter is inadequate to characterize interactions between components. Therefore, dynamic elasticity (G’) and viscosity ( a ’ ) were measured as a function of frequency, shear viscosity ( a ) was measured as a function of shear rate, andyield stress and recovery time were measured as functionsof strain and shearrate. The primary instrument used in this investigation was acone and plate rheometer which ensured that a constant shear rate was exerted on the Actin and actin-binding proteins are thoughtbetoessential sample at every point. The mechanical propertiesof profilin are of interest fortwo in maintaining the structural integrity of cytoplasm (1-3). To test this hypothesis, we (4) and others (5-8) have begun to reasons. First, profilin itself is a major cytoplasmic protein quantitate the mechanical properties of actin andsome actin- and may determine some mechanical properties of cytoplasm. actin polymerization and affects binding proteinsby using rheological techniques developed to Second,profilinregulates characterize complex materialswith solid- and liquid-like the filamentous structure of cytoplasm by binding to monomers and perhaps filaments as well. We were surprised to properties (called viscoelastic materials).Thisapproach should enable us to understand better how actin mechanical find that equilibrated samples of profilin alone (25-200 p ~ ) hadsignificantelasticityand viscositycompared to other properties are modulatedby actin-binding proteins. proteins. Furthermore, high concentrations of profilin also In a previous paper (4),we showed that actin filaments alone form a shear-sensitive viscoelastic solid and, remarka- changed the phaseof‘nonfilamentous actinfrom aviscoelastic bly, that nonfilamentous actin also formsa viscoelastic solid. solid to a liquid. Therefore, profilin not only limits the extent These results indicated that actin filaments and nonfilamen-of actin polymerization but also changes themechanical contousactin species weakly interactwitheachotherunder tributions of the nonfilamentous actin bound to the profilin. physiological conditions and that both filaments and nonG’ and 7’ are parameters measured in oscillatory experifilamentous species contribute to the viscoelastic properties ments in which minutedeformationsare used(14, 15). A of cytoplasm. Newtonian liquid, such as water, subjected to oscillatory deThe present study considers how Acanthamoeba profilin formation(strain) will exhibit the maximum stress at the affects the mechanical properties of actin. Profilins are small greatest shear rate. Therefore, the stress of a Newtonian liquid (12,000 daltons) soluble proteins that bind to actin monomerswill oscillate -90” out of phase with the input sinusoidal and inhibit actin polymerization into fiIaments (reviewed in strain. On the other hand, the stress of a Hookean solid, such as steel, will oscillate in phase with the input strain. The * This work was supported by National Institutes of Health Re- stress of a viscoelastic material will oscillate somewhere besearch Grants GM 26338 and GM 26132 (to T. D. p.) and a postdoc- tween 0” and -90” depending on how solid- or liquid-like it toral fellowship from the Muscular Dystrophy Association (to M. S.). is. Therefore, the dynamic elasticity is the in-phase compoThe costs of publication of this article were defrayed in part by the response. The dynamicviscospayment of page charges. This article must therefore be hereby nent of the sinusoidal material response -90” out of phase with the applied marked “advertisement” in accordance with 18 U.S.C. Section 1734 ity is the material solely to indicate this fact. strain. By definition, Newtonian liquids do not have a G’ and
10701
Downloaded from www.jbc.org at Yale University on July 23, 2009
We investigated the mechanical properties of two abundant, cytoplasmic proteins from Acantharnoeba, profilin and actin, and found that while both profilin and nonfilamentous actin alone behaved as solids, mixtures of the two proteins were viscoelastic liquids. When allowed to equilibrate, profilin formed a viscoelastic solid with mechanical properties similar to filamentous and nonfilamentous actin. Consequently, profilin itself may contribute significantly to the elasticity and viscosity of cytoplasm. The addition of profilin to nonfilamentous actin caused a phase transition from gel (viscoelastic solid) to sol (viscoelastic liquid) when the concentration of free actinbecame too low to form a gel. In contrast, profilin had little effect on the rigidity and viscosity of actin filaments. We speculate that nonfilamentous actin and profilin, both of which form shear-sensitive structures, can bemodeled as flocculant materials. We conclude that profilin may regulate the rigidity (elasticity) of the cytoplasm not only by inhibiting polymerization of actin, but also by modulating the mechanical properties of nonfilamentous actin.
