Visualizing Arp2/3 complex activation mediated by binding of ATP and WASp using structural mass spectrometry

Abstract

Actin-related protein (Arp) 2/3 complex nucleates new branches in actin filaments playing a key role in controlling eukaryotic cell motility. This process is tightly regulated by activating factors: ATP and WASp-family proteins. However, the mechanism of activation remains largely hypothetical. We used radiolytic protein footprinting with mass spectrometry in solution to probe the effects of nucleotide- and WASp-binding on Arp2/3. These results represent two significant advances in such footprinting approaches. First, Arp2/3 is the most complex macromolecular assembly yet examined; second, only a few picomoles of Arp2/3 was required for individual experiments. In terms of structural biology of Arp 2/3, we find that ATP binding induces conformational changes within Arp2/3 complex in Arp3 (localized in peptide segments 5–18, 212–225, and 318–327) and Arp2 (within peptide segment 300–316). These data are consistent with nucleotide docking within the nucleotide clefts of the actin-related proteins promoting closure of the cleft of the Arp3 subunit. However, ATP binding does not induce conformational changes in the other Arp subunits. Arp2/3 complex binds to WASp within the C subdomain at residue Met 474 and within the A subdomain to Trp 500. Our data suggest a bivalent attachment of WASp to Arp3 (within peptides 162–191 and 318–329) and Arp2 (within peptides 66–80 and 87–97). WASp-dependent protections from oxidation within peptides 54–65 and 80–91 of Arp3 and in peptides 300–316 of Arp2 suggest domain rearrangements of Arp2 and Arp3 resulting in a closed conformational state consistent with an “actin-dimer” model for the active state.

Arp2/3 complex is an ensemble of seven conserved protein subunits, including two actin-related proteins Arp2, and Arp3, and five subunits named ARPC1 p40, ARPC2 p34, ARPC3 p21, ARPC4 p20, and ARPC5 p16 (1–5). The Arp2/3 complex is thought to initiate actin polymerization as a branch on the side of an existing actin filament by forming a dimer of Arp2 and Arp3 arranged like subunits in an actin filament at the slow-growing pointed end of the new actin filament (1, 6, 7). The apo-complex is inactive (7) and requires ATP as well as nucleation-promoting cofactors such as SCAR/WASp family proteins and preexisting actin filaments to initiate growth of new filaments (8–10). This tight regulation of Arp2/3 complex activity is crucial to the temporal and spatial regulation of filament assembly at the leading edge of motile cells.

The molecular mechanism leading to activation of the Arp2/3 complex lacks comprehensive experimental support. The high-resolution crystal structure of the apo-form of the complex shows a large separation between the Arp2 and Arp3 subunits, so formation of a dimer of Arps similar to two consecutive subunits in an actin filament requires substantial conformational rearrangements (7). It has been postulated that nucleation-promoting factors may induce major conformational rearrangements in the Arp2/3 complex to bring the two Arps closer together (7). However, the current model of Arp2/3 activation remains primary hypothetical due to the lack of a high-resolution structure of an activated conformation as well as the lack of high-resolution, solution based structural approaches to address these questions.

ATP binding to both Arp2 and Arp3 is thought to drive at least part of this conformational change (7, 11). A crystal structure of the apo-form of the Arp2/3 complex showed the nucleotide pockets of Arp subunits 2 and 3 to be empty and open (7). In crystals of Arp2/3 complex soaked in ATP or ADP, nucleotides bound the clefts of both Arps (12). Nucleotide binding was accompanied by local rearrangements around the nucleotide clefts, but lattice contacts may have limited more extensive rearrangements of the molecule (11, 12).

The WASp family of proteins, which interact with preexisting filaments to stimulate full nucleating activity of the Arp2/3 complex, contain a conserved C-terminal segment (70 aa) known as WASp VCA(13). It consists of three distinct functional regions: an actin-binding verprolin-homology domain (V), a central region (C) that interacts with both the Arp2/3 complex and monomeric actin, and the acidic segment (A) that is thought to interact directly with the complex. It has been proposed that binding of the CA segment of WASp proteins to the complex may favor or stabilize a conformation where Arp subunits are in contact with each other and with an actin monomer associated with the V region (7). This actin monomer becomes the first subunit in the new actin filament branch (7). EM and single-particle averaging of yeast and bovine Arp2/3 complex using negative stain (≈25-Å image resolution) showed WASp-mediated closure of the gap between Arp2 and Arp3 apparently priming the complex for actin nucleation (14).

