Autoinhibition and Signaling by the Switch II Motif in the G-protein Chaperone of a Radical B12 Enzyme*

Background: MeaB is a G-protein chaperone of methylmalonyl-CoA mutase (MCM). Results: Mutations in the canonical switch II motif disrupt signaling in the MeaB-MCM complex. Conclusion: The switch II loop is autoinhibitory for the intrinsic GTPase activity of MeaB. Significance: Signaling in the MeaB-MCM complex is achieved via nucleotide-dependent conformational coupling between switches II and III. MeaB is an accessory GTPase protein involved in the assembly, protection, and reactivation of 5′-deoxyadenosyl cobalamin-dependent methylmalonyl-CoA mutase (MCM). Mutations in the human ortholog of MeaB result in methylmalonic aciduria, an inborn error of metabolism. G-proteins typically utilize conserved switch I and II motifs for signaling to effector proteins via conformational changes elicited by nucleotide binding and hydrolysis. Our recent discovery that MeaB utilizes an unusual switch III region for bidirectional signaling with MCM raised questions about the roles of the switch I and II motifs in MeaB. In this study, we addressed the functions of conserved switch II residues by performing alanine-scanning mutagenesis. Our results demonstrate that the GTPase activity of MeaB is autoinhibited by switch II and that this loop is important for coupling nucleotide-sensitive conformational changes in switch III to elicit the multiple chaperone functions of MeaB. Furthermore, we report the structure of MeaB·GDP crystallized in the presence of AlFx− to form the putative transition state analog, GDP·AlF4−. The resulting crystal structure and its comparison with related G-proteins support the conclusion that the catalytic site of MeaB is incomplete in the absence of the GTPase-activating protein MCM and therefore unable to stabilize the transition state analog. Favoring an inactive conformation in the absence of the client MCM protein might represent a strategy for suppressing the intrinsic GTPase activity of MeaB in which the switch II loop plays an important role.

MeaB from Methylobacterium extorquens and its human ortholog CblA play critical roles in the docking of coenzyme-B 12 (or 5Ј-deoxyadenosyl cobalamin (AdoCbl) 2 ) into the active site of the client enzyme methylmalonyl-CoA mutase (MCM) (1)(2)(3)(4). MCM belongs to the class of AdoCbl-dependent mutases that catalyze 1,2-rearrangement reactions (5). Dietary cobalamin is assimilated into AdoCbl and delivered to MCM in a complex trafficking pathway (6 -8). In humans, MCM functions in the mitochondrial catabolism of branched-chain amino acids, odd-chain fatty acids, and cholesterol by converting methylmalonyl-CoA to succinyl-CoA. Mutations in the auxiliary proteins or in MCM itself give rise to methylmalonic acidemia, an inborn error of metabolism that is inherited as an autosomal recessive disease (9 -13). Nearly 30 pathogenic mutations have been described in CblA (also known as MMAA) leading to lower AdoCbl levels and consequent impairment of MCM (11).
Our understanding about the function of CblA is derived primarily from biochemical studies on MeaB, which belongs to the G3E family of SIMIBI phosphate-binding loop (P-loop) G-proteins (14). Several members of this family of NTPases serve as metallochaperones. These include CooC, HypB, and UreG, accessory proteins involved in the maturation of nickelcontaining enzymes. CooC is a Ni 2ϩ -binding ATPase that undergoes metal-and nucleotide-dependent dimerization and catalyzes the insertion of nickel into carbon-monoxide dehydrogenase (15). UreG is a GTPase that is needed for the insertion of nickel into urease (16). HypB is activated by GTP-dependent dimerization and is needed for nickel insertion into the [Ni-Fe] hydrogenase (17). Some metallochaperones bind the transition metal and transfer it directly to the target protein, whereas others facilitate metal insertion but do not directly bind the cofactor. MeaB is an example of the latter class as it gates the transfer of AdoCbl from adenosyltransferase (ATR) to MCM (3) but does not itself bind the metal-containing cofactor. GTP hydrolysis by MeaB is required for transfer of AdoCbl from ATR to MCM.