10702
Actin and Profilin Mechanical Properties
the leading portion of the profilin I peak and the trailing portion of the profilin I1 peak. The final actin purification step was gel filtration on Sephadex G150 in 2 mM imidazole, pH 7.0, 0.2 mM ATP, 0.5 mM dithiothreitol, 0.2 mM CaCIZ, 0.5 mM azide. Separate peak fractions of profilin and actin were dialyzed at 5 "C into Buffer P (2 mM Pipes, pH 7.0, 0.1 mM MgCl2, 1mM EGTA, 0.2 mM ATP, 1 mM dithiothreitol, 0.5 mM azide) changed daily. Profilin and actin samples were mixed just before loading onto stationary rheometer platens. For filamentous ~ in Buffer P was added with the actin actin-profilin samples, 1 0 salt and profilin to make Buffer P-2 (10 mM Pipes, pH 7.0, 1mM MgClz, 1 mM EGTA, 100 mM KCl, 0.2 mM ATP, 1 mM dithiothreitol, 0.5 mM azide). Cytochrome c (horse heart type VI, Sigma) was gelfiltered on Sephadex G-75 in the manner of profilin. Rheometry-A R18 Weissenberg rheogoniometer (Sangamo Controls, Ltd., Bognor Regis, Sussex, England) was used in both continuous shear and small amplitude fo.rced oscillation modes. Sat0 et al. (4) describe the cone and plate geometry and the equations used to derive the mechanical parameters. Samples were equilibrated on stationary platens overnight a t 25 'C before experimentation (4). Assays-The profilin concentration was determined by absorbance a t 280 nm ( E = 1.2 cm2/mg, 1.4 X 10' M" cm") in 10 mM imidazole, pH 7.5, 1.5 mM azide (10). The actin concentration was determined by absorbance a t 290 nm ( E = 0.63 cm2/mg, 2.66 X lo' M" cm") or with the Bradford assay (18) using actin as the standard. Cytochrome c concentration was determined by absorbance at 550 nm ( E = 29.5 X lo3 M" cm"). Polyacrylamide gel electrophoresis was performed by the method of Laemmli (19). Rheometer samples which had been incubated for about 30 h at 25 f 0.01 "C were prepared for electron microscopy by spreading onto carbon-coated, glow-discharged grids and staining for 5 s in 1% uranyl acetate. Actin grids were also pretreated with 0.02% cytochrome c in 0.1% isoamyl alcohol (20).
12
9
I
I
I
I
I
a
I
I
*1
,A
$+*itt t t i-
"MH*tt 0
I
I
I
I
I
'
115
230
345
460
575
Time (rnin)
FIG. 1. Dynamic elasticity and dynamic viscosity of profilin as a function of time. Dynamic elasticity (G' = 0) and dynamic viscosity (7' = +) of 85 GM profilin in Buffer P ( A ) and Buffer P-2 The abbreviations used are: Pipes, 1,4-piperazinediethanesulfonic ( B ) at 0.6 Hz, 25 & 0.01 "C. The extreme right values of G' and q' correspond to equilibrium values reached after 1000 min. acid; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraaceticacid.
