Although nuclear actin and Arps (actin-related proteins) are often identified as components of multi-protein chromatin-modifying enzyme complexes, such as chromatin remodeling and histone acetyltransferase (HAT) complexes, their molecular functions still remain largely elusive. Here, we investigated the role of human Arp4 (BAF53, also known as actin-like protein 6A) in Brg1-containing chromatin remodeling complexes. Depletion of Arp4 by RNA interference impaired the integrity of these complexes and accelerated the degradation of Brg1, indicating a crucial role in their maintenance, at least in certain human cell lines. We further found that Arp4 can form a heterocomplex with β-actin. Based on structural similarities between conventional actin and Arp4, and the assumption that actin–Arp4 binding might mimic actin–actin binding, we introduced a series of mutations in Arp4 that might be expected to impair its interaction with β-actin. Some of them indeed caused reduced binding to β-actin. Interestingly, such mutant Arp4 proteins also showed reduced incorporation into Brg1 complexes, and their interaction with Myc-associated complexes as well as Tip60 HAT complexes were also impaired. Based on these findings, we propose that β-actin–Arp4 complex formation might be a crucial feature in some chromatin-modifying enzyme complexes, such as the Brg1 complex.

Actin–related proteins (Arps) are conserved throughout eukaryotes (Poch and Winsor, 1997; Muller et al., 2005). Among the documented examples, Arp4 to Arp9 are mainly detected in the cell nucleus and often function as components of multi-subunit chromatin-modifying enzyme complexes, such as the chromatin remodeling and histone acetyltransferase (HAT) complexes that play crucial roles in controlling gene expression, DNA replication and other chromatin transactions (Boyer and Peterson, 2000; Olave et al., 2002a; Blessing et al., 2004). For instance, the yeast INO80 chromatin remodeling complex contains Arp4, Arp5, Arp8 and actin, and among these proteins Arp4 and actin are essential for viability (Shen et al., 2000). Arp4 and Arp8 can bind to core histones, suggesting that they could facilitate the interaction of the complexes with nucleosomes (Harata et al., 1999; Shen et al., 2003). Consistent with this idea, mutant INO80 complexes lacking Arp5 or Arp8 show a reduction in ATPase and chromatin remodeling activities in vitro (Shen et al., 2003). In the case of Arp8 deficiency, Arp4 and actin also disappear from the complexes, indicating that Arp4 and actin are dispensable for INO80 complex assembly. The yeast SWI/SNF chromatin remodeling complex, and the highly related RSC (remodels the structure of chromatin) complex, contain Arp7 and Arp9, two essential Arps (Cairns et al., 1998; Peterson et al., 1998). In contrast to the Arps in INO80, Arp7 and Arp9 show no affinity for nucleosomes (Szerlong et al., 2003). The in vitro ATPase activity of mutant RSC complexes lacking Arp7 and Arp9 is partially impaired compared with the wild type (Szerlong et al., 2003; Szerlong et al., 2008). Actin itself is a slow ATPase and the central actin fold is important for ATP binding and hydrolysis (Wertman and Drubin, 1992; Kabsch and Holmes, 1995). Although Arps are very similar to actin in the central actin fold, it has been suggested that Arp7 and Arp9 lack ATP binding and hydrolysis properties (Cairns et al., 1998). Arp4 is also a component of the NuA4 HAT complex in budding yeast and, in this case, is required for integrity of the complex (Galarneau et al., 2000). Overall, molecular functions of nuclear Arps remain largely elusive.

Actin itself has also been recently implicated in important nuclear metabolism processes (Bettinger et al., 2004). Actin has been characterized as a component of several chromatin-modifying enzyme complexes (Boyer and Peterson, 2000; Olave et al., 2002a; Blessing et al., 2004), almost all of which simultaneously contain nuclear Arps, implying a close relationship. One study has addressed the functional role of actin in chromatin remodeling complexes using human SWI/SNF family BAF (for Brg- or Brm-associated factor) complexes. The BAF complex is based on human Brm or Brg1, close relatives to yeast SWI2, and contains one Arp, human Arp4 (BAF53, also known as actin-like protein 6A), and β-actin (Zhao et al., 1998; Kuroda et al., 2002; Olave et al., 2002b). Zhao et al. have shown that the DNA-dependent ATPase activity of the BAF complex is inhibited by the actin inhibitor latrunculin-B and, based on this finding, suggested that actin plays a crucial role in the chromatin remodeling process (Zhao et al., 1998). It has been also proposed that the BAF complex binds to actin filaments for localization to the nuclear matrix (Rando et al., 2002). More recently, several studies have shown that actin is also involved in transcription by RNA polymerase I (Fomproix and Percipalle, 2004; Philimonenko et al., 2004), II (Hofmann et al., 2004; Kukalev et al., 2005) and III (Hu et al., 2004). However, molecular functions of nuclear actin also remain largely elusive.

In this study, we found that depletion of Arp4 by RNA interference destabilized Brg1 chromatin remodeling complexes, accelerating the degradation of Brg1 and BAF170 (also known as SMARCC2), the two core subunits, in several human cell lines. Interestingly, we also found that β-actin and human Arp4/BAF53 can form a heterocomplex, probably a heterodimer. Therefore, we prepared a series of Arp4 mutants based on the structural similarity between actin and Arp4 and examined their ability to bind to actin and be incorporated into Brg1 complexes. Taken together, our data suggest that interactions between β-actin and Arp4 might be required for integrity of Brg1 complexes. During the preparation of the manuscript, Fenn et al. reported structure of yeast Arp4 and interaction between the Arp4 and actin (Fenn et al., 2011). Discussion on our and their results is also described.

Depletion of human Arp4/BAF53 impairs the integrity of Brg1 complexes and accelerates the degradation of Brg1 and BAF170 in several human cell lines

In human cells, several chromatin-modifying complexes containing β-actin and Arp4 have been identified, including BAF chromatin remodeling complexes (Zhao et al., 1998), Tip60 HAT complexes (Cai et al., 2003) and Myc-associated transcriptional activator complexes (Park et al., 2002). In this study, we first attempted to clarify the effects of Arp4 depletion on Brg1 chromatin remodeling complexes.

HeLa cells were infected with recombinant retrovirus carrying a control vector or an shRNA expression vector directed toward Arp4, selected with puromycin and examined at 5 days after infection. Whole-cell lysates were prepared and subjected to immunoblotting. As shown in Fig. 1A,B, levels of Arp4 protein were reduced to ∼20% by the shRNA expression. Cell growth was significantly inhibited by Arp4 depletion (Fig. 1D), suggesting that Arp4 is essential for cell growth. Arp4 has been implicated in many transcription processes. Therefore, we performed DNA microarray analyses. The microarray used contained ∼17,800 cDNAs, and only genes induced or repressed more than 2.5-fold were counted. Arp4 depletion in HeLa cells resulted in the induction of 87 genes and repression of 35 genes (supplementary material Table S1). Among them, a remarkable decrease in expression of CRYAB (encoding alpha B-crystallin) was noted. Since it has been identified as a gene that is activated by Brg1 (Liu et al., 2001), this can be taken to support the accuracy of the analysis. Thus far, it has been unclear how Arp4 depletion leads to growth suppression in HeLa cells.

Fig. 1.

Silencing of BAF53/human Arp4 leads to acceleration of degradation of Brg1 in HeLa and HCK1/T cells. HeLa cells were infected with retroviruses expressing Arp4 shRNAs, drug selected and examined at 5 days after infection (A–G). (A,B) Whole cell lysates were immunoblotted with the indicated antibodies. The membranes were also stained with CBB to confirm equal loading. Representative data are shown in A. The signal intensities of the bands were quantified and the means±s.d. of three independent experiments are shown, compared with control cells, set at 100 in B. (C) HeLa cells were infected with retroviruses expressing other shRNAs targeting Arp4 (Arp4 shRNA-3 and -4) and then analyzed as in A. (D) Growth rates of parental, control virus-infected, and Arp4-depleted HeLa cells (2000 cells per well in 96-well plates) were estimated with the MTS assay. (E) Changes in expression of Brg1 complex-related genes in Arp4-depleted cells. Expression of the indicated genes was investigated with DNA microarray analysis. Fold change of expression in Arp4-depleted cells compared with the control cells was calculated. (F,G) The half-lives of Brg1, BAF170, Arp4, and actin proteins were investigated. Cycloheximide (50 µg/ml) was added to the medium, and cells were sequentially harvested at the indicated time points for immunoblot analysis. Representative data are shown in F. The results of quantification are shown in G. The means±s.d. for data from two independent experiments are shown. (H) HeLa cells were transfected with the indicated siRNAs. Forty-eight hours after the transfection, whole cell lysates were subjected to immunoblotting. The signal intensities of the bands were quantified and are shown below the blot, with control siRNA-treated cells set at 100. (I) HCK1/T cells (normal human cervical keratinocytes immortalized with telomerase) were transfected and analyzed as in H. (J) HeLa cells, treated as in H, were immunoblotted with anti-PARP1 antibody. Bleomycin-treated HeLa cells (16 µM, 48 hr) were also analyzed.

Fig. 1.

Silencing of BAF53/human Arp4 leads to acceleration of degradation of Brg1 in HeLa and HCK1/T cells. HeLa cells were infected with retroviruses expressing Arp4 shRNAs, drug selected and examined at 5 days after infection (A–G). (A,B) Whole cell lysates were immunoblotted with the indicated antibodies. The membranes were also stained with CBB to confirm equal loading. Representative data are shown in A. The signal intensities of the bands were quantified and the means±s.d. of three independent experiments are shown, compared with control cells, set at 100 in B. (C) HeLa cells were infected with retroviruses expressing other shRNAs targeting Arp4 (Arp4 shRNA-3 and -4) and then analyzed as in A. (D) Growth rates of parental, control virus-infected, and Arp4-depleted HeLa cells (2000 cells per well in 96-well plates) were estimated with the MTS assay. (E) Changes in expression of Brg1 complex-related genes in Arp4-depleted cells. Expression of the indicated genes was investigated with DNA microarray analysis. Fold change of expression in Arp4-depleted cells compared with the control cells was calculated. (F,G) The half-lives of Brg1, BAF170, Arp4, and actin proteins were investigated. Cycloheximide (50 µg/ml) was added to the medium, and cells were sequentially harvested at the indicated time points for immunoblot analysis. Representative data are shown in F. The results of quantification are shown in G. The means±s.d. for data from two independent experiments are shown. (H) HeLa cells were transfected with the indicated siRNAs. Forty-eight hours after the transfection, whole cell lysates were subjected to immunoblotting. The signal intensities of the bands were quantified and are shown below the blot, with control siRNA-treated cells set at 100. (I) HCK1/T cells (normal human cervical keratinocytes immortalized with telomerase) were transfected and analyzed as in H. (J) HeLa cells, treated as in H, were immunoblotted with anti-PARP1 antibody. Bleomycin-treated HeLa cells (16 µM, 48 hr) were also analyzed.