Most G-proteins including the G3E GTPases are predicted to signal via conserved sequence motifs known as switch I and switch II (18 -22). Switch I and switch II function as loaded springs that interact with the Mg 2ϩ ion in the GTPase site and signal to target proteins by undergoing conformational changes in response to nucleotide binding, hydrolysis, and exchange. In G3E G-proteins, the EXXG peptide defines the minimal consensus sequence for the switch II region (14). The switch I sequence is not strongly conserved with the exception of the nearly ubiquitously present threonine and glutamate residues in Ras-like GTPases (23). The ␥-phosphate of GTP makes direct hydrogen bonding contacts with the main-chain amides of the conserved threonine and glycine residues in the switch I and II sequences, respectively. The carboxylate of the switch II glutamate residue forms a water-mediated contact with the Mg 2ϩ ion that is also coordinated by the ␤and ␥-phosphate oxygen atoms. In Ras proteins, the water molecule that serves as a nucleophile in the GTPase reaction is activated via interaction with a conserved glutamine residue in the switch II consensus motif, DXXGQ. In G3E proteins such as HypB, NifH, and SRP, the residue that activates the nucleophilic water is predicted to be a conserved aspartate residue at position 1 in the switch I sequence (24 -26). However, the corresponding aspartate in MeaB, Asp-92, is not positioned for catalysis. It is displaced 11-14 Å from the active site in all reported structures (27,28). CblA and MeaB comprise seven central and parallel-stranded ␤-sheets in the G-domain and flanking N-and C-terminal ␣-helical extensions (27,29). The C-terminal extension forms a dimerization arm. In addition to the switch I (residues 92-108) and switch II (residues 154 -162) loops, two other signature G-protein motifs are present in MeaB: the P-loop and the base specificity loop that extend between residues 62 and 70 and residues 200 and 204, respectively. The P-loop has a GXXGXGKST consensus sequence that interacts with the phosphate moiety of nucleotides.
The switch II motif in MeaB-like GTPases has the following consensus sequence: ETVGVGQSE. In a subset of SIMIBI G-proteins, GTP-dependent conformational changes in the switch regions expose dimerization surfaces that are essential for their biological function (30). In Ras-like GTPases, GTP hydrolysis elicits a conformational change in the switch I region that exposes an effector domain involved in recognition of downstream target proteins (22). In the G3E G-proteins, the functions of switch I and switch II motifs in the context of their metallochaperone functions have not been addressed. In MeaB, an additional conformationally dynamic region, switch III, plays a critical role in bidirectional signal transduction with MCM (28). Although unrelated at a primary sequence level, the switch III region in the G␣ subunits of heterotrimeric G-proteins is used for signal transmission (31)(32)(33)(34).
MeaB exhibits similar affinities for GTP and GDP and is expected to be predominantly GTP-loaded in the cell (2). The MeaB⅐GTP complex prevents assembly of the MCM active site with inactive precursors of AdoCbl (3). Comparison of the apo, GDP-, and GMPPNP-bound structures of MeaB with other G3E family members suggests that switch I and switch II are not in their catalytically active conformations (Fig. 1). The association of MeaB with MCM results in an ϳ100-fold rate enhancement of the GTPase reaction and increases the affinity for GTP (4). Thus, MCM functions as a GTPase-activating protein (GAP) for MeaB. Although other G3E subfamily proteins exhibit low intrinsic GTPase activity, GAP activators have not been reported for them to the best of our knowledge. Two models for GAP activation of MeaB by MCM that are not mutually exclusive include: (i) contribution in trans of one or more catalytic residues by MCM to the active site of MeaB and (ii) induction of a conformational change in the MeaB⅐MCM complex that brings the active site residues of MeaB into catalytic register (35,36).
In addition to gating the transfer of AdoCbl from ATR to MCM, MeaB also functions to protect MCM from inactivation during catalytic turnover and to rescue MCM that is inactivated (2,3). Thus, in the presence of either GDP or GTP, MeaB slows the rate of MCM inactivation by 30-fold (28), and the exchange of GDP for GTP drives the ejection of inactive cofactor from the active site of MCM. The limited biochemical studies support a role for CblA that is analogous to the better characterized bacterial MeaB in protecting MCM from inactivation and rescuing inactive MCM (37). Furthermore, human MCM also exhibits GAP activity with respect to CblA (29).
