Dissecting Requirements for Auto-inhibition of Actin Nucleation by the Formin, mDia1*

The mammalian formin, mDia1, is an actin nucleation factor. Experiments in cells and in vitro show that the N-terminal region potently inhibits nucleation by the formin homology 2 (FH2) domain-containing C terminus and that RhoA binding to the N terminus partially relieves this inhibition. Cellular experiments suggest that potent inhibition depends upon the presence of the diaphanous auto-regulatory domain (DAD) C-terminal to FH2. In this study, we examine in detail the N-terminal and C-terminal regions required for this inhibition and for RhoA relief. Limited proteolysis of an N-terminal construct from residues 1–548 identifies two stable truncations: 129–548 and 129–369. Analytical ultracentrifugation suggests that 1–548 and 129–548 are dimers, whereas 129–369 is monomeric. All three N-terminal constructs inhibit nucleation by the full C terminus. Although inhibition by 1–548 is partially relieved by RhoA, inhibition by 129–548 or 129–369 is RhoA-resistant. At the C terminus, DAD deletion does not affect nucleation but decreases inhibitory potency of 1–548 by 20,000-fold. Synthetic DAD peptide binds both 1–548 and 129–548 with similar affinity and partially relieves nucleation inhibition. C-terminal constructs are stable dimers. Our conclusions are as follows: 1) DAD is an affinity-enhancing motif for auto-inhibition; 2) an N-terminal domain spanning residues 129–369 (called DID for diaphanous inhibitory domain) is sufficient for auto-inhibition; 3) a dimerization region C-terminal to DID increases the inhibitory ability of DID; and 4) DID alone is not sufficient for RhoA relief of auto-inhibition, suggesting that sequences N-terminal to DID are important to RhoA binding. An additional finding is that FH2 domain-containing constructs of mDia1 and mDia2 lose >75% nucleation activity upon freeze-thaw.

Formin proteins are emerging as regulators of many cellular actin-based structures (1,2). Biochemically, formins exert several effects on actin polymerization dynamics, including acceleration of filament nucleation from monomers, inhibition of barbed end elongation rate, inhibition of complete barbed end capping by heterodimeric capping protein, and filament severing (3)(4)(5)(6)(7)(8)(9)(10). These in vitro activities are generally considered to result from the ability of formins to bind at or near the filament barbed end and to move processively with the barbed end as it elongates (11,12). Essential to these properties is the formin homology 2 (FH2) 1 domain, a 400-residue region generally found in the C-terminal half of the protein. Biochemical and structural studies show that the FH2 domain is dimeric for several formins (10,13,14), although longer constructs of the budding yeast formin, Bni1p, can tetramerize (6).
Mammals possess 15 formin genes, in seven distinct phylogenetic groups (15). For one mammalian formin, mDia1, the mechanisms regulating effects on actin have begun to be elucidated. The in vitro nucleation activity of the FH2-containing C terminus of mDia1 is inhibited potently by inclusion of a separate polypeptide containing the mDia1 N terminus (9), suggesting an auto-inhibitory regulatory mechanism. Cellular experiments and two-hybrid interactions implicate a short sequence C-terminal to FH2, known as the diaphanous autoregulatory domain (DAD), as a critical binding site for the N terminus in both mDia1 and the related protein, mDia2 (16). Direct effects of DAD on auto-inhibition of actin nucleation in vitro have not been examined.
Cellular studies suggest that binding of the Rho family GTPase, RhoA, to the N terminus of mDia1 can relieve autoinhibition (17). Biochemical studies on actin nucleation support these findings (9), with the caveat that RhoA does not relieve completely the auto-inhibitory effect of the N terminus of mDia1. This incomplete relief by RhoA might imply that a second, non-RhoA dependent, auto-inhibitory interaction between the N and C terminus might exist.