We examined protein levels of Brg1, BAF170 and BAF47/Ini1/hSNF5 (Fig. 1A,B), which are thought to be core subunits of the Brg1 complex (Phelan et al., 1999). Interestingly, Brg1 and BAF170 protein levels were reduced to ∼40% in Arp4-depleted HeLa cells. BAF47 protein levels were also reduced, but to a lesser extent. Actin protein levels were not affected. Similar results were obtained with other shRNA expression vectors targeting Arp4 (Fig. 1C). One possible reason for the significant reduction in Brg1 and BAF170 protein levels might be that their transcription was downregulated by Arp4 depletion. However, the microarray analysis revealed that their transcription levels were virtually unchanged (Fig. 1E). On the other hand, the levels of BAF47 transcripts appeared to be somewhat decreased (Fig. 1E), which might explain the slight decrease in BAF47 protein in Arp4-depleted cells. As expected, the microarray analysis detected the decrease in Arp4 transcript levels upon the shRNA expression (Fig. 1E). Taken together, these data indicate that Brg1 and BAF170 proteins are downregulated at the post-transcription level upon Arp4 depletion. We next investigated the half-lives of these proteins. In Arp4-depleted HeLa cells, the half-lives of Brg1 and BAF170 were significantly decreased (Fig. 1F,G). In contrast, the half-lives of actin and Arp4 were not significantly changed by Arp4 depletion. These data suggest that degradation of Brg1 and BAF170 might be increased in Arp4-depleted cells.

Using a different experimental system, we further assessed whether Arp4 depletion affects the stability of Brg1. HeLa cells were transfected with two different siRNAs towards Arp4 and analyzed at 48 hr post-transfection. As shown in Fig. 1H, the levels of Arp4 protein were reduced to ∼15% by both siRNAs, and Brg1 protein levels were concomitantly decreased. The levels of SNF2H, another chromatin remodeler not associated with Arp4, were unaffected. When the experiments were repeated with normal human cervical keratinocytes immortalized with telomerase (Tatsumi et al., 2006), the steady-state levels of Brg1 were again reduced upon Arp4 depletion (Fig. 1I). Similar results were obtained with normal human fibroblasts immortalized with telomerase (data not shown). We also investigated whether Arp4 depletion induces apoptosis by detecting PARP1 cleavage by caspases (i.e. appearance of the 89 kDa cleaved PARP1) as a readout of apoptosis (Oliver et al., 1998). Bleomycin-treated HeLa cells were used as a positive control. The data indicate that Arp4 depletion does not induce apoptosis in HeLa cells (Fig. 1J).

We then characterized the Brg1 complex in Arp4-depleted HeLa cells. Nuclear extracts were prepared from Arp4-depleted and control HeLa cells, and subjected to immunoblotting. As expected, the levels of Brg1 and BAF170 proteins in nuclear extracts from the Arp4-depleted cells were significantly decreased (Fig. 2A,B). The nuclear extracts were then subjected to immunopurification with anti-Brg1 antibody beads. In these experiments, the amounts of the nuclear extracts subjected to the immunopurification were adjusted to recover the Brg1 complex at the same levels both from control and Arp4-depleted cells (i.e. four times more nuclear extracts from Arp4-depleted cells than those from control cells were subjected to immunoprecipitation). The precipitates were immunoblotted with antibodies against Brg1, BAF170, Arp4, BAF47 and actin, and the signal intensities of the bands were quantified. If defective Brg1 complexes lacking Arp4 or Arp4 plus some other subunit(s) existed in Arp4-depleted HeLa cells, then their levels relative to Brg1 protein in the immunoprecipitates would be reduced. However, no significant difference was found in the levels of the subunits examined between the immunoprecipitates from control and Arp4-depleted cells (Fig. 2C). The data suggest that, if any, there is limited defective Brg1 complex present in Arp4-depleted cells. We also examined the DNA-dependent ATPase activity of the purified Brg1 complexes (Fig. 2D). In control immunoprecipitates from nuclear extracts of Arp4-depleted cells, relatively high nonspecific ATP hydrolyzing activity was observed. This might be due to more nuclear extract being subjected to immunoprecipitation in the case of Arp4-depleted cells. However, the levels of specific DNA-dependent ATPase activity for Brg1 complexes prepared from Arp4-depleted cells were essentially the same as those from control cells (Fig. 2D).

Fig. 2.

Arp4 silencing reduces the amount of Brg1 complexes in HeLa cells. Arp4-silenced cells were prepared with Arp4 shRNA retroviruses as in Fig. 1. (A,B) Nuclear extracts were prepared and immunoblotted with the indicated antibodies. Representative data are shown in A. The signal intensities of the bands were quantified and the mean±s.d. of three independent experiments are shown, compared with control cells set at 100, in B. (C) Brg1 complexes were immunopurified from the nuclear extracts and immunoblotted with the indicated antibody. For Arp4-depleted cells, four times more nuclear extracts than those from control cells were subjected to immunoprecipitation. The signal intensities of the bands were quantified, and shown, with those of control cells set at 100. (D) DNA-dependent ATPase activities of the immunopurified Brg1 complexes. Antibody beads with immunopurified Brg1 complexes were subjected to the ATPase assay. Activated calf thymus DNA was included as indicated. ATPase activities are shown as the percentage of ATP hydrolysis in each reaction.

Fig. 2.

Arp4 silencing reduces the amount of Brg1 complexes in HeLa cells. Arp4-silenced cells were prepared with Arp4 shRNA retroviruses as in Fig. 1. (A,B) Nuclear extracts were prepared and immunoblotted with the indicated antibodies. Representative data are shown in A. The signal intensities of the bands were quantified and the mean±s.d. of three independent experiments are shown, compared with control cells set at 100, in B. (C) Brg1 complexes were immunopurified from the nuclear extracts and immunoblotted with the indicated antibody. For Arp4-depleted cells, four times more nuclear extracts than those from control cells were subjected to immunoprecipitation. The signal intensities of the bands were quantified, and shown, with those of control cells set at 100. (D) DNA-dependent ATPase activities of the immunopurified Brg1 complexes. Antibody beads with immunopurified Brg1 complexes were subjected to the ATPase assay. Activated calf thymus DNA was included as indicated. ATPase activities are shown as the percentage of ATP hydrolysis in each reaction.

Arp4 can form a heterodimer with β-actin in a bacterial expression system

As a way to address the molecular functions of Arp4, we focused on its relationship with β-actin. The Arp2/3 complex regulates the formation of branched actin filaments; it has been suggested that the minus end (pointed end) of actin monomers binds to the faces of the Arp2 or Arp3 equivalent to the plus end (barbed end) of actin and then actin filaments are elongated (Welch et al., 1997; Mullins et al., 1998; De La Cruz and Pollard, 2001; Robinson et al., 2001; Nolen et al., 2004). Nuclear Arp7 and Arp9 may also form a heterodimer in yeast RSC chromatin remodeling complexes (Szerlong et al., 2003). Recently, it has been shown that nuclear Arps and actin interact with chromatin remodelers and modifiers through conserved helicase-SANT-associated (HSA) domains and yeast Eaf1-HSA–Arp4–actin and Swr1-HSA–Arp4–actin ternary complexes have been isolated (Szerlong et al., 2008). However, whether conventional actin and nuclear Arp form dimers (or heterocomplexes) without the help of the HSA domain has not been clarified, especially in a mammalian cell system. To address this issue, purified monomeric actin and bovine serum albumin, as a negative control, were immobilized on Sepharose beads and binding of in-vitro-translated human Arp4 was examined. We found significant Arp4 binding to actin monomers but no binding to control BSA (Fig. 3A).

Fig. 3.

β-actin and Arp4 form heterocomplexes. (A) Binding of Arp4 to monomeric actin in vitro. HA–Arp4 was synthesized by in vitro transcription–translation with a rabbit reticulocyte lysate, mixed with G-actin-coated beads or control BSA-coated beads, and incubated in the presence or absence of 100 µM ATP. After washing the beads, bound proteins were immunoblotted with anti-HA antibody. (B) Isolation of β-actin–Arp4 heterocomplexes produced by an in vitro transcription–translation system with bacterial extracts. Co-synthesized His–β-actin and HA–Arp4 were first subjected to immunopurification with anti-HA antibody and eluted with HA peptides (HA eluate). The eluate was further processed in a Nickel affinity column. After washing, His–β-actin was eluted with buffer containing 250 mM imidazole (1 ml three times). Each fraction was analyzed by silver staining and by immunoblotting with anti-HA and anti-His antibodies. Flow through: the flow-through fraction of the Nickel affinity column. (C) Fractionation of β-actin–Arp4 heterocomplexes on sucrose gradient centrifugation. Arp4–2HA and His–β-actin were co-produced in bacteria, purified on Nickel affinity gel, and then immunoprecipitated with anti-HA affinity matrix. The eluate was layered over 5–20% linear sucrose gradients and fractionated. Twenty-eight fractions were collected and analyzed by SDS-PAGE followed by silver staining. The signal intensities of the bands were quantified and are shown with the maximum values for each set at 100 (lower-left panel). (D) Re-fractionation of β-actin–Arp4 heterocomplexes. Fractions 6 and 7 from C were collected and re-subjected to 5–20% linear sucrose gradient centrifugation as above. Only the data for fractions 1–14 are presented. Non-specific bands apparent at 50, 60 and 70 kDa are due to contamination with marker proteins used for SDS-PAGE.

Fig. 3.

