ARPC1/Arc40 mediates the interaction of the actin-related protein 2 and 3 complex with Wiskott-Aldrich syndrome protein family activators.

The actin-related protein 2 and 3 (Arp2/3) complex is a seven-subunit protein complex that nucleates actin filaments at the cell cortex. Despite extensive cross-linking, crystallography, genetic and biochemical studies, the contribution of each subunit to the activity of the complex remains largely unclear. In this study we characterized the function of the 40-kDa subunit, ARPC1/Arc40, of the yeast Arp2/3 complex. We showed that this subunit is indeed a stable component of the Arp2/3 complex, but its highly unusual electrophoretic mobility eluded detection in previous studies. Recombinant Arc40 bound the VCA domain of Wiskott-Aldrich syndrome protein family activators at a K(d) of 0.45 mum, close to that of the full complex with VCA (0.30 microm), and this interaction was dependent on the conserved tryptophan at the COOH terminus of VCA. Using a newly constructed Delta arc40 yeast strain, we showed that loss of Arc40 severely reduced the binding affinity of the Arp2/3 complex with VCA as well as the nucleation activity of the complex, suggesting that Arc40 contains an important contact site of the Arp2/3 complex with VCA. The Delta arc40 cells exhibited reduced growth rate, loss of actin patches, and accumulation of cables like actin aggregates, phenotypes typical of other subunit nulls, suggesting that Arc40 functions exclusively within the Arp2/3 complex.

The Arp2/3 1 complex is a highly conserved actin regulator that nucleates branched actin network associated with regions of the plasma membrane (1,2). This complex contains seven subunits, including two actin-related proteins Arp2 and Arp3 and five novel polypeptides named ARPC1-5 (3)(4)(5). Purified Arp2/3 complex is inactive but stimulates rapid actin polymerization in the presence of activator proteins, the most potent type being the VCA domain-containing Wiskott-Aldrich syndrome protein (WASP) family members (1,2). It is hypothesized that the Arp2/3 complex binds VCA and an existing actin filament in a cooperative manner, inducing a major conforma-tional change, which brings Arp2 and Arp3 to an actin dimerlike arrangement and allows recruitment of an actin monomer tethered through the VC region of VCA, leading to formation of a new filament (6).
Chemical cross-linking, genetic, yeast two-hybrid, and reconstitution experiments and crystallography have converged on a unified model of subunit organization within the complex: ARPC2 and ARPC4 form a central scaffold, and the rest of the subunits are organized around this core structure (6). Despite this wealth of data, we currently still lack direct insights into the nature of the conformation change induced by contacts with VCA and F-actin and how various subunits contribute to this change. Chemical cross-linking experiments showed that three subunits, Arp2, Arp3, and ARPC1, are in close contact with VCA (7,8). However, these results do not necessarily imply that all three subunits contain binding sites for VCA or indicate the extent to which each subunit contributes to the interaction between the complex and VCA. Yeast two-hybrid studies identified ARPC3 as a subunit that could bind VCA; however, it has not been confirmed whether this observed interaction is direct (9). ARPC1 has been a particularly enigmatic subunit. Whereas this subunit is present in the Arp2/3 complex from various organisms, Arc40, the budding yeast ARPC1, was apparently absent from most preparations of the purified yeast complex (5). Since the yeast complex preparations demonstrate a VCAdependent actin nucleation activity comparable to that of the bovine Arp2/3 complex, the lack of ARPC1 implied that this subunit may not be required for the activity of the Arp2/3 complex. Despite the apparent lack of ARPC1/Arc40 in the purified yeast complex, it was clear that Arc40 can associate with the Arp2/3 complex in yeast (10) and appeared to be present in at least some preparations of the yeast complex (11). Moreover electron cryomicroscopy reconstruction showed that the yeast complex was nearly identical to the bovine complex, and the two reconstructions clearly did not show any difference as large as a missing 40-kDa subunit (12). Another unexpected result on Arc40 came from genetic analysis of yeast mutants bearing deletions of each of the Arp2/3 subunits (10). Unlike the other Arp2/3 subunit nulls, a previous effort to recover viable Arc40 null cells was unsuccessful, raising the possibility that the function of Arc40 may extend beyond that of the Arp2/3 complex.
