• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Nat Struct Mol Biol. Author manuscript; available in PMC Aug 4, 2011.
Published in final edited form as:
PMCID: PMC3150215

Structural characterization of a capping protein interaction motif defines a family of actin filament regulators


Capping protein (CP) regulates actin dynamics by binding the barbed ends of actin filaments. Removal of CP may be one means to harness actin polymerization for processes such as cell movement and endocytosis. Here we structurally and biochemically investigated a CP interaction (CPI) motif present in the otherwise unrelated proteins CARMIL and CD2AP. The CPI motif wraps around the stalk of the mushroom-shaped CP at a site distant from the actin-binding interface, which lies on the top of the mushroom cap. We propose that the CPI motif may act as an allosteric modulator, restricting CP to a low-affinity, filament-binding conformation. Structure-based sequence alignments extend the CPI motif–containing family to include CIN85, CKIP-1, CapZIP and a relatively uncharacterized protein, WASHCAP (FAM21). Peptides comprising these CPI motifs are able to inhibit CP and to uncap CP-bound actin filaments.

Actin polymerization provides force and organization to drive and shape many cellular processes. Protrusion of membranes in animal cells, for example, is proposed to be driven by the addition of actin subunits to free barbed ends of actin filaments. The barbed end is kinetically and thermodynamically favored over the pointed end for polymerization. Thus, the innate ability of actin polymerization to create force requires that, in general, actin-filament barbed ends should be capped to prevent unwanted polymerization, which would lead to the exhaustion of the actin monomer pool. Under these capped-filament conditions, polymerization can be initiated through three mechanisms: de novo nucleation, filament severing or filament uncapping. Nucleation and severing have been studied in depth, but the molecular details of uncapping mechanisms remain largely unexplored.

The universal actin-capping protein in eukaryotic cells is capping protein (CP), a heterodimer of structurally related α- and β-subunits. CP is found in a range of cellular actin-containing structures and has an affinity of 0.1–1 nM for filaments1,2. CP is present in relatively stable structures such as the sarcomere of striated muscle, where CP defines and anchors the barbed end of thin filaments at the Z-disc, leading to the alias CapZ. Similarly, CP appears to bind to the end of the actin-related protein 1 (Arp1) minifilament in the mammalian dynactin complex (reviewed in ref. 3). CP is also involved in highly dynamic actin structures. Lamellipodia of moving cells contain a mixture of uncapped and CP-capped actin filaments close to the leading edge4. The uncapping of these filaments and their subsequent elongation may provide force during lamellipodia protrusion.

The structural basis of actin-filament capping by CP has been established at low resolution. First, the X-ray structure of CP revealed a mushroom-shaped molecule comprising two symmetrically and structurally related subunits5. The stalk of the mushroom is formed from the α-helical N termini from both subunits. The cap of the mushroom consists of a common intersubunit β-sheet, across which the C-terminal regions form antiparallel helices terminating in unrelated structures termed ‘tentacles’. Both α- and β-tentacles are well ordered in the structure; however, the β-tentacle is bound to a neighboring symmetry-related molecule and is likely to be mobile in solution. Second, the CP crystal structure was oriented into 23-Å resolution cryo–electron microscopy data obtained from capped filaments6. This resulted in a capping model in which the α-tentacle and β-subunit of the mushroom cap provide the majority of the interaction surface with the base of the filament, physically impeding the addition of actin monomers. The β-tentacle is speculated to reorient to provide a second actin contact6.

Protein and lipid inhibitors of CP actin-filament capping are manifest. The protein CP ARp2/3 myosin I linker (CARMIL) is a multifunctional actin regulator consisting of an N-terminal domain that includes leucine-rich repeats and a 51-residue region near the C terminus, which comprises a CP-binding and uncapping motif that includes the CARMIL homology domain 3 (CAH3, residues 965–1038)710. CARMIL has also been proposed to contain a WH2 domain, an acidic domain and a polyproline region. CARMIL-like CP binding motifs exist in a range of proteins that have diverse modular architectures (Fig. 1a). CK2-interacting protein 1 (CKIP-1) consists of an N-terminal pleckstrin homology domain followed by the CP binding motif. CKIP-1 is able to inhibit CP from interacting with an actin filament11,12. A 22-residue peptide corresponding to a fragment of the CARMIL CAH3 (residues 984–1005) competes with CKIP-1 in binding CP and has a direct effect on the activity of CP. CD2-associated protein (CD2AP) and its homolog CIN85 both comprise three SH3 domains, the CP binding motif and a C-terminal coiled-coil domain1315. The CD2AP CP binding motif, consisting of residues 474–513, is sufficient to bind CP with Kd = 6 nM and has some uncapping activity. The CP binding motif is defined as LXHXTXXRPK(6X)P (ref. 13).

Figure 1
Domain organization of CPI motif–containing proteins. (a) Previously known CPI family members. L, leucine-rich repeat; SH3, src homology domain 3; PH, pleckstrin homology domain; CC, coiled coil. The green bars above CD2AP, CIN85 and CKIP-1 signify ...

Actin-filament capping by CP is also inhibited by the proteins CapZ-interacting protein (CapZIP) and V-1 (myotrophin). CapZIP is a 416-residue protein of unknown structure that is subject to stress-induced phosphorylation. The phosphorylation leads to the dissociation of the CapZIP–CP complex, suggesting a possible role in regulating actin remodeling16. V-1, a protein involved in the differentiation of muscle, is an ankyrin repeat protein that binds to CP, inhibiting the interaction with actin. V-1 has been proposed to interact through two loops, which protrude from the ankyrin-repeat back bone, with a CP binding site that includes the β-tentacle17. The CP-interacting motif on CapZIP is currently unknown. The phospholipid phosphatidylinositol 4,5-bisphosphate (PIP2) has also been reported to inhibit CP18,19. The PIP2 binding site overlaps with the actin binding site and includes Lys256 and Arg260 from the α-subunit and Arg225 from the β-subunit. Conflicting data exist as to whether PIP2 can remove CP from an actin filament18,19.