β-actin and Arp4 form heterocomplexes. (A) Binding of Arp4 to monomeric actin in vitro. HA–Arp4 was synthesized by in vitro transcription–translation with a rabbit reticulocyte lysate, mixed with G-actin-coated beads or control BSA-coated beads, and incubated in the presence or absence of 100 µM ATP. After washing the beads, bound proteins were immunoblotted with anti-HA antibody. (B) Isolation of β-actin–Arp4 heterocomplexes produced by an in vitro transcription–translation system with bacterial extracts. Co-synthesized His–β-actin and HA–Arp4 were first subjected to immunopurification with anti-HA antibody and eluted with HA peptides (HA eluate). The eluate was further processed in a Nickel affinity column. After washing, His–β-actin was eluted with buffer containing 250 mM imidazole (1 ml three times). Each fraction was analyzed by silver staining and by immunoblotting with anti-HA and anti-His antibodies. Flow through: the flow-through fraction of the Nickel affinity column. (C) Fractionation of β-actin–Arp4 heterocomplexes on sucrose gradient centrifugation. Arp4–2HA and His–β-actin were co-produced in bacteria, purified on Nickel affinity gel, and then immunoprecipitated with anti-HA affinity matrix. The eluate was layered over 5–20% linear sucrose gradients and fractionated. Twenty-eight fractions were collected and analyzed by SDS-PAGE followed by silver staining. The signal intensities of the bands were quantified and are shown with the maximum values for each set at 100 (lower-left panel). (D) Re-fractionation of β-actin–Arp4 heterocomplexes. Fractions 6 and 7 from C were collected and re-subjected to 5–20% linear sucrose gradient centrifugation as above. Only the data for fractions 1–14 are presented. Non-specific bands apparent at 50, 60 and 70 kDa are due to contamination with marker proteins used for SDS-PAGE.

To gain further insight into potential heterodimer formation by β-actin and Arp4, we sought to isolate the recombinant heterocomplex produced by an in vitro transcription–translation system with bacterial extracts. We used the bacterial system to exclude the possibility that any observed β-actin–Arp4 binding was mediated by other eukaryotic proteins. For these experiments, β-actin was tagged with six histidines (His) at its N-terminus, and Arp4 was tagged with the hemagglutinin (HA) epitope at its N-terminus. His–β-actin and HA–Arp4 were co-synthesized and subjected to immunopurification with anti-HA antibody. As shown in Fig. 3B, His–β-actin was co-eluted with HA–Arp4 by excess HA peptides. The eluate was further incubated with Nickel affinity gel and then eluted with imidazole. We finally obtained purified (His–β-actin)–(HA–Arp4) heterocomplexes (Fig. 3B). The ratio of His–β-actin to HA–Arp4 appeared to be stoichiometric in the purified complexes, as estimated from the silver staining intensity.

We then tried to determine the molecular mass of reconstituted β-actin–Arp4 heterocomplexes. In these studies, Arp4 was tagged with the tandem HA at its C-terminus and, together with His–β-actin, was bacterially expressed and purified. Then, the complexes were subjected to sucrose gradient centrifugation. They appeared to fractionate somewhat widely, from an apparent molecular mass of ∼100 kDa (Fig. 3C, fraction 6) to ∼400 kDa (fraction 10). In the higher molecular mass fractions, a ∼75 kDa protein was also co-eluted (fractions 8–10), which turned out to be a bacterial chaperon, DnaK, as determined by immunoblot analysis (data not shown). One possible interpretation for the result is while β-actin and Arp4 can form heterodimers, they readily bind to DnaK, leading to an increase in the molecular mass. In this context, it is notable that actin has structural similarity to Hsp70 (Kabsch and Holmes, 1995). It also remains possible that β-actin and Arp4 can form heteropolymers. To further clarify whether β-actin and Arp4 can form heterodimers in this experimental setting, we collected fractions 6 and 7, containing the putative dimers (Fig. 3C), and re-subjected them to sucrose gradient centrifugation. They re-fractionated from fraction 5 (∼80 kDa) to fraction 9 (∼200 kDa), with a peak in fraction 7 (Fig. 3D). The data provide further support for the notion that β-actin and Arp4 can form heterodimers. When human cells were transfected with HA–Arp4 and then subjected to immunoprecipitation with anti-HA antibodies, endogenous Arp4 was not co-precipitated (data not shown). Therefore, in human cells, there may be no complex that contains multiple Arp4 molecules. Taken together, the data suggest that Arp4 may form heterocomplexes, possibly heterodimers, with β-actin.

Creation of a series of Arp4 mutants based on structural similarities with β-actin

To assess the biological significance of Arp4–β-actin complex formation for the integrity of Brg1 complexes, we examined the influence of Arp4 mutations. For this, we utilized structural similarities between conventional actin and Arp4. When aligned using CLUSTALW program (http://align.genome.jp/clustalw/), Arp4 has overall ∼30.1% identity and ∼55.3% similarity with β-actin in the amino acid sequences (Fig. 4C). At least three relatively long insertion sequences could be identified in Arp4; amino acids K62–T72, E231–R248 and V278–H286 (Fig. 4C). High-resolution structures from crystals of actin monomers bound to other proteins or small molecules that prevent polymerization of actin have been published (Kabsch et al., 1990; Kabsch and Holmes, 1995; Hertzog et al., 2004; Nolen et al., 2004). Fig. 4A shows a ribbon diagram of human β-actin monomer drawn by the SWISS model ver.36.0003 (http://swissmodel.expasy.org//SWISS-MODEL.html) based on the above structures. Extensive mutational analyses featuring replacement of charged residues with alanine have been performed with yeast actin, identifying crucial residues located on the surface of the monomer whose mutation lead to lethality (Wertman and Drubin, 1992; Wertman et al., 1992). In principle, strong contacts are made between the minus and plus ends of the monomers during actin polymerization (Wertman and Drubin, 1992; De La Cruz and Pollard, 2001). Thus, it has been suggested that some of the identified residues may be directly involved in actin–actin binding. We reasoned that β-actin–Arp4 interactions might mimic this process. If so, it is conceivable that some of the crucial residues in actin may be conserved in Arp4.

Fig. 4.

Creation of a series of Arp4 mutants based on structural similarities with β-actin. (A) The deduced tertiary structure of the human β-actin monomer drawn by ribbon diagram, as determined with the SWISS-Model program and visualized with the ViewerLite 5.0 program. The positions of amino acids whose mutations cause recessive lethality (D286A/D288A, E334A/R335A/K336A) and dominant lethality (E205A/R206A/E207A, K326A/K328A) in budding yeast are indicated in red and blue, respectively. Note that these do not represent all mutations that lead to lethality. (B) The predicted tertiary structure of human Arp4, deduced by Rovetta server and visualized with the ViewerLite 5.0 program. Predicted positions of the amino acid residues we mutated in this study are indicated in green. Predicted regions of the three human Arp4 specific insertion sequences are indicated in orange. (C) Alignment of amino acid sequences of human β-actin and human Arp4 using the CLUSTALW program. The colored sequences correspond to those in A and B. (D) Amino acid stretches conserved between Arp4 and β-actin. The colored sequences correspond to those in A and B.

Fig. 4.

Creation of a series of Arp4 mutants based on structural similarities with β-actin. (A) The deduced tertiary structure of the human β-actin monomer drawn by ribbon diagram, as determined with the SWISS-Model program and visualized with the ViewerLite 5.0 program. The positions of amino acids whose mutations cause recessive lethality (D286A/D288A, E334A/R335A/K336A) and dominant lethality (E205A/R206A/E207A, K326A/K328A) in budding yeast are indicated in red and blue, respectively. Note that these do not represent all mutations that lead to lethality. (B) The predicted tertiary structure of human Arp4, deduced by Rovetta server and visualized with the ViewerLite 5.0 program. Predicted positions of the amino acid residues we mutated in this study are indicated in green. Predicted regions of the three human Arp4 specific insertion sequences are indicated in orange. (C) Alignment of amino acid sequences of human β-actin and human Arp4 using the CLUSTALW program. The colored sequences correspond to those in A and B. (D) Amino acid stretches conserved between Arp4 and β-actin. The colored sequences correspond to those in A and B.

E205A, R206A and E207A are dominant lethal mutations in yeast actin (Wertman and Drubin, 1992; Wertman et al., 1992) and a similar amino acid stretch (A224SKEAVR230) to those surrounding these amino acids (A204EREIVR210) was found in human Arp4 (Fig. 4C,D). Thus, we created Arp4 M1 with the K226A/E227A mutations (Fig. 4C,D). K326A/K328A are also dominant lethal mutations in yeast actin and a similar amino acid stretch (M376RLKLIA382) to those surrounding these amino acids (M325KIKIIA331) was found in human Arp4 (Fig. 4C,D). Thus, we created Arp4 M2 with R377A/L378A/K379A mutations (Fig. 4C,D). Similarly, we constructed Arp4 M3 with E388A/R389A/R390A and Arp4 M4 with I340A/R341A mutations (Fig. 4C,D). Using the Robetta server (http://robetta.bakerlab.org/), we obtained the predicted tertiary structure of human Arp4 (Fig. 4B). In this model, while the structures of subdomains 1 and 3 (equivalent to the actin plus end) appear very similar to actin, subdomains 2 and 4 (equivalent to the minus end) appear different, probably due to specific insertion sequences. The amino acid residues we mutated are all positioned in similar domains to those containing the corresponding amino acids of actin (Fig. 4A,B).

In addition, we also introduced other mutations in Arp4 where it would be expected that an interaction with β-actin would not be affected. Arp4 has a unique N-terminal six-amino-acid stretch that is lacking in β-actin (Fig. 4C). Arp4 M7 has an Y6A mutation (Fig. 4C,D), which might be associated with some transcriptional function of Arp4 (Lee et al., 2005). S14 in actin is an important residue for constitution of the ATP-binding pocket and the S14A mutation in yeast actin results in 40- to 60-fold reduction of ATP affinity (Chen et al., 1995). Although the mutation in the corresponding serine residue, S23A, in yeast Arp4 does not leads to any detectable phenotypic change (Görzer et al., 2003), we also prepared Arp4 M8 with a S20A mutation.

