RhoB can adopt a Mg2+ free conformation prior to GEF binding


  • Meera Soundararajan,

    1. Structural Genomics Consortium, University of Oxford, Old Road Campus Research Building, Old Road Campus, off Roosevelt Drive, Headington OX3 7DQ, United Kingdom
    Search for more papers by this author
  • Andrew Turnbull,

    1. Structural Genomics Consortium, University of Oxford, Old Road Campus Research Building, Old Road Campus, off Roosevelt Drive, Headington OX3 7DQ, United Kingdom
    Search for more papers by this author
  • Oleg Fedorov,

    1. Structural Genomics Consortium, University of Oxford, Old Road Campus Research Building, Old Road Campus, off Roosevelt Drive, Headington OX3 7DQ, United Kingdom
    Search for more papers by this author
  • Catrine Johansson,

    1. Structural Genomics Consortium, University of Oxford, Old Road Campus Research Building, Old Road Campus, off Roosevelt Drive, Headington OX3 7DQ, United Kingdom
    Search for more papers by this author
  • Declan A. Doyle

    Corresponding author
    1. Structural Genomics Consortium, University of Oxford, Old Road Campus Research Building, Old Road Campus, off Roosevelt Drive, Headington OX3 7DQ, United Kingdom
    • Structural Genomics Consortium, University of Oxford, Old Road Campus Research Building, Old Road Campus, off Roosevelt Drive, Headington OX3 7LD, UK===

    Search for more papers by this author


RhoB regulates vital cellular function including apoptosis,1 cell cycle2 and cell shape, and migration.3 It is upregulated by a variety of external stimuli and toxins4–6 implicating its specific role in stress signaling pathways. Expression of RhoB is regulated by Ras via Akt/PKB kinases that regulates cellular growth, proliferation, survival, and metabolism.7 RhoB has also been shown to display stage-specific function in regulating endothelial cell survival during vascular development.8 This small GTPase plays a vital role in cell death program and endosomal traficking through downstream effectors mDia.9–11 In addition, RhoB has been identified as a dynamic component of the signaling pathways that coordinate Src activation and endosome-mediated translocation to transmembrane receptors.12

As for almost all other small GTPases, RhoB activates the signaling pathways by switching from an inactive GDP-bound form to the active GTP-bound state. The ability of GTPases to function as molecular switches is regulated through the differential action of guanine nucleotide exchange factors (GEFs) and GTPase activating proteins (GAPs) along with the essential cofactor Mg2+. GEFs convert the GTPases to the active state by replacing bound GDP with GTP, whereas GAPs reverse this process and down regulate the signaling pathway by enhancing the GTPase activity, thus converting the enzyme into the inactive GDP-bound state. Structural studies on GTPases over the years have clearly shown significant conformational changes between the two states, mainly in the distinct regions of the switch I and switch II loops.13–16 All of these studies show that the conformation and stability of the switch regions depends on the association of the GTPase with nucleotide, magnesium, GAP, and GEFs at any given time.

The role of magnesium as an essential cofactor is well established in GTPase function through structural and functional studies. GTPases need Mg2+ for both high-affinity nucleotide binding and maximal GTP hydrolysis. The role of Mg2+ in the GTP hydrolysis of Rho GTPase is thought to stabilize the switch I effector loop at Tyr34, which in turn positions the Mg2+ and γ-phosphate optimally for hydrolytic attack.17 Binding of guanine nucleotides to GTPases in the presence of Mg2+ is higher due to the conformation stability aided by the cation-mediated octahedral coordination.18, 19 Removal of Mg2+ has been shown to completely abolish nucleotide binding capacities of Ras,20 Rab3A,21 Sec4p,22 and Ral1A.23

The structural information for the RhoA family of Rho GTPases was limited to only RhoA, which includes the only other structure of the Mg2+ free form of a human Ras GTPases. However, this Mg2+ free form of RhoA is a dominantly active mutant in which Gly14 is replaced by Val.24 Hence, even though there are abundant structures of GTPases in the presence of nonhydrolyzable GTP analogues or GDP, our knowledge of the apo forms and the native Mg2+ free forms of these enzymes is limited. Our crystal structure of the Mg2+ free form of RhoB addresses this issue, increasing the structural knowledge of this subfamily of GTPases, but in addition shedding light on the GTPase cycle as the structure appears to represent a conformation of the GTPase prior to GEF binding and nucleotide exchange. This along with the functional and biophysical data supports the idea that for this GTPase removal of the Mg2+ ion occurs prior to GEF binding.