Here we have investigated the poorly understood mechanism of CP-filament uncapping by the CP binding motifs from CARMIL and CD2AP. Sequence-alignment and truncation experiments led to the confirmation of a short CPI motif that retained some uncapping activity. We report the crystal structures of CARMIL and CD2AP CPI motifs (Homo sapiens) bound to chicken CP (Gallus gallus) and a longer fragment of CARMIL that also includes a second unique CP-interacting motif. Structural features and sequence comparisons allow expansion of the CPI motif–containing family from CARMIL, CD2AP, CIN85 and CKIP-1 to also include CapZIP and the newly characterized protein FAM21 (here termed WASHCAP, WASH- and capping protein–associated protein). In vitro assays confirm that peptides comprising each CPI motif possess the predicted CP inhibition and uncapping activities. The structures suggest consideration of an allosteric mechanism of filament uncapping by the CPI motif as a potential alternative to direct competition for actin-binding residues on CP.


Functional definition of the CARMIL CPI motif

To verify a minimal binding and uncapping fragment of CARMIL, we constructed a series of CARMIL truncations and tested them in pyreneactin polymerization assays. In this assay, we inferred an increase in bulk-solution polymeric actin from an increase in pyrene fluorescence20. First, in capping inhibition assays, we simultaneously added CARMIL truncations (Fig. 1) and CP to growing actin filaments that were nucleated from spectrin-actin seeds, which have free barbed ends and capped pointed ends. CP binding region 71 (CBR71) was slightly less effective than CBR115 in preventing CP from capping the barbed end of the filaments (Fig. 2a and Supplementary Fig. 1a). CBR49 showed a markedly reduced, yet measurable, anti-capping activity. We observed a similarly reduced level of activity for CBR37, a peptide derived from sequence-alignment analysis and later shown to comprise the CP binding region determined by crystallography (see below) flanked by two extra residues at each terminus.

Figure 2
Functional definition of the CPI motif. (a) Inhibition of CP capping by histidine (His)-tagged CBR115 and CBR37 as monitored by the pyrene-actin polymerization assay. (b) Uncapping of CP-capped actin filaments. Data for His-CBR71 and His-CBR49 are included ...

Second, in uncapping assays, we added CARMIL truncations to actin filaments that were capped at both ends, at the pointed end by spectrin seeds and at the barbed end by CP, in the presence of excess free actin subunits. The CARMIL truncations all showed concentration-dependent increases in pyrene fluorescence, in line with CP-uncapping activity (Fig. 2b and Supplementary Fig. 1b). The active concentrations for the CARMIL truncations in this uncap-ping assay were similar to those in the CP-binding assay (Fig. 2a and Supplementary Fig. 1a), suggesting that the properties of sequestration and uncapping are manifested in the same interaction between CARMIL and CP. We further substantiated the validity of the assay by showing that CBR37 is able to uncap CP-bound actin filaments in the absence of spectrin seeds. In this experiment, we included profilin in order to prevent the excessive formation of CP-stabilized actin nuclei (Supplementary Fig. 1c). We also removed a CP truncation that lacked the β-tentacle (CPΔβTent) and had reduced capping activity from the filaments by CBR37 (Supplementary Fig. 1d). Hence, the minimal peptide CBR37 shows both CP sequestration and uncapping activities and is referred to henceforth as the CP interaction (CPI) motif. The CAH3 domain from CARMIL includes peptide sequences that flank the CPI motif, leading to stronger uncapping activity10. However, an even larger region leads to the most potent uncapping observed in this study (CBR115). There is no clear homology between CARMIL, CD2AP and CKIP-1 outside the CPI motif; hence, the CPI motif is a more general CP interaction region that exists within CAH3 domain, which includes a more extensive CP binding region.

To assess the affinity of the CPI motif in binding CP, we labeled the CBR37 peptide (which contains a C-terminal cysteine residue) with the fluorescent probe Alexa Fluor 488 C5 maleimide. We titrated Alexa488-CBR37 with CP and observed a concentration-dependent increase in fluorescence anisotropy consistent with slowed molecular tumbling upon association with CP (Fig. 2c). Curve-fitting analysis suggests a Kd of 0.16 μM for this interaction, with 95% confidence interval for the range 85–244 nM. CPΔβTent and CP showed indistinguishable association kinetics with respect to Alexa488-CBR37, as did wild-type CP in the presence of G-actin (1:1 ratio) in low-salt buffer (Supplementary Fig. 2a,b). Hence, the CPI motif interacts with CP at a site that does not include the β-tentacle. This differentiates CP binding of the CPI motif from the CP interaction proposed for V-1 (ref. 17). The Alexa488-CBR37–CP complex could be dissociated by competition with unlabeled CBR37 (Kd = 0.10 μM, 95% confidence interval for 10–186 nM) or by F-actin but not by G-actin as monitored by a decrease in anisotropy (Fig. 2d,e).

