INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Myofibrils are highly differentiated forms of actin cytoskeleton that are specialized for muscle contraction, but the mechanisms by which these complex and precise structures are assembled and maintained are largely unknown. Actin, a major component of thin filaments, has an inherent tendency to polymerize into filaments in vitro. However, the assembly of actin in developing muscle is regulated, and consequently, about 40% of actin is present in a monomeric form (1). In embryonic chicken skeletal muscle, proteins that bind to G-actin to prevent them from polymerization have been identified as profilin (2), actin depolymerizing factor (ADF)¹ (3), and cofilin (4). Quantitative analysis has shown that the concentrations of these three proteins are sufficient for sequestering most of G-actin at a late stage of embryonic muscle (5), suggesting that they are responsible for regulating actin filament assembly.

ADF and cofilin are highly conserved proteins, are members of an ADF/cofilin family having 25-71% homology, and are found in diverse organisms. ADF/cofilin binds to both G- and F-actin at a stoichiometry of 1:1 and regulates the rate of actin polymerization (reviewed in Ref. 6). Recently, ADF/cofilin has been shown to affect the on/off rates at both ends of F-actin, which results in the enhancement of treadmilling (7). This function is necessary for the actin-based motility of Listeria monocytogenes (7, 8) and for actin turnover in cortical actin patches in yeast (9).

A muscle-specific function for ADF/cofilin has been suggested by two examples. These are a mammalian muscle-specific cofilin (M-cofilin) (10) and two ADF/cofilin homologues encoded by the Caenorhabditis elegans unc-60 gene (11). Mammalian M-cofilin is predominantly expressed in skeletal and cardiac muscles (10), but its exact function is unknown. Mutations in the unc-60 gene result in slow moving or paralyzed nematodes (11-13). By electron microscopy, unc-60 mutant muscle has large accumulations of thin filaments especially at the ends of muscle cells but only a few thin filaments scattered among thick filaments that are present in normal numbers and roughly organized into A-bands (12). Thus, unc-60 is required for proper positioning and the correct number of thin filaments in nematode striated muscle. The unc-60 gene has been shown to encode two transcripts, sharing only a single exon encoding the initiator methionine, and two homologous proteins of the ADF/cofilin family (11). These proteins, called UNC-60A and UNC-60B, are 165 and 152 amino acids long, respectively, and are 36% identical and 72% similar. Biochemical studies on members of the ADF/cofilin family in other organisms suggest that the UNC-60 proteins regulate actin polymerization. But, analysis of the primary structures of the UNC-60 isoforms does not allow us to predict how their biochemical properties might be different. We need to know the precise biochemical properties of the UNC-60 proteins to understand the role of unc-60 in muscle development. To address this question, we studied the biochemical characteristics of two UNC-60 proteins, and found that they regulate actin filament dynamics in different manners. The results suggest that two UNC-60 proteins have physiologically distinct functions.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Construction of Expression Vectors for UNC-60A and B-- cDNAs for UNC-60A and UNC-60B were amplified from a C. elegans cDNA library (kindly provided by Dr. R. Barstead, Oklahoma Medical Research Foundation) by polymerase chain reaction with Pfu DNA polymerase (Stratagene, Inc.). Forward and reverse primers for UNC-60A were 5-GATCCATATGAGTTCCGGTGTCATGGTCG and 5-GATTGGATCCCGTGTATCTAGTGATCTCC. Forward and reverse primers for UNC-60B were 5-GATCCCATGGCTTCCGGAGTCAAAGTTG and 5-CTAGGGATCCTTAGATTCTTTGGTTGGACATC. The amplified products for A and B were respectively digested by NdeI-BamHI and NcoI-BamHI, at the sites introduced by the primers, and then cloned into pET-3a and pET-3d (Novagen). The sequences of the inserts were verified by DNA sequencing not to contain any polymerase chain reaction-induced errors.