In this study, we interrogated the role of the switch II motif in mediating the chaperone functions of MeaB and in transmitting the GAP function of MCM. To this end, we used alaninescanning mutagenesis at positions in switch II that are conserved between MeaB and CblA. We demonstrate that switch II mutations compromise regulation of AdoCbl transfer from ATR to the MCM⅐MeaB complex, reduce protection of MCM from inactivation during turnover, and impair ejection of inactive cofactor from the active site of MCM. Remarkably, a subset of switch II mutations enhances the intrinsic GTPase activity of MeaB while disrupting MCM-dependent GAP activation. We also report the structure of MeaB crystallized in the presence of GDP and AlF x Ϫ . Our results suggest a strategy for switch II-dependent autoinhibition of the intrinsic GTPase activity of MeaB, which exists predominantly in an inactive conformation. Complexation with MCM is predicted to stabilize MeaB in a conformation in which its GTPase function is activated.

EXPERIMENTAL PROCEDURES
Materials-AdoCbl, GMPPNP, ATP, GTP, methylmalonic acid, coenzyme A, and other reagent grade chemicals were purchased from Sigma. Trifluoroacetic acid was purchased from Aldrich. Methylmalonyl-CoA was synthesized using malonyl-CoA synthetase and purified as described previously (38). The HPLC column used to quantify guanosine nucleotides was a Bondapak TM NH 2 10-m, 125 Å, 3.9 ϫ 300-mm column purchased from Waters.
Crystallization of MeaB-Protein samples were concentrated to 11 mg ml Ϫ1 in 50 mM HEPES buffer at pH 8.0, 2.5 mM MgCl 2 , 5 mM GDP-AlF x Ϫ . GDP-AlF x Ϫ was prepared by mixing GDP, AlCl 3 , and potassium fluoride at a 1:5:15 ratio and used after gel filtration. The MeaB⅐GDP crystals were grown by the sitting drop vapor diffusion method at 20°C by mixing 2 l of protein solution with 2 l of reservoir solution, which contained 20% (w/v) PEG 3350 and 0.2 M sodium sulfate. Crystals were cryoprotected for a few minutes before being flash frozen in liquid N 2 by transfer to a solution of 20% glycerol, 15% PEG 3350 (w/v), 0.15 M sodium sulfate, 5 mM GDP, and 1 mM AlF x Ϫ in 25 mM HEPES, pH 7.5. Crystals of MeaB⅐GDP were of space group P2 1 (a ϭ 63.7, b ϭ 78.6, c ϭ 69.6, ␤ ϭ 108.6) with two monomers in the asymmetric unit.
Data Collection and Structure Determination-Diffraction data for MeaB⅐GDP crystallized in the presence of AlF x Ϫ were collected at 100 K on beamline GM/CA-CAT 23-ID-B at the Advanced Photon Source, Argonne National Laboratory (Argonne, IL). Data were recorded on a Mar300 detector and processed with XDS (40) to 1.8-Å resolution (Table 1). EPMR (41) was used to determine initial phases for MeaB⅐GDP⅐AlF x Ϫ through molecular replacement using a single monomer of the MeaB⅐GDP structure (Protein Data Bank code 2QM7 (27)) as a search model. Loops containing residues 62-67, 95-100, 181-186, and 225-231 were removed from the search model to eliminate bias. Initial density allowed for the ligand to be modeled and added. REFMAC (42) of the CCP4 suite (43) was subsequently used for restrained refinement of the model using isotropic individual B-factors to a final R work of 0.180 and R free of 0.219. Model building and modification were performed with Coot (44), and the geometric quality of the models and their agreement with the structure factors were assessed with Mol-Probity (45). Crystallographic information and refinement statistics are provided in Table 1. Figures were generated with PyMOL (46).
Thermodynamics of GMPPNP Binding-Isothermal titration calorimetry experiments were performed at 10°C in Buffer A using a 300-l injection syringe and a 1.43-ml injection cell. Samples were prepared by filtration through a 0.2-m filter and then degassed under vacuum at 4°C using a ThermoVac sample degasser. Each titration was performed at least in duplicate. GMPPNP (10-l injections of 150 -400 M) was added to 10 -25 M MeaB. The data were analyzed using a two-site binding model using the MicroCal Origin program. Values for the dissociation constant at sites 1 and 2 were then compared with the values obtained from the solution of the Gibbs free energy equation: ⌬G 0 ϭ ϪRTln(K A ).