Downloaded from www.jbc.org at Yale University on July 23, 2009
have a frequency-independent q' (14). Elastic solids have a RESULTS frequency-independent G' called a n equilibrium elastic modDynamicElasticity and Viscosity of Profilin-Similar to ulus (G,) and an infiniteviscosity 7'. actin (4), profilin (85 gM) required incubation times greater 7 is defined asthefrictionalresistance of a sampleto than 600 min before the G' and 7' stabilized in Buffer P (Fig. constant flow (shear strain rate) (15, 16). The shearviscosity or P-2 (Fig. 1B) when oscillated at 0.6 Hz with a maxi1A) of a Newtonian liquid is independent of the shear rate (16). ) actin For biological materials, this parameter is a complex function mum strainof 0.083. Mixtures of profilin (5-85 p ~ and (24 PM) in either Buffer P or P-2 required similar times for of shear rate,which is determinedby the rheometergeometry equilibration. The final values of G' and 7 ' were independent and the resultant velocity gradient. Shear viscosity is independent of the extent of displacement (strain) which is not of the duration of oscillation at 0.6 Hz as measured for an additional 1500 min and were reproducible to &50% for five thecase for 7' for biological materials(4).Thisandthe different batchesof profilin. following techniques are disruptive analyses which should not The G' and 7' of the equilibrated samples of profilin were be confused with material conditions used in oscillatory analhighly dependentonthe oscillationfrequencywhichwas ysis (4). varied from 0.0002 to 1.896 Hz (see Fig. 2). For profilin (85 Yield stress and recovery time are empirical parameters usedtoconfirmtheexistence of structurein a complex WM) in Buffer P or P-2, log 7' was an inverse curvilinear material. Yield stress is defined as the maximum stress a solid function of log frequency; log G' increased slightly with log can sustain without disruption by continued deformation at frequency(Fig.2A). The dynamic viscosity was five times a constant shear rate (16).Recovery time is a measure of the higher for profilin in Buffer P-2 than in Buffer P, but the time required for self-healing of the structure to 90% of the dynamic elasticitywas unaffected by the higher salt condition presheared dynamic elasticity. Recovery time is loosely re- (Fig. 2A). At a concentration of 200 PM profilin, log 7' showed a slight lated to the material relaxation time (14,16). Both parameters are dependent on the rate and extent of deformation (strain). upwardcurvature,and log G' leveled off to afrequencyNewtonian liquids have neither a yield stress nor recovery independent value of 5.5 dyne/cm2 below 0.0012 Hz (Fig. 2B). This value corresponds to the equilibrium elastic modulus time. and is characteristic of aviscoelastic solid. Atfrequencies MATERIALS ANDMETHODS greater than 0.01 Hz, the values of G' and 7' were approxiProtein Purification-Profilin and actin were prepared from a mately the same for profilin concentrations between 25 and sucrose extract of Acanthamoeba (Neff strain) following published 200 PM in Buffer P. procedures (9, 12, 17). The final purification step for conventional The mechanical properties of conventional profilin, profilin profilin (9) was gel filtration on Sephadex G-75 equilibrated with 10 I, and profilin I1 were approximately the same. The magnimM imidazole, pH 7.5, 1.5 mM azide. We will use "profilin" to refer to this mixture of isoforms in this paper. Profilins I and I1 were tudes of G' and 7' for profilin I and I1 were up to two times separated by isocratic cation exchange chromatography on a 1 X 51- that for conventional profilin, and at low frequencies had G, cm column of carboxymethyl-agarose (Bio-Rad CM Bio-Gel A, 100- values = 10 dyne/cm2 in Buffer P (Fig. 2C). The data for 200 mesh) in 10 mM Pipes,' pH 6.5 (12). Fractions were pooled from profilin I1 were taken within 2 days of the final purification
Actin and Profilin Mechanical Properties
10703
O t
-0
20
46
uM
80
68
106
PROFILIN
FIG. 3. Dependence of the dynamic elasticity of nonfilamentous actin on the profilin concentration. Dynamic elasticity at 0.0012Hz for 24 p~ actin in Buffer P (G’ = 0) is plotted as a . function of increasing profilin concentration (0-85 p ~ ) Predicted values for the mixture (G’ = A) are the sums of the elasticities for the individual proteins. These values are expected fortwo ideally noninteracting materials. Each data point corresponds to the calculated mean from 2 to 7 experiments. 01
0
-1
J
3
3 2 1
0 -1
-
-4
-3
-2
Log Frequency
-1
0
1
1
(Hz)
FIG. 2. Dynamic elasticity and dynamic viscosity of profilin as a function of frequency. A , 85 p~ profilin in Buffer P (G’ = 0, 7’ = 0)and Buffer P-2 (G’ = W, q’ = 0 ) .B, 200 b~ profilin in Buffer P (G’ = 0, TJ’= 0). C, 85 &M profilin I (G’ = 0, q’ = 0)and profilin I1 (G’ = W, q’ = 0 ) in Buffer P. The horizontal line is the viscosity (q ’ = 4.03 poise) of the oil (Cannon Instruments, College Park, PA) used to calibrate the rheometer. In these experiments, a constant oscillation amplitude of 0.0046 radians for a maximum strain of 0.083 was used for the frequency range of 0.0002-2 Hz.