In this study, we reinvestigated the involvement of Arc40 in the function of the Arp2/3 complex through biochemical and genetic analysis. We present evidence that ARPC1/Arc40 is indeed a stable component of the Arp2/3 complex and mediates a major interaction between the complex and WASP family activators.

EXPERIMENTAL PROCEDURES
Media and Genetic Manipulations-Yeast cell culture and genetic techniques were carried out by methods described in Ref. 13. YPD contained 2% glucose, 1% yeast extract, and 2% Bactopeptone (Difco Laboratories).
Antibodies-Polyclonal antibody against yeast Arp3 (yG-18) was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). The anti-yeast Arp2 polyclonal antibody was raised in rabbits against an Arp2 peptide as described previously (14). The anti-yeast ARPC1/Arc40 antibody (YAR40) was raised in rabbits as described previously (10). The antiyeast ARPC3/Arc18 antibody (YAR18) was raised in rabbits against a GST-Arc18 recombinant protein. The anti-yeast ARPC5/Arc15 antibody (YAR15) was raised in rabbits against a GST-Arc15 recombinant protein. The anti-HA antibody (12CA5) was a gift from Dr. Frank McKeon (Harvard Medical School).
Gene Disruption of ARC40 -The ARC40 gene was disrupted by replacement with the TRP1 gene. The ARC40 gene was amplified from genomic DNA and cloned into pSKϩ to generate pDW9. A 685-bp XcmI-XbaI fragment (corresponding to amino acids 67-295 of Arc40) was replaced with a 900-bp SmaI-NheI fragment containing the TRP1 gene. The ⌬arc40 strain was generated by transforming the diploid strain RLY141 with the XhoI-NotI ⌬arc40::TRP1 fragment. Gene disruption was verified by PCR amplification and restriction enzyme analysis (data not shown), and haploid ⌬arc40 cells were obtained by sporulation and tetrad dissection.
Preparation of Yeast Extracts-Yeast extracts used for binding studies, gel filtration analysis, and pyrene-actin assays were prepared essentially as described previously (15). Unless specified otherwise, extracts were prepared in 1ϫ UBA buffer (50 mM KHepes, pH 7.5, 100 mM KCl, 3 mM MgCl 2 , 1 mM EGTA, 0.2 mM ATP, and 1 mM dithiothreitol). For the pyrene-actin assay, yeast extracts were fractionated by an additional ammonium sulfate precipitation (at 55%). The precipitate was resuspended and dialyzed overnight against 1ϫ UBA buffer and clarified by centrifugation at 100,000 ϫ g.
Recombinant Protein Expression and Purification-GST-tagged yeast Bee1-VCA (16), GST-tagged yeast Myo5-CA (17) (called Myo5A in Ref. 17), GST-tagged N-WASP-VCA (18), and GST-N-WASP-VCA W503A mutant (8) were expressed and purified from bacteria. To generate recombinant Arc40, a baculovirus expressing His 6 -HA-Arc40 was constructed using the Bac-to-Bac system (Invitrogen). 10 ml of insect cell lysate in 1ϫ UBA buffer was mixed with 2 ml of Ni-NTA resin and incubated overnight at 4°C. The resin was washed in a column and eluted with a gradient of 0 -1 M imidazole in 1ϫ UB buffer, and 1-ml fractions were collected. The Arc40 peak fractions were pooled, further cleaned through a UnoQ1 anion exchange column (BioRad), and dialyzed against 1ϫ UBA buffer.
Purification of Yeast Arp2/3 Complex and Rabbit Muscle Actin-Purification of the yeast Arp2/3 complex using either Ni-NTA resin or VCA affinity columns was described previously (5,15). Tandem affinity purification (TAP)-tagged (19) yeast Arp2/3 complex was purified as follows. Strains expressing a COOH-terminal TAP-tagged Arp2 or Arp3 as the sole source of Arp2 or Arp3, respectively, were generated. The tagged complex was purified from yeast extracts on IgG beads and eluted from the beads with tobacco etch virus protease. The complex was further purified on a UnoQ1 column (Bio-Rad) and stored in 50 mM Hepes, pH 7.5, 100 mM KCl, 3 mM MgCl 2 , 1 mM EGTA, 0.1 mM ATP, 0.2 M sucrose at Ϫ80°C.
