Cell Motility and the Cytoskeleton 21:25-37 (1992)

Evidence for a Gelsol n-Rich, Labile F-Actin Pool in Human Polymorphonuclear Leukocytes Raymond G. Watts and Thomas H. Howard Division of Hematology-Oncology, Department of Pediatrics (R.G.W., T.H.H.), and Department of Cell Biology (T. H.H.), The University of Alabama at Birmingham Filamentous (F) actin is a major cytoskeletal element in polymorphonuclear leukocytes (PMNs) and other non-muscle cells. Exposure of PMNs to agonists causes polymerization of monomeric (G) actin to F-actin and activates motile responses. In vitro, all purified F-actin is identical. However, in vivo, the presence of multiple, diverse actin regulatory and binding proteins suggests that all F-actin within cells may not be identical. Typically, F-actin in cells is measured by either NBDphallacidin binding or as cytoskeletal associated actin in Tritonextracted cells. To determine whether the two measures of F-actin in PMNs, NBDphallacidin binding and cytoskeletal associated actin, are equivalent, a qualitative and quantitative comparison of the F-actin in basal, non-adherent endotoxin-free PMNs measured by both techniques was performed. F-actin as NBDphallacidin binding and cytoskeletal associated actin was measured in cells fixed with formaldehyde prior to cell lysis and fluorescent staining (PreFix), or in cells lysed with Triton prior to fixation (PostFix). By both techniques, F-actin in PreFix cells is higher than in PostFix cells (54.25k3.77 vs. 23.523.7 measured as mean fluorescent channel by NBDphallacidin binding and 70.3?3.5% vs. 47.223.6% of total cellular actin measured as cytoskeletal associated actin). These results show that in PMNs, Triton exposure releases a labile F-actin pool from basal cells while a stable F-actin pool is resistant to Triton exposure. Further characterizations of the distinct labile and stable F-actin pools utilizing NBDphallacidin binding, ultracentrifugation, and electron microscopy demonstrate the actin released with the labile pool is lost as filament. The subcellular localization of F-actin in the two pools is documented by fluorescent microscopy, while the distribution of the actin regulatory protein gelsolin is characterized by immunoblots with antigelsolin. Our studies show that at least two distinct F-actin pools coexist in endotoxin-free, basal PMNs in suspension: 1) a stable F-actin pool which is a minority of total cellular F-actin, Triton insoluble, resistant to depolymerization at 4"C, gelsolin-poor, and localized to submembranous areas of the cell; and 2) a labile F-actin pool which is the majority of total cellular F-actin, Triton soluble, depolymerizes at 4"C, is gelsolin-rich, and distributed diffusely throughout the cell. The results suggest that the two pools may subserve unique cytoskeletal functions within PMNs, and should be carefully considered in efforts to elucidate the mechanisms which regulate actin polymerization and depolymerization in non-muscle cells. Key words: cytoskeleton, human neutrophils, actin binding proteins, cytochalasins, ultracentrifugation

INTRODUCTION

Received June 18, 1991; accepted August 12, 1991.

The basic unit Of the microfilamentous cytoskeleton in PolYmoFhonuclear leukocytes (PMNs) and other non-muscle cells are actin filaments (F-actin)

Address reprint requests to Dr. Raymond G. Watts, Pediatric Hematology-Oncology, The University of Alabama at Birmingham, 1600 7th Avenue South, CHT 651, Birmingham, AL 35233.

0 1992 Wiley-Liss, Inc.

26

Watts and Howard

which are in dynamic equilibrium with monomeric (G) actin and are constantly remodeled through the processes of actin polymerization and depolymerization [Stossel, 1987; Southwick and Stossel, 19831. In vitro, all purified F-actin, regardless of its isotype composition, is identical with respect to sensitivity to monovalent and divalent cations and to ability to interact with actin bindinghegulatory proteins, including force generating, structural, capping, and severing proteins [Pollard and Cooper, 1986; Stossel, 19881. However, in vitro, the higher orders of F-actin organization such as filament length, state of bundling, and network formation via filament crosslinking vary when filaments are exposed to filament severing proteins (e.g., gelsolin) bundling proteins (e.g., alpha actinin), or crosslinking proteins (e.g., filamin) [Pollard and Cooper, 1986; Howard et al., 1990al. Variations in these higher orders of filament organization in vitro lead to the possiblity that F-actin in cells, though of basically identical tertiary and quaternary structure, assume unique and diverse physiochemical characteristics through interactions with actin binding proteins which subserve unique functional roles. Consistent with this notion, F-actin in fibroblasts may appear as F-actin bundles in stress fibers or as a meshwork of filaments in the subcortical regions of undulating lamellipodia [Heath and Holifield, 1991; Byers and Fujiwara, 19821. Furthermore, recent studies in the chemotactic factor activated PMN [Cassimeris et al., 19901and squid axons [Fath and Lasek, 19881 suggest that, even in cells which lack stress fibers, all F-actin is not identical. Adherent, chemotactic factor activated PMNs possess two pools of F-actin which differ in location and in sensitivity to temperature and the cytochalasins, while squid axons contain large numbers of Triton-extractable, short, oligomeric F-actin as well as longer, non-extractable filaments. Identification and characterization of distinct F-actin pools in nonmuscle cells such as those recognized in squid axon and adherent chemotactic factor activated PMNs may be crucial to a more complete understanding of the role of microfilamentous cytoskeletal dynamics in non-muscle cell motile functions. Combination of the unique Triton extractability of an F-actin pool in squid axons and identification of pools of F-actin in substrate adherent, chemotactic factor activated PMNs suggest that the methods classically applied to non-adherent PMNs for quantifying actin polymerization and depolymerization may actually measure distinct, functionally significant pools of F-actin and raise questions as to whether such pools are present in non-activated or basal PMNs. Typically, the state of actin polymerization in basal and ligand activated PMNs and other non-muscle cells is quantified either as F-actin content determined by quantitative fluorescent phallatoxin binding [Howard and Meyer, 19841 or as the amount of actin retained in the