Arp4 mutants with reduced binding to β-actin show reduced incorporation into Brg1 complexes

To characterize properties of the prepared mutant Arp4 proteins, they were transiently overexpressed in HEK293T cells, and immunoprecipitated from prepared nuclear extracts with antibodies against the HA tag present at the C-terminus of Arp4. As shown in Fig. 5A, β-actin co-precipitated with wild-type Arp4–2HA. Under the experimental conditions, the efficiency of enrichment of Brg1 with anti-Brg1 antibodies appeared similar to that of Arp4 (Fig. 2C, Fig. 5C), suggesting that a large proportion of endogenous Arp4 proteins are associated with Brg1 in the nucleus, with the rest being in complex with other chromatin modifying proteins such as Tip60. In addition, only a small amount Brg1 co-precipitated with Arp4–2HA (data not shown). Furthermore, while Arp4–2HA was detectable in Flag–Brg1-HSA immunoprecipitates prepared from cells co-transfected with Arp4–2HA and Flag–Brg1-HSA (Fig. 5B), the Flag–Brg1-HSA was undetectable in Arp4-2HA immunoprecipitates (Fig. 5D). Therefore, the observed β-actin co-precipitation with overexpressed Arp4–2HA in Fig. 5A may mainly represent direct binding in the cells. Interestingly, co-precipitation of β-actin with the Arp4 M1 to M4 mutants was significantly decreased whereas that with Arp4 M7 and M8 was not (Fig. 5A), suggesting that the conserved residues are involved in the actin–Arp4 interaction.

Fig. 5.

Arp4 mutants with reduced binding to β-actin show reduced incorporation into the Brg1 complex. (A) Arp4 M1–M4 mutants show reduced binding to β-actin. HEK293T cells were transfected with the indicated Arp4–2HA expression vectors and nuclear extracts were prepared. After immunoprecipitation with anti-HA antibodies, the immunoprecipitates (IP) were blotted with the indicated antibodies (upper panels). Four percent of the input for HA–Arp4 and 0.07% of the input for β-actin were loaded. To estimate the co-precipitation efficiency of β-actin with Arp4–2HA, the signal intensities of the bands of co-precipitated β-actin were normalized to those of the precipitated Arp4–2HA. The means±s.d. of data from two independent experiments are shown with β-actin co-precipitation with the wild type Arp4–2HA set at 100 (lower panel). (B) Mutations M1–M4 in Arp4 compromise interactions of Arp4–2HA with Flag–Brg1-HSA. HEK293T cells were co-transfected with the indicated Arp4–2HA expression vectors along with Flag–Brg1-HSA expression vector and whole-cell extracts were prepared. After immunoprecipitation with anti-Flag antibodies, the precipitates were blotted with the indicated antibodies (upper panels). Four percent of the inputs were also loaded. To estimate the co-precipitation efficiency of Arp4–2HA with Flag–Brg1-HSA, the signal intensities of the bands of co-precipitated Arp4–2HA were normalized to those of the precipitated Flag–Brg1-HSA. The means±s.d. of data from two independent experiments are shown with co-precipitation with the wild type Arp4–2HA set at 100 (lower panel). (C) Arp4 mutants with reduced binding to β-actin show reduced incorporation into the Brg1 complex. HEK293T cells were transfected with the indicated Arp4–2HA expression vectors, and nuclear extracts were prepared. After immunoprecipitation with anti-Brg1 antibodies, the precipitates were blotted with the indicated antibodies (upper panels). Seventeen percent of the input for Brg1, BAF170, endogenous Arp4 and Arp4–2HA or 0.3% of the input for β-actin were loaded. To estimate the co-precipitation efficiency of Arp4–2HA with the Brg1 complex, the signal intensities of the bands of co-precipitated Arp4–2HA were normalized to those of the precipitated endogenous Arp4. The means±s.d. of data from two independent experiments are shown with the wild-type HA–Arp4 co-precipitation with the Brg1 complex set to 100 (lower panel). (D,E) Estimation of interaction of endogenous Arp4 and transfected Arp4–2HA with endogenous Brg1 complexes or exogenous Flag–Brg1-HSA. (D) As in B, HEK293T cells were transfected with the wild-type Arp4–2HA expression vectors along with the Flag–Brg1-HSA expression vector and whole-cell extracts were immunoprecipitated with anti-Flag or anti-HA antibodies. Precipitates were blotted with the indicated antibodies. (E) As in C, HEK293T cells were transfected with the wild-type Arp4–2HA expression vector and nuclear extracts were immunoprecipitated with anti-Brg1 antibodies and blotted as indicated.

Fig. 5.

Arp4 mutants with reduced binding to β-actin show reduced incorporation into the Brg1 complex. (A) Arp4 M1–M4 mutants show reduced binding to β-actin. HEK293T cells were transfected with the indicated Arp4–2HA expression vectors and nuclear extracts were prepared. After immunoprecipitation with anti-HA antibodies, the immunoprecipitates (IP) were blotted with the indicated antibodies (upper panels). Four percent of the input for HA–Arp4 and 0.07% of the input for β-actin were loaded. To estimate the co-precipitation efficiency of β-actin with Arp4–2HA, the signal intensities of the bands of co-precipitated β-actin were normalized to those of the precipitated Arp4–2HA. The means±s.d. of data from two independent experiments are shown with β-actin co-precipitation with the wild type Arp4–2HA set at 100 (lower panel). (B) Mutations M1–M4 in Arp4 compromise interactions of Arp4–2HA with Flag–Brg1-HSA. HEK293T cells were co-transfected with the indicated Arp4–2HA expression vectors along with Flag–Brg1-HSA expression vector and whole-cell extracts were prepared. After immunoprecipitation with anti-Flag antibodies, the precipitates were blotted with the indicated antibodies (upper panels). Four percent of the inputs were also loaded. To estimate the co-precipitation efficiency of Arp4–2HA with Flag–Brg1-HSA, the signal intensities of the bands of co-precipitated Arp4–2HA were normalized to those of the precipitated Flag–Brg1-HSA. The means±s.d. of data from two independent experiments are shown with co-precipitation with the wild type Arp4–2HA set at 100 (lower panel). (C) Arp4 mutants with reduced binding to β-actin show reduced incorporation into the Brg1 complex. HEK293T cells were transfected with the indicated Arp4–2HA expression vectors, and nuclear extracts were prepared. After immunoprecipitation with anti-Brg1 antibodies, the precipitates were blotted with the indicated antibodies (upper panels). Seventeen percent of the input for Brg1, BAF170, endogenous Arp4 and Arp4–2HA or 0.3% of the input for β-actin were loaded. To estimate the co-precipitation efficiency of Arp4–2HA with the Brg1 complex, the signal intensities of the bands of co-precipitated Arp4–2HA were normalized to those of the precipitated endogenous Arp4. The means±s.d. of data from two independent experiments are shown with the wild-type HA–Arp4 co-precipitation with the Brg1 complex set to 100 (lower panel). (D,E) Estimation of interaction of endogenous Arp4 and transfected Arp4–2HA with endogenous Brg1 complexes or exogenous Flag–Brg1-HSA. (D) As in B, HEK293T cells were transfected with the wild-type Arp4–2HA expression vectors along with the Flag–Brg1-HSA expression vector and whole-cell extracts were immunoprecipitated with anti-Flag or anti-HA antibodies. Precipitates were blotted with the indicated antibodies. (E) As in C, HEK293T cells were transfected with the wild-type Arp4–2HA expression vector and nuclear extracts were immunoprecipitated with anti-Brg1 antibodies and blotted as indicated.

Recently, it was shown that when Flag–Brg1-HSA and HA–Arp4 are co-transfected and immunoprecipitated with anti-Flag antibodies, (Flag–Brg1-HSA)–(HA–Arp4)–β-actin ternary complexes can be detected (Szerlong et al., 2008). Here, we investigated the effects of the Arp4 mutations on this ternary complex formation. HEK293T cells were co-transfected with Flag–Brg1-HSA and Arp4–2HA and subjected to immunoprecipitation with anti-Flag antibodies. Wild-type Arp4–2HA and endogenous β-actin specifically co-precipitated with Flag–Brg1-HSA (Fig. 5B,D). We found that co-precipitation of Arp4–2HA was impaired by mutations M1 to M4, whereas mutations M7 and M8 were without influence (Fig. 5B). Previous extensive mutagenic analyses of yeast Sth1-HSA–Arp7–Arp9 complexes suggested that Arp7 and Arp9 may interact with the HSA domain as a dimer and that their binding might be mediated by many contact sites along the HSA (Szerlong et al., 2008). Thus, our data could be interpreted as Arp4–β-actin complex formation being a prerequisite for its association with Brg1-HSA. This appears consistent with the fact that incorporation of yeast Arp7 and Arp9 into the RSC complex is interdependent (Szerlong et al., 2003). As shown in Fig. 5D, in HEK293T cells overexpressing Flag–Brg1-HSA and Arp4–2HA, the relative amount of Flag–Brg1-HSA and endogenous Arp4–β-actin complexes appeared higher than that of Flag–Brg1-HSA and exogenous Arp4–2HA–β-actin complexes. Therefore, the effects of the Arp4 mutations on β-actin co-precipitation were not observed in Flag–Brg1-HSA immunoprecipitates from HEK293T cells overexpressing Flag–Brg1-HSA and Arp4–2HA (data not shown).

The above data suggest that the mutations M1 to M4 may compromise Arp4 incorporation into genuine Brg1 complexes. We therefore precipitated Brg1 complexes with anti-Brg1 antibodies from nuclear extracts of HEK293T cells overexpressing Arp4–2HA. Exogenous wild-type Arp4–2HA was incorporated into Brg1 complexes although inefficiently compared with the endogenous Arp4 (Fig. 5C,E). Interestingly, the Arp4 M1 to M4 mutants that have a reduced affinity for β-actin also showed reduced incorporation into Brg1 complexes (Fig. 5C). In these experiments, efficiencies of co-precipitation of endogenous Arp4 were all equal. In addition, the Arp4 M7 mutant showed only a partial reduction in incorporation into the complexes, whereas M8 mutants were incorporated with comparable efficacy to the wild type. Similar effects of the Arp4 mutations on the Arp4–β-actin interaction and incorporation into Brg1 complexes were also observed with HeLa cells (data not shown).