We used oxidative protein footprinting and mass spectrometry to probe the effect of ATP and WASp on the conformation of the Arp2/3 complex in solution with resolution at the level of single side chains (15–17). In protein complexes, the reactive residues buried between interfaces are protected from oxidation, making it possible to detect changes in surface accessibility. However, allosteric changes in protein conformation induced by ligand binding can also give rise to either protections or enhancements of oxidation. Thus, confirmation of the proposed interface should include additional biochemical or structural data. This approach has been successfully applied to large proteins and complexes such as gelsolin (85 kDa) (18), the transferrin–transferrin receptor complex (a complex with total molecular mass of 330 kDa) (19), and F-actin (a megadalton filament with a monomer size of 40 kDa) (20); these prior experiments were carried out at 10–40 μM concentrations of protein with hundreds of picomoles of material. Arp2/3 has the greatest complexity of any sample examined to date using this method (220 kDa of unique sequence) and typical experiments were carried out with a few picomoles of material, this was necessitated in part by the need fully occupy the WASp binding sites on Arp2/3 by using a 1 μM concentration of Arp2/3.

We analyzed the conformation of the apo-form of the Arp2/3 complex and the conformational changes upon binding to ATP in the presence and absence of WASp. Data support the idea that WASp binding crosslinks Arp3 (within peptide segments 162–191 and 318–329) and Arp2 (within peptide segments 66–80 and 87–97). We also identified the sites within WASp that bind Arp2/3 complex. Overall, these data provide firm experimental support to previously hypothetical mechanisms of Arp2/3 activation and illustrate the emerging power of structural mass spectrometry approaches to provide solution based structural information on large complexes.

Results and Discussion

Consistency of Footprinting and Crystallographic Data for the Arp2/3 Complex.

Arp2/3 complex at a concentration of 1 μM prepared in either the absence or presence of 1 mM ATP was exposed to the X-28C x-ray white beam for intervals from 0 to 80 ms; on these time scales, oxidative modifications dominate the chemistry compared with cross-linking events and cleavage (21). Moreover, because the solvent-accessible surfaces of the most reactive side chains are primarily involved in the oxidation, the effects on the global protein structure are minimal (22). After digestion of the protein complex with trypsin, >200 peptides, covering ≈80% of the sequence in Arp2/3, were analyzed quantitatively by mass spectrometry coupled to high performance liquid chromatography. The identities of these peptides were verified by MS/MS analysis. Many short tryptic peptides (covering ≈20% of the Arp2/3 sequence) were not detected due to their poor retention on the C18 reverse phase column used. More than 160 peptides were clearly identified, but their oxidation was not detected, and they are considered to be not modified in these experiments. The extent of the side chain modification for the 30 peptides that were found to be reproducibly oxidized for both the nucleotide-free and nucleotide-bound Arp2/3 complex was derived from total ion currents for each modified peptide at each exposure time. Table 1 lists the first-order rate constants derived from the plots (dose–response curves). The measured fractions of unmodified peptide varied <5% in their absolute values in triplicate experiments.

Table 1.

Rate constants for the oxidation of Arp2/3 peptides at 0 mM ATP, 1 mM ATP, and 10 μM WASp in the presence of 1 mM ATP