Preparation of Recombinant UNC-60 Proteins-- Basically, UNC-60A and UNC-60B were produced and purified by the same method. The Escherichia coli strain BL21 (DE3) pLysS carrying an expression vector was grown in LB medium containing 50 µg/ml ampicillin at 37 °C until the A₆₀₀ reached 0.6. Protein expression was induced by adding 0.4 mM isopropyl -D-thiogalactopyranoside for 2 h at 37 °C. The cells were harvested by centrifugation and disrupted by a French pressure cell in a buffer containing 0.1 M NaCl, 1 mM EDTA, 20 mM Tris-HCl, pH 7.5, 1 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride. The lysate was clarified by centrifugation at 20,000 × g for 15 min and fractionated by adding solid ammonium sulfate (50% saturation, 291 g/l). The soluble fraction was thoroughly dialyzed against 20 mM Tris-HCl, pH 8.0, plus 0.2 mM dithiothreitol, applied to DEAE-cellulose (DE-52, Whatman) which had been equilibrated with the same buffer, and eluted with a linear NaCl gradient (0-0.4 M). The fractions containing UNC-60 proteins were concentrated with Centricon 10 or Centriprep 10 (Amicon) and applied to a Sephacryl S-200 column that had been equilibrated with 0.1 M KCl, 10 mM HEPES-NaOH, pH 7.5, 1 mM dithiothreitol. The fractions containing pure UNC-60 proteins were concentrated with Centricon 10 or Centriprep 10.

Assay for F-actin Binding and Depolymerization by Pelleting-- F-actin was prepared from rabbit muscle acetone powder as described (14) and suspended in 0.1 M KCl, 20 mM imidazole, pH 7.0, at a concentration of 10-20 mg/ml. F-actin at 10 µM was incubated with various concentrations of UNC-60 proteins in a buffer containing 0.1 M KCl, 2 mM MgCl₂, 20 mM HEPES-NaOH, pH 7.5, 1 mM dithiothreitol, incubated for 1 h at room temperature, and ultracentrifuged in a Beckman Airfuge at 140,000 × g for 20 min. The supernatants and the pellets were adjusted to the same volume and analyzed by 15% SDS-polyacrylamide gel electrophoresis followed by densitometry.

Light Scattering Measurements-- F-actin (5 µM) in 0.1 M KCl, 2 mM MgCl₂, 20 mM HEPES-NaOH, pH 7.5, 1 mM dithiothreitol was mixed with a final concentration of 5 µM UNC-60A or UNC-60B, and then light scattering at an angle of 90° and a wavelength of 500 nm was measured with a fluorescence spectrophotometer (Perkin-Elmer LS50B).

Cross-linking Assay-- Mixtures of F-actin (10 µM) and UNC-60A or UNC-60B (10 µM) were incubated with or without 15 mM of a zero-length cross-linking reagent, 1-ethyl-3-[3-(dimethyl-amino)propyl]carbodiimide (EDC) for 2 h at room temperature. As controls, F-actin, UNC-60A, or UNC-60B alone were incubated with EDC. They were analyzed by 15% SDS-polyacrylamide gel electrophoresis.

Electron Microscopy-- Samples were negatively stained with 1% uranyl acetate aqueous solution on carbon-supported Formvar-coated grids and observed with a JEOL 100CX electron microscope at an accelerating voltage of 80 kV.

Assay for Actin Polymerization-- G-actin was prepared from rabbit muscle acetone powder as described (4). The time course of actin polymerization was monitored by measuring the absorbance at 237 nm; increased absorbance reflects the G- to F-actin transformation (15, 16). The assay was performed at 25 °C with an Ultrospec 3000 spectrophotometer (Pharmacia Biotech Inc.) equipped with a water-heated 6-cell changer. G-actin (5 µM) was mixed with various concentrations of UNC-60A or B proteins in a buffer containing 0.2 mM ATP, 20 µM CaCl₂, 0.2 mM dithiothreitol, 2 mM Tris-HCl, pH 8.0, and incubated for 15 min. The polymerization was started by adding salt and buffer to a final concentration of 0.1 M KCl, 2 mM MgCl₂, 1 mM EGTA, 20 mM HEPES-NaOH, pH 7.5, and then monitoring the A₂₃₇.

Other Procedures-- Protein concentrations were determined by the Bradford assay (17) using -globulin as a standard. SDS-polyacrylamide gel electrophoresis was performed using 15% gels as described (18).