Effects of the Arp4 mutations on interactions with other chromatin-modifying enzyme complexes

We further examined the effects of the Arp4 mutations on interactions with other chromatin-modifying enzyme complexes. We first transiently overexpressed the Arp4–2HA mutants in HeLa cells stably expressing Flag–Tip60 at near-endogenous levels and carried out immunoprecipitation with anti-Flag antibodies. As expected from a previous report (Cai et al., 2003), wild-type Arp4–2HA specifically co-precipitated with Flag–Tip60 (Fig. 6A). Although neither the mutation M7 nor M8 affected incorporation of Arp4–2HA into Tip60 HAT complexes, this was partially impaired by the mutations M1 to M4, with ∼50% inhibition by the M3 mutation (Fig. 6A).

Fig. 6.

Effects of the Arp4 mutations on the interactions with Tip60 HAT complexes and Myc-associated complexes. (A) HeLa cells stably expressing Flag–Tip60 were transfected with the indicated Arp4–2HA expression vectors and nuclear extracts were prepared. In FLAG-Tip60 (−) lanes, parental HeLa cells were used. After immunoprecipitation with anti-Flag antibodies, the precipitates were blotted with the indicated antibodies (upper panels). Four percent of the input was loaded. To estimate the co-precipitation efficiency of Arp4–2HA with Flag-Tip60, the signal intensities of the bands of co-precipitated Arp4–2HA were normalized to those of the precipitated Flag–Tip60. The means±s.d. of data from two independent experiments are shown with co-precipitation of the wild-type Arp4–2HA with Flag–Tip60 set at 100 (lower panel). (B) HEK293T cells were transfected with the indicated Arp4–2HA expression vectors, together with Flag–Myc expression vector, and nuclear extracts were prepared. After immunoprecipitation with anti-Flag antibody, the precipitates were blotted with the indicated antibodies (upper panels). Four percent of the input was loaded. To estimate the co-precipitation efficiency of Arp4–2HA with Flag–Myc, the signal intensities of the bands of co-precipitated Arp4–2HA were normalized to those of the precipitated Flag-Myc. The means±s.d. from two independent experiments are shown with co-precipitation of the wild-type Arp4-2HA with Flag–Myc set at 100 (lower panel).

Fig. 6.

Effects of the Arp4 mutations on the interactions with Tip60 HAT complexes and Myc-associated complexes. (A) HeLa cells stably expressing Flag–Tip60 were transfected with the indicated Arp4–2HA expression vectors and nuclear extracts were prepared. In FLAG-Tip60 (−) lanes, parental HeLa cells were used. After immunoprecipitation with anti-Flag antibodies, the precipitates were blotted with the indicated antibodies (upper panels). Four percent of the input was loaded. To estimate the co-precipitation efficiency of Arp4–2HA with Flag-Tip60, the signal intensities of the bands of co-precipitated Arp4–2HA were normalized to those of the precipitated Flag–Tip60. The means±s.d. of data from two independent experiments are shown with co-precipitation of the wild-type Arp4–2HA with Flag–Tip60 set at 100 (lower panel). (B) HEK293T cells were transfected with the indicated Arp4–2HA expression vectors, together with Flag–Myc expression vector, and nuclear extracts were prepared. After immunoprecipitation with anti-Flag antibody, the precipitates were blotted with the indicated antibodies (upper panels). Four percent of the input was loaded. To estimate the co-precipitation efficiency of Arp4–2HA with Flag–Myc, the signal intensities of the bands of co-precipitated Arp4–2HA were normalized to those of the precipitated Flag-Myc. The means±s.d. from two independent experiments are shown with co-precipitation of the wild-type Arp4-2HA with Flag–Myc set at 100 (lower panel).

We also investigated effects of the mutations on the Arp4–Myc interaction. HEK293T cells were co-transfected with Flag–Myc and Arp4–2HA and subjected to immunoprecipitation with anti-Flag antibodies. As reported previously (Park et al., 2002), wild-type Arp4–2HA specifically co-precipitated with Flag–Myc (Fig. 6B). Although neither mutation M7 nor M8 affected interaction of Arp4–2HA with Myc, this was impaired by the mutations M1 to M4, with ∼60% inhibition by the M3 mutation (Fig. 6B).

Overall, these data appear consistent with the notion that the Arp4 mutants with reduced affinity for β-actin also exhibit impaired incorporation into Tip60 HAT and Myc-associated transcription complexes, although the extent may differ depending on the target.

Human Arp4 possesses core histone-binding activity

It is known that budding yeast Arp4 and Arp8 have histone binding properties (Harata et al., 1999; Shen et al., 2003). To investigate whether human Arp4 also possesses similar activity, we purified GST–Arp4 and examined binding to core histones by pull-down assay. Core histones strongly bound to GST–Arp4 but not to control GST (Fig. 7A) and some binding remained even after washing with buffer containing 1M NaCl. We then compared the histone binding capability of each mutant Arp4 protein. The mutations we introduced had only a little effect on histone binding, with ∼25% inhibition by M3 (Fig. 7B). It will be intriguing to identify Arp4 mutants that severely lose the histone binding activity. Critical residues might be positioned in the Arp4 specific insertion sequences.

Fig. 7.

Human Arp4 possesses core histone binding activity. (A) GST–Arp4 binds to core histones in vitro. GST–Arp4 or GST were incubated with purified core histones. The bound proteins were first eluted with buffer containing 1 M NaCl, followed by elution with SDS. The eluted proteins were analyzed by CBB staining. Five percent of the input was loaded. (B) Effects of the Arp4 mutations on histone binding. A GST pulldown assay was performed as above. Five percent of the input was loaded. To estimate the co-precipitation efficiency of core histones with GST-Arp4, the signal intensities of the bands of co-precipitated core histones were normalized to those of the precipitated GST–Arp4. The means±s.d. from three independent experiments are shown, with histone binding to the wild-type GST–Arp4 set at 100 (lower panel).

Fig. 7.

Human Arp4 possesses core histone binding activity. (A) GST–Arp4 binds to core histones in vitro. GST–Arp4 or GST were incubated with purified core histones. The bound proteins were first eluted with buffer containing 1 M NaCl, followed by elution with SDS. The eluted proteins were analyzed by CBB staining. Five percent of the input was loaded. (B) Effects of the Arp4 mutations on histone binding. A GST pulldown assay was performed as above. Five percent of the input was loaded. To estimate the co-precipitation efficiency of core histones with GST-Arp4, the signal intensities of the bands of co-precipitated core histones were normalized to those of the precipitated GST–Arp4. The means±s.d. from three independent experiments are shown, with histone binding to the wild-type GST–Arp4 set at 100 (lower panel).

Our analyses indicated that Arp4 forms heterocomplexes with β-actin (Fig. 3; Fig. 5A). At least in a bacterial expression system, they can form heterodimers (Fig. 3), reminiscent of the Arp7–Arp9 heterodimer (Szerlong et al., 2003). However, Arp4–β-actin heterodimers may be unstable under physiological condition, as follows. When (Arp4–Flag)–β-actin complexes immunopurified from Arp4–Flag-transfected HEK293T cells were analyzed by sucrose gradient centrifugation, many of them fractionated around their monomeric mass (data not shown). In addition, when purified monomeric G-actin proteins were mixed with recombinant Arp4 proteins and then the mixtures were subjected to gel filtration chromatography, again the dimeric complexes were hardly detectable (data not shown). The instability of Arp4–β-actin complex appears consistent with the reported finding that yeast Arp4 may interact with monomeric G-actins but cannot efficiently sequester them during the actin polymerization (Fenn et al., 2011). Under physiological condition, Arp4–β-actin dimers may be formed only transiently as intermediates for complete chromatin-modifying complexes. Indeed, it has been suggested that the HSA domains can form stable complexes with actin and Arp4 (Szerlong et al., 2008). It is possible that the relatively stable (Arp4–HA)–(His–β-actin) dimers produced in bacteria (Fig. 3) may have some biases in the structure compared with native ones. Further studies on this point would be required for an in-depth understanding.

To investigate the biological significance of Arp4–β-actin complex formation for the integrity of Brg1 complexes, we searched for Arp4 mutations that impaired the interaction with β-actin. The findings obtained with the Arp4 M2 (R377A/L378A/K379A) and M4 (I340A/R341A) mutants appear especially intriguing. Both the mutations change residues that may lie on the surface of Arp4 subdomain 3 (Fig. 4), and significantly reduced Arp4–β-actin interaction in human cells (Fig. 5A). Thus, these amino acids may be critical for Arp4–β-actin interactions. Incorporation of these Arp4 mutants into the Brg1 complexes was remarkably reduced (Fig. 5C), and their interaction with the Brg1-HSA domain was also impaired (Fig. 5B). To a certain extent, this was also the case for Tip60 HAT and Myc-associated complexes (Fig. 6), suggesting that Arp4–β-actin interactions may also be involved in efficient formation of these complexes. Binding of Arp4 to core histones was unchanged by these mutations (Fig. 7), suggesting that effects are not nonspecific. Arp4 M3 mutations (E388A/R389A/R390A) change residues that also may lie on the surface of Arp4 subdomain 3 but are somewhat distant from those changed by M2 and M4 mutations (Fig. 4). They significantly reduced Arp4–β-actin interaction in human cells (Fig. 5A). Interaction of the Arp4 M3 mutant with Brg1 HSA domain was inhibited (Fig. 5B) and, in line with this, its incorporation into Brg1 complexes was significantly reduced (Fig. 5C). Interaction of the Arp4 M3 mutant with the Tip60 HAT complex and Myc-associated complex was also partially impaired (Fig. 6). These results indicate a role for these residues in Arp4–β-actin interactions and formation of chromatin-modifying enzyme complexes. Binding of Arp4 to core histones was reduced only by ∼25% with the M3 mutations (Fig. 7). Arp4 M1 (K226A/E227A) mutations introduced into residues likely to lie on the surface of subdomain 4 (Fig. 4) significantly reduced the Arp4–β-actin interaction in human cells (Fig. 5A). Incorporation of the Arp4 M1 mutants into the Brg1 complexes was remarkably reduced (Fig. 5B,C) and that into Tip60 HAT complex and Myc-associated complex was also partially impaired (Fig. 6). Binding of Arp4 to core histones was reduced only by ∼20% with the M1 mutations (Fig. 7). For the Arp4 M7 mutation (Y6A), which alters a N-terminal residue (Fig. 4), associations with β-actin and Brg1 complex in human cells were only partially impaired (Fig. 5), whereas those with Tip60 and Myc were not affected at all (Fig. 6). It also did not affect Arp4 histone binding (Fig. 7). For the Arp4 M8 mutation (S20A), which alters a residue that may be associated with a potential ATP-binding pocket (Fig. 4), no change was observed in the phenotypes we tested (Figs 5, 6, 7). In aggregate, our mutational analyses suggest that Arp4–β-actin heterocomplex formation may be a crucial feature for incorporation of Arp4 into large chromatin-modifying enzyme complexes, although some allele-specific and target-specific phenotypes were also observed. This appears consistent with the fact that yeast Arp7 and Arp9 can form heterodimers and that their incorporation into the RSC complex is interdependent (Szerlong et al., 2003). Although our mutational analyses indicate the importance of Arp4–β-actin heterocomplex formation, it remains unclear how these two proteins interact. In this regard, the data obtained by Fenn et al. suggest that yeast Arp4 may bind to the plus end of monomeric actin (Fenn et al., 2011). It will be interesting to clarify the structure of Arp4–actin or the HSA–Arp4–actin complex.