Enzyme Inactivation Assays-Inactivation of MCM during steady-state turnover was monitored by UV-visible spectroscopy to follow conversion of MCM-bound cob(II)alamin to aquocob(III)alamin (H 2 OCbl) at 20°C in 0.1 M potassium phosphate at pH 7.5 containing 10 mM MgCl 2 . The reactions and sample preparations were performed in the dark to avert spurious H 2 OCbl formation by photolysis of AdoCbl followed by oxidation. Samples were prepared by the addition of reaction components in the order described below. MCM (25-30 M) was reconstituted with an equimolar concentration of AdoCbl. A slight molar excess (35-40 M) of MeaB was added to the MCM holoenzyme to generate an MCM-MeaB complex. GMPPNP was then added to the reaction mixtures to a final concentration of 1-2 mM. The reaction was initiated by the addition of methylmalonyl-CoA to a final concentration of 4.5-5 mM. The rates of inactivation were determined by plotting the change in absorbance at 351 nm, corresponding to H 2 OCbl formation, as a function of time. The kinetic traces were best fit by a single exponential equation: ⌬A t ϭ A 0 Ϫ ⌬A p e (Ϫkt) where ⌬A t is the absorbance at 351 nm as a function of time, A o is the initial absorbance of cob(II)alamin, ⌬A p is the reaction phase amplitude for H 2 OCbl formation, and k is the observed rate constant for MCM inactivation.
Assay for Transfer of AdoCbl from ATR to MCM-The ATPdependent transfer of AdoCbl from ATR to the MCM-MeaB⅐GMPPNP complex was performed in the dark at 20°C and monitored by UV-visible spectroscopy. Two equivalents of AdoCbl were added to 1 eq of ATR in Buffer A to generate holo-ATR. The apo-MCM-MeaB⅐GMPPNP complex was were mixed and incubated for 10 min at 20°C before addition of ATP to a final concentration of 5 mM. Release/transfer of AdoCbl from ATR was calculated using a ⌬⑀ 525 ϭ 6.69 mM Ϫ1 cm Ϫ1 . Bound versus free cofactor was separated using an Amicon centrifuge filter (10-kDa cutoff, 20 min, 4°C, 16,000 ϫ g). The concentration of free AdoCbl in the filtrate was calculated using ⑀ 525 ϭ 8.0 mM Ϫ1 cm Ϫ1 .
Release of Cob(II)alamin-MCM (30 -40 M) was mixed with 45-60 M wild-type, E154A, T155A, Q160A, S161A, E162A, or Q160A/E162A MeaB in Buffer A at 20°C under strictly anaerobic conditions such that [MCM]:[MeaB] was 1:1.5. Cob(II)alamin was generated by reduction of H 2 OCbl with tris(2-carboxyethyl)phosphine hydrochloride and added to a final concentration equal to that of the MCM-MeaB (wildtype or mutant) complex. The reaction mixture was incubated for 10 min at 20°C. GMPPNP in anaerobic Buffer A was added to a final concentration of 2 mM. The mixture was then incubated for 20 min at 20°C. Subsequently, the sample was made aerobic by air oxidation for 2 h and then filtered through a Centricon YM10 filter (10-kDa cutoff) to separate free from bound H 2 OCbl. Cob(II)alamin (but not H 2 OCbl) is released from MCM and is subsequently oxidized to H 2 OCbl, which was quantified using ⑀ 350 ϭ 26.5 mM Ϫ1 cm Ϫ1 .
GTPase Activity of MeaB-The steady-state kinetic parameters for the GTPase activity of wild-type or switch II mutants of MeaB were determined using 20 M MeaB or 5 M MeaB in the presence of 10 M MCM (to generate the MCM-MeaB complex). The reactions were performed in Buffer A containing 25 mM MgCl 2 at 20°C and initiated by the addition of GTP to a final concentration ranging from 0.01 to 10 mM. The reactions were quenched using 2 M trichloroacetic acid, and the GTPase activity of MeaB alone or the MCM-MeaB complex was determined using an HPLC assay as described previously (4). The concentration dependence for the observed rate of GTP hydrolysis was plotted versus GTP concentration, and the data were fit to a Michaelis-Menten equation to determine values for k cat and K m .