operationally defined Triton insoluble cytoskeleton (TICS) and quantified as cytoskeletal associated actin by gel scans of TICS [Phillips et al., 1980; White et al., 19831. The measurement of F-actin in cells as cytoskeleta1 associated actin (CAA) is based on the premise that cell membrane solubilization with Triton X- 100 defines a sedimentable fraction of the cell as a TICS, the actin retained in this operationally defined cytoskeleton as Factin, and the actin released from the cell into the Triton soluble fraction, Le., the Triton soluble supernatant (TSS), as G-actin. The F-actin in TICS is quantified by gel scans as CAA [Phillips et al., 1980; White et al., 1983; Yassin et al., 19851. In contrast, the measurement of F-actin by quantitative fluorescent phallatoxin binding is based upon saturation binding of phallatoxin to PMNs [Howard and Meyer, 19841 as quantified by flow cytometry. The fluorescent phallatoxin binding assay requires three steps: 1) cell fixation, 2) cell permeabilization, and 3) exposure to the fluorescent phallatoxin NBDphallacidin. To date, careful quantitative comparison of the F-actin measured by these two methods and studies to determine whether the two measures are equivalent are lacking. Such comparisons may be important for defining the mechanism(s) of actin polymerizatioddepolymerization in PMNs and the role of microfilamentous cytoskeletal dynamics in cellular functions. To determine whether the two measures of F-actin in PMNs, NBDphallacidin binding and CAA, are equivalent, a qualitative and quantitative comparison of the F-actin in basal, non-adherent PMNs measured by both techniques was performed. Our studies show that F-actin content in PMNs measured by these two techniques is not identical, suggesting that the two techniques measure distinct F-actin pools: 1) a stable F-actin pool which is a minority of total cellular F-actin, Triton insoluble, resistant to depolymerization at 4"C, gelsolin-poor, and localized to submembranous areas of the cell, and 2) a labile F-actin pool which is the majority of total cellular F-actin, Triton soluble, depolymerizes at 4"C, is gelsolin-rich, and distributed diffusely throughout the cell. The labile and stable-F-actin pools may be similar to the oligomeric pools and stable filaments in squid axons, and to the F-actin pools suggested in adherent, chemotactic factor activated PMNs. Results of these studies not only define unique F-actin pools in PMNs, but also describe a simple yet powerful technique for quantifying the F-actin pools in PMNs independently. The results of gelsolin partitioning and subcellular localization suggest that the two F-actin pools may subserve unique cytoskeletal functions within PMNs, and both should be considered carefully in efforts to elucidate the mechanisms which regulate actin polymerization and depolymerization in non-muscle cells.

Gelsolin-Rich, Labile F-Actin Pool in Human PMNs

MATERIALS AND METHODS

Standard stock chemicals were purchased from J.T. Baker Chemical Co. (Phillipsburg, NJ), Fisher Scientific (Fair Lawn, NJ), or Sigma Chemical Co. (St Louis, MO). Triton X-100 was obtained from Mallinckrodt, Inc. (Paris, KY). NBDphallacidin was supplied by Molecular Probes (Eugene, OR).

27

Cell Suspensions

Pre-Fix

*Fis-stainLyse

Isolation and Purification of Human PMNs

Post-Fix Lyse-Fix k stain Low speed (15,900 g x 2 min)

Human PMNs were isolated under endotoxin-free conditions (ETF) over Percoll gradients as previously described [Howard et al. , 1990bl. Cells were studied in suspension in HankdHEPES buffer (25 mM HEPES, 50 mM phosphate, 150 mM NaCl, 4 mM KCI, pH 7.15) to a final cell concentration of 2 x lo6 cells/experiment. High speed (366,000 g x 5 mins)

Quantification of F-Actin by Flow Cytometry

F-actin content as NBDphallacidin by flow cytometry was determined in fixed PMNs as prevously described [Howard and Meyer, 19841 with the modification that fixation (3.7% formalin) , cell lysis (1% Triton), and NBD staining (NBDphallacidin 1.7 X lo-' M) were performed in three separate steps. All cells were basal ETF human PMNs in suspension. Following fixation, lysis, and staining, 5,000 cells were analyzed on a Becton Dickinson Facstar flow cytometer to produce a quantitative measure of F-actin content, the mean fluorescence channel (MFC), as previously described [Howard and Oresajo, 1985bl. Cells in which fixation preceded lysis are termed PreFix; when lysis preceded fixation, PostFix. Quantification of F-Actin as Cytoskeletal Associated Actin

F-actin was also determined as CAA by a modification of the technique of Phillips et al. [ 1980; White et al., 19831. The CAA is operationally defined as F-actin and is that F-actin retained in the low speed pellet (15,900g x 2 min) following cell lysis with Triton X100 [1% in imidazole 10 mM, KC140 mM, and EGTA 10 mM, pH 7.15, with 7 mM diisopropylfluorophosphate (DFP) to prevent proteolysis] (see Fig. 1). The supernatant remaining following low speed centrifugation is the TSS. In some experiments, the procedure is modified to add fixation with formalin (3.7%) either prior to lysis (PreFix) or following lysis (PostFix). The CAA is then separated and the TSS is concentrated by precipitation with 20% trichloroacetic acid [Peterson, 19831. Samples were solubilized in a Tris buffer (0.625 M in 2% SDS, 10% glycerol, and 5% 2 mercaptoethanol) and evaluated by gradient (5-15%) SDS-polyacrylamide gel electrophoresis (SDS-PAGE) [Laemmli, 19701 utilizing the Mini Protean I1 system (Biorad, Rich-

A

HSP

HSS

Fig. 1 . Schematic of sequential centrifugation technique for separation of cellular fractions. PMNs in suspensions are Triton lysed either prior to (PostFix) or following (PreFix) fixation with formaldehyde. Initial low speed centrifugation (15,900g X 2 min) separates the Triton insoluble cytoskeleton (TICS) from the Triton soluble supernatant (TSS). The TSS is then subjected to high speed centrifugation (366,0008 X 5 min) to separate the high speed pellet (HSP) from the high speed supernatant (HSS).

mond, CA). Protein bands were quantified by densitometric gel scans of Coomassie blue stained gels via laser densitometry (LKB Ultrascan XL, Bromma, Sweden). Purification of Actin