We observed the effect of CP and the CPI motif on single actin filaments using total internal reflection fluorescence (TIRF) microscopy. We polymerized BODIPY tetramethylrhodamine C5 maleimide–labeled actin (red actin) in the presence of excess profilin and then capped it with CP. We simultaneously added CBR115, CBR37 or the CD2AP CPI motif with BODIPY FL C1 iodoacetamide–labeled G-actin (green actin) under polymerizing conditions while maintaining constant profilin and CP concentrations. Maintenance of profilin levels prevented the formation of CP-stabilized actin nuclei, and preserving a constant CP concentration was necessary to ensure competent filament capping during solution exchange due to its fast off rate. In these experiments, we observed red filaments acting as nuclei for green filaments in the presence of CBR115, CBR37 or CD2AP CPI motif or in the absence of CP (Fig. 3 and Supplementary Videos 1–4). However, green actin alone could not elongate from the red-actin CP-capped filaments in the absence of CARMIL truncations (Fig. 3e and Supplementary Video 5). Taken together, these data show that the CPI motif from CARMIL binds CP (Kd ~ 0.1–0.2 μM) and partially retains the uncapping and sequestration activities of CBR115 (Kd ~ 1.5 nM)8. Single-molecule TIRF microscopy experiments suggest that CPI motifs will form ternary complexes with CP at the barbed end of filaments10. This behavior is consistent with the efficiency of uncapping by CBR37 (Fig. 2b) relative to the CBR37–CP dissociation constant (Kd ~ 0.1–0.2 μM). Hence, CPI motifs may have a role in protein localization, in targeting either CP or the CPI motif–containing protein to specific cellular regions, as well as modulating the affinity of CP for actin filaments.

Figure 3
CARMIL- and CD2AP-induced uncapping of CP-capped actin filaments monitored by TIRF microscopy. (ac) CBR115 (200 nM) (a), CBR37 (250 μM) (b) and CD2AP CPI peptide (200 μM) (c) support green actin filament growth from the previously ...

Structures of the CARMIL and CD2AP CPI complexes with CP

To understand the nature of the uncapping interaction, we determined the crystal structure of the CPΔβTent–CARMIL CPI motif complex (Fig. 4a–d). The CPI motif wraps around the stalk at the underside of the cap of the mushroom-shaped CP. The CPI motif is in an extended conformation, which half encircles the α-helices of the CP stalk. The N-terminal one-third of the CPI motif contacts the α-subunit of CP and the C-terminal two-thirds contact the β-subunit. The binding position for the CPI motif is distant to the actin binding interface (viewer's perspective in Fig. 4) and to the PIP2 binding interface18. The structure of CPΔβTent in this complex is essentially identical, with the exception of the lack of the β-tentacle, to that when CP was crystallized alone5. A second structure of the CD2AP CPI motif bound to CPΔβTent showed the common interaction of the CPI motifs with CP (Fig. 5a). The C-terminal portion, which includes the LXHXTXXRPK(6X)P motif, shares an interaction mode with CARMIL in binding the CP β-subunit13; however, the N-terminal region takes a divergent path across the CP α-subunit.

Figure 4
The structure of the CPΔβTent–CARMIL CPI motif complex. (ac) Three views of the CPΔβTent–CARMIL CPI complex. CARMIL CPI motif is shown as a surface (yellow) and CP α-subunit (red) and β-subunit ...
Figure 5
Structural differences between CARMIL and CD2AP in binding CP. (a) The structure of the CPΔβTent–CD2AP CPI motif complex with CARMIL CPI motif structure overlaid. The C-terminal half of the CD2AP CPI motif (cyan) generally follows ...

Unique interactions of CARMIL with CP

In order to understand the interactions that distinguish CBR115 from CBR37, the CBR115–CP structure was elucidated (Fig. 5b). This structure reveals two ordered CARMIL regions that lie within the CAH3965–1038 (ref. 10), which roughly correspond to the CBR71964–1034 region studied here that has similar activity to CBR115 (Fig. 2a,b and Supplementary Fig. 1a,b). The first ordered CARMIL region is the CPI motif, Ile971–Cys1004, which is essentially identical to that observed in the CPΔβTent–CARMIL CPI motif complex (Fig. 5a). The second ordered CARMIL region, Arg1021–Thr1035, binds to the under-side of the CP mushroom cap on the opposite side of the CP mushroom stalk to which the CPI motif binds. This region appears not to be present in other CPI-motif proteins from sequence analyses and is termed here the CARMIL-specific interaction (CSI) motif. Together, these two regions resemble a finger (CPI) and thumb (CSI) encircling the stalk on the underside of the mushroom cap. The CSI motif has two prominent phenylalanine residues (Phe1029 and Phe1030) that are buried at the mushroom cap–stalk junction. We observed no interpretable electron density for the β-tentacle.

Sequence contributions to the interactions

The extended nature of the CARMIL CPI and CSI motifs leads to multiple interactions with CP (Figs. 4a and 5 and Supplementary Fig. 3). In these conformations, the CPI and CSI motifs have no domain structure. Hence, sequence conservation across species in the CPI (Fig. 4e) and CSI motifs generally correlate with critical CP-interacting residues. To determine the contribution of charged residues within to the uncapping process, alanine mutations of structurally important interacting residues within the CPI–CP complex were shown to have complementary lowering of activity in the pyrene-actin polymerization assay (Supplementary Fig. 4a–d). The widespread distribution of the interacting residues throughout the CPI motif suggests that these residues play a role in the association between the CPI motif and CP rather than providing individual elements to the uncapping process.