Overall RhoB structure

The sequence similarity between RhoB in comparison with RhoA and RhoC is 83 and 85%, respectively; the main differences occurring over the C-terminal regions. Full-length RhoB is a 196-amino acid polypeptide which is four amino acids longer than RhoA and RhoC. The structure presented here comprises residues Ile4–Asn187 with GDP bound in the nucleotide binding pocket, providing information on the core GTPase domain. Seven residues from Glu32 to Val38 are positionally disordered and were therefore not modeled. In many cases the disorder of the switch I loop is not unexpected in a GDP complex of a GTPase. In the active GTP-bound state a highly conserved Thr side chain from Switch I complexes with the magnesium, while other Switch I side chains interact with the γ-phosphate of the GTP thus helping to stabilize the Switch I loop conformation. These interactions are lost in the inactive form hence leading to positional variation of the Switch I loop as observed. The protein adopts a classical small GTPase nucleotide binding fold consisting of a six-stranded β-sheet flanked by five α-helices. The five α-helices (H1–H4 and H6) and the characteristic insert helix (H5) that is present only in Rho GTPases and six β-strands (B1–B6) connect with a B1-H1-B2-B3-H2-B4-H3-B5-H4-B6-H5-H6 topology similar to other Rho GTPases.

An unusual feature of the RhoB structure was that even though the nucleotide GDP was clearly present and the protein was purified in the presence of MgCl2, no Mg2+ ion was bound in the nucleotide binding site [Fig. 1(A)]. As seen before in the RhoA Mg2+ free structure,24 the Ala61 side chain [Fig. 1(A)] partially occupies the position that the Mg2+ ion would occupy, and hence emphasizing the importance of this side chain in either preventing Mg2+ binding or helping during the removal process.

Figure 1.

Conformational differences and similarities between the RhoB Mg2+ free structure and the various conformational states of RhoA. (A) 2Fo−Fc map focused on the nucleotide binding site of the RhoB structure. The known position for a bound Mg2+ ion is shown as a green sphere. (B) Comparison of switch (SW) I and II loops of Mg2+ free RhoB (2FV8) in red with active GTP and Mg2+ bound RhoA (1A2B; all-atom RMSD 2.09 Å) in green and inactive GDP and Mg2+ bound RhoA (1FTN; all-atom RMSD 1.67 Å) in blue. (C) Superimposition of Mg2+ free RhoB (2FV8) in red with the Mg2+free GEF bound RhoA (1LB1; all-atom RMSD 1.32 Å) structure in green and the Mg2+ free GDP bound RhoA (1DPF; all-atom RMSD 2.90 Å) shown in blue.

The functional state of the Mg2+ free RhoB structure

As the conformation of the switch loops and the presence of a particular nucleotide structurally define the functional state of small GTPases, the initial structural observations implied that RhoB was in the inactive state particularly due to the presence of bound GDP within the nucleotide binding site. RMSD values between RhoB and RhoA in the active state (1LB1) and inactive states (1FTN) of 2.1 and 1.7 Å, respectively, also pointed toward this interpretation. However, superimposition of the structures of RhoA in the active25 (PDB codes: 1A2B,26 1XCG,27 and 1LB1) and inactive (PDB codes: 1DPF24 and 1FTN28) states clearly illustrates that this RhoB structure adopts an alternative conformation with the largest differences observed over the Switch II loop regions [Fig. (1B)]. To reconfirm this observation three independent datasets were analyzed, each showing the same result that there was no electron density for a Mg2+ ion. Additionally, the switch I loop which was partially disordered with additional positive electron density indicating the direction of the conformation of switch I to be away from the nucleotide but continuing in the same direction as for the inactive GDP/Mg2+ bound RhoA structure. This is important as the highly conserved Thr37 side chain in the switch I loop is vital for the octahedral coordination of Mg2+ in the active GTP bound form,18 and thus helps to define the switch I conformation during effector-mediated interactions.16 Even though the switch I loop resembles more that for the inactive GDP RhoA conformation, the lack of bound Mg2+ and the differences in conformation of the switch II loops did not correlate with this structure representing an inactive state of the protein.