Size Exclusion Chromatography-Samples were prepared in Buffer A containing 0.5 mM GDP, 80 -90 M wild-type or mutant MeaB Ϯ equimolar MCM, and 0.5 mM AdoCbl in a total volume of 150 l. The samples were loaded onto a Superdex-200 HR 10/30 column equilibrated with Buffer A. The protein complexes were eluted at 4°C in the dark to minimize photolysis of AdoCbl.

Characterization of a New MeaB⅐GDP Crystal Structure-MeaB crystallized in the presence of GDP-AlF 4
Ϫ , a potential transition state mimic, in the P2 1 space group and a unit cell similar to that reported for MeaB crystallized in the presence of GDP (Table  1). However, AlF 4 Ϫ was not observed in any of the more than 10 structures that were solved. Furthermore, despite the presence of Mg 2ϩ in the crystallization solution, the cation was not observed in the new MeaB⅐GDP crystal structure (Fig. 2a), which appeared to be very similar to that of MeaB⅐GDP published previously (27). However, closer inspection revealed two significant differences as described below.
Superposition of the new MeaB⅐GDP and previously determined MeaB⅐GDP structures yielded root mean square deviation values of 2.31 Å for the C␣ atoms of all residues. Superimposition of the MeaB dimers and alignment with respect to the nucleotides revealed no differences in the nucleotide binding a Statistics for the highest resolution shell are enclosed in parentheses. b R sym ϭ Α͉I Ϫ I͉/ΑI where I is the observed intensity and I is the average intensity obtained from multiple measurements. c R cryst ϭ ΑʈF obs ͉ Ϫ ͉F calc ʈ/Α͉F obs ͉ where F calc and F obs are the calculated and observed structure factor amplitudes, respectively. d R free , R-factor based on 5% of the data excluded from refinement. e Root mean square. OCTOBER 25, 2013 • VOLUME 288 • NUMBER 43 sites between the new MeaB⅐GDP and previously reported MeaB⅐GDP structures. However, Arg-108 in the previously reported MeaB⅐GDP structure (Fig. 2b, yellow) is solvent-exposed and facing away from the active site cavity, whereas it is swung in and engages in a salt bridge contact with the carboxylate of Glu-154 in the new structure of MeaB⅐GDP (Fig. 2b,  blue). To assess whether the observed structural differences were due to slightly different crystallization conditions, the structure of MeaB⅐GDP was redetermined under the conditions used to obtain the new MeaB⅐GDP structure crystallized in the presence of AlF x Ϫ . The new and published (Protein Data Bank code 2QM7) structures of MeaB⅐GDP were identical.

Switch II Signaling in MeaB
Additional differences between the new MeaB⅐GDP and previously determined MeaB⅐GDP structures are observed in the switch III region (residues 178 -188) and its interaction with conserved residues Gln-160 and Glu-162 in switch II (Fig. 2c). In the new structure of MeaB⅐GDP reported here, the side chain of Gln-160 (Fig. 2c, blue) is swung ϳ25°from its position in the previously reported MeaB⅐GDP structure (yellow). The electron density for the side chain of Lys-188 in the new MeaB⅐GDP structure allows it to be modeled in two configurations. In one configuration, Lys-188 contacts switch II via an ionic interaction with the carboxylate of Glu-162 and a single hydrogen bond with the carbonyl oxygen of Gly-159. In the other configuration, the interactions between Lys-188 and switch II are disrupted. In contrast, hydrogen bond contacts between Lys-188 of switch III and Glu-162 or any other part of switch II are not observed in the previous MeaB⅐GDP structure (Fig. 2c, yellow). The contacts formed between switch II and switch III in the MeaB⅐GDP structure reported here are also distinct from those captured in the structure of MeaB⅐GMPPNP in which the side chain of Lys-188 interacts with the carboxylate of Glu-162 and the amide side chain of Gln-160 (Fig. 2c, gray).