The CPI motif family of proteins

The structural and functional definition of the CPI motif prompted reinvestigation of sequence databases to determine the extent of the CPI motif–containing family. The known family members (CARMIL, CD2AP, CIN85 and CKIP-1) show homology within the CPI motif to CapZIP and a recently characterized protein, FAM21 (NCBI NP_001005751) (Figs. 1b and 6a6a). We observed no obvious homo-logy between the CP binding regions of CARMIL and either V1 or the actin monomer– and CP-binding protein twinfilin. We subjected peptides comprising the CPI motifs from each protein to the pyrene-actin uncapping and inhibition assays. All six peptides showed uncapping and inhibitory activities, functionally verifying the extended CPI-motif family (Fig. 6b and Supplementary Fig. 4e). Sequence alignment reveals that FAM21 and CapZIP have highest residue identity with CD2AP (Fig. 6a), whereas CKIP-1 is more similar to CARMIL in the N-terminal one-third of the CPI motif. Notably, full-length CapZIP showed a similar level of uncapping activity toward CP as measured for the isolated CapZIP CPI motif (Supplementary Fig. 1e and Fig. 6b). This starkly contrasts to the CPI motif from CARMIL, which shows a markedly reduced uncapping activity in comparison to that of CBR115, which has a second CP binding region in the CSI motif (Figs. 2b and 5b5b). Considering this variation in uncapping activities, the balance between CP capping and uncapping as the predominant mode of action may differ among the CPI-motif proteins.

Figure 6
The CPI motif family of proteins. (a) Structure-based sequence alignment of the CPI motif family. α and β refer to the interacting subunit of CP. (b) Pyrene-actin assays monitoring CP uncapping induced by the six members of the CPI motif ...

Direct competition model of CP uncapping

To determine the mechanism by which the CPI motif removes CP from an actin filament, we constructed a series of models. The cryo–electron microscopy structure of the CP-capped actin filament (Fig. 7a) provided the template for superimposing the CPI motif onto the capped filament6 (Fig. 7a,b). There is no direct steric clash between the CPI motif and the terminal actin protomer in the model (Fig. 7a,b). However, at the resolution of the electron microscopy structure (23 Å), errors in the CP placement or nonmodeled conformational changes in either CP or the terminal actin require that this mechanism be considered. The side chains of Lys980, Glu983 and Phe985 from the CARMIL CPI motif approach the side chains of Ile287 (7.5 Å), Arg290 (9 Å) and Met283 (8 Å) from actin, respectively. That these residues are not well conserved either within the CARMIL family (Fig. 4e) or, more generally, in the CPI motif (Fig. 6a) argues that direct competition for close or overlapping binding sites may not be the mechanism of uncapping, particularly given the high affinity of CP for F-actin (Kd = 0.1–1 nM). The longer, more potent CP binding fragment of CARMIL CAH3 (ref. 10) includes the CSI motif that is also distant from the actin-binding surface (Fig. 7b). The disordered residues between the CPI and CSI motifs may be sufficient in length to reach toward the actin binding site on the topside of the CP mushroom cap. However, this region shows lower sequence identity across species relative to the CPI and CSI motifs. In particular, chicken CARMIL contains a two-residue deletion. CBR49, which includes eight residues from the disordered region, shows little change in uncapping activity with respect to CBR37 (Fig. 2 and Supplementary Fig. 1b). Thus, the disordered region may not be involved in direct competition.

Figure 7
Models of CPI motif uncapping. (a) Model of a CP-capped actin filament6. The cyan and orange helices and arrow denote the proposed repositioning of the β-tentacle. The CARMIL CPI (yellow surface) and CSI (green surface) motifs are superimposed ...

β-tentacle sequestration model of CP uncapping

The current model of CP binding to an actin filament involves the movement of the β-tentacle to contact the terminal actin protomer between subdomains 1 and 3 (ref. 6) (cyan versus orange in Fig. 7a). The CARMIL CPI motif binds the α-subunit of CP in close proximity to the β-tentacle (compare Fig. 7a, cyan, to Fig. 7b, yellow). Hence, a second possible mechanism for CPI motif–induced uncapping may be through the creation of a β-tentacle binding site at the interface of the α-subunit of CP and the CARMIL CPI motif to compete with the actin–β-tentacle interaction (Fig. 7c). To test this possibility, we assayed either the β-tentacle alone or the β-tentacle as a glutathione-S-transferase (GST) fusion protein by gel filtration, pulldown and fluorescence anisotropy studies for interaction with CPΔβTent in the presence of CBR37 or CBR115. We observed no interaction between the β-tentacle and the CPΔβTent complexes for either the β1-subunit and β2-subunit isoforms of CP, which differ by four residues in the tentacle region (Supplementary Figs. 5 and 6). Moreover, CBR37 showed uncapping activity for CPΔβTent (Supplementary Fig. 1d), and β-tentacle was not ordered in the CBR115–CP complex structure, effectively ruling out competition for the β-tentacle as a sole mechanism for uncapping.

Allosteric model of CP uncapping

A final possibility for an uncapping mechanism is that the CPI motif locks CP in a less favorable conformation for interaction with a filament. This mechanism requires that CP undergo a conformational change in binding to the actin filament (Fig. 7d). In support of this mechanism, the two molecules in the asymmetric unit of the original CP structure5 shows variation in the curvature of the common β-sheets. When the alternate conformation was included in the capped filament model (Fig. 7e), we observed discernable differences at the actin interface. Hence, we propose that the CPI motif may function as a clasp around the stem of mushroom-shaped CP, restricting the conformation of the filament-binding cap and preventing flexing of the common β-sheet. Thus, upon binding CP at the barbed end of a capped filament, the CPI motif may restrict the CP conformation, stabilizing it in the structure determined here, which is less compatible with filament interaction, leading to the allosteric uncapping of the filament. The uncapping activity of CBR37 is approximately three orders of magnitude lower than that of CBR115 (Fig. 2b). The superior binding of CBR115 (1.5 nM (ref. 8)) to CP through the CSI motif relative to that of CBR37 (0.1–0.2 μM) accounts, at least in part, for the increased uncapping activity. This suggests that the allosteric mechanism should also be considered for CAH3 uncapping activity as a possible alternative to direct competition. High-resolution structures of the actin–CP complex are required to categorically distinguish between the direct competition and allosteric mechanisms of CPI motif–induced uncapping.