	Peptide	Sequence	SA, Å2	Oxidized residue(s)	Rate of the oxidation, s−1
					0 mM ATP	1 mM ATP	1 mM ATP/Arp/WASp
Arp3	5–18	LPACVVDCDTGYTK	Y-103	Y16	0.6 ± 0.1	0.1 ± 0.01	0.1 ± 0.01
	80–91	HGIVEDWDLMER	H-154, W-34, M-11	M89	1.0 ± 0.1	0.8 ± 0.1	0.4 ± 0.04
	92–99	FMEQVIFK	M-6	M93	1.0 ± 0.1	1.0 ± 0.1	1.1 ± 0.1
	103–123	AEPEDHYFLLTEPPLNTPENR	Y-46	Y109	0.4 ± 0.1	0.5 ± 0.1	0.5 ± 0.1
	162–191	TLTGTVIDSGDGVTHVIPVAEGYVIGSCIK	H-57, Y-61	Y184	5.9 ± 0.8	5.0 ± 0.5	2.1 ± 0.2
	199–209	DITYFIQQLLR	Y-100	Y202	1.1 ± 0.1	1.3 ± 0.1	1.3 ± 0.1
	212–225	EVGIPPEQSLETAK	V-134, P-107	V213, P217	0.5 ± 0.03	0.2 ± 0.03	0.2 ± 0.03
	318–329	NIVLSGGSTMFR	M-58, F-24	M327	3.6 ± 0.3	1.8 ± 0.2	1.2 ± 0.2
	410–418	HNPVFGVMS	M-174	M417	2.2 ± 0.3	1.6 ± 0.1	1.7 ± 0.2
Arp2	54–65	DLMVGDEASELR	M–	M56	2.1 ± 0.2	1.9 ± 0.3	1.4 ± 0.2
	66–80	SMLEVNYPMENGIVR	M–, Y–, M–	M67, Y72, M74	6.1 ± 0.7	5.7 ± 0.5	2.2 ± 0.3
	87–97	HLWDYTFGPEK	W–, Y–, F–	W89	1.4 ± 0.2	1.3 ± 0.2	0.3 ± 0.1
	107–118	ILLTEPPMNPTK	M–	M114	2.9 ± 0.4	4.8 ± 0.6	4.0 ± 0.4
	233–250	LALETTVLVESYTLPDGR	L-101, Y-23	Y244	2.1 ± 0.2	1.9 ± 0.2	1.9 ± 0.2
	300–316	HIVLSGGSTMYPGLPSR	M-130	M309	5.5 ± 0.4	3.6 ± 0.4	1.4 ± 0.2
	369–375	DNFWMTR	M–	M373	2.9 ± 0.4	2.8 ± 0.3	3.2 ± 0.4
P40	75–82	NAYVWTLK	Y-20	Y77	0.7 ± 0.1	0.7 ± 0.1	0.6 ± 0.2
	119–135	VISICYFEQENDWWVCK	F-57, W-120	W131	0.8 ± 0.1	0.8 ± 0.1	0.8 ± 0.1
	189–216	MPFGELMFESSSSCGWVHGVCFSANGSR	M-6, M-9, F-18	M195	2.8 ± 0.2	2.6 ± 0.2	2.7 ± 0.3
	342–360	CSQFCTTGMDGGMSIWDVR	M-17	M350	1.6 ± 0.2	1.6 ± 0.2	1.9 ± 0.3
P34	1–9	MILLEVNNR	M-33	M1	1.1 ± 0.1	1.0 ± 0.1	1.3 ± 0.1
	79–106	VYGSYLVNPESGYNVSLLYDLENLPASK	Y-54, Y-26	Y83	1.3 ± 0.2	1.2 ± 0.2	1.0 ± 0.1
	107–117	DSIVHQAGMLK	M-26	M115	1.2 ± 0.1	1.4 ± 0.1	1.6 ± 0.3
	148–157	DDETMYVESK	M-2	M152	0.8 ± 0.1	0.8 ± 0.1	0.7 ± 0.1
P21	2–14	PAYHSSLMDPDTK	Y-40, M-31, P-118	M9	0.7 ± 0.2	0.5 ± 0.1	0.5 ± 0.1
	38–50	DTDIVDEAIYYFK	Y-94, Y-33	Y47	0.3 ± 0.02	0.4 ± 0.04	0.3 ± 0.03
	94–119	EMYTLGITNFPIPGEPGFPLNAIYAK	M-2, Y-80, P-122	M95, Y96	4.6 ± 0.4	5.1 ± 0.6	5.2 ± 0.4
P20	107–128	KPVEGYDISFLITNFHTEQMYK	Y-22, F-78, Y-36, M-2	M126, Y127	3.1 ± 0.3	3.4 ± 0.3	3.0 ± 0.4
	131–144	LVDFVIHFMEEIDK	M-10	M139	0.5 ± 0.1	0.6 ± 0.1	0.5 ± 0.04
P16	48–60	QGNMTAALQAALK	M-33	M51	5.1 ± 0.5	5.5 ± 0.4	5.2 ± 0.4

Boldface indicates potentially modifiable residues in the inactive Arp2/3 structure.