Additional structural comparisons of the Mg2+ free RhoB structure with the structures of RhoA in complex with GEFs or GAP proteins indicated that RhoB most closely resembles the structure of RhoA in complex with the GEF proteins Dbs exchange factor or PDZRhoGEF26, 27 [Fig. 1(C)]. The all-atom RMSD value for this Mg2+ free RhoA structure with RhoB is 1.3 Å indicating that, of the structures examined here, these two most resemble one another. This is particularly noticeable for the switch II regions where even the side chains adopt the same conformation which is significantly different from either the inactive or active RhoA structures. One highly significant residue on the switch II loop is a Gln (RhoB Gln63), which plays a crucial role in GTP hydrolysis by positioning the catalytic water molecule for nucleophilic attack.18, 29 In this structure, the side chain for this residue is turned away from the active site and plays no role in nucleotide binding similar to the Mg2+ free RhoA [Fig. 1(C)].

In the RhoA Mg2+ free form, the switch I loop is distant from the nucleotide binding pocket [Fig. 1(C)]. This lead to the hypothesis that the structure represented an open conformation with the switch I loop peeled away in readiness for nucleotide exchange.24 Closer examination of the RhoA Mg2+ free crystal structure along with its symmetry-related molecules showed that the residues in the switch I loop form a major protein–protein contact within the crystal, which calls into question the interpretation that it represented an open conformation [Fig. 2(A,B)]. For our RhoB Mg2+ free form, the switch I loop does not form any symmetry-related protein–protein contacts hence the structure more likely represents an undistorted native conformation [Fig. 2(C,D)]. This is further backed up by the close structural similarity between this RhoB Mg2+ free form and the Mg2+ free RhoA in complex with PDZRhoGEF or Dbs [Fig. 1(C)]24, 26, 27 and the Mg2+ free plant Rop4 GTPase in complex with its GEF, RopGEF.30 This similarity with the Rop4/RopGEF complex even extends as far as positional disorder of the switch I loop over almost the same equivalent residues as this RhoB structure.

Figure 2.

(A) Crystal packing view of the Mg2+ free RhoA structure (1DPF). RhoA is colored according to secondary structural elements with the closest symmetry-related molecules to the Switch I loop colored either blue or purple. The residues that make up the Switch I loop are drawn in ball-and-stick mode and colored according to atom type. (B) View of Mg2+ free RhoA at 90° to (A). (C) and (D) are similar views of the crystal packing for the Mg2+ free RhoB (2FV8). Representations and coloring as for RhoA. Seven residues of the Switch I loop are disordered so the connection between the ordered regions of Switch I is represented by a dashed line.

Effect of Mg2+ on GTPase activity

The effect of Mg2+ on the intrinsic GTPase activity was measured to confirm the influence of the cation on substrate hydrolysis. The assays clearly show that the enzyme is active in both the presence and absence of the cation [Fig. 3(A)]. In the absence of Mg2+, GTP hydrolysis was detected albeit at a comparatively slow rate of 0.067 min−1. In the presence of excess Mg2+, the release of γPi increased with a turnover rate of 0.514 min−1. The determined rate of RhoB hydrolysis is similar to RhoA with very slow hydrolysis under steady-state conditions in the absence of GAP proteins, unlike the high rates of hydrolysis determined for Cdc42 which is believed to be enhanced by an internal arginine finger.31 This correlates with the previous studies reporting the influence of Mg2+ on nucleotide binding and hydrolysis where RhoA, Cdc42, and Rac1 have Mg2+-independent basal GTP hydrolysis, but require Mg2+ for physiological activity and activation in the presence of GAPs.32 More importantly it also demonstrates that our RhoB protein preparation is capable of catalysis and that the structure adopted in the crystal form is not a result of misfolded protein.

Figure 3.

Hydrolysis of RhoB in the presence and absence of Mg2+. (A) Intrinsic GTP hydrolysis of the protein shown as a function of RhoB concentration in the presence (red filled squares) or absence (blue filled circles) of MgCl2. The rate of inorganic phosphate release (nM/min) was plotted against the enzyme concentration (μM). (B) Influence of Mg2+ on stabilization thermal denaturation curves of RhoB by itself, with the addition of GDP and in the presence and absence of Mg2+. Mid point values of the denaturation curves of RhoB under the various given conditions. (C) Repeat experiment as for (B) but using the buffer 50 mM ammonium citrate pH 5.5 plus 150 mM NaCl. Each thermal denaturation curve is an average of three repeats.