The CPI motif family of proteins

Here we have structurally and functionally characterized the CPI motifs from CARMIL and CD2AP, and we have shown that CIN85 and CKIP-1 contain bona fide uncapping motifs in addition to those reported for CARMIL and CD2AP. The CPI motif family has been extended to include the known CP binding partner CapZIP and the largely uncharacterized protein FAM21 (Fig. 1). The architectures of this family of proteins place the CPI motif in a wide range of multiple-domain settings. Members of this family of proteins are known to integrate signaling pathways to control actin dynamics through interactions with CP and other actin regulators. CKIP-1 links the protein kinase CK2, which is involved in regulating cell polarity, morphology and differentiation, to actin remodeling11,12. The homologous proteins CD2AP and CIN85 are known to be associated with endocytic components, CP and cortactin21. CIN85 also binds WIP, WIRE, N-WASP and CARMIL, whereas CD2AP is indirectly linked to N-WASP through PSTPIP1 (refs. 2123). Thus, CD2AP and CIN85 tie endocytosis and cadherin-based junctions to actin remodeling and tissue patterning24. Duboraya, a zebrafish homolog of CapZIP, has been shown to be needed for actin organization and cilia formation in Kupffer's vesicle for proper left-right patterning16,25. CARMIL integrates actin-filament uncapping, monomer binding, Arp2/3–mediated filament nucleation and myosin binding within its own architecture for use at the leading edge and in pinocytosis7. These data suggest that the CPI motif is likely to be a general actin-filament regulator that is incorporated into diverse proteins in order to harness actin polymerization for a wide variety of cellular processes. The precise role of any CPI motif will be dependent on the larger protein architecture. Indeed, flanking sequences around the CARMIL CPI motif that include the CSI motif, which does not appear to be conserved in the other CPI motif members, greatly increased the uncapping potency (Figs. 2a and 5b5b), whereas CapZIP appeared to solely rely on the CPI motif (Fig. 6b and Supplementary Fig. 1e). This situation is analogous to that of the unstructured Tβ4/WH2 motif, which contributes to G-actin sequestering and actin-filament nucleating machineries in different protein settings2628.

FAM21, a protein with four isoforms, has recently been characterized at the cellular level. In two reports, FAM21 (ref. 29) (also referred to as KIAA0592 (ref. 30)) has been found to be a component of the WASH (Wiskott-Aldrich syndrome protein and SCAR homolog) complex that activates Arp2/3 nucleation of actin filaments. FAM21-WASH was shown to locate to endosomes, where it controls fission and retrograde trafficking through Arp2/3–controlled actin polymerization and microtubule interactions29,30. The larger WASH complex contains WASH, FAM21, CP, tubulins, coiled-coil domain containing 53 proteins, strumpellin (KIAA0196 (ref. 31)), uncharacterized protein KIAA1033 and hsp70s. Here we have shown a direct interaction between CP and FAM21, which is likely to be reproduced in the WASH complex. The WASH complex showed reduced filament capping with respect to CP, which is in line with the properties of the FAM21 CPI motif. Hence, we propose that FAM21 be renamed WASHCAP (WASH- and capping protein–associated protein) to reflect its known interacting partners (Fig. 1b). Components of the WASH complex have also been implicated in pathogen entry into cells. WASHCAP was identified as the virus penetrating factor that is critical for fluid-phase vaccinia virus cellular uptake through endocytosis32. Similarly, WASH has been implicated in Salmonella typhimurium invasion33.

WASHCAP has been observed to be highly phosphorylated by MS-based proteomics and to be present both in the cytoplasm and in the nucleus (http://www.phosida.com/). The acidic nature (pI 4.8) of this poly-LF(D/E)nLF–containing protein (Fig. 1b) is enhanced through phosphorylation. Several of the phosphorylation sites show consensus sequences for CK1 or CK2, and approximately half are sensitive to epidermal growth factor stimulation. The poly-LF(D/E)4–6LF motifs are followed by basic regions. Reoccurring sequences are suggestive of repetitive structure and/or multiple binding motifs. The acidic stretches in WASHCAP are reminiscent of Arp2/3–binding ‘A’ motifs and potentially may help locate Arp2/3 to the WASH complex. Similarly, WASHCAP basic regions may be useful in targeting the WASH complex to membranes. Yeast two-hybrid screens have identified two transcriptional repressors, ataxin 1 and TRIM27, as potential binding partners (http://www.h-invitational.jp/). Hence, WASHCAP may be a phosphorylation- dependent actin regulator with a transcriptional feedback role.