The rate of oxidation of a reactive peptide depends on the reactivity of the side-chain groups to hydroxyl radical attack and their solvent accessibility. The most reactive amino acids are usually sulfur-containing and aromatic residues. We calculated surface accessibility to identify which primary target residues were likely to be modified in the apo-Arp2/3 complex (Table 1). Several segments of Arp2 are disordered in the crystal structures, so we could not calculate their surface accessibility. Of the 50 Met residues in the complex, 52% are at least partially solvent accessible (≈5 Ų or greater) and thus candidates for modification based on our previous studies (22). Of these 50 methionine residues, 20 were detected to be oxidized in these experiments (≈40%), whereas the other methionine residues were not modified due to minimal surface accessibility; ≈10% of the methionine residues were on peptides that were never identified (Table 1). The remaining residues observed to be oxidized were primarily tyrosine and tryptophan, with a few observed oxidations of proline. In the case of the aromatics, only residues that exhibited solvent accessibilities in excess of 20 Ų were expected to suffer oxidation (22) and overall, this was the case.

Fig. 1B shows the 30 peptides observed to be oxidized along with the reactive residue(s) that were detected within the apo-Arp2/3 complex. The observed modified sites are distributed throughout the Arp2/3 complex providing multiple probes for each subunit.

Fig. 1.

Schematic representation of the structure of the Arp2/3 complex apo-form with the modeled subdomains 1 and 2 of Arp2. (A) Ribbon diagram of Arp2/3 apo-form with the color codes for subunits: codes for subunits: Arp3, orange; Arp2, red for subdomains 3 and 4, pink for the actin backbone model of subdomains 1 and 2; ARPC1/p40, green; ARPC2/p34, light blue; ARPC3/p21, magenta; ARPC4/p20, dark blue; and ARPC5/p16, yellow. (B) Peptides that were reactive in the absence of ATP are color coded as per A; oxidized side chains are also indicated. (C) Reactive peptides whose oxidation rate decreased in the presence of 1 mM ATP including 5–18, 212–225, and 318–329 (Arp3), and 300–316 (Arp2) are indicated; peptide 107–118 (Arp2), whose oxidation rate increased, is also shown. (D) Peptides identified within Arp2/3 complex, whose oxidation rate decreased on binding of WASp and 1 mM ATP including 80–91, 162–191, and 318–329 (Arp3), and 54–65, 66–80, 87–97, and 300–316 (Arp2).

ATP Binding to Arp2/3 Complex Shows Occupancy at the Nucleotide-Binding Clefts of Arp2 and Arp3.

In the presence of 1 mM ATP, the oxidation rates for 25 of the 30 reactive peptides were identical (within error) to the rates observed in the absence of ATP (Table 1). However, the binding of ATP modifies the oxidation rates of five peptides (Figs. 1C and 2A). ATP protects four of the five peptides from oxidation (i.e., modification rate decreases) and increases the rate of oxidation of segment 107–118 of Arp2 2-fold. Substantial decreases in the oxidation rate ranging from 2- to 6-fold were observed for peptide segments 5–18, 212–225, and 318–329 in Arp3. Peptide 300–316 in Arp2 showed modest protection of 1.5-fold. The protected peptides are clustered within Arp3 and Arp2 without any substantial changes in p40, p34, p21, p20, or p16.

Fig. 2.

Histogram plots of modification rates for oxidized peptides within Arp2/3 complex in the absence of ATP (blue) and in the presence of 1 mM ATP (red) (A), and for oxidized peptides within Arp2/3 in the presence of 1 mM ATP (red) and WASp and for 1 mM ATP only (off-white) (B).