The new MeaB⅐GDP and previously reported MeaB⅐GDP structures also differ in the interaction of the switch III region with a flexible loop from the adjacent monomer extending between residues 224 and 232. In one of the monomers of the new MeaB⅐GDP structure, the backbone carbonyl oxygen of Glu-183 is hydrogen-bonded to Ser-230 (Fig. 2d, blue). In contrast, the conformation of switch III in the previously solved MeaB⅐GDP structure precludes Glu-183 from contacting the 224 -232 loop (Fig. 2d, yellow). The only contact formed between switch III and the 224 -232 loop in the previously reported MeaB⅐GDP structure is a single hydrogen bond between the side chain of His-224 and the carboxylate of residue Asp-182. At the same subunit interface of the new MeaB⅐GDP structure, Asp-182 is disordered. In the opposing subunit interface of the MeaB⅐GDP structure reported here, Asp-182 does not engage in a hydrogen bond interaction with the 224 -232 loop and is displaced ϳ9 Å compared with its position at the corresponding subunit interface in the previously reported MeaB⅐GDP structure. The relative orientations of the switch I, II, and III loops in the new MeaB⅐GDP structure are shown in Fig. 3.
Effect of Switch II Mutations on MCM-MeaB Complex Formation-Gel filtration chromatography was used to qualitatively assess complex formation between MCM and MeaB in the presence of a molar excess of GDP and AdoCbl. Wild-type and mutant MeaBs and MCM migrated with molecular masses corresponding to 68 and 142 kDa, respectively, consistent with their calculated masses. With each MeaB mutant, formation of an MCM-MeaB complex with a molecular mass of 243 kDa was observed, consistent with the formation of a 1:1 complex between the two proteins (data not shown).

GMPPNP Binding to Switch II MeaB Mutants-The
MeaB homodimer binds nucleotides with negative cooperativity ( Table 2). GMPPNP binds to sites 1 and 2 with K D values of 0.8 Ϯ 0.5 and 9.5 Ϯ 1.9 M, respectively. The K D values for GMPPNP binding range from 0.16 Ϯ 0.01 to 1.3 Ϯ 0.3 M at site 1 and from 0.48 Ϯ 0.07 to 9.6 Ϯ 1.4 M at site 2. Overall, the switch II mutations do not have a substantial impact on GMP-PNP binding with the exception of the E162A and Q160A/ E162A mutants in which the affinity for the nucleotide increases ϳ5-20-fold. GMPPNP binds to each switch II mutant with negative cooperativity.
GTPase Activity of Switch II Mutants-The intrinsic and GAP-stimulated GTPase activities of the single and double switch II mutants were compared with wild-type MeaB (Table 3). Surprisingly, the E154A, Q160A, and Q160A/E162A mutants exhibited ϳ5-10-fold activation of the intrinsic GTPase activity compared with wild-type MeaB. The GAP function of MCM was suppressed in the E154A, Q160A, and Q160A/E162A mutants compared with the ϳ100-fold rate enhancement observed with wild-type MeaB. The S161A mutation had a negligible impact on the intrinsic GTPase rate, which was activated in the presence of MCM. The E162A mutation led to a modest reduction in the intrinsic and GAP-stimulated rates of GTP hydrolysis but resulted in the same -fold GAP activation as seen for wild-type MeaB. The T155A mutant exhibited a modest 2.5-fold increase in the intrinsic GTPase activity and a mild impairment in GAP activation (37versus 100-fold for wild-type MeaB). The Michaelis-Menten analysis yielded a K m value of 0.414 mM for the wild-type MeaB-MCM complex, which was ϳ2-fold higher than the K m value for wildtype MeaB alone (Table 3). Compared with wild-type MeaB, the switch II mutations led to a ϳ0.5-3.5-fold change in the K m value for GTP in isolated MeaB and an ϳ0.5-1.7-fold change in the MCM-MeaB complex.