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Effects of UNC-60 Proteins on F-actin-- Recombinant UNC-60A and B proteins without any extra sequences were bacterially produced, purified (Fig. 1), and used in the following experiments. First, we found that their effects on F-actin were quite different. The actin-binding activities of UNC-60A and B were examined by a co-pelleting assay with preassembled F-actin (Fig. 2). In the presence of UNC-60A, the amount of unpolymerized actin was increased in the supernatant, whereas UNC-60A did not co-sediment with F-actin (Fig. 2A), indicating that UNC-60A primarily depolymerized actin filaments and bound to G-actin. Nearly complete depolymerization was observed when a two molar excess or more of UNC-60A was added (Fig. 2A, graph). In contrast, UNC-60B co-sedimented with F-actin but did not increase the amount of unpolymerized actin (Fig. 2B), showing that UNC-60B bound to F-actin but did not have actin-depolymerizing activity. UNC-60B alone precipitated only at a negligible amount (less than 3% of total protein, data not shown). The binding of UNC-60B to F-actin was saturated at a molar ratio of 1.2:1 (Fig. 2B, graph), suggesting that the stoichiometry of the binding was 1 to 1. ADF/cofilins in vertebrates, yeast, and plants have been shown to have a pH-dependent F-actin-binding/actin-depolymerizing activity; they bind to F-actin at pH 6.5-7.1 and depolymerize actin filaments at pH 7.3-8.3. Therefore, we performed the co-pelleting assay at pH 6.8, 7.5, and 8.0. However, the apparent activities of UNC-60A and B were not changed under the conditions examined, although the actin-depolymerizing activity of UNC-60A was stronger at higher pH (data not shown).

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Purified recombinant UNC-60A and B proteins and an alignment of their sequences. Bacterially expressed UNC-60A (A) and UNC-60B (B) were purified as described under "Experimental Procedures," and 5 µg of each protein was separated by SDS-PAGE with a 15% gel. Molecular mass markers in kDa are indicated on the left. Below is an alignment of UNC-60A and B amino acid sequences performed by ALIGN at the GeneStream Ssearch network server (CRBM, Montpellier, France). Double and single dots indicate identical and similar amino acids, respectively. They are 35.5% identical and 72.3% similar. The underline emphasizes the eight extra residues found in UNC-60A that, based on the yeast cofilin structure, are likely to reside in a -sheet exposed on the surface of the protein.

View larger version (35K): [in this window] [in a new window]   Fig. 2.   Effects of UNC-60A and B on F-actin examined by a pelleting assay. F-actin (10 µM) was incubated with the indicated concentrations of UNC-60A (A) or UNC-60B (B) for 1 h at room temperature. The mixtures were ultracentrifuged, and the supernatants (s) and pellets (p) were analyzed by SDS-PAGE. The position of actin is indicated by an arrowhead. Molecular mass markers in kDa are indicated on the left. Quantitative analysis of the gels are shown on the right. The concentrations of each protein that was contained in the pellets were plotted versus total concentrations of UNC-60A or UNC-60B.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Effects of UNC-60A and B on light scattering of F-actin. F-actin (5 µM) was incubated without (A) or with 5 µM each of UNC-60A (B) or UNC-60B (C), and the changes in the intensities (Int) of light scattering were measured. The scattering of actin alone (A) and actin with UNC-60B (C) was stable within the range of ± 2 for 1 h, while that of actin with UNC-60A kept decreasing to 10 for 30 min (data not shown).

View larger version (92K): [in this window] [in a new window]   Fig. 4.   Effects of UNC-60A and B on F-actin examined by electron microscopy. F-actin (5 µM) was incubated without (A) or with 5 µM each of UNC-60A (B) or UNC-60B (C), negatively stained with uranyl acetate, and observed by electron microscopy. Only short filaments were observed in the presence of UNC-60A (B), while UNC-60B maintained long filaments (C) as well as actin alone (A). Bar, 0.2 µm.

View larger version (55K): [in this window] [in a new window]   Fig. 5.   Chemical cross-linking of actin with UNC-60A or B. A cross-linking assay of actin with UNC-60A and B was performed using EDC as a cross-linker as described under "Experimental Procedures." When actin and UNC-60A (lane 3) or UNC-60B (lane 6) were incubated with EDC, bands of 60 kDa (arrow), which correspond to the sum of the two proteins, appeared. These bands were not detected by incubating actin (lane 1), UNC-60A (lane 2), or UNC-60B (lane 5) alone with EDC or without EDC (lanes 4 and 7). Molecular mass markers in kDa are indicated on the left.