Actin was purified from rabbit skeletal muscle by the technique of Spudich and Watt [1971] and further purified in a chromatographic step over a Sepharose 2B column as previously described [Howard and Lin, 19791. High Speed Sedimentation Technique

TSS following low speed centrifugation is separated into high speed pellet (HSP) and high speed supernatant (HSS) by ultracentrifugation in a TL 100 ultracentrifuge (Beckman, Palo Alto, CA) at either 108,OOOg X 50 min or 366,0008 X 5 min (see Fig. 1). These speeds have been reported to sediment F- but not G-actin [Grazi and Magri, 1987; Miller et al., 19881. To confirm that sedimentation conditions used pellets F-, but not G-actin at the supernatant concentration of our cellular system and buffer, G-actin purified from rabbit skeletal muscle was placed into a 1:1 Hanks/lysis buffer at a concentration of 1 X lo-'' M (calculated concentration for all proteins in TSS of PostFix samples-see below). One

Watts and Howard

28

KD

NBDphallacidin Binding

18

To qualitatively identify the filaments in the HSP as F-actin, the specific binding of NBDphallacidin to the HSP and HSS were compared. The fractions were stained with 3.3 X lop7 M NBDphallacidin for I hr in the dark. The NBD was then extracted into methanol [Howard and Oresajo, 1985a1 and the presence of fluorescence at 510-535 nm evaluated by photon counting on a photon counting spectrofluorometer (SLM 8000C, SLM Industries, Inc., Urbana, IL) following excitation at 465 nm. Results are shown as actual emission spectra. The presence of F-actin as evidenced by NBDphallacidin binding is noted by a fluorescent peak at 510-535 nm.

15

Fluorescent Microscopy

71

28

F

G

STD

Fig. 2. Demonstration of the ultracentrifugation assay for F-actin. Non-polymerized (G-actin) and polymerized samples (F-actin) of purified actin in the cellular buffer system were centrifuged at 366,OOOg X 5 min, the pellets solubilized and evaluated by 5-15% SDS-PAGE. The F-actin sample sedimented while the G-actin remained in sohtion. MW = molecular weight standards; F = F-actin; G = G-actin; STD = 1.5 pg actin standard on gel.

sample of G-actin was induced to polymerize by addition of 150 mM KC1 and 2 mM MgC1, overnight at room temperature [Pollard and Cooper, 19861. The G and F samples were then centrifuged at 366,OOOg X 5 min, the pellet solubilized and run on SDS-PAGE. As shown in Figure 2, the F-actin sedimented while the G-actin did not. This result confirms the specificity of the centrifugation conditions for F-actin. Electron Microscopy For electron microscopy, the HSP was suspended in distilled water, gently vortexed, and stained with 1% uranyl acetate for transmission electron microscopy on a Phillips 301 TEM. Magnifications are noted in figure legends. lmmunoblot Analysis The partitioning of gelsolin between cellular fractions was determined by immunoblot techique [Towbin et al., 19791 in a Mini-Protean transblot system (Biorad) utilizing an anti-human gelsolin monoclonal antibody as primary antibody [Chaponnier et al., 19811 and peroxidase tagged mouse anti-IgG as secondary antibody (Kirkegaard and Perry, Inc., Gaithersburg, MD). Actin was similarly identified using a mouse antiactin monoclonal antibody (Oncogene Science, Manhasset, NY).

For fluorescent microscopy, cell suspensions of PreFix and PostFix samples were stained with 3.3 X M NBDphallacidin, cytospun onto glass slides (1,000 rpm X 5 min; Shandon Cytospin 2, Shandon Southern Products, Ltd., Cheshire, UK), and viewed by fluorescent microscopy (Leitz Vario Orthomat 2, Federal Republic of Germany). RESULTS Comparison of F-Actin Content in Basal PMNs Measured as CAA and by NBDphallacidin Binding

Assay of F-actin content by quantitative NBDphallacidin binding involves cell fixation (with 3.7% formaldehyde), cell permeabilization (with 1 % Triton X- 100 or lysophosphatidylcholine), and staining with NBDphallacidin. To determine the effect on F-actin content in PMNs , the fixatiodlysis sequences was altered to mimic the sequence utilized to measure F-actin as CAA and F-actin content was measured as NBDphallacidin binding. As shown in Figure 3, the F-actin content of basal human PMNs lysed prior to fixation and staining (PostFix cells; 23.5k3.57 mean fluorescent channel), a procedure similar to that used to define CAA in TICS [Phillips et al., 19801, is approximately 50% less than the F-actin content of cells which are fixed prior to lysis and staining (PreFix cells; 54.25k3.77 mean fluorescent channel), the technique utilized to measure F-actin content by NBDphallacidin binding (PreFix cells). The results show that initial fixation retains a quantity of Factin which is lost if cells are lysed prior to fixation and staining. The F-actin lost following exposure to Triton is hereafter referred to as a labile F-actin pool. This labile F-actin pool comprises 50-60% of the total F-actin content of basal PMNs. To assure that a unique labile F-actin pool is released by Triton treatment prior to fixation, and to directly compare measurements of F-actin by NBDphalla-

Gelsolin-Rich, Labile F-Actin Pool in Human PMNs

29

stable F-actin pool; the other, a Triton soluble, labile F-actin pool. The Labile F-Actin Pool Is Released From PMNs as F-Actin

PreFix

PostFix

Fig. 3. F-actin content in basal PMNs measured by NBDphallacidin binding: effect of fixatiodlysis sequence. Shown is the F-actin content of PreFix (fixation with formaldehyde prior to lysis and staining) and PostFix (Triton lysis prior to fixation and staining) basal, endotoxinfree PMNs in suspension measured by NBDphallacidin binding and expressed as mean fluorescence channel of 2 x lo6 cells. Results represent the mean2SD of six experiments. Note that PreFix PMNs have twice the F-actin content of PostFix PMNs indicating that, in basal PMNs, cell lysis prior to fixation results in net loss of F-actin.