step because precipitates were observed in the solution by 8 days after isolation. For comparison, a protein of the same size as profilin, cytochrome c (12,000 daltons), was analyzed in Buffer P and found to be a Newtonian liquid with a viscosity of 0.03 poise with no measurable G‘. Paraffin oil and Buffer P alone were also Newtonian liquids with viscosities of 1.5 and about 0.01 poise, respectively.
1 O t
-4
-3 -2 -1 0 i o g F ~ P C C ~ C Ii ’d ? )
1
FIG. 4. Dynamic elasticity and dynamic viscosity of actin and mixtures of actin and profilin in buffer P. A , 24 p~ actin in buffer P (G’ = 0, q’ = 0).The equilibrium elastic modulus, G,, at Mechanical Properties of Mixtures of Actin a n d Profilinlow frequency was 8 dyne/cm*. No actin filaments were observed in The rheological measurements for mixtures of nonfilamen- this sample by electron microscopy. B, 24 p M actin plus 25 p M profilin tous actin and profilin confirmed previous evidence that the in Buffer P (G’ = 0, q’ = 0).At low frequency, G. = 8 dyne/cm2. C, two proteins interact; the dynamic elasticity and dynamic 24 p M actin plus 56 pM profilin in Buffer P (G’ = 0, TJ’= 0).Log G’ viscosity of the mixturewere much less than the sum predictedis linear with log frequency as measured to 0.0003 Hz. 24 pM actin from the two materials alone (see Figs. 3 and 4). This was plus 85 p~ profilin in Buffer P (G’ = W, q’ = 0).G’ for this mixture true for all concentrations of profilin tested (25-85 p M , see curves downward toward zero with decreasing frequency. The observed tendency of G’ to decrease at low frequency with added profilin Fig. 3), and above 56 p~ profilin the dynamic elasticity tendedwas reproduced for two separate batches of profilin and actin. toward G‘ = 0 a t low frequencies (Fig. 4 0 , indicating that
the mixture was now a viscoelastic liquid rather than a solid like actin alone (Fig. 4A) (4). Thus, high concentrations of profilin (56-85 p M ) caused actin to undergo a phase transition from a gel to a sol. The absence of actin filaments in these samples was confirmed by electron microscopy of negatively stained specimens (20). Profilin had no significant effect on the mechanical properties of actin filaments. That is, mixtures of filamentous
actin (24 p M in Buffer P-2) with 0-85 p M profilin did not show a phase transition behavior from a gel to a sol in our frequency range. The variation in the profilin concentration randomly shifted these plots along the vertical axis by less than a factor of 2. All of these mixtures formed viscoelastic solids with G, that were within *loo% of actin alone (see Fig. 5) and contained numerous actin filaments as observedin
Downloaded from www.jbc.org at Yale University on July 23, 2009
4r
0
10704
Actin and Profilin Mechanical Properties
0
g 2
-4
-3
-2
-1
LogFreauency
0
B
1
(Hz!
FIG.5. Dynamic elasticity and dynamic viscosity of 24 I.LM actin plus 86 p M profilin in Buffer P-2 (G'= 0, q' = 0).G. = 11 dyne/cm2 for this case as compared with G . = 16 & 6 dyne/cm* for filamentous actin alone.