Effects of Inorganic Phosphate on the Interactions of UNC-60 Proteins with Actin-- Both the actin-depolymerizing activity of UNC-60A and the F-actin binding activity of UNC-60B were inhibited by inorganic phosphate (P_i) (Fig. 6). Addition of P_i to mixtures of F-actin and UNC-60A decreased the amount of actin in the supernatant (Fig. 6A). Similarly, in the presence of P_i, the amount of UNC-60B that was pelletable with actin was decreased (Fig. 6B). As a control, increasing the potassium concentration by adding potassium chloride did not affect the activities of the UNC-60 proteins (data not shown). P_i has been shown to bind to F-actin with an affinity of several millimolar (23) and create the state of ADP-P_i-actin which is the intermediate in the hydrolysis of ATP into ADP. Therefore, these results suggest that both UNC-60A and B preferentially bind to ADP-bound actin within the filaments in agreement with the properties of actophorin and plant ADF (7, 22).

View larger version (29K): [in this window] [in a new window]   Fig. 6.   Inhibition of UNC-60A and B activities by inorganic phosphate. F-actin (10 µM) was incubated with 10 µM UNC-60A (A) or UNC-60B (B) in the presence of the indicated concentrations of potassium phosphate (Pi) for 1 h at room temperature. Then, the mixtures were examined by a pelleting assay. Both the supernatants (s) and pellets (p) were analyzed by SDS-PAGE. Molecular mass markers in kDa are indicated on the left. The positions of actin are indicated by arrowheads. Increasing amounts of Pi decreased the amount of actin, which was depolymerized by UNC-60A (A), and the amount of UNC-60B, which was co-sedimented with F-actin (B).

Effects of UNC-60 Proteins on Actin Polymerization-- UNC-60A and B affect the kinetics of actin polymerization in different manners (Fig. 7). Fig. 7A shows that, in the presence of UNC-60A, the rate of polymerization was slowed and the amount of polymerized actin was decreased in a concentration-dependent manner. The early phase of actin polymerization was strongly inhibited by equal molar or 2 molar excess of UNC-60A, suggesting that UNC-60A bound to G-actin, so that it inhibited the nucleation step of actin polymerization and sequestered actin monomer to prevent polymerization. However, two molar excess of UNC-60A transiently enhanced the polymerization rate (from 1,000 to 2,000 s), which may be due to a weak filament severing activity similar to that reported for actophorin (22) and ADF (24, 25).

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Effects of UNC-60A and B on actin polymerization. G-actin (5 µM) was incubated with the indicated molar ratios of UNC-60A (A) or UNC-60B (B) for 15 min at 25 °C, then polymerization was started by adding salt, and the change in the absorbance at 237 nm was measured as described under "Experimental Procedures."

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

The present study shows that the two ADF/cofilin proteins that are generated by alternative splicing of the unc-60 gene have differential functions in regulating actin dynamics. This is the first clear demonstration of functional diversity for the ADF/cofilin family in a single organism. So far, 3 and 2 ADF/cofilin proteins have been found in mammals and chickens, respectively, but their functional differences are not clear. Previously, ADF (or destrin) and cofilin had been characterized as functionally distinct proteins. ADF primarily depolymerizes F-actin and does not bind to F-actin (3, 21, 26, 27), whereas cofilin binds to both G- and F-actin and depolymerizes F-actin in a pH-sensitive manner (4, 16, 28). However, recent refined biochemical studies on the activities of human and chicken ADF have shown that ADF, like cofilin, binds to F-actin at pHs between 6.5 and 7.1 and shows increasing actin-depolymerizing activity at pHs between 7.1 and 8.0 (24, 25).