We used tandem MS analysis to identify specific side chains sensitive to ATP binding (Table 1). Specifically, residues within the nucleotide clefts of Arp2 and Arp3, including Tyr-16, Val-213, Pro-217, and Met-327 from Arp3, as well as Met-309 from Arp2, were less reactive with bound nucleotide. For Arp3, the protections within the cleft between subdomains 1 and 4 indicate partial cleft closure (23). These localized changes in the solution structure of Arp2/3 complex are consistent with differences in the conformations of Arp3 in crystals with and without bound nucleotide (12). In addition, within Arp2, protection from oxidation was observed in the interface of subdomains 3 and 4 (peptide 300–316), whereas probe sites within subdomains 1 and 2 did not change their reactivity upon ATP binding. This represents entirely new experimental information, because subdomains 1 and 2 of Arp2 are not observed in any crystal structures (12). FRET data of Goley et al. (11) demonstrate that nucleotide binding induces conformational changes in Arp2/3 in solution; our results indicate that the primary sites of conformational change are located adjacent to five probe sites within Arp2 and 3 as twenty five additional probe sites located throughout the complex did not change their reactivity. Furthermore, >160 additional peptides that were routinely detected did not show oxidation in the presence of nucleotide, thus the surface areas of their potentially reactive residues are not exposed to solvent as a result of ATP binding.

WASp VCA Binding Induces Allosteric Conformational Rearrangements.

To investigate the effect of WASp binding on Arp2/3 complex, 10 μM WASp VCA and 1 μM Arp2/3 were incubated in the presence of 1 mM ATP. This nearly saturated Arp2/3 complex with WASp so that any ligand-induced conformational changes could be examined (see Methods). Seven of the 30 peptides that exhibited oxidation had reduced oxidation rates in the presence of WASp VCA compared with ATP binding alone (Figs. 1D and 2B). These protected peptides are located in Arp3, including segments 80–91, 162–191, and 318–329, and in Arp2 including segments 54–65, 66–80, 87–97, and 300–316. The observed decreases in oxidation rate for Arp3 segment 162–191 and Arp2 segments 66–80 and 87–97 were 3- to 5-fold; prior footprinting experiments suggest protections of this magnitude are characteristic of burial in a well protected interface (15, 19, 24). Tandem MS analysis identified specific residues that are protected from oxidation in the Arp/WASp complex. These side-chain residues include Met-89, Tyr-184, and Met-327 in Arp3 and Met-56, Met-67, Tyr-72, Met-74, Trp-89, and Met-309 in Arp2 (Fig. 1D). Binding of WASp VCA did not change the rate of oxidation of other subunits in the complex.

Several of the reactive residues we identified were predicted by homology modeling to be potential interaction sites where WASp binds Arp2/3. For example, sequence analysis and modeling studies (25) predicted that Met-327 and Phe-328 on subdomain 3 of Arp3 should have altered solvent accessibility upon nucleotide binding and form a site for WASp binding. We see that Met-327 is protected upon nucleotide binding and exhibits additional protection upon WASp binding. In addition, conserved residues within sequences 154–161 are adjacent to Tyr-184 that shows WASp-dependent protection (25). Probes on the back of the Arp2 subdomain 2 (e.g., Arg-42, Leu-55, Met-56, Leu-64, Ser-66, Asn-71, and Asp-90) were also suggested as candidates for binding WASp. Note that the Arp2 residues Met-56, Met-67, Tyr-72, Met-74, and Trp-89, exhibited a protection from oxidation on WASp VCA binding.

A number of biochemical and biophysical approaches, including cross-linking (26–28), NMR (27,29), fluorescence (30), and FRET studies (11) have been used to examine the contact sites of WASp on Arp2/3 complex. Specifically, chemical cross-linking studies (26) favor binding of the A segment of WASp to Arp3. Furthermore, recently published NMR and cross-linking data suggest that a bivalent interaction of CA regions with the Arp2/3 complex stabilize activated Arp2/3 complex through interactions with Arp2, Arp3, ARPC1 and ARPC3 (27). However, site-directed spin labeling data only indicate that the regions of CA are within 25 Å of ARPC3, and not necessarily at the binding site (27). Modeling studies predicted that a patch of 46 conserved residues on ARPC1/p40 might be a potential binding site for WASp (25), however our experiments do not observe any conformational changes as function of WASp binding in this region.