In this study, we examine mDia1 auto-inhibition in more detail. Deletion experiments show that auto-inhibition can be uncoupled from RhoA relief, suggesting that mDia1's N-terminal binding sites for RhoA and for the mDia1 C terminus are not identical. In addition, deletion of DAD from the C terminus does not affect nucleation but decreases the inhibitory potency of the N terminus 20,000-fold, suggesting that DAD mediates a high affinity interaction important for potent auto-inhibition. A synthetic DAD peptide binds the N terminus and partially disrupts auto-inhibition, supporting the role of DAD.

EXPERIMENTAL PROCEDURES
DNA Constructs-Constructs of mouse mDia1 (accession number U96963) were generated by reverse transcription-PCR and cloned into pGEX-KT, as described previously (9). Deletion constructs were generated by PCR from longer constructs, using Pfu DNA polymerase (Stratagene). The mouse mDia2 521-1171 construct in pGEX-KT was a kind gift from Dr. Arthur Alberts (Van Andel Research Institute).
Protein Preparation and Purification-All proteins were expressed and purified through the thrombin cleavage step following the procedure described in detail in Refs. 9 and 10. After elution of thrombincleaved protein from glutathione-Sepharose, further purification varied as follows. N-terminal constructs (1-548, 129 -548, and 129 -369) were purified by fast protein liquid chromatography on a SourceS15 5/5 or 10/10 column (Amersham Biosciences) and then concentrated on Q Sepharose Fast Flow (Amersham Biosciences) and dialyzed into the following buffer: 2 mM NaPO 4 , pH 7.0, 50 mM NaCl, 0.1 mM MgCl 2 , 0.1 mM EGTA, 0.5 mM DTT. N-terminal constructs could be frozen in aliquots of Ͻ50 l in liquid nitrogen and subsequently stored at Ϫ70°C with no loss of inhibitory activity or detectable aggregation. C-terminal constructs (748 -1255, 748 -1203, 748 -1175, and mDia2-(521-1171)) were purified by fast protein liquid chromatography on a SourceQ15 5/5 or 10/10 column (Amersham Biosciences), concentrated on SP Sepharose Fast Flow (Amersham Biosciences), and dialyzed into the following buffer: 2 mM NaPO 4 , pH 7.0, 150 mM NaCl, 0.1 mM EGTA, 0.5 mM DTT. C-terminal constructs lost Ͼ75% of their nucleation activity when frozen. C-terminal constructs were stored at concentrations Ͻ10 M at 4°C for 2 weeks or with 1 volume of glycerol at Ϫ20°C for two months with no loss of activity. Protein concentrations were determined by two methods: from absorbance at 280 nm (extinction coefficient calculated by protein sequence using us.expasy.org/cgi-bin/protparam) and by Coomassie-stained SDS-PAGE using known amounts of actin as standards. Rabbit skeletal muscle actin was purified from acetone powder (18) and labeled with pyrenyl iodoacetamide (19). Both unlabeled and labeled actin were gel-filtered on S200 (20), which was crucial to obtain reproducible polymerization kinetics.