Nuclear actin and Arps play multiple roles in a context-dependent manner, for example, yeast and human Arp4 can bind to histones (Harata et al., 1999; Shen et al., 2003) (Fig. 7) and murine neuron-specific Arp4/BAF53b is required for recruitment of BAF complex to specific promoters (Wu et al., 2007). In the present study, we found a new aspect to human Arp4 functions; silencing of Arp4 by RNAi impairs the integrity of Brg1 chromatin remodeling complexes and accelerates the degradation of Brg1 and BAF170 in several human cell lines, including normal cells (Fig. 1). In this regard, the previous demonstration that Brg1 stability is affected by BAF47 and BAF155 subunits is of interest (Cui et al., 2004; Sohn et al., 2007). The requirement of human Arp4 for Brg1 complex integrity is reminiscent of the previous finding that mutations in Arp4 disrupt NuA4 HAT complex in budding yeast (Galarneau et al., 2000). On the other hand, Arp4 and actin appear to be dispensable for the INO80 complex assembly (Shen et al., 2003). In this regard, it is also notable that Brg1, BAF170, BAF155 and BAF47 can form stable complexes without the help of actin and Arp4 when overexpressed in insect cells (Phelan et al., 1999). Lee et al. have reported that silencing of Arp4 by siRNA did not decrease the steady-state levels of Brg1 proteins in NIH3T3 cells (Lee et al., 2007) and it has also been reported that neural BAF complexes can be assembled in the absence of the neuron-specific Arp4/BAF53b in murine neural tissues (Wu et al., 2007). Thus, it is possible that the impact of Arp4 on Brg1 complex integrity is dependent on cell type and/or cellular condition.

There are several possibilities to explain why Arp4 depletion influences the integrity of Brg1 complexes in certain human cells. One is that Arp4 depletion might reduce the expression of some gene(s) that are required for the complex integrity. For example, transcription of some other subunit(s) than Arp4 might be reduced. However, this was ruled out by our microarray analysis (Fig. 1E). It is also possible that canonical chaperon proteins like Hsp70 and Hsp40 are involved in the folding of certain subunit(s) and/or in the complex assembly and their expression is repressed. However, no significant change was found in transcription levels of canonical chaperon genes in the microarray analysis (data not shown). We also must consider the possibility that Arp4 depletion primarily induces degradation of Brg1 and BAF170. Since degradation of actin and Arp4 was unchanged, it seems unlikely that protein degradation is nonspecifically accelerated in Arp4-silenced HeLa cells. Brg1 and BAF170 proteins are structurally unrelated although incorporated into the same complex. Therefore, if this scenario were the case, two specific degradation pathways must be activated by Arp4 depletion, which seems very unlikely. At least in our microarray analysis, no significant increase was found in the expression of genes that are known to be associated with specific protein degradation pathways such as ubiquitination-associated E1, E2 and E3 components (data not shown). Therefore, at this time, our preference lies with the possibility that Arp4 proteins itself may play a crucial role in the Brg1 complex assembly and/or maintenance of the complex stability. The unincorporated Brg1 and BAF170 proteins may be more easily brought to degradation.

Production of retrovirus vector expressing Arp4 shRNA

For silencing Arp4, the 19-nucleotide sequence corresponding to Arp4 cDNA nucleotides 194–212 (underlined) was expressed as an shRNA as follows. Two oligonucleotides (Arp4-S1, 5′-GATCCCCGAGATGACGGAAGCACATTTTCAAGAGAAATGTGCTTCCGTCATCTCTTTTTGGAAA-3′; and Arp4-AS1, 5′-AGCTTTTCCAAAAAGAGATGACGGAAGCACATTTCTCTTGAAAATGTGCTTCCGTCATCTCGGG-3′) were annealed and introduced into the pSI-MSCVPuro-H1R retroviral expression vector (Tatsumi et al., 2006). The sequences of other shRNAs for Arp4 (Arp4 shRNA-3 and -4) are available upon request. HeLa cells were infected with the retroviruses and selected with 0.5 µg/ml puromycin for 2 days.

DNA microarray analysis

Total RNA was isolated as described previously (Fujita et al., 2003). Preparation of cRNAs labeled with Cy3 or Cy5 CTP and hybridization to DNA microarrays were performed according to Agilent protocol. The microarray contains ∼17,800 cDNAs (Human1A Ver.2 Oligo Microarray, Agilent Technologies).

siRNA experiments

HeLa cells were transfected with siRNAs using HiPerFect (Qiagen) and HCK1/T cells (human cervical keratinocytes immortalized with telomerase) (Tatsumi et al., 2006) using Lipofectamine RNAiMAX (Invitrogen). siRNA were synthesized (IDT) with the following sequences (sense strands): Brg1-1, 5′-GGAAGAUUACUUUGCGUAUCGCGdGdC-3′; Brg1-2, 5′-CCAGAGCUGAGAUGGCAUAGGCCdTdT-3′; Arp4-2, 5′-GGCAGUGUAAUAGUGGCAGGAGGdAdA-3′; Arp4-3, 5′-GGUAGAAAGAGAUGACGGAAGCAdCdA-3′; and control scrambled, 5′-CUUCCUCUCUUUCUCUCCCUUGUdGdA-3′.

Purification and ATPase assay of the Brg1 complexes

The Brg1 complexes were immunoprecipitated directly from soluble nuclear extracts (125 µl for control HeLa cells and 500 µl for Arp4-depleted HeLa cells) with anti-Brg1 antibody beads (15 µl). The beads were washed twice with ATPase buffer (20 mM Tris-HCl pH 7.4, 10 mM MgCl2, 2 mM DTT, 1 mM ATP) and resuspended in 30 µl of the same buffer containing 10 µCi [γ-33P]ATP. When indicated, 3 µg of activated calf thymus DNA (Amersham) was included. The mixtures were incubated at 37°C for 90 min. After the reaction, an aliquot of the mixture was spotted onto a PEI cellulose plate (Merck), and chromatography was carried out in 0.5 M formic acid and 1 M LiCl. The plates were analyzed on a Bio-Imaging analyzer BAS-2500 (Fuji Film).

In vitro binding assay for actin and Arp4

HeLa cell cytoplasmic G-actin was purified essentially as described previously (Ayscough et al., 1997). Purified G-actin (0.5 mg) was coupled to 0.3 ml CNBr-activated Sepharose 4 Fast Flow in actin coupling buffer (0.1 M NaHCO3 pH 8.3, 0.5 mM CaCl2, 0.5 mM phenylmethylsulfonyl fluoride). Plasmid pcDNA3-HA-Arp4/BAF53 (kindly provided by Masahiko Harata, Tohoku University, Sendai, Japan), with which HA-tagged human Arp4/BAF53 cDNA is transcribed from the T7 promoter, was previously described (Kuroda et al., 2002). HA-tagged Arp4 proteins were synthesized by in vitro transcription–translation with a rabbit reticulocyte lysate (TNT T7 quick coupled transcription–translation system, Promega). HA–Arp4 produced by in vitro translation in 20 µl of reaction mixture was diluted with 280 µl binding buffer (20 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.1% Triton X-100, 0.5 mM MgCl2, 0.1 mM DTT, 0.1 mM ATP), mixed with 20 µl actin-beads or control BSA-beads, and incubated at 4°C for 60 min. The beads then were washed with binding buffer. Bound proteins were eluted with 30 µl SDS sample buffer (62.5 mM Tris-HCl pH 6.8, 2% SDS, 5% β-mercaptoethanol, 10% glycerol, 0.01% Bromophenol Blue).

Isolation of β-actin–Arp4 heterocomplexes

For Fig. 3B, human Arp4 cDNA was inserted into pIVEX2.6d (Roche) and β-actin cDNA (Fujita et al., 2003) was inserted into pIVEX2.4a (Roche). HA-tagged Arp4 and His-tagged β-actin were co-synthesized from pIVEX2.6d-Arp4 and pIVEX2.4a- β-actin using an in vitro transcription–translation system with bacterial extracts (RTS 500, Roche), and bound to 0.3 ml Anti-HA Affinity Matrix (Roche). After washing with NET gel buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.1% Triton X-100, 1 mM EDTA), the bound proteins were eluted with 2 ml NET gel buffer containing HA peptide (Roche) at 1 mg/ml. The eluate was then loaded onto 1 ml Nickel affinity gel (HIS-Select, Sigma). The beads were washed with 15 ml buffer H (50 mM sodium phosphate pH 8.0, 250 mM NaCl) containing 10 mM imidazole and eluted with 3 ml buffer H containing 250 mM imidazole.

For Fig. 3C,D, Arp4–2HA and His–β-actin were simultaneously expressed in Escherichia coli strain BL-21 Lys S using pERDuet-1 vector (Novagen). First, Arp4 cDNA was inserted into pERDuet-1 and then the C-terminal S-tag was replaced with tandem HA-tag. Finally, β-actin cDNA was inserted into pETDuet-1-Arp4. HA–Arp4 and His–β-actin were co-produced from pETDuet-1-β-actin-Arp4, purified on Nickel affinity gel and then immunoprecipitated with Anti-HA Affinity Matrix. One milliliter of the eluate was layered over 5–20% linear sucrose gradients (9 ml) containing 0.05% Triton X-100 and centrifuged at 28,000 rpm for 20 h at 4°C in a P40ST swing rotor (Hitachi Koki Co., Ltd.). Twenty-eight fractions (0.4 ml each) were collected and analyzed.