Effect of Mg2+ on stability of RhoB

The effect of Mg2+ on RhoB stability was tested in the presence and absence of Mg2+ ions with the addition of GDP [Fig. 3(B)]. The midpoint of the temperature denaturation curve (Tm) of the apo protein was determined to be 54.5°C. This protein preparation is likely to contain both GDP and Mg2+, however it is expected that not all of the available GDP and Mg2+ sites are fully occupied. This appeared to be true for GDP as the addition of 200 μM GDP shifted Tm in the positive direction by 7.7°C. Maximum stabilization by 64.7°C was observed in the presence of GDP and excess Mg2+. Fully occupying the GDP sites and removing all Mg2+ ions with the addition of EDTA destabilizes the RhoB by 7.2°C, shifting Tm to 47.3°C. Significantly, under these conditions RhoB is maximally stabilized in the presence of Mg2+ and nucleotide and that complete removal of Mg2+ causes a very large destabilization in comparison with the fully loaded Mg2+ complex by 17.4°C even when all of the GDP sites are fully occupied.

The RhoB structure was crystallized at pH 5.5 and in the presence of ammonium citrate. It is possible that these conditions could alter the behavior of RhoB hence the temperature denaturation assay was repeated but this time using the following buffer 50 mM ammonium citrate pH 5.5, 150 mM NaCl, and 2 mM DTT. Figure 3(C) shows the denaturation curves for 100 μM RhoB plus RhoB in the presence of 5 mM EDTA, 200 μM GDP, or 2 mM MgCl2. Even though the maximum stabilization provided by Mg2+ is lower at 14.7°C as compared with the earlier experiment at pH 7.5, it is important to note that the order and general trend are the same indicating that the protein's behavior is not dramatically altered at the pH values studied here.

Concluding remarks

The main finding of this structural study is that this Mg2+ free RhoB structure is structurally similar to small GTPase in a GEF-bound conformation. Additional structural information to back this interpretation is provided by the Rop4/GDP/PRONE8 structural complex.30 The mechanism that prevents Mg2+ binding appears to be the same as for the Rop4/GDP/PRONE8 structure in that an Ala side chain (RhoB Ala61) from the Switch II loop flips around from its normal position, where the methyl side chain points away from the Mg2+ binding site, to a position in which the side chain overlaps with the Mg2+ binding site [Fig. 1(A)]. Now that Mg2+ cannot bind the nucleotide binding site is less stable and therefore primed for the next step of GDP removal. The new feature for this mechanism that stems from this research is that this can occur without the presence of a bound GEF protein. The next step is the removal of the bound GDP. Within this structure a highly conserved Glu side chain of the Switch II loop is present directly after the conserved catalytic Gln, which in the active state plays no apparent role. In the Rop4 GEF complex, this Glu side chain interacts with the P-loop Lys that normally binds and helps to stabilize the β-phosphate of the GDP. In doing so it decreases the affinity of the GDP thus providing a partial mechanism by which GDP is released.30 These same sets of interactions exist in the Mg2+ free form of RhoB. In this case, Glu64 of RhoB forms a salt bridge with the P-loop Lys18 adopting a similar conformation as Rop4 and by extension plays a similar role as Glu65 of Rop4 in decreasing the affinity of the GDP. Significantly, however, RhoB was never in the presence of any of its known GEF proteins therefore this GTPase is able to undergo Mg2+ release without a GEF protein facilitating the removal of the cation.

The substrate hydrolysis study of RhoB supports the finding that Mg2+ is not essential for basal GTPase activity but that the intrinsic activity is enhanced approximately eightfold upon binding to the cation. Moreover, the thermal denaturation lends credence to the idea that Mg2+ ions play a central role in the hydrolytic process by stabilizing a structural conformation that allows nucleophilic attack of the γ-phosphate but also its removal destabilizes the overall RhoB structure potentially aiding nucleotide exchange once bound to a GEF protein.