Pyrene-Actin Polymerization Assay-Pyrene-actin assays were performed essentially as described in Ref. 23. Ca 2ϩ -G-actin was converted into Mg 2ϩ -G-actin before polymerization by a brief incubation in exchange buffer (10 mM EGTA and 1 mM MgCl 2 ). Actin polymerization was initiated by the addition of 1.5 M G-actin (10% pyrene-labeled) in KMET buffer (50 mM KCl, 1 mM MgCl 2 , 1 mM EGTA, 10 mM Tris, pH 7.0) to Arp2/3 complex and VCA and monitored by continuous pyrene fluorescence measurements ( ex ϭ 347 nm, em ϭ 386 nm) in a Cary Eclipse fluorescence spectrophotometer (Varian). Determination of the concentration of the barbed ends formed in these reactions was carried out as described previously (24).
In Vitro Protein Binding Assays and K d Determination-Affi-Gel coupled with GST-N-WASP-VCA (340 ng of protein/l of beads), GST-N-WASP-VCA W503A (460 ng of protein/l beads), GST (350 ng of protein/l of beads), yeast Myo5-CA (400 ng of protein/l of beads), or GST-Bee1-VCA (320 ng of protein/l of beads) was equilibrated in 1ϫ UBA buffer. For binding His 6 -HA-Arc40, ϳ40 l of GST-N-WASP-VCA-, GST-N-WASP-VCA W503A -, and GST protein-bound beads (1.75 M) and 1.5, 3, and 4.5 M purified Arc40 were mixed in a total volume of 200 l; or 40 l of yeast Myo5-CA-or GST-Bee1-VCA-bound beads (1.75 M) and 1.5 M purified Arc40 were mixed in a total volume of 200 l. For binding Arp2/3 complex from yeast extract, 5 l of GST-VCA beads were mixed with variable amounts of high speed extracts or purified Arp2/3 complex in a total volume of 200 l. K d was determined as described previously (8). Briefly GST-VCA-coated Affi-Gel beads (0 -8 M) were incubated with 250 nM purified His 6 -HA-Arc40 or 200 nM Arp2/3 complex in 1ϫ UBA buffer ϩ 2 mg/ml bovine serum albumin. GST-coated Affi-Gel beads were added to the reaction mixtures to equalize the total bead volumes in each binding reaction. The samples were incubated for 1 h at 4°C with agitation. Supernatants were collected by low speed centrifugation. Beads were washed twice in 1ϫ UBA buffer with 1% Nonidet P-40. Free and bound fractions were detected by Western blotting after resolving the proteins by SDS-PAGE. The bands were quantified using ImageQuant software (Amersham Biosciences). The data were plotted and fitted with a single rectangular hyperbola equation: Other Methods-Rhodamine-phalloidin staining of fixed cells and electron microscopy were carried out as described previously (25). Gel filtration chromatography of yeast extracts or purified yeast Arp2/3 complex was carried out on a Bio-Rad SE1000/17 or Amersham Biosciences Superose 12 column, respectively, at 0.5 ml/min. Tryptic digestion and the identification of proteins were done as described previously (5) at the Taplin biologic mass spectrometry facility by liquid chromatography/mass spectrometry.

RESULTS
The Budding Yeast Arp2/3 Complex Always Contains ARPC1/Arc40 -Previously we used two different methods to purify the Arp2/3 complex from budding yeast, one involving Ni-NTA affinity and conventional columns and the other using an affinity column to which the CA domain was covalently coupled (5,15). CA domain is the VCA region lacking the G-actin-binding V motif. To determine whether the lack of ARPC1/Arc40 was due to the purification procedures, we developed a purification of the complex using the TAP tag, which was faster and gentler (see "Experimental Procedures"). However, like the CA affinity-purified complex, the complexes purified using the TAP tag still lacked Arc40 as visualized by SDS-PAGE and Coomassie Blue staining (Fig. 1, A and B). As a positive control, bovine Arp2/3 complex purified using the CA affinity column contained the 40-kDa ARPC1 (Fig. 1A). We noticed, however, that the yeast complexes purified by both methods had a smear at the top of the gel (Ͼ200 kDa) (Fig. 1, A and B, large black arrows) and occasionally an additional band around 85 kDa. To identify the proteins present in the smear, the region of the gel (boxed) was excised and subjected to mass spectrometry analysis. The peptide sequences obtained for the band exclusively matched the sequence of Arc40 and cover 36% of the open reading frame (Fig. 1C, bold). Immunoblot analysis using a polyclonal anti-yeast Arc40 antibody (10) further confirmed the identity of the high molecular weight smear and the occasional 85-kDa bands to be Arc40 and also revealed Arc40 species at monomer size (see Fig. 2, B and C). This result suggests that the yeast Arc40 copurifies with the complex but exhibits abnormally high and heterogeneous molecular weights.
