A G-protein editor gates coenzyme B12 loading and is corrupted in methylmalonic aciduria

Abstract

The mechanism by which docking fidelity is achieved for the multitude of cofactor-dependent enzymes is poorly understood. In this study, we demonstrate that delivery of coenzyme B(12) or 5'-deoxyadenosylcobalamin by adenosyltransferase to methylmalonyl-CoA mutase is gated by a small G protein, MeaB. While the GTP-binding energy is needed for the editing function; that is, to discriminate between active and inactive cofactor forms, the chemical energy of GTP hydrolysis is required for gating cofactor transfer. The G protein chaperone also exerts its editing function during turnover by using the binding energy of GTP to elicit release of inactive cofactor that is occasionally formed during the catalytic cycle of MCM. The physiological relevance of this mechanism is demonstrated by a patient mutation in methylmalonyl-CoA mutase that does not impair the activity of this enzyme per se but corrupts both the fidelity of the cofactor-loading process and the ejection of inactive cofactor that forms occasionally during catalysis. Consequently, cofactor in the incorrect oxidation state gains access to the mutase active site and is not released if generated during catalysis, leading, respectively, to assembly and accumulation of inactive enzyme and resulting in methylmalonic aciduria.

Fig. 1.

Schematic representation of AdoCbl conformations in ATR and MCM and their principal spectral features.

Fig. 2.

Gating of coenzyme B₁₂ transfer. (A) ATP-dependent AdoCbl release/transfer from holo-ATR (8 μM final concentration) in the presence of 100 μM MCM:MeaB·GTP (●) or MCM:MeaB·GMPPNP (■) in Buffer A containing 15 mM MgCl₂ at 20 °C. After each addition of ATP, MCM activity was assayed using the radiolabeled assay (○,MCM:MeaB·GTP; □, MCM:MeaB·GMPPNP). (B) MeaB inhibits ATP-dependent AdoCbl transfer from holo-ATR to wild-type but not to R585C MCM. Cofactor transfer from 10–12 μM holo-ATR to varying molar ratios of [MeaB]/[MCM] ± 2 mM nucleotides. The amount of cofactor transferred from holo-ATR to wild-type or R585C MCM in the absence of MeaB was 38–48%. (C) Enzyme-monitored turnover of MCM loaded with AdoCbl from holo-ATR in the presence of MeaB·nucleotide. The spectrum of 10 μM holo-ATR in the presence of a 10-fold molar excess of apo-MCM:MeaB·GTP (black) and after addition of 7 mM ATP (gray) and 7.2 mM (R and S)-methylmalonyl-CoA (dashed). The shift in the absorption maxima from 458 nm (black) to 525 nm (gray) corresponds to the conversion of base-off AdoCbl·ATR to base-off/His-on AdoCbl·MCM. Addition of substrate results in the accumulation of cob(II)alamin (increase at 474 nm and a decrease at 525 nm). (Inset) Quantification of cob(II)alamin present during turnover when MCM:MeaB·nucleotide (52–184 μM, with 2.5 mM GTP or GMPPNP) was reconstituted with holo-ATR (19–20 μM in AdoCbl) (n = 4).

Fig. 3.

Editing function of MeaB during cofactor assembly. EPR spectroscopic monitoring of cob(II)alamin transfer from ATR (40 μM in cob(II)alamin, A and B, i) to 1.6 equivalent MCM (A, ii) or 1.2 equivalent R585C MCM (B, ii) in the presence of MeaB (A and B, ii), or MeaB with 2.5 mM GDP (A and B, iii), GTP (A and B, iv), or GMPPNP (A and B, v). For the mutant, a molar ratio [MeaB]/[R585C MCM] of 2.1 was used in comparison to a stoichiometric ratio for wild-type MCM. The EPR spectra are the average of two scans.

Fig. 4.

MeaB rescues inactive MCM. (A) EPR spectra of 50 μM cob(II)alamin bound to MCM in the presence of 1.5 molar excess of MeaB (i), MeaB·GDP (ii), MeaB·GTP (iii) or MeaB·GMPPNP (iv), and free cob(II)alamin (v) in Buffer A. The asterisks point to the hyperfine structures in the absorption feature when cob(II)alamin is bound to MCM but not when it is in solution. (B) Comparison of the percentage of cofactor remaining bound to MCM following addition of 1.5 (to wild-type MCM) or 2.5 (to R585C MCM) molar excess of MeaB ± the indicated nucleotides. (C) Comparison of the percentage of cofactor remaining bound to MCM in the presence of 7.8 mM adenosine or 9 mM 5′-deoxyadenosine, and 14 mM (R and S)-methylmalonyl-CoA and a 1.5 molar excess of MeaB ± the indicated nucleotides.

Fig. 5.

Model for gating and editing of cofactor transfer and its corruption by a pathogenic mutation. (A) Location of the arginine residue (corresponding to R616 in the human and R585 in the M. extorquens sequences) in the crystal structure P. shermanii MCM (PDB: 1REQ). AdoCbl (red) and the arginine residue of interest (navy) are shown in ball representation. The two subunits of MCM are differentiated by shades of blue. (B) Dependence on GTP concentration of the rate of GTP hydrolysis by MeaB (□) MeaB:MCM (●) or MeaB:R585C MCM (○) in Buffer A at 37 °C. Michaelis-Menten analysis of the rate of GTP hydrolysis yielded the kinetic parameters reported in Table 1. (C) Schematic model for the gating and editing functions of MeaB.