Overall, these results, combined with sequence analysis and modeling studies indicate that WASp VCA mediates interactions between the two Arp subunits by crosslinking Arp3 and Arp2. In particular the level of protection seen for Arp3 residue M327 and Y184, and Arp2 residues M67, Y72, M74, and W89 suggest they are critical buried sites in the Arp2/3 WASp binding interface. Furthermore, we suggest that WASp-dependent protections from oxidation for segments 80–91 (Arp3) and segments 54–65 and 300–316 (Arp2) are related to physical contact between the Arps related to dimer formation.

Residues of WASp VCA That Interact with Arp2/3 Complex.

To identify residues in the WASp VCA domain that interact with Arp2/3 complex, we carried out oxidation experiments with 10 μM Arp2/3 complex and 1 μM WASp VCA (residues 432–502) to saturate most of the WASp with Arp2/3 complex. Samples were exposed for 0–80 ms to the x-ray beam, subjected to proteolysis and analysis by mass spectrometry. Three peptides of WASp including 447–472 and 447–476 located in the C region and 480–502 (mostly) from the A region were oxidized based on our LC-MS analysis (Table 2 and Fig. 3). Peptide 447–472 was oxidized at approximately the same rate in the presence of excess Arp 2/3 complex as free WASp VCA. Arp2/3 complex decreased the rate of oxidation of segment 447–476 by 83% and of peptide 480–502 by 33%. These results are consistent with previous biochemical (30) data and NMR line broadening results (29) indicating that both the C and A regions of WASp bind Arp2/3 complex. Specifically, Arp2/3 binding caused consistent broadening of resonances predominantly in the distal A sequence and throughout ≈15 continuous residues (465–483 of WASp) of the C region (29).

Table 2.

Rate constants for the oxidation of WASp peptides in the absence and presence of the Arp2/3 complex

Peptide	Sequence	Oxidized residue(s)	Rate of the oxidation, s−1
			ATP/WASp	WASp/Arp/ATP
447–472	TPGAPESSALQPPPQSSEGLVGALMH	M471	1.0 ± 0.2	1.5 ± 0.3
447–476	TPGAPESSALQPPPQSSEGLVGALMHVMQK	M471, M474	170 ± 30	29 ± 3
480–502	AIHSSDEGEDQAGDEDEDDEWDD	W500	5.8 ± 0.6	4.1 ± 0.2

Boldface indicates potentially modifiable residues in the inactive Arp2/3 structure.

Fig. 3.

Histogram plots of modification rates for VCA peptides within WASp in the presence and absence of Arp2/3 including 447–472, 447–476 (A), and 480–502 (B).

CID MS/MS analysis showed that Arp2/3 complex protected Met-474 and Trp-500 but not Met-471. Substitution of Met-475 within residues 467–481 of N-WASp (equivalent to Met-474 of WASp) to Ala leads loss of nucleation promoting activity (29). Our results confirm the burial of Met-474 (C region) in the WASp interaction with Arp2/3.

Binding to Arp2/3 complex has a very interesting effect on the oxidation of Trp-500. Oxidation of WASp VCA produced both +16 and +32 Da products in the absence of Arp2/3 complex, but only the +32 Da product in the presence of Arp2/3 complex. The overall modification rate constant for oxidation of isolated WASp was 5.8 ± 0.6 s⁻¹. This value can be deconvolved into individual rates of 0.9 ± 0.07 s⁻¹ for the +16 Da product and 4.6 ± 0.7 s⁻¹ for the +32 Da product so, the primary product is the +32 Da species. In the presence of Arp2/3 complex, the modification rate for the C-terminal peptide was 4.1 ± 0.2 s⁻¹ and only the +32 Da species was produced. Thus, Arp2/3 complex selectively suppressed the production of the +16 oxidation species.