High Molecular Weight Species of Arc40 Are Formed during SDS-PAGE-Since most of the Arc40 associated with purified yeast Arp2/3 complex was seen as high molecular weight aggregates on SDS-PAGE, we performed gel filtration fractionation to determine whether these aggregates were present in the native complex. If the large aggregates were formed before gel electrophoresis, the Arp2/3 complex should elute faster than expected of a 220-kDa globular complex and exhibit size heterogeneity, but if Arc40 associates with the complex as a monomer, it should cofractionate with the other subunits with ARPC1/Arc40 Functions in Budding Yeast expected elution rate. A Superose 12 gel filtration column with calibrated separation range from 15 to 669 kDa was used. As shown in Fig. 2, A and B, the large aggregates cofractionated with the complex in the expected fractions (peak at fraction 26). Immunoblot analysis confirmed that the large aggregates in these fractions were indeed Arc40. This result suggests that Arc40 associates with the Arp2/3 complex normally but forms large aggregates during SDS-PAGE. Attempts to prevent or solubilize these aggregates were unsuccessful. Serendipitously we analyzed by SDS-PAGE purified Arp2/3 complex that was bound to CA-coated beads. The beads were directly boiled in the sample buffer before gel loading. We found that binding to CA strongly reduced Arc40 aggregates and increased the proportion of monomeric Arc40 in the gel (Fig. 2C, arrow).
Arc40 Directly Binds VCA-The above result led us to suspect that Arc40 directly interacts with VCA, and this interaction somehow prevents Arc40 aggregation in the gel. To test this hypothesis, recombinant baculovirus-expressed His 6 -HAtagged Arc40 was used to test the interaction with GST-VCA from Bee1 and N-WASP and CA from yeast Myo5. Purified VCA and CA fragments were cross-linked to Affi-Gel beads and tested for interaction with purified Arc40 by bead pull-down assays (see "Experimental Procedures"). Both VCA fragments demonstrated similar levels of binding to recombinant Arc40, while CA from yeast Myo5 also showed strong binding, whereas the GST control did not bind (Fig. 3A). For the following analysis, we used the GST-VCA and GST-VCA W503A peptides from N-WASP because the mutant and wild-type N-WASP-VCA peptides were characterized previously for binding to the Arp2/3 complex (26). Mutations affecting the conserved tryptophan in the A region were previously shown to disrupt the interaction of VCA with the Arp2/3 complex (8,26). As shown in Fig. 3A, GST-N-WASP-VCA bound Arc40 in a dose-dependent manner, and GST-N-WASP-VCA W503A mutant had significantly reduced affinity to Arc40. Quantification of the amount of Arc40 bound to GST-N-WASP-VCA and GST-N-WASP-VCA W503A at 1.5 M Arc40 showed that binding to the mutant was decreased by 21-fold compared with the wild type, suggesting that Arc40 binds directly to VCA and this interaction specifically requires the conserved tryptophan in the A region. The K d values measured for Arc40-VCA and Arp2/3-VCA interactions were 0.45 Ϯ 0.16 and 0.30 Ϯ 0.12 M, respectively (Fig. 3B). The similarity in binding affinities of these interactions suggests that Arc40 is a major contributor of Arp2/3 complex binding to VCA. This conclusion was further supported by analysis of the Arc40 null complex (see below).