DAD Peptide Synthesis, Fluorescent Labeling, and Fluorescence Anisotropy-A peptide containing amino acids 1177-1200 of mouse mDia1 (DETGVMDSLLEALQSGAAFRRKRG) was synthesized by the W. M. Keck Small Peptide Synthesis Facility at Yale University, with the following modifications: N-terminal acetylation; N-terminal cysteine residue; C-terminal amidation. The crude peptide was dissolved in water to 14 mg/ml in 100 mM NaPO 4 , pH 7.0. Peptide was labeled with fluorescein-5Ј-maleimide (FITC-maleimide, Molecular Probes F150) by mixing 1.4 mg/ml peptide with 1 mM FITC-maleimide at 20°C for 1 h. FITC-peptide was separated from unincorporated FITC-maleimide by Sephadex G-10 gel filtration (Amersham Biosciences) in 2 mM NaPO 4 , pH 7.0, 50 mM NaCl and then dialyzed for 48 h in two 4-liter changes of the same buffer, using SnakeSkin 3,500 molecular weight cut-off tubing (Pierce). Electrospray mass spectrometry (Dartmouth Proteomics Core Facility) of the purified peptide revealed a single peak at 3179 daltons, the expected mass of FITC-DAD, and no lower mass peak was observed (consistent with apparent complete labeling and unincorporated FITCmaleimide removal). FITC-DAD concentration was determined using the extinction coefficient supplied by Molecular Probes (83,000 M Ϫ1 cm Ϫ1 at 492 nm in 50 mM Tris-HCl, pH 9.0). Fluorescence anisotropy measurements were taken in an PC1 spectrofluorimeter (ISS Inc., Champaign IL), using 10 nM FITC-DAD in 10 mM imidazole, pH 7.0, 50 mM KCl, 1 mM MgCl 2 , 1 mM EGTA, 0.2 mM ATP, 0.5 mM DTT.
Limited Proteolysis of mDia1 N Terminus-The 1-548 fragment (100 M, 6 mg/ml) was digested with 300 nM chymotrypsin (Sigma C6423), trypsin (Sigma T8658), or proteinase K (Sigma P2308) in 10 mM Tris-Hcl (pH 8.0), 250 mM NaCl, 0.5 mM EDTA, 0.5 mM DTT for up to 1 h at 4°C (chymotrypsin or trypsin) or at 20°C (proteinase K). Digestion was terminated by the addition of phenylmethylsulfonyl fluoride/diisopropyl fluorophosphate (Calbiochem) to 1 mM each (from 50 mM stock of each in ethanol) and incubation on ice for 15 min. Upon restoration of eyesight, residual inhibitor was inactivated by the addition of dithiothreitol (Calbiochem) to 10 mM from a 1 M stock in water (freshly made) and further incubation for 15 min on ice. The major proteolytic fragment from each digestion was isolated by fast protein liquid chromatography on a SourceQ15 5/5 column (Amersham Biosciences) and then dialyzed in the same buffer used for mDia1 1-548 storage (described above), frozen in Ͻ50-ml aliquots in liquid nitrogen, and stored at Ϫ70°C. Masses of proteolytic products were determined by matrixassisted laser desorption ionization spectroscopy, and N-terminal sequences were determined by Edman degradation, both carried out in the Dartmouth Proteomics Core Facility.
Protein Size Analysis Techniques-Gel filtration chromatography was conducted using a Superdex 200 10/30 column (Amersham Biosciences) calibrated with both high and low molecular weight standards. Elution volumes (in ml) of standards were: void, 8.8; included, 21.7; thyroglobulin, 9.2; ferritin, 10.3; catalase, 12.2; aldolase, 13.7; albumin, 14.1; ovalbumin, 15.4; chymotrypsinogen, 17.7; and RNase A, 18.1. Stokes radii were calculated following the manufacturer's instructions. Gel filtration was conducted in sizing buffer (150 mM NaCl, 1 mM MgCl 2 , 1 mM EGTA, 10 mM NaPO 4 , pH 7.0, 0.5 mM DTT). Analytical ultracentrifugation was conducted using a Beckman Proteomelab XL-A and an AN-60 rotor. In all ultracentrifugation analysis, the proteins used were the peak fraction from the above gel filtration, and thus were used in the sizing buffer. For velocity analytical ultracentrifugation, protein was centrifuged at 35,000 rpm and 20°C, and UV absorbance was monitored every 2 min by continuous scan at 0.003-cm steps. For most experiments, protein was detected at 220 nm, with a starting absorbance of 0.5 absorbance units. For N-terminal fragments, this absorbance was equivalent to 1.25, 1.45, and 2 M for mDia1 1-548, 129 -548, and 129 -369, respectively. For mDia1 C-terminal fragments, this absorbance was equivalent to ϳ1.4 M in all cases (mDia1 748 -1255, 748 -1203, and 748 -1175) and to 1.2 M for mDia2 521-1171. For mDia1 129 -548 at higher concentrations, absorbance at 280 nm was used, with a starting concentration of 0.5 absorbance units (32 M). Protein partial specific volume, buffer density, and buffer viscosity were determined using Sednterp. 