cidin binding and CAA, the effect of alterations in fixatiodlysis sequence on F-actin content of PMNs was also assessed by determination of CAA. Cell suspensions of PreFix or PostFix basal PMNs were prepared, the operationally defined TICS was separated, and the Factin content in TICS was quantified as CAA by scans of Coomassie blue stained SDS polyacrylamide gels. As shown in Figure 4a, when F-actin content is measured as CAA the F-actin content of PreFixed cells (70.3+3.5% of total cellular actin, n = 6) was consistently greater than the F-actin content of PostFixed cells (47.223.5% of total cellular actin). Reciprocally, 25% more of the total cellular actin was released into the TSS of PostFixed cells (52.8+3.6% of total cellular actin) than into the TSS of PreFix cells (29.7?3.5% of total cellular actin). As shown in Figure 4b, the amount of actin released into the TSS of PostFixed cells depends upon the duration of Triton lysis, is significant after as little as 15 sec of lysis, and rapidly approaches a maximal value which is 9095% attained by 2-5 min after Triton exposure and 100% by 30 min of lysis. To assure that observations on F-actin content determined as CAA or by NBDphallacidin binding are not explained by stabilization of F-actin by the phallatoxin, measurements of F-actin content as CAA were made in the presence and absence of 1.7 X 10-7M NBDphallacidin in PreFix and PostFix cells. NBDphallacidin had no effect on F-actin content as measured by CAA (data not shown). The results show that lysis of PMNs with Triton causes loss of filamentous actin from the TICS. The loss of F-actin is apparent if F-actin content is measured either by NBDphallacidin binding or as CAA from gel scans. The results show that two pools of F-actin exist in basal PMNs-one is a Triton insoluble,

Triton exposure of PMNs results in extensive dilution of the PMN cytosol and release of the labile F-actin pool from the total cellular F-actin pool. Since F-actin is a dynamic polymer, the labile F-actin pool could be lost either as monomer, if the labile F-actin pool represents F-actin which depolymerizes upon dilution, or as oligomers or filaments. To distinguish between these possibilities, the sedimentation behavior, the morphology, and the NBDphallacidin binding characteristics of the actin released into the TSS of Triton-treated PMNs were determined. To determine the sedimentation behavior of the actin in the TSS of PMNs, TSS of PreFixed and PostFixed cells were subjected to ultracentrifugation at speeds demonstrated to sediment F- but not G-actin [Grazi and Magri, 1987; Miller et al., 19881 (see also Fig. 2). High speed centrifugation sedimented a large, firm pellet from the TSS of PostFix but no pellet from the TSS of PreFix cells (Fig. 5 ) . As shown in Figure 6, SDS-PAGE analysis of the proteins in the HSP from the TSS of PostFix cells shows the pellet is enriched for a 43 kd protein which is actin as evidenced by immunoreactivity with anti-actin antibody on immunoblots (not shown). Quantitatively, 63.7&2.2% of the actin in the TSS of PostFix cells sediments in the HSP. In contrast, no actin sediments from the TSS of PreFix PMNs. Results are identical when a sedimentation force of 366,0008 X 5 min or 108,OOOg X 50 min is utilized and results are similar if M cytochalasin D (CD) is added to the TSS to prevent actin polymerization in the supernatant following lysis (see Fig. 9 below). To determine the morphology of the actin sedimented into the HSP from the TSS of PostFix samples, the HSP was examined by transmission electron microscopy. As shown in Figure 7, filaments and bundles of filaments are identified in the HSP of PostFix cells. In contrast, the extremely small HSP from PreFix cells contains no identifiable filamentous structures (not shown). Furthermore, the NBDphallacidin binding assays of the actin sedimented in the HSP from the TSS of PostFix PMNs demonstrates that the pelleted actin is F-actin. As determined by the NBDphallacidin extraction assay [Howard and Oresajo, 1985a1, NBDphallacidin bound quantitatively to the HSP of PostFix PMNs but not to the HSP of PreFix PMNs or to the HSS of either Pre- or PostFix PMNs (Fig. 8). Since NBDphallacidin binds specifically to F-actin, but not to G-actin, this result shows that the actin released into the TSS from PostFix

30

Watts and Howard KD

106

71 Actin

44

28

---

MW

TICS

TSS

Pre Fix

TICS

TSS

Whole Cell

TICS

TSS

STD

Post Fix

C A

WJ

c 0

Y

E WJ

. I

Y c

-

d

f 0

t

2

-3

2 I=

. I

Y

V

cp

CL

b”

0s

15s 30s 45s 60s 120s 5 m 15m30m Duration of Lysis Prior to Fixation

Fig. 4. a: Analysis of proteins in the TICS and TSS of PreFix and PostFix PMNs by gel electrophoresis. Shown is a representative 5 15% gradient gel of proteins separated from the TICS and the TSS of 2 X lo6 PMNs. Cell suspensions were fixed with formalin before (PreFix) or after (PostFix) exposure to 1% Triton and the TICS separated from the TSS by low speed centrifugation as shown in Figure 1. Proteins in the TSS were precipitated with 20% TCA prior to SDSPAGE analysis. Note that cell lysis prior to fixation (PostFix) results in a loss of actin from the TICS into the TSS when compared to PreFix PMNs. In the whole cell control which is neither lysed nor fixed, no

protein separation occurs with centrifugation. The actin standard (STD) is 1.0 pg. b: Effect of duration of exposure to Triton on quantity of F-actin released from the TICS. Shown is the amount of actin released from the TICS into the TSS of PostFix samples as a function of duration of Triton lysis prior to fixation. The F-actin released is expressed as a percentage of total F-actin lost measured by NBDphallacidin binding. Note that Triton exposure results in a rapid loss of actin from the TICS into the TSS which reaches a maximum by 2-5 min at 25°C. Similar results are obtained if the actin released from the TICS is measured by laser densitometry.

PMNs is F-actin. In contrast, actin in the HSS does not bind NBDphallacidin. Finally, quantitative studies show that the quantity of F-actin released by Triton from the TICS of PostFix cells as labile F-actin (23.1-30.1 % of total cellular actin)

is similar to the amount of actin pelleted by high speed centrifugation to form the HSP (34% of total cellular actin). The results demonstrate that the labile F-actin pool is a true quantifiable entity and the labile F-actin pool is released into the TSS as filaments. However,

Gelsolin-Rich, Labile F-Actin Pool in Human PMNs

Fig. 5 . Comparison of HSP sedimented from the TSS of PreFix and PostFix PMNs. Shown are photographs of HSP sedimented from the TSS of 1 X lo7 PostFix (A) or PreFix (B) PMNs by centrifugation at 366,0008 X 5 minutes at 25°C. Note that lysis prior to fixation (A) results in a visible pellet with high speed centrifugation, while fixation prior to lysis (B) results in no visible pellet at high speed.