Implications for the biological role of the CPI motif

The data presented here indicate that the CPI motif may be important in recruiting CP or CPI motif–containing proteins and in changing the dynamics of capped filaments. Superimposition of the X-ray structures onto the electron microscopy model of a CP-capped actin filament6 suggests that the CPI and CSI motifs may regulate the affinity of CP for actin filaments through an allosteric mechanism (Fig. 7). In cells, CP dissociates from the actin-filament network relatively quickly34 compared to the dissociation rate seen with pure proteins in vitro35, which suggests that uncapping may occur in cells. Four of the six CPI motif–containing proteins (CARMIL, CD2AP, CIN85 and WASHCAP) are used in endocytotic processes and link CP to Arp2/3 through diverse architectures or complexes. The functional linkage of Arp2/3 to CP has been shown to increase actin-based force generation through enhancing filament nucleation36. Physical association of CP and Arp2/3 raises a possible scenario whereby Arp2/3 complexes may be targeted, through the CPI motif, to the ends of CP-capped actin filaments. Such interactions have the potential to uncap the barbed end of a mother filament and simultaneously nucleate a daughter filament.


Methods and any associated references are available in the online version of the paper at http://www.nature.com/nsmb/.

Supplementary Material




Supplementary Table 1. Methods for the TIRF assays and for the supplementary figures are given in Supplementary Methods.


We thank the Biomedical Research Council of A*STAR for support to R.C.R. and the National Synchrotron Radiation Research Center, a facility supported by the National Science Council of Taiwan for provision of beam time and assistance in data collection. The Synchrotron Radiation Protein Crystallography Facility is supported by the National Research Program for Genomic Medicine. J.A.C. acknowledges US National Institutes of Health grant GM 38542 for support.


Plasmids, proteins and peptides

We expressed chicken full-length CP (α1/β1) and CPΔβTent (α1/β1 lacking the final 34 residues) in BL21(DE3) E. coli using pET-3d-βI′/βII, kindly provided by T. Obinata (Chiba University) and purified them as described previously37 with the following modifications. We replaced the cation-ion exchange chromatography step with size-exclusion chromatography using a Superdex 200 HiLoad 16/60 column (GE Healthcare). We supplemented all the buffers except the gel-filtration buffer with 0.2 mM PMSF, 0.5 mM benzamidine and 20 units per ml aprotinin. We treated protein samples with 4 mM Pefabloc SC accompanied by the protector solution (Roche) before injection onto the gel-filtration column.

We amplified the CBR115 fragment of human CARMIL1a (GenBank FJ009082) residues Glu964–Ser1078 from cDNA by PCR and cloned it into the pGEX-6P-3 vector (GE Healthcare). We expressed the GST fusion protein in BL21(DE3) E. coli and purified it with glutathione fast-flow Sepharose resin (GE Healthcare). We grew and induced cultures with IPTG at 23 °C. After elution from the glutathione resin, we mixed GST-CBR115 with PreScission protease (GE Healthcare) and dialyzed the mixture into S-Sepharose buffer (10 mM Tris-HCl, pH 8.0, 10 mM KCl, 0.1 mM EDTA, 0.5 mM DTT, 1 mM NaN3) overnight (16 h). To purify CBR115 from GST, we applied the mixture to an S-Sepharose column and eluted it with a 10- to 700-mM KCl gradient. For storage, we dialyzed CBR115 into 10 mM Tris-HCl, pH 8.0, 40 mM KCl, 0.1 mM EDTA, 0.5 mM DTT, 1 mM NaN3 and kept it on ice. We made His-tagged CBR115 residues Glu964–Ser1078 by cloning the region of interest into the pRSFDuet-1 vector (Novagen) and purified it using the Talon metal affinity resin (Clontech). We determined concentrations of CBR115 fragments based on the far UV absorbance at A280 or A215 and A225.

We purified actin from rabbit skeletal muscle38. We cloned human profilin 1 into the pSY5 vector, a modified version of pET-21d(+), and expressed it with an 8×His tag39. We purified the protein with a HisTrap column followed by gel filtration with a Superdex 75 HiLoad 16/60 column (GE Healthcare). We purchased CBR52, the reported fragment of the CARMIL domain that inhibits CP and uncaps actin filaments (Leu961–Asp1012)9, CBR37 (Ser968–Cys1004), defined from the structural analysis of the CPΔβTent–CPI motif complex, CPI peptides (Fig. 6a) from CD2AP, CIN85 and CKIP-1 (Auspep Pty. Ltd) and WASHCAP and CapZIP CPI peptides (Neo peptide) at >95% purity.

Actin polymerization assays

We prepared spectrin-actin seeds (spectrin-band 4.1-actin complex) as described40 with minor modifications. We collected 100 ml of fresh whole blood in acid citrate/dextrose. We extracted the red blood cell ghost pellet with 3–5 ml (per 10 ml of original cells) of extraction buffer (0.3 mM Na2HPO4) at 37 °C for 10 min. We centrifuged the extract for 1 h at 125,000g, made the supernatant in 2 mM DTT and added an equal volume of ethylene glycol for storage at −20 °C.

We performed pyrene-actin polymerization assays as described2,8 with minor modifications. For a total volume of 200 μl, we Mg2+-exchanged actin subunits (5% pyrene labeled) at 1.5 μM in G-buffer (2 mM Tris-HCl, pH 7.2, 0.2 mM ATP, 0.5 mM DTT, 1 mM NaN3 and 0.2 mM CaCl2). Then, we added 10 μl of 20× KME buffer (200 mM Tris-HCl, pH 7.5, 1 M KCl, 20 mM MgCl2, 20 mM EGTA) and 20 μl of spectrin-actin seeds. For capping assays, we added 10 nM CP to the mixture at time zero. We varied the concentration of CBR fragments and CPI peptides, which were also added at time zero. For the uncapping assays followed by pyrene-actin polymerization, we polymerized actin from seeds in the presence of 10 nM CP for 200 s, added the CBR fragments of CARMIL and CPI peptides and then polymerized them for 300 s.