Modeling of tertiary structures of β-actin and Arp4

The predicted tertiary structure of human Arp4 was deduced using the Rovetta server (http://robetta.bakerlab.org/).

Construction of mammalian expression vectors

For mammalian expression of Arp4–2HA, the Arp4 cDNA was excised from pETDuet-1-Arp4 and inserted into pCMV/Myc/nuc (Invitrogen). To generate Arp4 mutants, oligonucleotide-directed mutagenesis (QuickChange Site-directed Mutagenesis Kit; Stratagene) was performed with each oligonucleotide and a complementary oligonucleotide as follows; Y6A, 5′-gagttatgagcggcggcgtggccggcggagatgaagttg-3′; S20A, 5′-ggagcccttgtttttgacattggagcctatactgtgagagctggttatg-3′; K226A/E227A, 5′-ccatatatgattgcatcagctgcagctgttcgtgaaggatctcc-3′; R377A/L378A/K379A, 5′-cagaaaactcctccaagtatggctgcagcattgattgcaaataatacaac-3′; E388A/R389A/R390A, 5′-gcaaataatacaacagtggctgcagcttttagctcatggattggcggc-3′; and I340A/R341A, 5′-gggatgtgtgatattgatgctgctccaggtctctatggcagtg-3′.

For expression of Flag–Myc, the Myc cDNA was inserted into p3xFLAG-CMV10 (Sigma). The expression vector for the Flag-tagged Brg1 HSA domain (pcDNA3-Flga-Brg1 HSA) was as described previously (Szerlong et al., 2008) and kindly provided by Bradley R. Cairns (University of Utah, Salt Lake City, UT).

Immunoprecipitation

The indicated plasmids (total 6 µg) were transiently transfected into 3×106 HEK293T cells in 100-mm culture dishes with TransIT-293 reagent (Mirus, Madison, WI). HeLa cells were transfected with HilyMax Transfection reagent (Dojin Chemicals, Kumamoto, Japan). Forty-eight hours after transfection, cells were lysed in 1 ml mCSK buffer containing 0.1% Triton X-100, 1 mM DTT, and multiple protease inhibitors. After centrifugation, the remnant unclear pellet was resuspended in 1 ml mCSK buffer containing 500 mM NaCl, 0.1% Triton X-100, 1 mM DTT and multiple protease inhibitors to obtain nuclear extract. For Fig. 5B, transfected cells were directly lysed in mCSK buffer containing 500 mM NaCl, 0.1% Triton X-100 and 1 mM DTT. After centrifugation, aliquots of the extracts were immunoprecipitated with anti-HA (3F10, Roche), anti-Flag (M2, Sigma), anti-Brg1 or control IgG (Dako) antibodies and protein-G–Sepharose beads (Amersham Bioscience). The beads were washed with the buffer and immunoprecipitates were eluted with SDS sample buffer.

GST–Arp4 pulldown assay for histone binding

Core histones were purified from HeLa cells essentially as described previously (Kundu et al., 1999). For bacterial expression of glutathione S-transferase (GST)–Arp4, Arp4 cDNA was introduced into pGEX6P-2 (Amersham Biosciences). Wild-type GST–Arp4 and GST–Arp4 mutants were produced in Escherichia coli strain BL-21 Lys S and purified as described previously (Sugimoto et al., 2008). GST–Arp4 (7 µg) was incubated with 5 µg of purified core histone in binding buffer (200 mM NaCl, 20 mM Tris-HCl pH 7.4, 1 mM DTT, 0.05% Triton X-100, 5% glycerol) at 4°C for 180 min and then collected on glutathione beads. The beads were washed three times with 1 ml binding buffer and bound proteins were eluted with buffer containing 1M NaCl, followed by complete elution with SDS sample buffer. The samples were finally analyzed by SDS-PAGE with Coomassie Brilliant Blue (CBB) staining.

Immunoblotting and antibodies

Immunoblotting and quantification of band signals were performed as described previously (Sugimoto et al., 2008). For production of anti-Brg1 antibodies, a peptide corresponding to the 50-amino-acid C-terminal region of Brg1 was synthesized and used for immunizing rabbits. The antibodies were affinity-purified. For immunopurification of Brg1 complexes, the anti-Brg1 antibodies were cross-linked to protein-G–Sepharose beads (Amersham) at 1mg/ml with dimethylpimelimidate. Anti-DnaK antibodies were provided by Masaaki Kanemori (Kanazawa University, Kanazawa, Japan). Other antibodies were purchased as follows: Flag-tag (M2, Sigma), HA-tag (HA.11, BAbCO and 3F10, Roche), His-tag (70796-3, Novagen), actin (MAB1501, Chemicon), human BAF53 (ab3882, abcam), BAF170 (E-6, Santa Cruz), BAF47/hSNF5/Ini1 (H-300, Santa Cruz), and PARP1 (#9542, Cell Signaling).

We thank Masahiko Harata (Tohoku University) and Bradley R. Cairns (University of Utah) for plasmids, Tomona Tsuji and Satoko Yoshida for technical assistance, and Akiko Noguchi for secretarial work. We also thank Masaaki Kanemori for providing antibodies.