I!
4L i
A
Downloaded from www.jbc.org at Yale University on July 23, 2009
electron micrographs of negatively stained material. Continuous Shear Experiments-Continuous shear experiments confirmed that profilin diminishes the rigidity of nonfilamentous actin but not of filamentous actin. When subjected to continuously increasing strain, mixtures of filamentous actin and profilin responded with a curvilinear increase in stress until the material was disrupted at the yield stress (in this case8.2 dyne/cm2) and thenbegan to flow as a liquid (Fig. 6A). This was seen as an asymptotic drop in stress toa constant value after material disruption. Filamentous actin ) similarly to the mixture (shownin Fig. alone (24p ~ behaved 6A) with a yield stress of 10.1 k 2 dyne/cm2. 200 PM profilin in Buffer P showed a small yield stress of2.2 dyne/cm2 in ) similar Buffer P (Fig. 6 B ) .Nonfilamentous actin (24 p ~ was to conventional profilin (Fig. 6B) but had higher yield stress a 0 ' of 6.5 dyne/cm2. In contrast, mixtures of nonfilamentous actin 0 5 10 15 20 and profilin at the phase transition concentration did not 5;%r S t r o l l - , showareproducible yield stress (Fig. 6C). Thestress for FIG.6. Shear stress as a function of strain. Equilibrated samNewtonian liquids, such as viscosity standard oil, responded ples were strained (deformed) from time 0 at a constant shearrate of as a step function of continuously increasing strain, as seen 0.546 s-' by rotating the bottom rheometer plate at 0.09 rpm. The material shear stresswas monitored by the freely suspended top cone. in Fig. 6D. After cessation of shearing, all samplesof actin andprofilin See Sat0 et al. (4) for details. A , 24 p M filamentous actin plus 85 WM recovered their equilibrium mechanical properties with time, profilin in Buffer P-2. The yield stress was 8.2 dyne/cm2. B , 200 p~ profilin in Buffer P. The yield stress was 2.2 dyne/cm2. C, 24 confirming that a shear-sensitive structure was formed re- nonfilamentous actin plus 55 p~ profilin in Buffer P. The complex versibly in all of these profilin and actin-profilin samples (Fig. stress response supports the conclusion from the frequency data that 7). The recovery time, however,was very sensitive to the this concentration of actin and profilin forms a transitional intershear strain rate and the total deformation (strain). 90% The mediate between solid (gel) and liquid (sol). D,Viscosity standard oil recovery times for profilin and actin-profilin samples were (Cannon Instruments).The stress of this andother Newtonian liquids greater than 40 min and very difficult to reproduce, showing responds as a step function of strain. that these sampleswere more shear-sensitive than actin alone (4). Shear Viscosity of Profilin and Actin-Like actin (4-6, 8), shear viscositiesforprofilin I,conventional profilin, and mixtures of actin and profilin were highly dependent on the shear rate (Fig. 8). Plots of log 7 versus log shear rate were curvilinear butlevelled off asymptotically to a constant value at shear rates >50 s-'. This behavior indicates that profilin and mixturesof actin andprofilin form shear-sensitive (shearthinning) structures which break apart into component subU units at shear rates greater than 50s" (Fig. 8, A and B). Shear viscosity values for 25, 85 and 200 WM profilin in shear Buffer P were nearly identical (Fig. 8 A ) , which parallels the 1 5 0 01 4 6 01 4 2 01 3 8001 3 4 0" ' similarity of dynamic parametersobserved in frequency plots TIME (rnin) for profilin from 25 to 200 pM. All three curves were asympFIG.7. Recovery of the mechanical properties of filamentotic to a shear rate-independent viscosity of 0.04 poise. By tous actin plus profilin after disruption by shearing. 24 p~ comparison, the shear viscosity for nonfilamentous actin (Fig. filamentous actin plus 85 WM profilin in Buffer P-2 (G' = 0).Dynamic elasticity (G') was monitored at 0.6 Hz before and after disruption at 8C) was asymptotic to 0.05 poise at high shear rates. Shear viscosities of nonfilamentous actin andprofilin were a shear rateof 0.546 s-' for 1min, as in Fig. 6A. The arrow indicates the point along the time course at which the sample was sheared. All half the values for profilin at low shear rates.However, above actin and/or profilin samples in Buffer P or P-2 self-healed with time shear ratesof 10 s-', the shearviscosity levelled off to a higher that was dependent on the extent of deformation (strain) and shear value of 0.07 poise, suggesting that larger components exist strain rate. In this case, 90% recovery occurred at about 40 min.