Peptide labeling and fluorescence anisotropy measurements

We labeled the C-terminal free sulfhydryl group of CBR37 with the thiol-specific fluorescence probe Alexa Fluor 488 C5 maleimide (Molecular Probes). We removed the excess of the probe with a PD-10 desalting column (GE Healthcare). We estimated the efficiency of labeling to be 50% based on an SDS-PAGE gel. We determined the concentration of the conjugated protein from absorption spectra assuming that the extinction coefficient of the probe at the lowest energy maximum (72,000 M–1 cm–1 at wavelength 493 nm) is not affected by conjugation41. We dispensed and equilibrated (30 min) the reaction mixture, which consisted of 250 nM fluorescent conjugate and increasing concentrations of protein or protein complex, in a 96-well, black flat-bottomed plate (Nunc). We measured fluorescence anisotropy at wavelength 518 nm after correcting for instrument polarization bias (G-factor) using a Safire2 fluorimeter (Tecan). We fit binding curves of anisotropy versus protein or protein complex concentrations to a single-site binding model42, and we fit competition curves as described43 using GraphPad Prism4 (http://www.graphpad.com/).

Crystallography and TIRF assays

Details of X-ray crystallographic experiments combined with data collection, merging and refinement statistics are detailed in


Accession codes. Protein Data Bank: The coordinates and merged reflection data for CARMIL and CD2AP CPIs bound to CPΔβTent and the CBR115–CP complex have been deposited with accession codes 3LK2, 3LK4 and 3LK3, respectively.

Note: Supplementary information is available on the Nature Structural & Molecular Biology website.


The authors declare no competing financial interests.