2 Scans 1-200 were analyzed using Sedfit87. For equilibrium analytical ultracentrifugation, multiple concentrations of protein, in sizing buffer, were centrifuged at 7000, 10,000, and 14,000 rpm for 20, 15, and 15 h, respectively. Three protein concentrations were used: the concentrations indicated above and two 2-fold serial dilutions of these concentrations. Scans at 220 or 280 nm and 0.001-cm steps were recorded every hour. Winmatch software 3 was used to confirm equilibrium, and Winreedit software 3 was used to trim the data. Winnonln software 3 was used to fit the data. First, individual concentrations at individual speeds were analyzed for speed-and concentration-dependent systematic variation. Next, data at multiple concentrations and speeds were fit to a single species model. If the systematic residual error pattern of this fit suggested non-ideality (21), the data from all speeds and concentrations were fit to a dissociating dimerization model, using a fixed corresponding to the monomer mass (determined by Sednterp). 2 Protein masses were confirmed by matrix-assisted laser desorption ionization.
Actin Polymerization by Fluorescence Spectroscopy-A detailed procedure is described in Ref. 22. Unlabeled and pyrene-labeled actin were mixed in G buffer (2 mM Tris, pH 8, 0.5 mM DTT, 0.2 mM ATP, 0.1 mM CaCl 2 , and 0.01% NaN 3 ) to produce an actin stock of the desired pyrenelabeled actin percentage (5% unless otherwise stated). This stock was converted to Mg 2ϩ salt by a 2-min incubation at 23°C in 1 mM EGTA, 0.1 mM MgCl 2 immediately prior to polymerization. Polymerization was induced by the addition of 10ϫ KMEI (500 mM KCl, 10 mM MgCl 2 , 10 mM EGTA, and 100 mM imidazole, pH 7) to a concentration of 1ϫ, with the remaining volume made up by G buffer. Added proteins were mixed together for 1 min prior to their rapid addition to actin to start the assay. Pyrene fluorescence (excitation 365 nm, emission 407 nm) was monitored in a PC1 spectrofluorimeter (ISS Inc.) or an LS50B spectrofluorimeter (PerkinElmer Life Sciences). The time between the mixing of final components and start of fluorimeter data collection was measured for each assay and ranged between 10 and 25 s.

Physical Properties of mDia1 N-terminal Constructs-We
have shown previously that an N-terminal fragment of mDia1 (residues 1-548, Fig. 1A) potently inhibits actin nucleation by FH2 domain-containing C-terminal constructs (9). In this study, we further dissected the N terminus. First, we characterized two proteolytically derived deletions of 1-548: a chymotrypsin-generated fragment from residues 120 -548 and a proteinase K-generated fragment from residues 129 -369 (Fig.  1A). For proteinase K, digestion at 20°C for up to 60 min resulted in no additional cleavage products (not shown). Given the promiscuous nature of proteinase K (24,25), resistance to digestion under these conditions suggests that the region from 129 to 369 is very stably folded. For chymotrypsin, digestion at 4°C for 1 h produced the 120 -548 fragment, whereas digestion at 20°C produced an additional product with similar SDS-2 Program by D. Hayes and T. Laue. 3 Program by D. Yphantis. PAGE mobility to that of the proteinase K-derived fragment (not shown). Trypsin digestion produced similar results to those with chymotrypsin (not shown).