Molecular dynamics simulations on a series of Mg2+-free GTPase structures demonstrated that these enzymes could adopt a conformation similar to a GEF-bound structure without the presence of a GEF protein to induce this conformation.33 Our structural data adds experimental proof to this analysis. Figure 1(C) shows the close conformational similarity of this RhoB structure with GEF-bound molecules of RhoA, Rac1, and Cdc42. Analysis of the effect of Mg2+ on GTP or GDP selectivity and GEF binding properties demonstrated that removal of Mg2+ increases the affinity of GEF for the GTPase and allows for preferential binding of GTP.32, 34 Hence using the available structural and functional data the major intermediates in the GTPase cycle between the active and GEF bound form can be represented using the following scheme:

equation image

The structural conformation of RhoB as presented here precedes GEF binding and is represented by GTPase.GDP in the above scheme.

Further studies will be required to uncover the mechanism by which the Mg2+ is removed that is the trigger that induces the switch II loop Ala side chain to knock out the Mg2+ and for the switch II Glu to destabilize the bound GDP as the previous assumption that this occurs as a result of GEF protein binding in this case is unnecessary. This structure therefore calls into question the role that the GEF protein now plays that is does the GEF only stabilize the GTPase in the nucleotide free state and/or does it actively contribute to the nucleotide exchange process? What is also unexplained for both this structure and the Rop4/PRONE8 complex is why these GTPases adopted these Mg2+ free forms when both were crystallized in the presence of excess amounts of Mg2+ ions. Even with these unanswered questions this Mg2+ free RhoB structure enhances our understanding of the structurally dynamic nature of small GTPases, it provides a clear picture of the steps involved in nucleotide exchange and thus how they function as efficient molecular switches.



A sequence containing the GTPase domain of RhoB (GenBank: gi:7661962) was amplified by PCR from DNA and subcloned into an in-house vector carrying kanamycin resistance, pNIC28-Bsa4, using ligation-independent cloning. The resulting plasmid expresses residues 4–289 of the GTPase with an N-terminal hexahistidine tag and TEV protease tag cleavage site (extension MHHHHHHSSGVDLGTENLYFQ*SM-). After digestion with TEV protease, the protein retains an additional serine and methionine on the N-terminus.

Expression and purification

The plasmid encoding RhoB polypeptide was transformed into BL21(DE3) competent cells, and the transformants used to inoculate 1 L of Terrific Broth media with 50 μg/ml kanamycin and grown at 37°C until an OD600 of 0.6 was reached. The protein was induced with 0.5 mM IPTG for 16 h at 18°C. The harvested cells were resuspended in lysis buffer (50 mM potassium phosphate pH 0.7.4, 500 mM NaCl, 5% glycerol, 10 mM imidazole). The cells were lysed using an Emulsiflex C5 high-pressure homogenizer (Avestin). The RhoB was extracted from the clarified supernatant by affinity tag purification using Ni-NTA resin (Qiagen). Protein containing supernatant was bound to the Ni-NTA and washed with a wash buffer (50 mM potassium phosphate pH 7.4, 500 mM NaCl, 5% glycerol, 30 mM imidazole). The protein was eluted with elution buffer (50 mM potassium phosphate pH 7.4, 500 mM NaCl, 5% glycerol, 250 mM imidazole). The protein was further purified by gel filtration chromatography (S200 16/60) in 50 mM Hepes pH 8.0, 150 mM NaCl, 2 mM MgCl2. The protein for GTPase assay was purified without MgCl2. Fractions containing the RhoB were identified using SDS-PAGE, and then pooled and concentrated to 18 mg/mL using a 10 kDa cutoff concentrator. The identity of RhoB was confirmed by mass spectrometry under denaturing conditions using Agilent LC-MS system with a reversed phase column (expected: 23599.5 Da and observed: 23,600 Da).