1. Kovar DR, Wu JQ, Pollard TD. Profilin-mediated competition between capping protein and formin Cdc12p during cytokinesis in fission yeast. Mol. Biol. Cell. 2005;16:2313–2324. [PMC free article] [PubMed]
2. Wear MA, Yamashita A, Kim K, Maeda Y, Cooper JA. How capping protein binds the barbed end of the actin filament. Curr. Biol. 2003;13:1531–1537. [PubMed]
3. Cooper JA, Sept D. New insights into mechanism and regulation of actin capping protein. Int. Rev. Cell Mol. Biol. 2008;267:183–206. [PMC free article] [PubMed]
4. Mejillano MR, et al. Lamellipodial versus filopodial mode of the actin nanomachinery: pivotal role of the filament barbed end. Cell. 2004;118:363–373. [PubMed]
5. Yamashita A, Maeda K, Maeda Y. Crystal structure of CapZ: structural basis for actin filament barbed end capping. EMBO J. 2003;22:1529–1538. [PMC free article] [PubMed]
6. Narita A, Takeda S, Yamashita A, Maeda Y. Structural basis of actin filament capping at the barbed-end: a cryo-electron microscopy study. EMBO J. 2006;25:5626–5633. [PMC free article] [PubMed]
7. Jung G, Remmert K, Wu X, Volosky JM, Hammer JA., III The Dictyostelium CARMIL protein links capping protein and the Arp2/3 complex to type I myosins through their SH3 domains. J. Cell Biol. 2001;153:1479–1497. [PMC free article] [PubMed]
8. Yang C, et al. Mammalian CARMIL inhibits actin filament capping by capping protein. Dev. Cell. 2005;9:209–221. [PMC free article] [PubMed]
9. Uruno T, Remmert K, Hammer JA., III CARMIL is a potent capping protein antagonist: identification of a conserved CARMIL domain that inhibits the activity of capping protein and uncaps capped actin filaments. J. Biol. Chem. 2006;281:10635–10650. [PubMed]
10. Fujiwara I, Remmert K, Hammer JA., III Direct observation of the uncapping of capping protein-capped actin filaments by CARMIL homology domain 3 (CAH3). J. Biol. Chem. 2010;285:2707–2720. [PMC free article] [PubMed]
11. Canton DA, et al. The pleckstrin homology domain-containing protein CKIP-1 is involved in regulation of cell morphology and the actin cytoskeleton and interaction with actin capping protein. Mol. Cell. Biol. 2005;25:3519–3534. [PMC free article] [PubMed]
12. Canton DA, Olsten ME, Niederstrasser H, Cooper JA, Litchfield DW. The role of CKIP-1 in cell morphology depends on its interaction with actin-capping protein. J. Biol. Chem. 2006;281:36347–36359. [PMC free article] [PubMed]
13. Bruck S, et al. Identification of a novel inhibitory actin-capping protein binding motif in CD2-associated protein. J. Biol. Chem. 2006;281:19196–19203. [PMC free article] [PubMed]
14. Gaidos G, Soni S, Oswald DJ, Toselli PA, Kirsch KH. Structure and function analysis of the CMS/CIN85 protein family identifies actin-bundling properties and heterotypic-complex formation. J. Cell Sci. 2007;120:2366–2377. [PubMed]
15. Hutchings NJ, Clarkson N, Chalkley R, Barclay AN, Brown MH. Linking the T cell surface protein CD2 to the actin-capping protein CAPZ via CMS and CIN85. J. Biol. Chem. 2003;278:22396–22403. [PubMed]
16. Eyers CE, et al. The phosphorylation of CapZ-interacting protein (CapZIP) by stress-activated protein kinases triggers its dissociation from CapZ. Biochem. J. 2005;389:127–135. [PMC free article] [PubMed]
17. Bhattacharya N, Ghosh S, Sept D, Cooper JA. Binding of myotrophin/V-1 to actin-capping protein: implications for how capping protein binds to the filament barbed end. J. Biol. Chem. 2006;281:31021–31030. [PMC free article] [PubMed]
18. Kim K, et al. Structure/function analysis of the interaction of phosphatidylinositol 4,5-bisphosphate with actin-capping protein: implications for how capping protein binds the actin filament. J. Biol. Chem. 2007;282:5871–5879. [PMC free article] [PubMed]
19. Kuhn JR, Pollard TP. Single molecule kinetic analysis of actin filament capping. Polyphosphoinositides do not dissociate capping proteins. J. Biol. Chem. 2007;282:28014–28024. [PubMed]
20. Cooper JA, Walker SB, Pollard TP. Pyrene actin: documentation of the validity of a sensitive assay for actin polymerization. J. Muscle Res. Cell Motil. 1983;4:253–262. [PubMed]
21. Dikic I. CIN85/CMS family of adaptor molecules. FEBS Lett. 2002;529:110–115. [PubMed]
22. Havrylov S, Rzhepetskyy Y, Malinowska A, Drobot L, Redowicz MJ. Proteins recruited by SH3 domains of Ruk/CIN85 adaptor identified by LC-MS/MS. Proteome Sci. 2009;7:21–39. [PMC free article] [PubMed]
23. Badour K, et al. The Wiskott-Aldrich syndrome protein acts downstream of CD2 and the CD2AP and PSTPIP1 adaptors to promote formation of the immunological synapse. Immunity. 2003;18:141–154. [PubMed]
24. Johnson RI, Seppa MJ, Cagan RL. The Drosophila CD2AP/CIN85 orthologue Cindr regulates junctions and cytoskeleton dynamics during tissue patterning. J. Cell Biol. 2008;180:1191–1204. [PMC free article] [PubMed]
25. Oishi I, Kawakami Y, Raya A, Callol-Massot C, Izpisá Belmonte JC. Regulation of primary cilia formation and left-right patterning in zebrafish by a noncanonical Wnt signaling mediator, duboraya. Nat. Genet. 2006;38:1316–1322. [PubMed]
26. Hertzog M, et al. The β-thymosin/WH2 domain; structural basis for the switch from inhibition to promotion of actin assembly. Cell. 2004;117:611–623. [PubMed]
27. Irobi E, et al. Structural basis of actin sequestration by thymosin-β4: implications for WH2 proteins. EMBO J. 2004;23:3599–3608. [PMC free article] [PubMed]
28. Dominguez R. The β-thymosin/WH2 fold: multifunctionality and structure. Ann. NY Acad. Sci. 2007;1112:86–94. [PubMed]
29. Gomez TS, Billadeau DDA. FAM21-containing WASH complex regulates retromer-dependent sorting. Dev. Cell. 2009;17:699–711. [PMC free article] [PubMed]
30. Derivery E, et al. The Arp2/3 activator WASH controls the fission of endosomes through a large multiprotein complex. Dev. Cell. 2009;17:712–723. [PubMed]
31. Valdmanis PN, et al. Mutations in the KIAA0196 gene at the SPG8 locus cause hereditary spastic paraplegia. Am. J. Hum. Genet. 2007;80:152–161. [PMC free article] [PubMed]
32. Huang CY, et al. A novel cellular protein, VPEF, facilitates Vaccinia virus penetration into HeLa cells through fluid phase endocytosis. J. Virol. 2008;82:7988–7999. [PMC free article] [PubMed]
33. Hänisch J, et al. Molecular dissection of Salmonella-induced membrane ruffling versus invasion. Cell. Microbiol. 2010;12:84–98. [PubMed]
34. Miyoshi T, et al. Actin turnover-dependent fast dissociation of capping protein in the dendritic nucleation actin network: evidence of frequent filament severing. J. Cell Biol. 2006;175:947–955. [PMC free article] [PubMed]
35. Schafer DA, Jennings PB, Cooper JA. Dynamics of capping protein and actin assembly in vitro: uncapping barbed ends by polyphosphoinositides. J. Cell Biol. 1996;135:169–179. [PMC free article] [PubMed]
36. Akin O, Mullins RD. Capping protein increases the rate of actin-based motility by promoting filament nucleation by the Arp2/3 complex. Cell. 2008;133:841–851. [PMC free article] [PubMed]
37. Soeno Y, et al. Generation of functional β-actinin (CapZ) in an E. coli expression system. J. Muscle Res. Cell Motil. 1998;19:639–646. [PubMed]
38. Burtnick LD, et al. Structure of the N-terminal half of gelsolin bound to actin: roles in severing, apoptosis and FAF. EMBO J. 2004;23:2713–2722. [PMC free article] [PubMed]
39. Wang H, et al. Helix straightening as an activation mechanism in the gelsolin superfamily of actin regulatory proteins. J. Biol. Chem. 2009;284:21265–21269. [PMC free article] [PubMed]
40. Casella JF, Maack DJ, Lin S. Purification and initial characterization of a protein from skeletal muscle that caps the barbed ends of actin filaments. J. Biol. Chem. 1986;261:10915–10921. [PubMed]
41. Rusinova E, et al. Alexa and Oregon green dyes as fluorescence anisotropy probes for measuring protein-protein and protein-nucleic acid interactions. Anal. Biochem. 2002;308:18–25. [PubMed]
42. Owen BA, Lang WH, McMurray CT. The nucleotide binding dynamics of human MSH2–MSH3 are lesion dependent. Nat. Struct. Mol. Biol. 2009;16:550–557. [PMC free article] [PubMed]
43. Vinson VK, De La Cruz EM, Higgs HN, Pollard TP. Interactions of Acanthamoeba profilin with actin and nucleotides bound to actin. Biochemistry. 1998;37:10871–10880. [PubMed]
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...