31

in the TSS does not alter the amount of actin pelleted from the TSS of PostFix PMNs into the HSP. This result suggests that the labile F-actin pool is actually released from PMNs as F-actin and is not released as monomer which subsequently polymerizes in the TSS to F-actin. Furthermore, the concentration of actin in the TSS of PostFix PMNs was determined and is far below the critical concentration (Ccr) for actin polymerization (0.37 pM) in the buffer system of the TSS [Southwick and Young, 19901. The total protein, total volume, and percent of total protein as actin in the TSS from 2 X lo6 PostFix PMNs are 84 pg, 400 p1, and 13.2% respectively. Therefore the molar concentration of actin in the TSS is no more than 2 X lo-" M [(8.4 X 10-6g X 0.13214.2 X lo3 g/mole)/4 x 11, a value well below the Ccr for filament assembly in the TSS buffer. These results show that the F-actin in the labile F-actin pool is not formed by polymerization of actin in the TSS following cell lysis. Paired with results on sedimentation, NBDphallacidin binding, and morphology (Fig. 68), these results establish that the labile F-actin pool is released into the TSS as oligomers or filaments.

KD

The Labile F-Actin Pool Is Cold Sensitive: The Stable F-Actin Pool Is Cold Insensitive

71

Actin 44

24

18

MW

TICS

HSP

I

HSS TICS

HSP

Pre Fix

HSS STD 1

I I

Post Fix

Fig. 6. Analysis of proteins in the HSP and HSS of PreFix and PostFix PMNs by SDS-PAGE. Shown is a 5-15% gradient gel of proteins from the TICS, HSP, and HSS fractions (see Fig. 1) of 2 X lo6 PreFix (lanes 1-3) and PostFix (lanes 4-6) PMNs. Note that no protein is sedimented in the HSP of PreFix samples, while an actin-enriched protein fraction is found in the HSP of PostFix samples. Lane 7 represents a 1.0 pg actin standard.

these results do not exclude the possibility that the labile F-actin pool is a dilutionally depolymerizable pool of F-actin which is lost as monomer into the TSS and subsequently polymerizes to form F-actin in the TSS. To exclude this possibility the effect of CD addition on quantity of actin pelleted by high speed centrifugation and the monomer concentration in the TSS relative to the critical concentration for actin polymerization in the TSS buffer system were determined. M CD an As shown in Figure 9, addition of inhibitor of actin polymerization [Fox and Phillips, 19811, prior to cell lysis and maintenance of lop5M CD

In chemotactic factor activated, adherent PMNs or PMNs in suspension, a significant portion of the total F-actin is depolymerized by exposure of PMNs to 4°C [Howard and Meyer, 1984; Cassimeris et al., 19901. To determine whether the labile and stable F-actin pools in basal PMNs are differentially sensitive to cold depolymerization, F-actin content was quantified in PreFix and Postfix cells at both 25 and 4°C. Cooling of PreFix PMNs to 4°C prior to fixation decreases F-actin content to 18.OOk3.9 mean fluorescence channel, n = 4 , which is 25% of the F-actin content of PreFix PMNs at 25°C (F-actin content 52.1422.9) yet similar to the F-actin content of basal PostFix PMNs at either 25 or 4°C (Fig. 10). Simlilar results are obtained when F-actin is quantified as CAA by gels and demonstrate a loss of actin in the HSP (labile F-actin pool) and a gain of actin in the HSS (G-actin pool) (not shown). The results show that the temperature-induced fall in F-actin content is the result of depolymerization of the labile F-actin pool and demonstrate differential susceptibility to cold depolymerization between the F-actin pools. The labile F-actin pool is cold sensitive; the stable F-actin pool is cold insensitive. The Labile F-Actin Pool Is Gelsolin-Rich: The Stable F-Actin Pool Is Gelsolin-Poor

Gelsolin is a 90 kd protein which in vitro in the presence of micromolar Ca2+ binds to, and caps, the barbed end of actin filaments, thus preventing monomer

32

Watts and Howard

Fig. 7 . Morphology of the HSP from the TSS of PostFix PMNs. Shown are transmission electron micrographs of the HSP sedimented from the TSS of PostFix PMNs. Samples were processed for transmission electron microscopy as described in Materials and Methods.

Note typical filaments of F-actin. Samples of PreFix HSP or HSS of either Pre- or PostFix cells failed to demonstrate any filaments. Bar = 100 nM. X 264,000.

addition at the barbed end. Gelsolin is believed to play an important regulatory role in actin polymerization of PMNs [Howard et al., 1990al. In basal PMNs, a significant fraction of gelsolin (85% in Ficoll-Hypaque purified PMNs and 35% in ETF PMNs) is bound to F-actin as a 1:l EGTA resistant complex [Pollard and Cooper, 1986; Janmey et al., 1985; Chaponnier et al., 1987; Howard and Chaponnier, 19901. To determine whether gelsolin is differentially partitioned between the labile and stable F-actin pools in PMNs the total F-actin pool, the labile F-actin pool, the stable F-actin pool, and the G-actin pool were analyzed for the presence of gelsolin by immunoblots of PreFix and PostFix PMN TICS, HSP, and HSS. As shown in Figure 11, gelsolin is present in PreFix PMN TICS and in the HSP and HSS of PostFix PMNs. In contrast, gelsolin is not found in the PostFix TICS and PreFix HSP. Since the stable F-actin pool is retained in PostFix TICS while labile F-actin is lost into the TSS and sediments in the HSP from the TSS of PostFix PMNs, the

result shows that gelsolin is present in the labile but not the stable F-actin pool. Differential partitioning of a given actin regulatory protein (such as gelsolin) with either the labile or stable F-actin pools implies a functional role for that protein in determining the effect or distribution of the actin pools. The results show that the labile and stable F-actin pools differ qualitatively with respect to presence or absence of gelsolin. The labile F-actin pool is gelsolin-rich while the stable F-actin pool is gelsolin-poor. Cellular Localization of the Stable and Labile F-Actin Pools in PMNs