Switch II Mutations Affect the Rate of Oxidative Inactivation of MCM-In the presence of wild-type apo-MeaB, the rate of oxidative inactivation of MCM is 8.8 ϫ 10 Ϫ3 min Ϫ1 (Fig. 4a). Addition of GMPPNP or GDP to the wild-type MCM-MeaB complex decreases the inactivation rate nearly 23-fold (3.9 ϫ 10 Ϫ4 min Ϫ1 ) (2). The 23-fold effect of wild-type MeaB⅐GMPPNP on MCM inactivation is slightly larger than the 15-fold effect that we have reported previously and may be due to buffer differences (2). Although the Q160A (2.8 ϫ 10 Ϫ4 min Ϫ1 ) and E162A (2.8 ϫ 10 Ϫ4 min Ϫ1 ) mutants do not appreciably impact the rate of MCM inactivation, E154A (2.0 ϫ 10 Ϫ4 min Ϫ1 ) and Q160A/E162A (2.1 ϫ 10 Ϫ4 min Ϫ1 ) lead to slightly enhanced protection. The S161A (4.4 ϫ 10 Ϫ3 min Ϫ1 ) and   T155A (6.8 ϫ 10 Ϫ3 min Ϫ1 ) mutations exhibit 11-and 17-fold higher rates of inactivation, respectively, than wild-type MeaB. Impaired GTPase-dependent AdoCbl Transfer from ATR to MCM by Switch II Mutants-The M. extorquens ATR binds 2 eq of AdoCbl per homotrimer, and binding of ATP to the vacant site initiates the transfer of a single equivalent of AdoCbl to the MCM-MeaB complex (3). In the presence of GMPPNP, AdoCbl transfer to the wild-type MCM-MeaB complex is blocked. Instead, 1 eq of the cofactor is released from ATR into solution (Fig. 4b). With GMPPNP bound to the MCM-E154A MeaB mutant complex, AdoCbl transfer becomes modestly uncoupled from nucleotide hydrolysis (0.82 eq released and 0.18 eq transferred). Cofactor transfer from ATR to the complexes of MCM and T155A, Q160A, S161A, or E162A MeaB in the presence of GMPPNP (0.46 -0.52 eq of AdoCbl is transferred) is more significantly impaired. The Q160A/E162A double mutation severely impacts GTPase gating of AdoCbl transfer from ATR, and 0.9 eq of the cofactor was transferred instead of being released into solution in the presence of GMPNP.
Rescue of MCM by Switch II Mutants-Our previous studies have shown that MeaB promotes expulsion of cob(II)alamin that has become uncoupled from 5Ј-deoxyadenosine during turnover in MCM (3). Here, we examined the role of switch II in exerting this chaperone function of MeaB. In the presence of stoichiometric cob(II)alamin (with respect to MCM) and an excess of each switch II mutant of MeaB, cob(II)alamin is bound to MCM. Following addition of an excess of GMPPNP to the wild-type MeaB-MCM⅐cob(II)alamin complex, ϳ97% of the inactive cob(II)alamin is detected in the filtrate, indicating that the inactive cofactor was ejected from the MCM active site (Fig. 4c). In contrast, in the presence of the MeaB mutant E154A, T155A, Q160A, or S161A, ϳ55-65% of cob(II)alamin remained associated with MCM. In the MCM-MeaB complex harboring the E162A single or the Q160A/E162A double mutation, ϳ80% of cob(II)alamin remained bound to MCM following addition of GMPPNP.

DISCUSSION
Although the roles of switch motifs in the catalytic and signaling mechanisms of many G-protein have been studied quite extensively (14,20,21,47,48), their role in cofactor delivery is not as well characterized. Unlike the structures of CooC and NifH, the crystal structures of MeaB obtained in the presence of P i , GMPPNP, and GDP (crystallized in the presence and absence of AlF x Ϫ ) have failed to capture a catalytically active GTPase conformation. This is also reflected in the absence of Mg 2ϩ in the active site of all available MeaB crystal structures, raising questions about the possible significance of these inactive conformations to MeaB function. In this study, we examined the role of the conserved switch II residues in suppressing the intrinsic GTPase activity of MeaB in the absence of MCM.
In HypB and other G3E GTPases, the general base that activates water is a conserved aspartate residue (Asp-69 in HypB; Fig. 1b) positioned N-terminally to the switch I sequence (26). The structure of HypB⅐GTP␥S shows Asp-69 positioned ϳ5-6 Å from the ␥-phosphinothioyl group of GTP␥S and interacting with water molecules, one of which presumably serves as the nucleophile. The residue corresponding to Asp-69 in HypB is Asp-92 in MeaB. The MeaB structures containing GDP (crystallized in the presence of AlF x Ϫ or GMPPNP) demonstrate that hydrophobic side chains of Val-156 and Val-158 in switch II impede access of Asp-92 to the nucleotide (Fig. 5, a-c). We speculate that the interaction between Glu-154 and Arg-108 is important for maintaining an autoinhibitory conformation in MeaB that is alleviated in the MCM-MeaB complex. In all MeaB structures with the exception of MeaB⅐GDP (Protein Data Bank code 2QM7), Glu-154 interacts with neighboring residues, e.g. Lys-68 from the P-loop and Arg-108 in switch I (Fig. 5, a and c). Some or all of these interactions are expected to be disrupted if Glu-154 serves as a Mg 2ϩ ligand in the MCM-MeaB complex. The postulated dual role of Glu-154 in autoinhibition and GAP activation is consistent with the ϳ10-fold increase in the intrinsic GTPase activity of the E154A mutant and its relative insensitivity to the GAP activity of MCM.