Upon determining the N-and C-terminal residues of these proteolytic fragments, we produced analogous bacterial expression constructs, with the only modification being that the chymotrypsin fragment-like construct started at residue 129 instead of 120. The bulk of our subsequent characterization was carried out using these bacterially expressed fragments. However, proteolytically generated proteins mimicked their bacterially expressed counterparts in all properties tested below. All protein preparations appeared homogenous by SDS-PAGE (Fig. 1B).
Before testing the effects of these N-terminal deletions on actin, we characterized their native molecular sizes. Similar to our previous study (9), 1-548 eluted with apparent Stokes radius of 65.7 Å by gel filtration chromatography, whereas 129 -548 and 129 -369 eluted at 58.1 and 28.4 Å, respectively ( Fig. 2A). Assuming that each protein is spherical, these Stokes radii suggest molecular masses of 380, 241, and 40.3 kDa for 1-548, 129 -548, and 129 -369, respectively. Sedimentation velocity analytical ultracentrifugation of the peak gel filtration fractions, at concentrations between 1 and 2 M, resulted in predominant sedimenting species of 4.6, 4.1, and 2.4 s for 1-548, 129 -548, and 129 -369, respectively (Fig. 2B, Table I). However, both 1-548 and 129 -548 contained second sedimenting species at lower s values. Since the samples had been gel-filtered immediately prior to velocity centrifugation, the secondary peaks may suggest a reversible dissociation.
Sedimentation equilibrium analytical ultracentrifugation at multiple protein concentrations (between 0.3 and 2 M) and multiple speeds suggested that 129 -369 was monomeric. In contrast, 129 -549 and 1-548 fit best to monomer-dimer equi-libria, with apparent dissociation constants of 328 and 46 nM, respectively ( Fig. 2C and Table I). For 1-548 and 129 -548, the large discrepancy between molecular mass estimates from equilibrium ultracentrifugation and gel filtration, as well as the large frictional ratios calculated for velocity ultracentrifugation (Table I), suggest that these proteins are elongated rather than spherical. When equilibrium experiments were conducted at 20-fold higher concentrations (32 M maximum concentration), 129 -548 exhibited properties indistinguishable from those of a dimer (Table I and Fig. 2D). These results suggest that the region between residues 370 and 548 confer dimerization potential, with dimerization equilibrium constants in the range of 100 nM.
Effects of N-terminal Deletions on Actin Nucleation-We characterized the effects of 1-548, 129 -548, and 129 -369 on actin polymerization kinetics using the pyrene-actin polymerization assay in the absence or presence of 2.5 nM mDia1 748 -1203 (Fig. 1A). This construct contains the FH2 domain and DAD and is a potent nucleator (Fig. 3). None of the Nterminal constructs alone affected actin polymerization kinetics at concentrations up to 20 M ((9) and not shown). In contrast, all N-terminal constructs inhibited nucleation by 748 -1203 (Fig. 3, A-C). The IC 50 values for each construct were ϳ2 nM for 1-548, 20 nM for 129 -548, and 200 nM for 129 -369 (Fig. 3D). When 748 -1255 was used as the nucleation factor, results were indistinguishable from those using 748 -1203 (not shown).
We then tested the ability of RhoA to relieve inhibition by the N-terminal constructs. In contrast to 1-548, the inhibition of which is partially relieved by RhoA (9), the inhibitory activities of 129 -548 and 129 -369 are unaffected by high concentrations of RhoA ( Fig. 4 and not shown). The result was identical with either GDP or GMP-PNP-bound RhoA.
Importance of DAD to Auto-inhibition-From the data in Fig.  3, a construct containing the FH2 domain and DAD (748 -1203) behaves similarly to a construct containing the entire C terminus (748 -1255 (9)) in that it is a potent nucleator and its nucleation ability is inhibited by N-terminal constructs. To determine the role of DAD in auto-inhibition, we expressed a shorter construct in which DAD was truncated (748 -1175, Fig.  1A). This construct was equally potent to 748 -1255 and 748 -1203 in actin nucleation (Fig. 5A). We next compared the ability of 748 -1175 to be inhibited by 1-548. Measurable inhibition required 1-548 concentrations Ͼ10 M (Fig. 5B), but complete inhibition was never attained at the concentrations tested (Fig.  5C). The estimated IC 50 of 1-548 for 748 -1175 is 40 M.