MS/MS analysis of the doubly protonated ion for peptide 480–502 indicated that Trp-500 was the only residue within this peptide modified in the radiolysis experiment [see supporting information (SI) Fig. 5]. All b_6–18 ions are observed to be unchanged with respect to the unmodified peptide fragment, whereas y_3–17 ions are shifted by +16 or +32 Da for one or two oxygen products, respectively. This finding indicates that oxidation took place at the C-terminal end of this peptide. Further examination of the ion series revealed that the doubly protonated b₂₀²⁺ ion remained unchanged, whereas the b₂₁²⁺ ion showed modified products of +16 and +32 Da.

Close examination of the chemical mechanism of tryptophan oxidation explained why the +16-Da product disappears when WASp VCA binds Arp2/3 complex (Fig. 4). During radiolysis, hydroxyl radicals attack tryptophan side chains at the benzene and pyrrole rings; these results in +16-Da mass changes in the first case and +32-Da mass changes in the latter case due to the high reactivity of the pyrrole ring to subsequent oxidation upon initial oxidative attack (31, 32). Interaction of the benzene ring of Trp-500 with Arp2/3 complex protects this part of the side chain from oxidation, while leaving the pyrrole ring susceptible to oxidation leading to the N-formylkynurenine product. This mechanism explains opposite effects of saturating concentrations of Arp2/3 on the +16- and +32-Da products as well as the modest decrease in oxidation rate. Overall, our data suggest that Arp2/3 complex interacts with WASp within the C region with Met-474 and A region at Trp-500 through direct burial of the benzene ring.

Fig. 4.

Characteristic oxidations of tryptophan amino acid with specific mass changes.

Conclusion

We demonstrate the ability of the hydroxyl radical-mediated protein footprinting combined with mass spectrometry to examine the nature of protein–protein interactions and allosteric changes in conformation induced by ligand binding in a large (220 kDa) and complex macromolecular assembly. Moreover, this approach allows the study of protein surfaces under a variety of solution conditions at micromolar concentrations of analyte with resolution at the level of single side chains. Data presented here suggest conformational changes within Arp2/3 complex on ATP binding in Arp3 (localized in peptides 5–18, 212–225, and 318–327) and Arp2 (within peptide 300–316). These data are consistent with nucleotide docking within the nucleotide clefts of the actin related proteins promoting a cleft closure for the Arp3 subunit. Our data also suggest a bivalent WASp interaction site that anchors Arp3 (within peptide segments 162–191 and 318–329) and Arp2 (within peptide segments 66–80 and 87–97). WASp-dependent protections from oxidation within other peptides suggest domain rearrangements of Arp2 and Arp3 resulting in a closed conformational state consistent with an “actin-dimer” model. Our conclusion also is very consistent with EM (14) and modeling data (25), as well as FRET experiments (11) that indicate that Arp2/3 complex undergoes a substantial conformational change and that Arp2 and Arp3 adopt an actin-short-pitch dimer configuration (33). In addition, the Arp2/3 complex contacts WASp within the C subdomain at residue Met 474 and within the A subdomain to Trp 500.

Methods

Arp2/3 Complex Preparation.

Nucleotide-free Arp2/3 complex was purified from bovine thymus as described by Higgs et al. (13, 34). It was further dialyzed against 10 mM cacodylic acid sodium salt trihydrate, 1 mM MgCl₂, and 25 mM KCl (pH 7) at 4°C for 24 h. Protein concentrations were determined by using the molar absorptivity for Arp2/3 complex (ε₂₉₀ = 139000 M⁻¹ cm⁻¹). Nucleotide-bound complex was made by adding ATP to a final concentration of 1 mM after dialysis against 10 mM cacodylic acid sodium salt trihydrate, 1 mM MgCl₂ (pH 7). Arp2/3 complex bound to ATP and WASp were prepared by adding 10 μM WASp to 1 μM Arp2/3 complex in 1 mM ATP. For all experiments without WASp, the Arp2/3 complex concentration was adjusted to 1 μM. A 5-μl portion of each protein solution was dispensed into a 0.7-ml tube for radiolysis. The K_d of WASP VCA for Arp2/3 complex is 0.9 μM (30). Thus, our biochemical conditions of complex formation (10× excess of one component) assure >90% occupancy of the binding sites of interest.