Funding

This work was supported in part by a grant to M.F. from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Ayscough
K. R.
,
Stryker
J.
,
Pokala
N.
,
Sanders
M.
,
Crews
P.
,
Drubin
D. G.
(
1997
).
High rates of actin filament turnover in budding yeast and roles for actin in establishment and maintenance of cell polarity revealed using the actin inhibitor latrunculin-A.
J. Cell Biol.
137
,
399
416
.
Bettinger
B. T.
,
Gilbert
D. M.
,
Amberg
D. C.
(
2004
).
Actin up in the nucleus.
Nat. Rev. Mol. Cell Biol.
5
,
410
415
.
Blessing
C. A.
,
Ugrinova
G. T.
,
Goodson
H. V.
(
2004
).
Actin and ARPs: action in the nucleus.
Trends Cell Biol.
14
,
435
442
.
Boyer
L. A.
,
Peterson
C. L.
(
2000
).
Actin-related proteins (Arps): conformational switches for chromatin-remodeling machines?
Bioessays
22
,
666
672
.
Cai
Y.
,
Jin
J.
,
Tomomori–Sato
C.
,
Sato
S.
,
Sorokina
I.
,
Parmely
T. J.
,
Conaway
R. C.
,
Conaway
J. W.
(
2003
).
Identification of new subunits of the multiprotein mammalian TRRAP/TIP60-containing histone acetyltransferase complex.
J. Biol. Chem.
278
,
42733
42736
.
Cairns
B. R.
,
Erdjument–Bromage
H.
,
Tempst
P.
,
Winston
F.
,
Kornberg
R. D.
(
1998
).
Two actin-related proteins are shared functional components of the chromatin-remodeling complexes RSC and SWI/SNF.
Mol. Cell
2
,
639
651
.
Chen
X.
,
Peng
J.
,
Pedram
M.
,
Swenson
C. A.
,
Rubenstein
P. A.
(
1995
).
The effect of the S14A mutation on the conformation and thermostability of Saccharomyces cerevisiae G-actin and its interaction with adenine nucleotides.
J. Biol. Chem.
270
,
11415
11423
.
Cui
K.
,
Tailor
P.
,
Liu
H.
,
Chen
X.
,
Ozato
K.
,
Zhao
K.
(
2004
).
The chromatin-remodeling BAF complex mediates cellular antiviral activities by promoter priming.
Mol. Cell. Biol.
24
,
4476
4486
.
De La Cruz
E. M.
,
Pollard
T. D.
(
2001
).
Structural biology. Actin' up.
Science
293
,
616
618
.
Fenn
S.
,
Breitsprecher
D.
,
Gerhold
C. B.
,
Witte
G.
,
Faix
J.
,
Hopfner
K-P.
(
2011
).
Structural biochemistry of nuclear actin-related proteins 4 and 8 reveals their interaction with actin.
EMBO J.
30
,
2153
2166
.
Fomproix
N.
,
Percipalle
P.
(
2004
).
An actin-myosin complex on actively transcribing genes.
Exp. Cell Res.
294
,
140
148
.
Fujita
M.
,
Ichinose
S.
,
Kiyono
T.
,
Tsurumi
T.
,
Omori
A.
(
2003
).
Establishment of latrunculin-A resistance in HeLa cells by expression of R183A D184A mutant β-actin.
Oncogene
22
,
627
631
.
Galarneau
L.
,
Nourani
A.
,
Boudreault
A. A.
,
Zhang
Y.
,
Héliot
L.
,
Allard
S.
,
Savard
J.
,
Lane
W. S.
,
Stillman
D. J.
,
Côté
J.
(
2000
).
Multiple links between the NuA4 histone acetyltransferase complex and epigenetic control of transcription.
Mol. Cell
5
,
927
937
.
Görzer
I.
,
Schüller
C.
,
Heidenreich
E.
,
Krupanska
L.
,
Kuchler
K.
,
Wintersberger
U.
(
2003
).
The nuclear actin-related protein Act3p/Arp4p of Saccharomyces cerevisiae is involved in transcription regulation of stress genes.
Mol. Microbiol.
50
,
1155
1171
.
Harata
M.
,
Oma
Y.
,
Mizuno
S.
,
Jiang
Y. W.
,
Stillman
D. J.
,
Wintersberger
U.
(
1999
).
The nuclear actin-related protein of Saccharomyces cerevisiae, Act3p/Arp4, interacts with core histones.
Mol. Biol. Cell
10
,
2595
2605
.
Hertzog
M.
,
van Heijenoort
C.
,
Didry
D.
,
Gaudier
M.
,
Coutant
J.
,
Gigant
B.
,
Didelot
G.
,
Préat
T.
,
Knossow
M.
,
Guittet
E.
et al.  (
2004
).
The β-thymosin/WH2 domain; structural basis for the switch from inhibition to promotion of actin assembly.
Cell
117
,
611
623
.
Hofmann
W. A.
,
Stojiljkovic
L.
,
Fuchsova
B.
,
Vargas
G. M.
,
Mavrommatis
E.
,
Philimonenko
V.
,
Kysela
K.
,
Goodrich
J. A.
,
Lessard
J. L.
,
Hope
T. J.
et al.  (
2004
).
Actin is part of pre-initiation complexes and is necessary for transcription by RNA polymerase II.
Nat. Cell Biol.
6
,
1094
1101
.
Hu
P.
,
Wu
S.
,
Hernandez
N.
(
2004
).
A role for β-actin in RNA polymerase III transcription.
Genes Dev.
18
,
3010
3015
.
Kabsch
W.
,
Holmes
K. C.
(
1995
).
The actin fold.
FASEB J.
9
,
167
174
.
Kabsch
W.
,
Mannherz
H. G.
,
Suck
D.
,
Pai
E. F.
,
Holmes
K. C.
(
1990
).
Atomic structure of the actin:DNase I complex.
Nature
347
,
37
44
.
Kukalev
A.
,
Nord
Y.
,
Palmberg
C.
,
Bergman
T.
,
Percipalle
P.
(
2005
).
Actin and hnRNP U cooperate for productive transcription by RNA polymerase II.
Nat. Struct. Mol. Biol.
12
,
238
244
.
Kundu
T. K.
,
Wang
Z.
,
Roeder
R. G.
(
1999
).
Human TFIIIC relieves chromatin-mediated repression of RNA polymerase III transcription and contains an intrinsic histone acetyltransferase activity.
Mol. Cell. Biol.
19
,
1605
1615
.
Kuroda
Y.
,
Oma
Y.
,
Nishimori
K.
,
Ohta
T.
,
Harata
M.
(
2002
).
Brain-specific expression of the nuclear actin-related protein ArpNα and its involvement in mammalian SWI/SNF chromatin remodeling complex.
Biochem. Biophys. Res. Commun.
299
,
328
334
.
Lee
J. H.
,
Lee
J. Y.
,
Chang
S. H.
,
Kang
M. J.
,
Kwon
H.
(
2005
).
Effects of Ser2 and Tyr6 mutants of BAF53 on cell growth and p53-dependent transcription.
Mol. Cells
19
,
289
293
.
Lee
K.
,
Kang
M. J.
,
Kwon
S. J.
,
Kwon
Y. K.
,
Kim
K. W.
,
Lim
J-H.
,
Kwon
H.
(
2007
).
Expansion of chromosome territories with chromatin decompaction in BAF53-depleted interphase cells.
Mol. Biol. Cell
18
,
4013
4023
.
Liu
R.
,
Liu
H.
,
Chen
X.
,
Kirby
M.
,
Brown
P. O.
,
Zhao
K.
(
2001
).
Regulation of CSF1 promoter by the SWI/SNF-like BAF complex.
Cell
106
,
309
318
.
Muller
J.
,
Oma
Y.
,
Vallar
L.
,
Friederich
E.
,
Poch
O.
,
Winsor
B.
(
2005
).
Sequence and comparative genomic analysis of actin-related proteins.
Mol. Biol. Cell
16
,
5736
5748
.
Mullins
R. D.
,
Heuser
J. A.
,
Pollard
T. D.
(
1998
).
The interaction of Arp2/3 complex with actin: nucleation, high affinity pointed end capping, and formation of branching networks of filaments.
Proc. Natl. Acad. Sci. USA
95
,
6181
6186
.
Nolen
B. J.
,
Littlefield
R. S.
,
Pollard
T. D.
(
2004
).
Crystal structures of actin-related protein 2/3 complex with bound ATP or ADP.
Proc. Natl. Acad. Sci. USA
101
,
15627
15632
.
Olave
I. A.
,
Reck–Peterson
S. L.
,
Crabtree
G. R.
(
2002a
).
Nuclear actin and actin-related proteins in chromatin remodeling.
Annu. Rev. Biochem.
71
,
755
781
.
Olave
I.
,
Wang
W.
,
Xue
Y.
,
Kuo
A.
,
Crabtree
G. R.
(
2002b
).
Identification of a polymorphic, neuron-specific chromatin remodeling complex.
Genes Dev.
16
,
2509
2517
.
Oliver
F. J.
,
de la Rubia
G.
,
Rolli
V.
,
Ruiz–Ruiz
M. C.
,
de Murcia
G.
,
Murcia
J. M.
(
1998
).
Importance of poly(ADP-ribose) polymerase and its cleavage in apoptosis. Lesson from an uncleavable mutant.
J. Biol. Chem.
273
,
33533
33539
.
Park
J.
,
Wood
M. A.
,
Cole
M. D.
(
2002
).
BAF53 forms distinct nuclear complexes and functions as a critical c-Myc-interacting nuclear cofactor for oncogenic transformation.
Mol. Cell. Biol.
22
,
1307
1316
.
Peterson
C. L.
,
Zhao
Y.
,
Chait
B. T.
(
1998
).
Subunits of the yeast SWI/SNF complex are members of the actin-related protein (ARP) family.
J. Biol. Chem.
273
,
23641
23644
.
Phelan
M. L.
,
Sif
S.
,
Narlikar
G. J.
,
Kingston
R. E.
(
1999
).
Reconstitution of a core chromatin remodeling complex from SWI/SNF subunits.
Mol. Cell
3
,
247
253
.
Philimonenko
V. V.
,
Zhao
J.
,
Iben
S.
,
Dingová
H.
,
Kyselá
K.
,
Kahle
M.
,
Zentgraf
H.
,
Hofmann
W. A.
,
de Lanerolle
P.
,
Hozák
P.
et al.  (
2004
).
Nuclear actin and myosin I are required for RNA polymerase I transcription.
Nat. Cell Biol.
6
,
1165
1172
.
Poch
O.
,
Winsor
B.
(
1997
).
Who's who among the Saccharomyces cerevisiae actin-related proteins? A classification and nomenclature proposal for a large family.
Yeast
13
,
1053
1058
.
Rando
O. J.
,
Zhao
K.
,
Janmey
P.
,
Crabtree
G. R.
(
2002
).
Phosphatidylinositol-dependent actin filament binding by the SWI/SNF-like BAF chromatin remodeling complex.
Proc. Natl. Acad. Sci. USA
99
,
2824
2829
.
Robinson
R. C.
,
Turbedsky
K.
,
Kaiser
D. A.
,
Marchand
J. B.
,
Higgs
H. N.
,
Choe
S.
,
Pollard
T. D.
(
2001
).
Crystal structure of Arp2/3 complex.
Science
294
,
1679
1684
.
Shen
X.
,
Mizuguchi
G.
,
Hamiche
A.
,
Wu
C.
(
2000
).
A chromatin remodelling complex involved in transcription and DNA processing.
Nature
406
,
541
544
.
Shen
X.
,
Ranallo
R.
,
Choi
E.
,
Wu
C.
(
2003
).
Involvement of actin-related proteins in ATP-dependent chromatin remodeling.
Mol. Cell
12
,
147
155
.
Sohn
D. H.
,
Lee
K. Y.
,
Lee
C.
,
Oh
J.
,
Chung
H.
,
Jeon
S. H.
,
Seong
R. H.
(
2007
).
SRG3 interacts directly with the major components of the SWI/SNF chromatin remodeling complex and protects them from proteasomal degradation.
J. Biol. Chem.
282
,
10614
10624
.
Sugimoto
N.
,
Kitabayashi
I.
,
Osano
S.
,
Tatsumi
Y.
,
Yugawa
T.
,
Narisawa–Saito
M.
,
Matsukage
A.
,
Kiyono
T.
,
Fujita
M.
(
2008
).
Identification of novel human Cdt1-binding proteins by a proteomics approach: proteolytic regulation by APC/CCdh1.
Mol. Biol. Cell
19
,
1007
1021
.
Szerlong
H.
,
Saha
A.
,
Cairns
B. R.
(
2003
).
The nuclear actin-related proteins Arp7 and Arp9: a dimeric module that cooperates with architectural proteins for chromatin remodeling.
EMBO J.
22
,
3175
3187
.
Szerlong
H.
,
Hinata
K.
,
Viswanathan
R.
,
Erdjument–Bromage
H.
,
Tempst
P.
,
Cairns
B. R.
(
2008
).
The HSA domain binds nuclear actin-related proteins to regulate chromatin-remodeling ATPases.
Nat. Struct. Mol. Biol.
15
,
469
476
.
Tatsumi
Y.
,
Sugimoto
N.
,
Yugawa
T.
,
Narisawa–Saito
M.
,
Kiyono
T.
,
Fujita
M.
(
2006
).
Deregulation of Cdt1 induces chromosomal damage without rereplication and leads to chromosomal instability.
J. Cell Sci.
119
,
3128
3140
.
Welch
M. D.
,
DePace
A. H.
,
Verma
S.
,
Iwamatsu
A.
,
Mitchison
T. J.
(
1997
).
The human Arp2/3 complex is composed of evolutionarily conserved subunits and is localized to cellular regions of dynamic actin filament assembly.
J. Cell Biol.
138
,
375
384
.
Wertman
K. F.
,
Drubin
D. G.
(
1992
).
Actin constitution: guaranteeing the right to assemble.
Science
258
,
759
760
.
Wertman
K. F.
,
Drubin
D. G.
,
Botstein
D.
(
1992
).
Systematic mutational analysis of the yeast ACT1 gene.
Genetics
132
,
337
350
.
Wu
J. I.
,
Lessard
J.
,
Olave
I. A.
,
Qiu
Z.
,
Ghosh
A.
,
Graef
I. A.
,
Crabtree
G. R.
(
2007
).
Regulation of dendritic development by neuron-specific chromatin remodeling complexes.
Neuron
56
,
94
108
.
Zhao
K.
,
Wang
W.
,
Rando
O. J.
,
Xue
Y.
,
Swiderek
K.
,
Kuo
A.
,
Crabtree
G. R.
(
1998
).
Rapid and phosphoinositol-dependent binding of the SWI/SNF-like BAF complex to chromatin after T lymphocyte receptor signaling.
Cell
95
,
625
636
.

Supplementary information