The Generation and Phenotypic Analysis of ⌬arc40 Cells-To directly test whether Arc40 is required for the activity of the Arp2/3 complex in vivo and in vitro, it would be useful to have a viable ⌬arc40 yeast strain. In previous work we found that 〈rc40 was the only Arp2/3 subunit that was essential for cell viability. With the newly annotated yeast genome data base, we realized that the original ARC40 disruption construct affected an adjacent small reading frame encoding an essential protein, Dad3 (27). Thus, we generated a new disruption construct that deletes amino acids 67-295 of Arc40 (see "Experimental Procedures"). The new ⌬arc40/ARC40 heterozygous diploid strain was sporulated and dissected on a YPD plate. The ⌬arc40 spores were viable but grew very slowly compared with wild-type colonies (Fig. 4A), and only 30% (from 60 tetrads) of ⌬arc40 spores were able to form colonies. The mutant cells exhibit highly aberrant morphology and size heterogene- 1. ARPC1/Arc40 is present in the Arp2/3 complex purified from yeast. A, comparison of bovine brain Arp2/3 complex (lane 1) with yeast Arp2/3 complexes (lanes 2 and 3). Bovine brain Arp2/3 complex was purified on a VCA affinity column. The yeast complex was purified from strains expressing Arp2-TAP (lane 2) or Arp3-TAP (lane 3) by using an IgG column and tobacco etch virus protease elution (see "Experimental Procedures"). Arrows in A and B denote the high molecular weight aggregates. After cleavage, the remaining calmodulin-binding peptide resulted in a molecular weight shift for the tagged proteins (arrowheads). B, yeast Arp2/3 complex isolated on a CA affinity column. C, Arc40 peptides (residues in bold) identified by tryptic digestion and tandem mass spectrometry of the high molecular weight band (boxed) excised from lane 3 in A.
ity (data not shown). Rhodamine-phalloidin staining of the mutant cells showed disruption of cortical actin patches and accumulation of aberrant actin cable-like structures (Fig. 4B). These phenotypes are entirely consistent with those of other Arp2/3 subunit null mutants (10,28) as well as with the disruption of BEE1 encoding the only WASP-like protein in yeast (25). Images from thin section electron microscopy revealed that ⌬arc40 cells accumulated a large number (74 Ϯ 12/section) of post-Golgi vesicles in the bud, whereas only a few (5 Ϯ 1/section) vesicles were observed in wild-type cell sections (Fig.  4C). A similar vesicle accumulation phenotype was also observed in ⌬bee1 cells (25). These results suggest that Arc40 functions exclusively in the Arp2/3 complex and further confirm that disruption of cortical actin patches affects a late step of exocytosis.
A Partial Arp2/3 Complex Is Present in ⌬arc40 Cells and Has Reduced Affinity with VCA-Since we were able to recover viable ⌬arc40 cells, we examined the effect of this mutation on the integrity of the Arp2/3 complex. Yeast extracts prepared from the wild-type or ⌬arc40 strain were analyzed by gel fil-tration chromatography. Fractions were resolved by SDS-PAGE, transferred, and blotted with antibodies against Arp2, Arp3, ARPC3/Arc18, and ARPC5/Arc15. In wild-type extracts, these subunits were mainly present in the complex fraction and were undetectable in the monomer fractions (Fig. 5A). In ⌬arc40 extracts, Arp3, Arc15, and Arc18 were still mainly present in the complex fractions, but ϳ60% of Arp2 shifted into the monomer fractions, and ϳ40% remained in the complex fractions (Fig. 5B). Since ARPC2/Arc35 and ARPC4/Arc19 are core subunits (6), these must be present in the complex in ⌬arc40 extracts for assembly of any subcomplex. Therefore, this result suggests that an Arp2/3 complex lacking only Arc40 exists in the mutant extract, although the affinity of Arp2 with the complex is reduced.
Because Arc40 alone binds to VCA with an affinity close to that between VCA and the purified Arp2/3 complex, we tested whether loss of Arc40 had any effect on the latter interaction. Equal amounts of high speed extracts from wild-type and ⌬arc40 cells were incubated with VCA-coated Affi-Gel beads.

FIG. 2. Gel filtration analysis of purified yeast Arp2/3 complex.