WASp VCA Preparation.

GST-WASp, which includes only the VCA region of bovine WASp (432–502), was prepared according to Marchand et al. (30) but without an N-terminal cysteine. The WASp was further dialyzed against 10 mM cacodylic acid sodium salt trihydrate, 1 mM MgCl₂ (pH 7). The molar absorptivity at 280 nm of 46200 M⁻¹ cm⁻¹ was used to determine the GST-WASp concentration.

Radiolysis.

Radiolysis experiments were performed at beamline X-28C of the National Synchrotron Light Source, Brookhaven National Laboratory. Exposure times were controlled by using an electronic shutter (Vincent Associates, Rochester, NY). All experiments were performed at a ring energy of 2.8 GeV and at a beam currents ranging between 195 and 225 mA according to published procedures (35, 36). All experiments were performed at ambient temperature.

Proteolysis and MS Analysis.

Radiolyzed protein complex was digested with sequencing grade modified trypsin at an enzyme to protein ratio of 1:50 (wt/wt) at 37°C for 12 h. The digestion reaction was terminated by freezing the samples. Using a column switching technique mediated by a Switchos module (LC Packings, Sunnyvale, CA), the resulting peptides (800 fmol) were loaded onto a 300 μm ID × 5 mm C18, PepMap RP trapping column to preconcentrate and wash away excess salts. The loading flow of the Switchos was set to 30 μl/min, with 0.1% formic acid as the loading solvent. Reverse phase separation was performed on a 75 μm ID × 15 cm C18, PepMap column using the UltiMate nano separation system (LC Packings). Buffer A (95% water, 5% acetonitrile, and 0.1% formic acid) and buffer B (5% water, 95% acetonitrile, and 0.1% formic acid) were used to design a gradient. Proteolitic peptides eluted from the column with the gradient of acetonitrile of 1% per min were directed to a Thermo Finnigan (San Jose, CA) quadrupole ion trap mass spectrometer (DecaXP Plus) equipped with a nano-spray ion source and with the needle voltage of 1.8 kV. All mass spectra were acquired in the positive ion mode. The detected ion currents were used to determine the extent of oxidation by separate quantitation of the unmodified proteolytic peptides and their radiolytic products. The sites of oxidation were determined from MS/MS experiment. MS/MS data of the peptide mixtures were searched against bovine database for the oxidation of the tryptic peptides from bovine Arp2/3 complex using BioWorks 3.2 software (Thermo Electron, San Jose, CA). In addition, detected product–ion spectra for modified peptides were manually verified and correlated with hypothetical product-ion spectra predicted for the proteolysis products of the Arp2/3.

Because methionine is very reactive, and comprises most of the probes in this study, its accurate quantitation is very important to the validity of the results. We have determined that after radiolysis up to 90% of methionine oxidation is mediated by H₂O₂ and other stable radical species produced during exposure (37). To eliminate these secondary oxidations, 10 mM Met-NH₂xHCl buffer (pH 7.0) is added immediately after irradiation to all samples to quench methionine oxidation not induced by primary hydroxyl radical attack.

Modification Rate Calculation.

The integrated areas of unmodified and modified peptide ions, obtained from the total ion current, were used to calculate the modification extent as described (35). The fraction unmodified peptide was fit to the equation Y = Y₀ e^−kt using the Origin 6.0 (Microcal Software, Northampton, MA) program, where Y and Y₀ are the fraction of unmodified peptide at a time t and 0 (seconds), respectively and k is a first-order rate constant. Dose–response curves are presented as unmodified fraction (plotted on a logarithmic scale) versus x-ray exposure time.

Side Chain Solvent Accessibility Calculation.

To verify the correlation of side-chain reactivity and solvent accessibility for modified residues for Arp2/3 complex, the solvent accessibility surface areas of all side-chains were calculated in Ų. The VADAR computer program (PENCE, University of Alberta, Edmonton, Canada); the crystal structure of the apo-form of the Arp2/3 complex (Protein Data Bank ID code 1K8K) was used for this analysis.