Conformational plasticity surrounding the active site of NADH oxidase from Thermus thermophilus

Abstract

Biotechnological applications of enzymes can involve the use of these molecules under nonphysiological conditions. Thus, it is of interest to understand how environmental variables affect protein structure and dynamics and how this ultimately modulates enzyme function. NADH oxidase (NOX) from Thermus thermophilus exemplifies how enzyme activity can be tuned by reaction conditions, such as temperature, cofactor substitution, and the addition of cosolutes. This enzyme catalyzes the oxidation of reduced NAD(P)H to NAD(P)(+) with the concurrent reduction of O2 to H2O2, with relevance to biosensing applications. It is thermophilic, with an optimum temperature of approximately 65°C and sevenfold lower activity at 25°C. Moderate concentrations (≈1M) of urea and other chaotropes increase NOX activity by up to a factor of 2.5 at room temperature. Furthermore, it is a flavoprotein that accepts either FMN or the much larger FAD as cofactor. We have used nuclear magnetic resonance (NMR) titration and (15)N spin relaxation experiments together with isothermal titration calorimetry to study how NOX structure and dynamics are affected by changes in temperature, the addition of urea and the substitution of the FMN cofactor with FAD. The majority of signals from NOX are quite insensitive to changes in temperature, cosolute addition, and cofactor substitution. However, a small cluster of residues surrounding the active site shows significant changes. These residues are implicated in coupling changes in the solution conditions of the enzyme to changes in catalytic activity.

Keywords: NMR titration; enzyme cofactor substitution; isothermal titration calorimetry; nuclear magnetic resonance; spin relaxation; thermophile; urea activation.

Figure 1

NOX cofactor structures. flavin adenine dinucleotide (FAD) (a) and flavin mononucleotide (FMN) (b).

Figure 2

Indole side-chain regions of NOX ¹H/¹⁵N NMR spectra. Wild-type NOX (a), and the mutants W47F (b), W131F (c), and W204F (d) recorded at 50°C and 18.8 T (800 MHz ¹H. Larmor frequency) in the absence of cofactor.

Figure 3

NOX cofactor titrations with FMN (a, b) and FAD (c–f). Raw differential power ITC isotherms for NOX with FMN (a) and FAD (c). Integrated ITC injection heats plotted as a function of the cofactor:monomer molar ratio for FMN (b) and FAD (d). Data were fit to a 1 set of identical sites binding model. FAD injections into buffer alone are shown in the inset to panel (e). NOX ¹H (e) and ¹⁵N (f) chemical shift displacements plotted as a function of the ligand:protein molar ratio. Lines correspond to fits according to Eq. 2.

Figure 4

NOX chemical shift perturbations by environmental variables. Backbone ribbon diagrams of NOX (PDB 1NOX), with the two subunits colored in blue and white and the FMN cofactor illustrated in green balls and sticks. In (a), dark purple spheres correspond to residues whose ¹H/¹⁵N peaks are missing in spectra of apo NOX but not in the presence of FMN or FAD. In (b–d) spheres correspond to residues with larger than average ¹H/¹⁵N chemical shift perturbations (ppm) in response to changing environmental variables: FMN vs. FAD [>0.25 dark purple; 0.17–0.25 purple; 0.11–0.17 dark pink; 0.10–0.11 light pink] (b) 25–50°C [|∂δ_N/∂T| > 17 ppb/K dark pink; ∂δ_H/∂T < −4.5 ppb/K light pink; both [|∂δ_N/∂T| > 17 ppb/K, and ∂δ_H/∂T < −4.5 ppb/K dark purple] (c) and 0 vs. 2M urea [>0.55 dark purple; 0.45–0.45 purple; 0.35–0.45 dark pink; 0.32–0.35 light pink] (d). Small white spheres on the white monomer indicate the locations of residues lacking crosspeaks in ¹H/¹⁵N correlation NMR spectra. Malleable residues listed in Table2 are labeled on the structures. Image generated with UCSF chimera.

Figure 5

Effect of temperature on NOX NMR spectra. Overlay of subregion of NOX ¹H–¹⁵N HSQC correlation spectra obtained over temperatures spanning 25°C (blue) to 50°C (red) (a) representative temperature-dependent changes in ¹H (b) and ¹⁵N (c) chemical shifts obtained for the backbone amide resonance of Val 7. Temperature coefficients for ¹H (d) and ¹⁵N (e) plotted as a function of residue number. ¹H temperature coefficients more negative than −4.5 ppb/K and ¹⁵N temperature coefficients greater in magnitude than 177 ppb/K are indicated by black bars. Secondary structure elements were taken as αA (10–16), αB (30–42), β1 (52–57), αC (60–69), αD (74–78), β2 (81–87), αE (89–98), αF (107–121), αG (125–149), β3 (153–157), αH (162–169), β4 (179–184), αJ (198–201).

Figure 6

Effect of urea on fast timescale motions in NOX. Lipari-Szabo order parameters, S², reflecting motions on the ns–ps timescale, obtained for NOX in the presence (black) and absence (gray) of 1.2M urea, plotted as a function of residue number.