Actin and Profilin Mechanical Properties
10705
3r/
TABLEI Calculation of the concentrations of profilin and actin species The total actin was constant at 24 p ~ and , the concentrations of the other species were calculated assuming a 1:1 complex of profilin with actin having a KDof 5 p ~ .
2
Total profilin
8.7 1.8
Free actin
Free profilin
FM
/IM
PM
PM
0 24 56 85
24.0
0 8.7 35.0 62.8
0
15.3 21.0 3.0 22.2
p * ~ ~ ~ ~ f n
= 16 +. 6 dyne/cm') (4).Second, shear experiments showed
Downloaded from www.jbc.org at Yale University on July 23, 2009
that profilin at equilibrium has a yield stress that compares with actin (4)and with living cytoplasm of Physarum plasmodium (21). Third, shearviscosity experiments showed that profilin is very shear-sensitive when mechanically disturbed but will self-heal with time. Effects of Profilin on the Mechanical Properties of ActinThe observed phase transition of nonfilamentous actin from a gel to a sol can be explained in terms of the low concentration of free actin at highprofilin concentration (Table I). About 10 p~ nonfilamentous actin was required to form an actin gel.' With 56 or 85 p~ profilin, the calculated concen. tration of nonfilamentous actinwas less than 5 p ~Similarly, 0 1 2 3 - 2 - 1 the calculated concentrations of free profilin in these samples Log Shear Rate (5.') ) less than concentrationsof profilin alone (35 and 63p ~ were . these reasons, the mixture FIG. 8. Shear viscosity (t~) as a function of shear rate. A , 85 that failed to gel (e.g. 85 p ~ ) For p M profilin I in BufferP (I)= 0 ) .85 p M conventional profilin (profilin liquified at the concentrationsof profilin that we tested. We I and 11) in Buffer P (I) = 0).B , 24 p~ nonfilamentous actin plus 85 expect that these mixtures may form a viscoelastic solid at p M profilin in Buffer P (I) = 0). C, 24 p M nonfilamentous actin in higher concentrations of profilin, since free profilinalone (200 Buffer P (g = A). 24 P M cytochrome c in Buffer P (7 = A,0.02 It 0.01 ) a gel. Additional datawill be necessary to establish poise). The shear rate independence of the cytochrome c viscosity p ~ forms the rheological properties of the complex of nonfilamentous shows that it is a Newtonian liquid that does not form higher order actin with profilin. The available data are inconclusive but structure. consistent with the complex having weak mechanical propin the mixture than in actin or profilin alone (Fig. 8B). For erties compared to free actin and free profilin. Profilin had no significant effect on the mechanical propcomparison, the shear viscosity of gel-filtered cytochrome c erties of filamentous actin because there was no concentrawas plotted asa function of shear rate in Fig. 8C. Cytochrome c (24 p ~ could ) not be distinguished from Newtonian liquids tion-dependent changein G', q', or q observed for the mixture . think this of filamentous actin and profilin (25-85 p ~ ) We of 0.01 and 0.033 poise on our instrument. interpretation is correct because these concentrations of profilin would reduce the actin filament contentby less than 8% DISCUSSION because the critical concentrationfor