Although the activities of UNC-60A and B are different, both of them are functionally related to members of the ADF/cofilin family that have been characterized previously. The actin-depolymerizing activity of UNC-60A is quite similar to that of Acanthamoeba actophorin (29), echinoderm depactin (20), and most of ADF/cofilins at alkaline pH. In addition, the profile of the effect of UNC-60A on actin polymerization kinetics is similar to that of chicken ADF (3). These activities can be explained by its strong G-actin binding ability that results in depolymerization of F-actin and sequestering G-actin to inhibit polymerization. The ability of UNC-60B to bind to F-actin has been observed for vertebrate ADF and cofilin at neutral pH (4, 16, 24, 25). However, its lack of ability to depolymerize actin is a unique feature of UNC-60B. The effect of UNC-60B on the kinetics of actin polymerization is consistent with that observed for actophorin (29), chicken cofilin (4), and plant ADF (7). Our results that substoichiometric amounts of UNC-60B accelerated, but that excess UNC-60B inhibited the rate of polymerization, can be interpreted as follows. UNC-60B binds to G-actin to inhibit polymerization, whereas it preferentially binds to F-actin which causes the enhancement of the rate of actin assembly. This effect might be due to the change in the on rate at the barbed end of actin filaments as has been shown for plant ADF (7). The mechanism of the acceleration of polymerization is still unclear. Recently, binding of cofilin to F-actin has been shown to change the twist of the filament (19). Likewise, UNC-60B may cause a conformational change in the actin filament leading to a more competent state for polymerization.

Our results that inorganic phosphate inhibit the activities of both UNC-60A and B strongly suggest that UNC-60A and B preferentially bind to ADP-actin rather than ATP-actin. Preferential binding to ADP-actin appears to be a common property of ADF/cofilin family members (7, 22, 35) and is likely to contribute to an acceleration of the off rate at the pointed end of F-actin (7). Probably, this is the case for UNC-60A because it rapidly depolymerizes F-actin. However, because UNC-60B does not depolymerize F-actin, it implies that UNC-60B-bound F-ADP-actin is physiologically significant. Therefore, it is of interest to examine whether UNC-60B-bound F-actin behaves differently from F-actin alone. Previously, porcine cofilin has been shown to inhibit the binding of tropomyosin and myosin to F-actin (16). Thus, a future goal will be to investigate the ability of UNC-60B-bound F-actin to bind to some other actin-binding proteins.

The differences in the activities of the UNC-60 proteins are likely to result from their structural differences. The sequence identity between UNC-60A and UNC-60B is 36%, which is considerably lower than the group of mammalian ADF/cofilins (three members are 70% identical). However, both UNC-60A and UNC-60B are about 30% homologous to all three members of mammalian ADF/cofilins, and no outstanding similarity to a particular protein was detected. One obvious difference is that UNC-60A is 13 amino acids longer than UNC-60B. The alignment of the two sequences shows that an extra eight amino acids exist in UNC60A from Ile-50 to Asp-57 (Fig. 1, shown by underline). The equivalent region of yeast cofilin consists of the outer most strand of -sheet and is exposed on the surface of the molecule (30), but the function of this region is unknown. In addition, this sequence contains four acidic residues, which are likely to affect the charge distribution on the molecular surface. Because charged amino acids on cofilin have been shown to be important in its actin-binding (31, 32), the extra sequence in UNC-60A may be responsible for its functional difference from UNC-60B.

The differential activities of UNC-60 proteins presented here strongly suggest that the two homologous proteins have physiologically distinct functions. The fine structure genetic map (13) and our preliminary genomic sequencing of viable unc-60 alleles has revealed that all the mutations are found within the coding region for unc-60B,² implying that UNC-60B has a specific role in thin filament assembly in muscle cells. Our results on the effect of UNC-60B on the polymerization kinetics in vitro suggests that, during muscle development, when actin concentration is low initially, UNC-60B inhibits actin polymerization, but later, when actin concentration is high (1), UNC-60B accelerates actin polymerization and thus the formation of thin filaments. This function of UNC-60B is directly relevant to that of ADF/cofilin in vertebrate muscle. ADF/cofilin is involved in the regulation of actin assembly in chicken embryonic muscles (3, 4). In addition, a muscle-type cofilin isoform is expressed in mammalian skeletal and cardiac muscles (10) although its function in muscle cells is not yet clear. UNC-60A may be widely involved in the many processes which require actin dynamics. ADF/cofilin has been shown to be essential for cytokinesis (33, 34) and endocytosis (9), which are universal events in a broad range of cells. Accordingly, C. elegans should have an ADF/cofilin which is expressed in a variety of cells in addition to a muscle-specific isoform. Currently, we are raising antibodies against UNC-60A and UNC-60B and plan to determine tissue distribution of both proteins by immunofluorescence microscopy.