Previous results in chemotactic factor activated PMNs suggested that the two pools of F-actin were differentially localized to unique anatomical and potentially functional regions of the PMN. To determine whether the unique F-actin pools which differ in retention of gelsolin in basal, non-adherent PMNs are also uniquely distributed within the PMNs, we localized the stable and labile

Gelsolin-Rich, Labile F-Actin Pool in Human PMNs 100

1

P

0

PreFix HSP

n 500

Buffer

p

525

550 Wavelength (nm)

I

575

600

Fig. 8. NBDphallacidin binding to the HSP of PostFix PMNs. Shown is the NBDphallacidin binding spectrum of PreFix and PostFix PMN fractions. The HSP (labile F-actin pool) of PostFix cells contains NBDphallacidin binding activity as evidenced by a fluorescent peak at 510-535 nm (excitation 465 nm). The HSS of PreFix or PostFix PMNs and the cell control (cell suspensions to which no NBD is added) show minimal non-specific fluorescence, while the buffer control shows no fluorescence. The NBDphallacidin binding of PreFix HSP (in which no pellet is evident) is similar to that of buffer.

F-actin pools in PMNs by comparing the distribution of F-actin in NBDphallacidin-stained PostFix and PreFix PMNs by fluorescence microscopy. Results are shown in Figure 12. In PMNs which retain both the stable and labile F-actin pools (PreFix PMNs), the F-actin is diffusely distributed throughout the basal, non-adherent PMN. In PostFix PMNs, which retain only the stable F-actin pool, F-actin is limited to the subcortical region of basal nonadherent PMNs. The results show that the stable F-actin pool is localized to subcortical regions of the PMN. In contrast, the labile F-actin pool is lost symmetrically from all areas of the cell, suggesting diffuse distribution of the labile F-actin pool within basal, non-adherent PMNs. DISCUSSION

The studies reported in this paper provide an improved understanding of the microfilamentous cytoskeleta1 organization in PMNs and have important implications for studies which seek to elucidate the mechanisms of regulation of actin polymerization in PMNs. Our findings show that in basal PMNs, actin exists in at least three states: stable F-actin, labile F-actin, and G-actin. The stable F-actin pool accounts for a minority of F-actin in a basal cell (10-30% of total cellular actin and 3540% of the total cellular F-actin after 5 min of exposure to Triton at 25"C), is retained in the TICS following cell membrane permeabilization, does not depolymerize at

33

4"C, is localized to submembranous areas of the cell and lacks gelsolin. The stable F-actin pool is analogous to CAA, a commonly utilized measure of F-actin content in PMNs and other cells [White et al., 1983; Southwick and Young, 1990; Yassin et al., 1985; Wallace et al., 19871. This stable F-actin pool likely represents the basic microfilamentous cytoskeletal core as demonstrated by electron microscopy of Triton-exposed PMNs and platelets [Ryder et al., 19841. The labile F-actin pool accounts for 40-45% of total cellular actin and 60-65% of total F-actin in basal cells (after 5 min of cell lysis at 25"C), is released from the TICS into the TSS as F-actin, is gelsolin-rich, is sensitive to depolymerization at 4"C, and is diffusely distributed throughout the basal cell. Presumably this labile F-actin pool consists of short oligomers or filaments similar to those released from Triton-treated squid axons. The remainder of the actin, G-actin, accounts for 30% of total cellular actin, is not sedimented by low or high speed centrifugation, and does not bind NBDphallacidin. F-actin pools of differing characteristics have previously been recognized in squid axons [Fath and Lasek, 19881 and in adherent, activated PMNs [Cassimeris et al., 19901. Analogous to squid axons, the microfilamentous cytoskeleton of basal PMNs is organized as a subcortical rim of F-actin (stable F-actin pool) and a much larger cytoplasmic pool of actin which exists predominantly as a labile pool of F-actin. These cytoplasmic actin pools are available for use in cytoskeletal remodeling following cellular activation. Two distinct F-actin pools were also reported in adherent, chemotactic factor activated PMNs [Cassimeris et al., 19901. Cassimeris et al. [1990] reported that a labile F-actin pool is localized to areas of filament growth or turnover following FMLP activation and is lost following removal of chemoattractant, addition of cytochalasin B, or cold exposure in activated, adherent PMNs, while F-actin content in basal, adherent PMNs is insensitive to cold exposure or cytochalasin B. Their results suggest that two pools of F-actin exist only in activated adherent cells. In contrast, studies reported here with basal, endotoxin free PMNs in suspension utilize a simple, direct measure of F-actin and gelsolin partitioning in Tritonized cells to show that two distinctive F-actin pools exist in non-adherent, basal PMNs in suspension, that the labile F-actin pool is lost as F-actin, and that the labile, but not the stable F-actin pool contains gelsolin. These observations extend the original description of the two F-actin pools and define a simplified and readily applicable method for direct quantification of the F-actin pools in PMNs. Presence of a labile F-actin pool in basal, non-adherent PMNs and the absence of a labile F-actin pool in basal, adherent PMNs [Cassimeris et al., 19901 suggest distinct differences may

34

Watts and Howard

KD 196

I06 71 44

ACTIN

28 18

MW TICS HSP HSS

- CD

+ CD

Fig. 9. F-actin in the HSP is not formed by polymerization of monomer: evidence with cytochalasin D. Shown is a 5-15% gradient SDSPAGE gel of proteins from the TICS, HSP, and HSS fractions (see Fig. 1) of 2 x lo6 PostFix PMNs lysed in the presence (+CD) or M cytochalasin D. Cytochalasin D is added to absence (-CD) of

60,

0

I

PreFix 25 PostFix 25

PreFix 4

TICS HSP HSS STD

PostFix 4 0

Fig. 10. The effect of cold exposure on the stable and labile F-actin pools. Shown is the F-actin content of 2 X lo6 basal PreFix or PostFix PMNs incubated at 25 or 4°C measured by NBDphallacidin binding and expressed as mean fluorescence channel. The results represent the mean? SD of four separate experiments. Note that exposure of PreFix PMNs to 4°C results in a fall in cellular F-actin content to the level of PostFix PMNs incubated at either 25 or 4°C. The result suggests that exposure of PreFix PMNs to 4°C results in depolymerization of the labile F-actin pool. The stable F-actin pool (PostFix PMNs) remains resistant to cold depolymerization.