actin polymerization is We characterized the effect of profilin on the mechanical PM in Buffer P-2. about 0.1 properties of actin to search for possible roles for profilin in Modeling-The molecular basis of the surprising mechanithecytoplasm.Towardthis goal, equilibratedsamples of amoeba actin and profilin were oscillated and sheared over a cal properties of profilin and nonfilamentous actin is not clear.Onone hand,electron micrographs of theseparate wide range (5 orders of magnitude) that included frequencies rheometer samplesin Buffer P were devoid of actin filaments and rates of deformation analogous to those expected inside and organized profilin structure. On the other hand, these a cell. We found that Acanthamoeba profilin alone is a comsamples behaved like viscoelastic solids (gels). plex viscoelastic material with significantrigidity (elasticity) Apossible explanation is that nonfilamentous rctin and suggesting that it contributes substantially to the mechanical profilin flocculate into branched,aggregate structures similar properties of cytoplasm.Remarkably, we found that even to other nonfilamentous materials including kaolin, polystythough both profilin and nonfilamentous actin alone formed rene beads, boehmite, proteins (casein), and DNA fragments viscoelastic solids, the mixture of the two underwent a phase (23-27). Thesematerials alsoform viscoelastic solids and transition into a viscoelastic liquid. have rheological properties similar to those presented here for Evidence for Profilin Structure-Several lines of evidence profilin and actin. Specifically, emulsions of particles 1-1000 show that Acanthamoebaprofilin formed a structure with nm in diameterform shear-thinning, amplitude-sensitivegels substantialmechanicalpropertiescomparedwithanother (flocs) capable of undergoing phase transitions (28-30). Casmallprotein,cytochrome c. First, at about physiological sein (26) and polystyrene dispersions (24) also exhibit me, conventionalprofilin formed a chanical independence to concentration around the minimum concentrations (200 p ~ (IO)), viscoelastic solid with a G, of 5.5 dyne/cm'. Purified profilin concentration necessary to form a gel. This typeof modelling I and I1 a t 85 I . ~ Malso formed similar viscoelastic solids with may prove useful in understanding the unit interactions inG, values = 10 dyne/cm'. This behavior and the values of these mechanical propertieswere similar to both nonfilamen- ' M. Sato, W. H. Schwarz, and T. D. Pollard, unpublished obsertous actin ( G , = 8 2 3 dyne/cm') and filamentous actin (G, vations.
10706
Mechanical Properties Profilinand Actin
Acknowledgments-We
gratefully thank G. Leimbach and D. Chan
for help in collecting some of the data.
REFERENCES 1. Korn, E. D. (1982) Physiol. Reu. 62,672-737 2. Pollard, T. D., and Cooper, J. A. (1986) Annu. Reu. Biochem. 5 5 , 987-1035 3. Stossel, T. P., Chaponnier, C., Ezzell, R., Hartwig, J. H., Janmey, P., Kuiatkowski, D., Lind, S., Smith, D., Southwick, F. S., Yin, H. L., and Zaner, K. S. (1985) A n n u Reu. Cell Biol. 1, 353-402 4. Sato, M., Leimbach, G., Schwarz, W. H., and Pollard, T. D.