The similar effect of the E154A, Q160A, and Q160A/E162A mutations on the intrinsic and GAP-stimulated GTPase activities is surprising ( Table 3). The modest impact of the E162A point mutation on the intrinsic and GAP-stimulated GTPase activities suggests that the kinetic phenotype of the Q160A/ E162A double mutant results primarily from the alanine substitution of Gln-160. Gln-160 is not positioned near the active site in any of the MeaB structures, is not conserved in other G3E GTPases (e.g. CooC, NifH, and HypB), and is not predicted to directly participate in catalysis. Rather, Gln-160 is highly mobile and engages in various interactions with residues in the switch II and switch III loops and with a conserved threonine (Thr-163) immediately C-terminal to switch II (Fig. 6). Thus, Gln-160 forms direct or water-mediated hydrogen bonds with Glu-161, Glu-162, Thr-163, or Lys-188. However, within individual subunits of the apo-MeaB (Fig. 6c) and previously reported MeaB⅐GDP (Fig. 6e) structures, Gln-160 is not engaged in hydrogen bonding interactions. Hence, although our kinetic data suggest that Gln-160 is important for autoinhibition, the structures of MeaB do not provide insights into a possible mechanism. The inability of the S161A and E162A mutations to mirror the Q160A phenotype suggests that the individual disruption of their interaction with Gln-160 is insufficient to disrupt the autoinhibited conformation of MeaB.
The switch II residues Thr-155, Ser-161, and Glu-162 are not conserved among G3E GTPases. Furthermore, obvious roles for Thr-155, Ser-161, and Glu-162 in GTP binding or in the GTPase reaction are not apparent from the MeaB structures that are available. The T155A and S161A mutations result in modest perturbations of the GTPase activity of isolated MeaB and in the MCM-MeaB complex (Table 3). These data indicate that neither Thr-155 nor Ser-161 plays a very important role in autoinhibition. In contrast, E162A enhances the affinity of iso-   OCTOBER 25, 2013 • VOLUME 288 • NUMBER 43 lated MeaB for GTP at sites 1 and 2 by 20-and 15-fold, respectively (Table 2). Thus, the effect of the E162A substitution on autoinhibition of the intrinsic GTPase activity might be exerted via enhanced affinity for and stabilization of the substrate, GTP.

Switch II Signaling in MeaB
We have speculated that switches II and III communicate for bidirectional signal transmission between MeaB and MCM (28). Lys-188 in switch III and Glu-162 and Gln-160 in switch II might be important for relaying information about nucleotide identity and hydrolysis between these switch motifs (Fig. 2c). A mutation in CblA corresponding to the Lys-188 residue in MeaB is pathogenic (49). We have shown that mutation of Lys-188 has pleiotropic consequences including loss of regulated GTPase-dependent AdoCbl loading into MCM and impaired repair of inactive MCM. In this study, we demonstrate that single mutations of the interacting partner residues Gln-160 and Glu-162 partially uncouple cofactor docking in MCM from GTP hydrolysis in MeaB (Fig. 4b) and ejection of inactive cofactor from MCM (Fig. 4c). The Q160A/E162A double mutant is more substantially impaired than either single mutant and more closely resembles the phenotype of the K188A/E mutants (28). Similar perturbations in MeaB function, albeit of varying magnitude, are also observed with other switch II mutants, consistent with a role for the entire switch II motif in signaling between MeaB and MCM via switch III.
G-proteins are engaged in a diverse array of regulatory processes (22,50,51). They are among most common and ancient regulatory proteins in nature (14). An increasing body of evidence demonstrates significant variations in the mechanisms of and factors used for GTP hydrolysis and signal transduction that are not readily predicted by sequence and structural homology (52). This complexity is also evident within the G3E metallochaperone family in that MeaB, which possesses an atypical active site, is stabilized in an autoinhibitory conformation via its switch II motif in the absence of the GAP protein MCM.