Approximately 100 g of yeast complex was loaded on a Superose 12 gel filtration column, and 0.5-ml fractions were collected. A, the fractions were precipitated with trichloroacetic acid, resolved by 15% SDS-PAGE, and stained with Coomassie Blue. B, the same fractions as shown in A were blotted with an antibody against yeast Arc40. C, VCA binding reduces formation of Arc40 aggregates. Lane 1, purified Arp2/3 complex. Lane 2, purified Arp2/3 complex bound to VCA beads. The proteins were resolved by 12.5% SDS-PAGE and blotted with an antibody against yeast Arc40. M, molecular mass markers; F, fraction. . Unbound Arp2/3 complex or His 6 -HA-Arc40 was visualized by immunoblotting using an antibody against Arp3 or HA, respectively. Bound and free VCA was quantified and plotted, and K d was calculated from each plot by nonlinear regression using the SigmaPlot software.

ARPC1/Arc40 Functions in Budding Yeast
Bound and free Arp2/3 complexes were resolved by SDS-PAGE and blotted with antibodies against Arp3. Fig. 6A shows that loss of Arc40 strongly reduced the affinity of VCA with the Arp2/3 complex. Furthermore little Arp2 in ⌬arc40 extract was bound to VCA beads, suggesting that Arp2 itself does not bind VCA (Fig. 6B). To further rule out that the reduction in VCA affinity observed in ⌬arc40 extract was due to partial loss of Arp2, we also tested ⌬arp2 extracts in the above experiment. The lack of Arp2 had only a slight effect on the interaction between Arp2/3 complex with VCA (Fig. 6A), suggesting that Arp2 does not contribute significantly to the binding affinity of Arp2/3 complex with VCA. Previous cross-linking results also implicated Arp3 in Arp2/3 interaction with VCA (7,8); however, the GST-VCA beads exhibited the same level of interaction with the Arp2/3 complex in ⌬arp3 extracts as that in wild-type extracts (Fig. 6C). These results strongly suggest that ARPC1/Arc40 mediates a major interaction between the Arp2/3 complex and VCA.
Arc40 Is Required for VCA-stimulated Actin Nucleation-Previous analysis of VCA from different WASP-like proteins and VCA mutants suggested that the ability of VCA to activate the Arp2/3 complex does not strictly correlate with the binding affinity between the complex and VCA (26,29). Therefore, although Arc40 contributes significantly to the interaction with the Arp2/3 complex, it was unclear whether Arc40 is required for the nucleation activity of the Arp2/3 complex. However, because ⌬arc40 cells grew extremely poorly, we were not able to obtain enough cells to purify the Arc40 null complex. To circumvent this problem, we developed a strategy to assay Arp2/3 complex activity in crude yeast extracts. In this assay, resolubilized and dialyzed 55% ammonium sulfate precipitates of high speed supernatants were used in the pyrene-actin polymerization assay (see "Experimental Procedures"). In the absence of VCA, the extract exhibited no stimulation of actin polymerization, whereas in the presence of VCA, actin nucle-ation was strongly stimulated as evidenced by the shortened lag phase and increased rate during the polymerization phase (Fig. 7). This nucleation activity showed dose dependence on extracts and VCA as expected for Arp2/3-based actin nucleation (Fig. 7, A-D).
Using this assay, we compared the activity of extracts prepared from a wild-type and a ⌬arc40 strain. Both extracts had the same concentration of Arp2/3 complex as determined by immunoblot analysis (Fig. 7E, inset). In contrast to wild-type extracts, extracts prepared from the mutant strain exhibited no nucleation activity at a wide range of extract concentrations (Fig.  7E). To determine whether the lack of activity in the ⌬arc40 extract was due to lack of Arc40 as opposed to indirect effects from unhealthy ⌬arc40 cells, we added back baculovirus-expressed His 6 -HA-Arc40 to the ⌬arc40 extract. The actin nucleation activity was significantly restored to the ⌬arc40 extract in a His 6 -HA-Arc40 concentration-dependent manner (Fig. 7F). This result suggests that Arc40 is required for the VCA-stimulated actin nucleation activity of the Arp2/3 complex.