exist in F-actin pools in adherent and non-adherent PMNs [Southwick et al., 19901. Since a Triton soluble, gelsolin-rich, labile F-actin pool, a Triton insoluble, gelsolin-poor, stable F-actin pool, and G-actin monomers all exist simultaneously in the PMN, actin polymerization can no longer be concep-

prevent polymerization of actin following cell lysis. Note that no significant difference in actin distribution is noted with or without CD. +CD = TICS 27%, HSP 48%, HSS 25% of actin; -CD = TICS 29%, HSP 42%, HSS 29%.

tualized as a simple equilibrium between a monomer and a single class of polymer. Now, existence of two distinct F-actin pools in equilibrium with monomer suggests additional unique mechanisms for actin polymerization and depolymerization in PMNs. These mechanisms include the possibility that both stable and labile F-actin pools may grow or shrink simultaneously or independently, with or without monomer addition, with or without net change in amount of free monomer, and with or without a net increase or decrease in the amount of total F-actin content within the cell. The distinct Triton solubility and insolubility of F-actin in the two F-actin pools imply both potential unique functions for the two F-actin pools and a unique and distinctive difference, not in the actin itself, but in the actin regulatory and binding proteins associated with the two F-actin pools. The observations on differential partitioning of gelsolin between the two Factin pools demonstrate asymmetric distribution of at least one such actin regulatory protein between the two F-actin pools. Ultrastructural evidence supports involvement of a gelsolin-associated F-actin pool in chemotactic factor mediated actin polymerization since gelsolin is localized by immunogold technique to the cytoplasm of resting macrophages and platelets in association with short actin filaments [Hartwig et al., 19891. With cellular activation, the gelsolin assumes a submembranous location suggesting transposition of the short actin oligomers (analogous to labile F-actin) to the membrane surface.

Gelsolin-Rich, Labile F-Actin Pool in Human PMNs

35

KD 89KD

106 71

44

28

18

MW

TICS

HSP

I

HSS

TICS I

HSP

Pre Fix

HSS

STD I

I

Post Fix

Fig. 11. Partitioning of gelsolin between the labile and stable F-actin pools. PMN fractions produced as described in Figure 1 were evaluated by immunoblot for reactivity with an anti-gelsolin monoclonal antibody as described in Materials and Methods. In Prefix samples, all of the gelsolin is found in the TICS, while in PostFix samples, gelsolin is excluded from the TICS and localizes to the HSP and HSS. STD is a 1.5 pg gelsolin standard.

Fig. 12. Subcellular localization of stable and labile F-actin pools. Shown are fluorescent photomicrographs of NBDphallacidin (3.3 X lo-’ M)-stained PreFix (stable and labile) and PostFix (stable) PMNs. Note that PreFix cells containing both stable and labile F-actin pools

demonstrate diffuse staining with NBDphallacidin. In contrast, PostFix cells containing only the stable F-actin pool show staining localized to the submembranous areas of the cell with a diffuse loss of fluorescence in the remainder of the cell.

Direct demonstration that the stable F-actin pool is analogous to CAA, i.e., F-actin in the TICS, has important implications for interpretation of experiments which seek to elucidate the mechanisms for regulation of actin

polymerization in PMNs without considering both pools of F-actin. For example, studies using CAA as a measure of total cellular F-actin may not be fully representative of the dynamics of actin polymerization and depolymeriza-

36

Watts and Howard

tion since, as classically measured, CAA contains only the stable F-actin to the exclusion of labile F-actin. Studies utilizing F-actin content in PMNs measured by NBDphallacidin binding or as CAA must therefore take into account the fact that the two measures of F-actin are not equivalent, that conditions and timing of fixation, lysis, and cell staining are critical to interpretation of NBDphallacidin binding results, and that CAA measurements as classically performed exclude a large F-actin pool released into the TSS. Finally, the techniques described here for quantification and characterization of the labile F-actin pool, stable F-actin pool, and G-actin pools are simple, reproducible, and avoid the pharmacologic manipulation with agents such as the cytochalasins required by other techniques of separation [Cassimeris et al., 19901. Concurrent quantification of labile F-actin, stable F-actin, and G-actin provides a powerful tool to study actin polymerization within specific pools and holds the potential for allowing improved understanding of the mechanisms for regulation of actin polymerization in PMNs. The assay allows the possibility of examining the response of specific pools to transmembrane signaling, and investigation of the role of specific actin binding and regulatory proteins in determining cytoskeletal organization in PMNs . ACKNOWLEDGMENTS

This work was supported in part by grants K11 HL02601-01 and NIAID A125214 from the National Institutes of Health and IRG 66-31 from the American Cancer Society. R.G.W. is a Dixon Fellow of the Dixon Foundation, Birmingham, Alabama. T.H.H. is an Established Investigator of the American Heart Association. REFERENCES Byers, H.R., Fujiwara, K. (1982): Stress fibers in cells in situ: Immunofluorescence visualization with antiactin, antimyosin, and anti-alpha-actinin. J. Cell Biol. 93:804-811. Cassimeris, L. McNeill, H., and Zigmond, S.H. (1990): Chemoattractant stimulated polymorphonuclear leukocytes contain two populations of actin filaments that differ in their spatial distributions and relative stabilities. J. Cell Biol. 110:1067-1075. Chaponnier, C., Janmey, P.A., and Yin, H.L. (1981): The actin filament-severing domain of plasma gelsolin. J. Cell Biol. 103: 1473-81. Chaponnier, C.H., Yin, H.L., and Stossel, T.P. (1987): Reversibility of gelsolin/actin interaction in macrophages: Evidence of Ca’ + independent pathways. J. Exp. Med. 165:97-106. Fath, K.R., and Lasek, R.J. (1988): Two classes of actin microfilaments are associated with the inner cytoskeleton of axons. J. Cell Biol. 107:613-621. Fox, J.E.B., and Phillips, D.R. (1981): Inhibition of actin polymerization in blood platelets by cytochalasins. Nature 292:650652.