To test directly an interaction between DAD and the N terminus, we synthesized a fluorescein-coupled 25-residue peptide containing the DAD sequence (FITC-DAD). By fluorescence anisotropy (26), FITC-DAD binds both 1-548 and 129 -548 (Fig. 6A), with a K d app of 0.25 M. The anisotropy change produced by 129 -369 was too small to determine affinity with accuracy, but the presence of 129 -369 consistently raised the anisotropy of FITC-DAD, suggesting that it bound as well. FITC-DAD was sufficient to disrupt the auto-inhibitory interaction between 748 -1255 and 1-548 (Fig. 6B). Similar to RhoA relief of auto-inhibition, FITC-DAD was unable to cause full recovery of 748 -1255 activity (Fig. 6C).
Effect of Freezing on mDia1 Nucleation Activity-Several laboratories have published work in which nucleation by mDia1 FH2 domain-containing constructs is examined (7,9,14,27,28), and the nucleation potency of mDia1 varies widely among these studies. Since this variability could influence mechanistic interpretations, we sought to determine factors that might cause these activity changes. We found that "flash" freezing resulted in the loss of Ͼ75% nucleation potency for mDia1 748 -1203 and mDia2 521-1171 (containing FH1, FH2, DAD, and C-terminal sequences), as well as for mDia1 748 -1255 and 748 -1175 ( Fig. 7 and not shown). Samples were frozen in 10-l aliquots in thin-walled PCR tubes by plunging into liquid nitrogen, conditions in which the sample was frozen in Ͻ2 s. This loss of activity occurred regardless of the presence of 50% glycerol. In contrast, similar freezing of mDia1 N-terminal constructs did not affect their abilities to inhibit nucleation (not shown). DISCUSSION In this study, we characterize the auto-inhibitory properties of mDia1 biochemically, with the following results. First, our work extends previous cellular findings of DAD function (16), suggesting that the role of DAD is to increase binding affinity to the N-terminal auto-inhibitory domains rather than to be part of the auto-inhibitory mechanism. Second, we identify a core inhibitory region, comprising residues 129 -369, which we refer to as DID (diaphanous inhibitory domain). Third, we show that auto-inhibition can be uncoupled from RhoA relief of autoinhibition since inhibition by the minimal DID or DID with a C-terminal extension is not relievable by RhoA. Fourth, we identify a region between DID and the FH1 domain that mediates dimerization.
We identify DID by limited proteolysis using proteinase K. The exceptional resistance of residues 129 -369 to proteinase K suggests that this region adopts a stable tertiary structure. In addi-  Table I. A, gel filtration chromatography of mDia1 1-548 (red), 129 -548 (green), and 129 -369 (blue). Lines above peaks indicate fractions used for subsequent analytical ultracentrifugation experiments. B, sedimentation velocity analytical ultracentrifugation of 1.25 M mDia1 1-548 (red), 1.45 M 129 -548 (green), and 2 M 129 -369 (blue). C, sedimentation equilibrium analytical ultracentrifugation of the same proteins. Curves were best fit to a monomer-dimer equilibrium (mDia1 1-548 and 129 -548) or to a single monomeric species (mDia1 129 -369), from nine data sets (three speeds and three concentrations). The protein concentrations were the same as in panel B, along with two serial dilutions of 2-fold. Curves depicted are for the middle concentrations at 10,000 rpm. Residuals are depicted below the curves. D, sedimentation equilibrium analytical ultracentrifugation of mDia1 129 -548 at higher concentration (32, 16, and 8 M). Data from nine conditions (three speeds and the three above concentrations) fit best to a single dimeric species. Curves depicted are for 16 M at 7000 (gray), 10,000 (black), and 14,000 rpm (red). Residuals are depicted below the curves.

