Extended Data Figure 1: Catabolic pathway featuring HODM and additional HODM characterization in vivo and in vitro.
From: An oxidative N-demethylase reveals PAS transition from ubiquitous sensor to enzyme
a, Trimethylamine degradation pathway with links to C1 metabolism in proteobacteria. 1, trimethylamine monooxygenase; 2, trimethylamine N-oxide demethylase; 3, haem-dependent oxidative demethylase (HODM), also known as secondary amine monooxygenase (SAMO); 4, methylamine dehydrogenase (EC 1.4.99.3); 5, methylamine oxidase (EC 1.4.3.21); 6, N-methylglutamate synthase (EC 2.1.1.21); 7, N-methylglutamate dehydrogenase (EC 1.5.99.5); 8, γ-glutamylmethylamide synthetase (EC 6.3.4.12); 9, γ- glutamylmethylamide-dissimilating enzyme; 10, methylene-H4F dehydrogenase/cyclohydrolase (EC 1.5.1.15); 11, formyl-H4F deformylase/ formyl-H4F synthetase (EC 3.5.1.10); 12, formate dehydrogenase (EC 1.2.1.2). Formaldehyde can be assimilated via the ribulose monophosphate (RuMP) pathway or, via methylene-tetrahydrofolate, through formation of serine. The inset shows an SDS–PAGE gel of purified recombinant HODM, revealing bands corresponding to all four subunits. Additional bands are visible that appear to correspond to multimeric forms or proteolytic fragments of HODM subunits as verified by mass spectrometry of tryptic digests. b, Formaldehyde leakage from HODM in vivo. Efficient detoxification of formaldehyde requires the γ subunit as well as tetrahydrofolic acid (THF). Cells containing wild-type and mutant HODM enzyme (Δγ, deletion of subunit gamma) from Methylobacillus flagellatus (strain KT) were grown in the presence of DMA (grey bars) or DMA and glycine (shaded bars), respectively. Fluorescence readings were subsequently corrected for background activity and scaled with the mutant enzyme set as 100%. The data show that glycine, as well as the γ subunit, have strong effects on enzymatic formaldehyde production. While the first increases intracellular levels of THF, the latter is required for THF binding. c, The HODM ferrous-oxy complex is long-lived at room temperature. The ferrous-oxy decay to the ferric state was monitored at 409 nm and plotted as a function of time. The decay curve observed was fitted using the exponential decay equation to derive a half-life of 50 ± 10 min with a decay rate of k = 0.014 ± 0.001 min−1. d, Determination of the redox potential of the HODM haem subunit. The main panel shows UV-visible spectra for the HODM haem domain during a redox titration. The oxidized enzyme (thick black line, spectrum recorded at −76 mV versus normal hydrogen electrode) has its Soret maximum at 409 nm. The fully reduced enzyme (thick line, spectrum recorded at −317 mV versus NHE) has its Soret maximum at 422 nm. Intermediate spectra are indicated in thin lines, and there is an isosbestic point at approximately 415 nm. The HODM haem domain also displays increased absorbance in the Q-band region (~520–620 nm) in the reduced (ferrous) state. The inset shows a plot of absorbance change (ΔA422 minus ΔA409) versus applied potential (versus NHE) fitted using the Nernst equation. This provides a midpoint reduction potential value for the HDOM haem domain FeIII/FeII couple of 128 ± 4 mV versus NHE. e, f, DMA binding to ferrous HODM (e) and DMA binding to ferrous-oxy HODM (f). Absorbance changes (as shown in the inset) as a function of DMA concentration are fitted using a hyperbolic function/Morrison equation, leading to a Kd = 1.7 ± 0.2 mM for ferrous HODM (ferric HODM has a Kd of 1.5 ± 0.1 mM, data not shown) and a Kd = 15 ± 3 μM for ferrous-oxy HODM (errors bars are s.e.m., n = 3).