DISCUSSION
Previously it was unclear whether Arc40 was a necessary component of the active Arp2/3 complex and whether the function of Arc40 exists beyond that of the Arp2/3 complex. Here we first showed by mass spectrometry and gel filtration that Arc40 is indeed a component of purified yeast Arp2/3 complex but has a strong tendency to form high molecular weight species when analyzed by SDS-PAGE. Attempts to reduce this tendency, including addition of fresh 0.1 M dithiothreitol or 10% ␤-mercaptoethanol to SDS-PAGE sample buffer and low temperature (37°C) ⌬arc40 cells (B). Yeast extracts were analyzed on a SE1000 gel filtration column, and 0.5-ml fractions were collected. Every other fraction of fractions 15-37 were resolved by 12.5% SDS-PAGE and immunoblotted using antibodies against yeast Arp2, Arp3, Arc18 (ARPC3), and Arc15 (ARPC5). *, a cross-reacting band with the anti-Arp2 antibody as it also appeared in ⌬arp2 extracts.

FIG. 4. ⌬arc40 cells exhibit phenotypes typical for Arp2/3 subunit null mutants.
A, tetrads dissected from a ⌬arc40 heterozygous diploid strain and grown for 5 days at room temperature. The large colonies were confirmed to be wild type (WT), and the tiny colonies were confirmed to be ⌬arc40. Only 30% of the ⌬arc40 spores were able to form colonies. B, rhodamine-phalloidin staining of wild-type and ⌬arc40 yeast cells. Bar, 5 m. C, thin section electron microscopy images of a representative wild-type and ⌬arc40 cell. Bar, 500 nm. heat denaturation, were unsuccessful. However, treatment with 6 M urea prior to the addition of sample buffer could enrich the fraction of Arc40 that ran with expected mobility (data not shown). One possible cause for the formation of the high molecular weight species may be the amino acid composition and structural characteristics of Arc40. The budding yeast Arc40 is composed of 384 amino acids, a large portion of which are hydrophobic (grand average of hydropathicity (GRAVY) ϭ Ϫ0.221). The crystal structure of bovine ARPC1 (30) revealed a seven-blade ␤-propeller composed of WD40 repeats. The top face of the ␤-propeller is normally in contact with the rest of the complex. It is possible that when the complex is denatured in the presence of SDS, the extensive ␤-sheets and high hydrophobicity led to formation of the aggregates.
In addition to confirming that ARPC1/Arc40 is a component of the Arp2/3 complex, our results also suggest that this subunit contains a major contact site between the Arp2/3 complex and VCA. Purified Arc40 bound to the activator VCA with a K d of 0.45 M, close to that of the yeast Arp2/3 complex with VCA (0.30 M). The latter was similar to the affinity (K d ϭ 0.9 M) measured for the bovine Arp2/3 with human WASP-VCA by fluorescence anisotropy (26) and the affinity (K d ϭ 0.76 M) for the bovine Arp2/3 complex with human N-WASP-VCA by pulldown assays (8). Loss of Arc40 from the Arp2/3 complex resulted in a 5-10-fold reduction of the affinity with VCA, and FIG. 6. Loss of Arc40 but not Arp2 and Arp3 strongly reduced the affinity between the Arp2/3 complex and VCA. A, a, yeast extracts prepared from wild-type, ⌬arp2, and ⌬arc40 strains were incubated with VCA beads. Unbound (supernatant (Sup)) and bound (Pellet) Arp2/3 complex was detected by immunoblot analysis using the anti-Arp3 antibody. b, the percent bound of the total (supernatant ϩ pellet) from wild-type and various mutant extracts was quantified by densitometry and shown as histograms. B, a, yeast extracts prepared from wild-type and ⌬arc40 strains were incubated with VCA beads, and unbound (Sup) and bound (Pellet) Arp2/3 complex was detected by immunoblot analysis using the anti-Arp2 antibody. b, the percent bound of the total (supernatant ϩ pellet) from wild-type and ⌬arc40 extracts was quantified by densitometry and shown as histograms. C, a, yeast extracts prepared from wild-type and ⌬arp3 strains were incubated with VCA beads. Unbound (Sup) and bound (Pellet) Arp2/3 complex was detected by immunoblot analysis using the anti-Arp2 antibody. b, the percent bound of the total (supernatant ϩ pellet) from wild-type and ⌬arp3 extracts were quantified by densitometry and shown as histograms. WT, wild type.