Crystallization and structure determination

Crystals of RhoB in complex with GDP were obtained by the sitting-drop method of vapor diffusion using a 96-well Greiner plate (Crystal Quick™ low profile) at 20°C and drops containing 150 nL of protein plus 150 nL of reservoir solution. These crystals grew optimally using 20% PEG3350 as the precipitant in 0.2M ammonium citrate, pH 5.0. Data were collected from a single, flash-frozen crystal (100 K) to 1.9 Å using a MAR225 imaging plate at beamline X10SA at the Swiss Light Source (Villigen, Switzerland). All data were reduced using the HKL2000 data processing system.35 This crystal belonged to space group C2 with unit cell parameters a = 137.5 Å, b = 42.2 Å, c = 33.9 Å, and β = 90.6°, and one molecule in the asymmetric unit based on a Vm value of 2.2 Å3/Da (assuming MW = 22,600 Da). The data collection statistics are presented in Table I. The structure of RhoB was determined by molecular replacement using the program PHASER36 implemented in the CCP4 suite (Collaborative Computational Project number 4, 1994). A homology model was constructed using SWISS-MODEL in the program's default settings37 based on the PDB coordinate sets, 1LB1 and 1 × 86. The derived model was used as the starting model for molecular replacement and gave a clear solution for a single molecule in the asymmetric unit. Subsequently, density improvement and automated model building was performed using ARP/wARP.38 Several rounds of iterative model building and REFMAC39 refinement followed and difference density in the map was assigned as the GDP moiety bound to the protein. The final model (comprising 184 amino acids, 112 water molecules and a GDP molecule), refined using data between 34.4 and 1.9 Å resolution, has an average B-factor of 38 Å2 and an R- and free R-factor of 21.0 and 25.3%, respectively, with good geometry. The refinement statistics are summarized in Table I. The coordinates and structure factors have been deposited in the Protein Data Bank with accession code 2FV8.

Table I. X-ray data collection statistics
  • a

    Rsym = ΣhklΣi|I − 〈Ii〉|/ΣhklΣiIi, where Ii is the intensity of a given measurement and the sums are over all measurements and reflections. Values in parentheses refer to the highest resolution shell.

  • b

    Rwork = Σ||F(obs)| − |F(calc)||/Σ|F(obs)| for the 95% of the reflection data used in the refinement.

  • c

    Rfree = Σ||F(obs)| − |F(calc)||/Σ|F(obs)| for the remaining 5%.

Resolution (Å)50.00–1.90 (1.97–1.90)
Total observations35,430 (2,789)
Unique observations14,890 (1,462)
Multiplicity2.4 (1.9)
Completeness (%)96.7 (97.5)
I/σ(I)〉15.4 (3.0)
Rsyma0.057 (0.341)
Refinement statistics
Resolution (Å)34.4–1.9
RMSD bond distances (Å)0.012
RMSD bond angles (°)1.413
Total number of non-H atoms1,387
Av. protein B value (Å2)38
Number of solvent molecules112
Av. solvent B value (Å2)41
Number of ligand atoms28
Av. ligand B value (Å2)36

GTPase activity assay

RhoB hydrolysis in the presence and absence of Mg2+ was measured using Malachite Green Assay. Briefly, 80-μL reaction mixture containing 5 μM GTP was incubated with 5–150 μM RhoB in 20 mM HEPES pH 7.5, 150 mM NaCl buffer, and 1 mM MgCl2 at 30°C. EDTA (5 mM) was supplemented depending on the assay. Control experiments were performed in each independent measurement to provide background measurements. The experiment was performed according to the manufacturer's instructions. The product formation was determined by recording the absorption at 650 nm using a SPECTRAmax spectrophotometer (Molecular Devices). Blanks containing substrate concentrations in the reaction buffer with malachite green solution were subtracted from each recording. The data were fitted to the Michaelis–Menten equation using nonlinear regression (Kaleidograph software). GTP was purchased from Sigma.

Thermal stability measurements

Thermal melting experiments were carried out using a real-time PCR machine Mx3005p (Stratagene, La Jolla, CA). Protein was buffered in 10 mM HEPES pH 7.5, 150 mM NaCl and assayed in a 96-well plate at a final concentration of 100 μM in 20 μL volume. SYPRO-Orange (Molecular Probes, Eugene, OR) was added as a fluorescence probe at a dilution of 1 in 1000. The plate was covered by optical foil, shaken gently for 10 min at room temperature, and centrifuged at 1000g for 30 s before starting the experiment. Excitation and emission filters for the SYPRO-Orange dye were set to 465 and 590 nm, respectively. Temperature was raised with a step of 1°C per 1.0 min from 25 to 96°C and fluorescence readings were taken at each interval. The temperature dependence of the fluorescence during the protein denaturation process was approximated by the equation

equation image

where yF and yU are the fluorescence intensity of the probe in the presence of completely folded and unfolded protein, respectively.40 The slope of the baselines of the fluorescent of the native and denatured state as a function of temperature was approximated by a linear fit. Data fitting was carried out using KaleidaGraph (Reading, PA).


We thank members of the Structural Genomics Consortium for assistance with plasmid preparation and diffraction data collection. The Structural Genomics Consortium is a registered charity that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co., Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research, and the Wellcome Trust.