Grazi, E., and Magri, E. (1987): Kinetic heterogeneity of F-actin polymers. Biochem. J. 248:721-25. Hartwig, J.H., Chambers, K.A., and Stossel, T.P. (1989): Association of gelsolin with actin filaments and cell membranes of macrophages and platelets. J. Cell Biol. 108:467-479. Heath, J.P., and Holifield, B.F. (1991): Cell locomotion: New research tests old ideas on membrane and cytoskeletal flow. Cell Motil. Cytoskeleton 18:245-257. Howard, T.H., and Chaponnier, C. (1990): Gelsolin regulates, but does not create, plus end nuclei during FMLP-induced actin polymerization in neutrophils (PMNs). J. Cell Biol. 111 (5)(Part 2):162a. Howard, T.H., and Lin, S . (1979): Specific interaction of cytochalasins with muscle and platelet actin filaments in vitro. J. Supramol. Struct. 11:283. Howard, T.H., and Meyer, W.H. (1984): Chemotactic peptide modulation of actin assembly and locomotion in neutrophils. J. Cell Biol. 98:1265-71. Howard, T.H., and Oresajo, C.O. (1985a): A method for quantifying F-actin in chemotactic peptide activated neutrophils: Study of the effect of tBOC peptide. Cell Motil. 5545-557. Howard, T.H., and Oresajo, C.O. (1985b): The kinetics of chemotactic peptide-induced change in F-actin content, F-actin distribution and the shape of neutrophils. J. Cell Biol. 101:107885. Howard, T., Chaponnier, C., Yin, H., and Stossel T. (1990a): Gelsolin-actin interaction and actin polymerization in human neutrophils. J. Cell Biol. 110:1983-1991. Howard, T.H., Wang, D., and Berkow, R.L. (1990b): Lipopolysaccharide modulates chemotactic peptide induced actin polymerization in neutrophils. J. Leukocyte. Biol. 47:13-24. Jammey, P.A., Chaponnier, C., Lind, S.E., Zaner, K.S., Tosssel, T.P., and Yin, H.L. (1985): Interactions of gelsolin and gelsolin-actin complexes with actin: Effects of calcium on actin nucleation, filament severing and end blocking. Biochemistry 24: 37 14-3723. Laemmli, U.K. (1970): Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680685. Miller, L., Phillips, M., and Reisler, E. (1988): Polymerization of G-actin by myosin subfragment I. J. Biol. Chem. 263:19962002. Peterson, G.L. (1983): Determination of total protein. In Colowick, S.P., and Kaplan, N.O. (eds.); “Methods of Enzymology, Vol 91.” San Diego; Academic Press, Inc., pp. 95-99. Phillips, D.R., Jennings, L.K., and Edwards, H.H. (1980): Identification of membrane proteins mediating the interaction of human platelets. J. Cell Biol. 86:77-86. Pollard, T.D., and Cooper, J.A. (1986): Actin and actin binding proteins. A critical evaluation of mechanisms and functions. Ann. Rev. Biochem. 55:987-1035. Ryder, M.I., Weinreb, R.N., and Niederman, R. (1984): The organization of actin filaments in human polymorphonuclear leukocytes. Anat. Rec. 209:7-20. Southwick, F.S., and Stossel, T.P. (1983): Contractile proteins in leukocyte function. Semin. Hermatol. 20:305-321. Southwick, F.S., and Young, C.L. (1990): The actin released from profilin-actin complexes is insufficient to account for the increase in F-actin in chemoattractant stimulated polymorphonuclear leukocytes. J. Cell Biol. 110:1965-73. Southwick, E.S., Dabiri, G.A., Paschetto, M., and Zigmond, S.H. (1990): PMN adherence induces actin polymerization by a transduction pathway which differs from that used by chemoattractants. J. Cell Biol. 109:1561-69.

Gelsolin-Rich, Labile F-Actin Pool in Human PMNs Spudich, J.A., and Watt, S . (1971): The regulation of rabbit muscle contraction: Biochemical studies of the interaction of the tropomyosin-troponin complex with actin and the proteolytic fragments of myosin. J . Biol. Chem. 246:4866-4871. Stossel, T.P. (1987): The molecular biology of phagocytes and the molecular basis of non-neoplastic phagocyte disorders. In Stamatoyannopoulas, G., Nienhuis, A. W., Leder, P., Majerus, P.W. (eds.); “The Molecular Basis of Blood Diseases.” Philadelphia; W.B. Saunders Co., pp. 499-533. Stossel, T.P. (1988): The mechanical responses of white blood cells. In Gallin J.I., Goldstein, I.M., Snyderman, R. (eds.); “Inflammation: Basic Principles and Clinical Correlates. ” New York: Raven Press, pp. 325-342. Towbin, H . , Staehelin, T., and Gordon, J . (1979): Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose

37

sheets: Procedure and some applications. PNAS U.S.A. 76: 4350-4354. Wallace, P.J., Packman, C.H., Wersto, R.P., and Lichtman, M.A. (1987): The effects of sulfiydryl inhibitors and cytochalasin on the cytoplasmic and cytoskeletal actin of human neutrophils. J. Cell. Physiol. 132:325-330. White, J.R., Naccache, P.H., and Sha’afi, R.I. (1983): Stimulation by chemotactic factor of actin association with the cytoskeleton in rabbit neutrophils. J. Biol. Chem. 258:14041-14047. Yassin, R., Shefcyk, J., White, J.R., Tao, W., Volpi, M., Molski, T.F.P., Naccache, P.H., Feinstein, M.P., and Sha’afi, R.I. (1985): Effects of chemotactic factors and other agents on the amounts of actin and a 65,000 molecular weight protein associated with the cytoskeleton of rabbit and human neutrophils. J. Cell Biol. 101:182-188.

Evidence for a gelsolin-rich, labile F-actin pool in human polymorphonuclear leukocytes.

Filamentous (F) actin is a major cytoskeletal element in polymorphonuclear leukocytes (PMNs) and other non-muscle cells. Exposure of PMNs to agonists ...
1MB Sizes 0 Downloads 0 Views