(1985) J . Biol. Chem. 260,8585-8592 5. Jen, C. J., McIntire, L. V., and Bryan, J. (1982) Arch. Biochem. B i o p h y ~216,126-132 . 6. Zaner, K. S., and Stossel, T. P. (1983) J. Biol. Chem. 258,1100411009 7. Schamus, E., Booth, S., Hallaway, B., and Rosenberg, A. (1985) J. Biol. Chem. 260,3724-3730 8. Opperman, W., and Jaberg, B. (1985) Rheol. Acta 2 4 , 525-529 9. Reichstein, E., and Korn, E. D. (1979) J. Biol. Chem. 254,61746179 10. Tseng, P. C., Runge, M. S., Cooper, J. A,, Williams, R. C., Jr., and Pollard, T. D. (1984) J. Cell Biol. 9 8 , 214-221 11. Ampe, C., Vandekerkove, J., Brenner, S. L., Tobacman, L., and Korn, E. D. (1985) J . Biol. Chem. 260,834-840 12. Kaiser, D. P., Sato, M., Ebert, R., and Pollard, T. D. (1986) J . Cell Biol. 1 0 2 , 221-226 13. Pollard, T. D.,and Cooper, J. A. (1984) Biochemistry 2 3 , 66316641 14. Ferry, J. D. (1970) ViscoelasticProperties of Polymers, pp. 11118, John Wiley & Sons, New York 15. Walters, K. (1968) Basic Concepts and Formulae for the Rheogo-
niometer, Sangamo Controls Ltd., Sussex, England 16. Van Wazer, J. R., Lyons, J. W., Kim, K. Y., and Colwell, R. E. (1963) Viscosity and Flow Measurement: A Laboratory Handbook of Rheology, Interscience Publishers, New York 17. Gordon, D. J., Eisenberg, E., and Korn, E. D. (1976) J. Biol. Chem. 251,4718-4786 18. Bradford, M. M. (1976) Anal. Eiochem. 7 2 , 248-254 19. Laemmli, U. K. (1970) Nature 227, 680-685 20. Jockusch, B. M., and Isenberg, G. (1981) Proc. Natl. Acad. Sci. U. S. A . 78,3005-3009 21. Sato, M., Wong, T. Z., and Allen, R. D. (1983) J . Cell Biol. 9 7 , 1089-1097 22. Porter, K. R., Beckerle, M., and McNiven, M. (1983) Mod. Cell Biol. 2,259-302 23. Michaels, A. S., and Bolger, J. C. (1962) I and EC Fundamentals 1,153-162 24. Tadros, T. F. (1984) ACS (Am. Chem. SOC.)Symp. Ser. 240, 411-431 25. Lih, M. (1961) J. Colloid Sci. 1 6 , 297-310 26. Tokita, M., Niki, R., and Hikichi, K. (1984) J. Physiol. SOC.Jpn. 53,480-482 27. Bloomfield, V. A. (1986) ACS (Am. Chem. Soc.) Polymer Preprints 2 7 , 249-250 28. Hirtzel, C. S., and Rajagopalan (1983) ResearchandResearch Needs in Colloidal Interactions, Office of Interdisciplinary Research, National Science Foundation, Wash., D. C. 29. Van de Ven. T. G. M.. and Hunter, R. J. (1979) J. Colloid Interface sci. 6 8 , i35-143 30. Parsegian, V. A. (1982) Adu. Colloid Interface Sci. 1 6 , 49-56
Downloaded from www.jbc.org at Yale University on July 23, 2009
volved in the actin and profilin system and may explain the similar mechanical independence to concentration observed for profilin from 25 to 85 pM in Buffer P. Aggregates (flocs) are generally composed of unit structures called floccules which may themselves beaggregates. Electron micrographs of negatively stained material suggest that actin floccules and profilin floccules may be larger than monomers (4). Structural detailsof the aggregate species are not clear at this time. A floccule is generally able to interact with another particle anywhere on its surface, but this may not be thecase with nonfilamentous actin and profilin. We think that mixtures of nonfilamentous actin and profilin have less than the expected mechanical properties because sites required for selfassociation between actin species and between profilin species were blocked in the complex. Details of these complex interactions are not clearat this time. Interpretation-Actin and profilin are two of the most (10). The prevalentproteinsin Acantharnoebacytoplasm rheological properties measured here should be relevant to living cytoplasm because we included deformations,shear rates, frequencies, and protein concentrations expectedfor the cell. Previously, it was known that profilin modulates the actin pool available to polymerize into filaments (1-3), but our results demonstrate two additional functions for profilin. First, profilin itself can forma viscoelastic solidat physiological conditions and may contribute to the structural integrity of the cytoplasm. Second, high concentrations of profilin can alter the phase state of nonfilamentous actin and so may regulate the rigidity of actin-rich cytoplasm ( 2 2 ) . That is, profilin may be used to make regions of the cytoplasm more liquid (such as the endoplasm) and thereforemore amenable to motility.