ARPC1/Arc40 Functions in Budding Yeast
this effect was not due to partial loss of Arp2 from the complex since Arp2 null complex had only slightly reduced binding to VCA.
A potential involvement of ARPC1 in activator interaction has been suggested by previous chemical cross-linking experiments, which showed that VCA can be cross-linked to ARPC1, Arp2, and Arp3 (7,8). Interestingly one study concluded that the COOH-terminal conserved tryptophan of VCA is involved in binding to Arp3 because truncating the DDW motif from the VCA COOH terminus reduced cross-linking to Arp3 by 81% and to ARPC1 by 37% (8). However, our results showed that the conserved tryptophan is required for the interaction with purified ARPC1/Arc40. These results are not necessarily contradictory because cross-linking data do not quantitatively reflect affinity. This is exemplified by the findings that whereas both Arp3 and Arp2 cross-link strongly with VCA, loss of each of these subunits had little effect on the interaction of the Arp2/3 complex with VCA. If binding between the VCA tryptophan and Arc40 contributes strongly to the affinity between VCA and Arp2/3 complex, a loss of this interaction could have a large effect on low affinity contacts at adjacent sites. However, we tested whether an excess of recombinant Arc40 could inhibit the Arp2/3 activity in the extract, but we did not detect any effect. This is not too surprising since the interaction between the Arp2/3 complex and VCA is strongly enhanced upon F-actin binding (31), and hence Arc40 is not necessarily an effective inhibitor in the nucleation assay.
The recent homology modeling of Arp2/3 complex did not explore ARPC1 as a potential site for VCA binding (32). The ␤-propeller structure, which contains extensive hydrophobic pockets (30), is an attractive target for docking the conserved tryptophan of VCA. Arc40 also has an extensive patch of basic residues at the solvent-exposed surface that faces away from the rest of the complex (30). Thus the interaction between Arc40 and VCA could be further enhanced by favorable electrostatic interactions between the basic patch and acidic A region. Future modeling, mutagenesis, and structural analyses could pinpoint the site in ARPC1 that binds the conserved tryptophan in the A region. Because this residue is at or near the very COOH terminus of VCA, locating its binding site in Arp2/3 should provide an important anchor point for reconstructing the extended interaction of the activator peptide with the Arp2/3 complex.
Through analysis using a new ARC40 deletion construct, we showed that Arc40 not only contributes to binding the activator peptide but also is required for the actin nucleation activity of the Arp2/3 complex. The role for Arc40 in the nucleation process is probably not restricted to activator binding. Since ARPC1 can directly bind F-actin (data not shown) and VCA, this subunit is in a unique position to mediate the cooperativity between the interactions of Arp2/3 complex with F-actin and with the activator. Additionally our data suggested that Arc40 is required for tight association of Arp2 with the rest of the complex, and therefore Arc40 could also have a strong influence on the orientation and/or conformation of Arp2. A function in VCA binding and in tethering Arp2 to the complex could both explain a loss of actin nucleation in ⌬arc40 extract. Our result is also largely consistent with the results from reconstitution experiments using human Arp2/3 subunit expressed in insect cells (33). Partial complex was purified from a reconstitution mixture that lacked both ARPC1 and ARPC5. Interestingly stoichiometric Arp2 was present in this partial complex, suggesting that ARPC1 in the human complex is not as important for Arp2 association with the complex as it in yeast. This partial complex also exhibited a severe defect in actin nucleation, although residual activity ( 1 ⁄85 of the wild-type) was observed. Together that and our study demonstrate a critical role for the ARPC1 subunit in Arp2/3 complex and VCA-mediated actin nucleation.
Phenotypically ⌬arc40 cells are indistinguishable from other subunit null cells except ⌬arc18, whose defects are less severe (10). This indicates that Arc40 functions entirely within the Arp2/3 complex unlike originally thought. The vesicle accumulation phenotype is intriguing. This was previously observed in ⌬bee1 cells, which also lack cortical actin patches (25). The actin patches in yeast are thought to be sites of endocytosis as indicated by phenotypes of mutants affecting a large number of actin patch components (34). The exocytic defects in ⌬bee1 and ⌬arc40 mutants may suggest that endocytosis and exocytosis are directly linked in yeast or that endocytic recycling is required for continuous activity of the machinery required for the last steps of exocytosis.