TABLE I Size determination of protein constructs used in this study
Constructs are listed by first and last amino acids (Fig. 1A). Masses are in kDa. MALDI, matrix-assisted laser desorption/ionization-time of flight. Mass, solution, mass as determined by sedimentation equilibrium analytical ultracentrifugation, using at least three protein concentrations and three speeds. Sedimentation coefficient and frictional coefficient are determined by sedimentation velocity analytical ultracentrifugation. Signal, of total mass recovered by sedimentation velocity in main peak centered at indicated sedimentation coefficient. Numbers given in parentheses for 1-548 and 129 -548 indicate values for second species present. tion, the ability of this region to inhibit nucleation by the mDia1 C terminus suggests that it mediates the major auto-inhibitory interaction. We predict that this region contains the DAD-interacting region. Sequence alignments demonstrate that many formins, including members of the metazoan DAAM and FRL groups, as well as budding yeast Bni1p, contain regions similar to DID (15). All of these proteins also possess putative DADs. Thus, auto-inhibitory regulation through DID/DAD binding may be common to many formins. Notable exceptions are proteins in the delphilin, FMN, and INF metazoan groups (15), which we predict to be regulated by different mechanisms.
In contrast, since neither this region nor the 129 -548 construct appears to interact productively with RhoA, motifs important for RhoA binding must lie N-terminal to residue 129. Previous two-hybrid studies found that RhoA bound an mDia1 construct containing residues 63-260 (29). Possibly, full affinity for RhoA requires residues both within DID and N-terminal to DID.
Inclusion of residues from 369 to 548 causes N-terminal fragments to dimerize with affinity in the 100 nM range. Since a coiled-coil region is strongly predicted from 470 to 550 (15), we hypothesize that this coiled-coil region mediates dimerization, raising at least two questions. First, does this region dimerize in a parallel or antiparallel orientation? Second, what is the multimeric state of full-length mDia1, which contains both this dimerization region and the dimeric FH2 domain? Answers to these questions will be necessary for a complete understanding of mDia1 regulation.
Our experiments clearly show that DAD is required for high affinity auto-inhibition. Deletion of DAD from the C terminus raises the IC 50 of the N-terminal 1-548 construct from 2 nM to 40 M, or 20,000-fold. Synthetic DAD peptide binds DID-containing constructs and partially relieves inhibition. Mutations that disrupt DAD function in cells also disrupt DAD function in our biochemical assays. 4 Since DAD is not required for nucleation, and inhibition of the DAD-less C terminus can be affected at high concentrations of N terminus, we hypothesize that the role of DAD is to supply a high affinity interaction that enables a second interaction between DID and the FH2 domain. This second interaction is inhibitory to nucleation. The fact that RhoA does not fully relieve auto-inhibition suggests that a second activator is required for full activation. This activator may disrupt the non-DAD-mediated interaction or further destabilize the DID/DAD interaction.
The nucleation activity of mDia1 and mDia2 FH2 domaincontaining constructs is strongly diminished by freezing. This fact might contribute to varying mDia1 nucleation potencies observed by different laboratories (7,9,14,27,28  tivity to freezing might imply that FH2 domain-containing constructs are sensitive to other storage conditions as well. We observe that mDia1 FH2 domain-containing constructs partially precipitate in a variety of buffer conditions when stored at high concentrations (Ͼ10 M) at 4°C for extended periods (Ͼ3 days). For these reasons, we store mDia1 Ͻ10 M at 4°C, in which case full activity is maintained for ϳ2 weeks. For longer storage, the addition of glycerol to 50% v/v and storage at Ϫ20°C maintains full activity for several months. This sensitivity to freezing might not be universal for FH2 domains since freezing FH2 domain-containing constructs of FRL1 (10) does not affect polymerization or severing activities. 5