The H-NOX protein structure adapts to different mechanisms in sensors interacting with nitric oxide

Abstract

Some classes of bacteria within phyla possess protein sensors identified as homologous to the heme domain of soluble guanylate cyclase, the mammalian NO-receptor. Named H-NOX domain (Heme-Nitric Oxide or OXygen-binding), their heme binds nitric oxide (NO) and O₂ for some of them. The signaling pathways where these proteins act as NO or O₂ sensors appear various and are fully established for only some species. Here, we investigated the reactivity of H-NOX from bacterial species toward NO with a mechanistic point of view using time-resolved spectroscopy. The present data show that H-NOXs modulate the dynamics of NO as a function of temperature, but in different ranges, changing its affinity by changing the probability of NO rebinding after dissociation in the picosecond time scale. This fundamental mechanism provides a means to adapt the heme structural response to the environment. In one particular H-NOX sensor the heme distortion induced by NO binding is relaxed in an ultrafast manner (∼15 ps) after NO dissociation, contrarily to other H-NOX proteins, providing another sensing mechanism through the H-NOX domain. Overall, our study links molecular dynamics with functional mechanism and adaptation.

Fig. 1. The H-NOX domain. (A) Overall shape of the human dimeric soluble guanylate cyclase in its activated state (PDB file 7D9S). Grey represents the α1 subunit, blue the β1 subunit and dark blue the H-NOX domain. (B) Tertiary structure of the homologous H-NOX protein from the bacteria Caldanaerobacter tengcongensis (PDB file 5JRV). (C) Close-up of the NO-bound distorted heme of Ct H-NOX. (D) Distortion of the NO-bound heme of Ct H-NOX (blue) is compared with the nitrosylated heme of myoglobin (PDB file 2FRK).

Fig. 2. Temperature dependence of the electronic absorption spectra of NO-liganded Cb H-NOX (A), Np H-NOX (B) and Ct H-NOX (C). [NO] = 200 μM. Note the difference in the temperature ranges. The spectra of all unliganded species were recorded at 20 °C. The Q-bands are displayed in ESI, Fig. S2. (D) Heme coordination states involved with corresponding Soret band wavelength.

Fig. 3. Time-resolved absorption data for NO interacting with Cb H-NOX at T = 4 °C and 40 °C. The data at 20 °C can be seen in ESI, Fig. S3. (A and D) Transient spectra at selected time delays after photodissociation of NO. (B and E) Spectral components from SVD analysis of the time-wavelength matrix. (C and F) SVD kinetic components describing the evolution of spectral components with fitted parameters in ESI Table S1.

Fig. 4. Spectral components of NO geminate rebinding obtained from SVD analysis of the matrix of transient spectra for Cb H-NOX (A) and Np (B) H-NOX sensors at different temperatures compared to bovine (Bos taurus) soluble guanylate cyclase (C) whose transient spectrum does not change as T is raised. Analysis of the full data set for Np is shown in ESI, Fig. S7 and Table S2.

Fig. 5. Time-resolved absorption data for NO interacting with Ct H-NOX at two different temperatures. (A and D) Transient spectra at selected time delays after NO photodissociation at 20 °C and 74 °C, respectively. (B and E) Spectral components from SVD analysis of the time-wavelength matrices data. (C and F) Associated kinetic components from SVD analysis. Parameters of the SVD kinetics fit in Table S3.

Fig. 6. (A) Transient spectra of the photo-excited 5c-His ferrous Ct H-NOX in the absence of NO. (B) Kinetics at particular wavelengths with fitted time constants of excited states decay. (C) Calculated spectrum of the relaxed heme 200 ps after photo-excitation (red) compared with the transient spectrum of the unliganded heme before excitation (−3 ps, orange) and steady-state spectrum of same sample (green). The sample was degased and placed in pure argon. The shift corresponds to ΔE = 214 cm⁻¹.

Fig. 7. Resonance Raman spectra of unliganded ferrous H-NOX and its corresponding nitrosylated analogs: (a) Np H-NOX deoxy; (b) Np H-NOX + NO; (c) Ct H-NOX deoxy; (d) Ct H-NOX + NO. Spectra were recorded with weak CW excitation at 442 nm. Raman bands assignment is given in ESI, Table S5.T = 20 °C. [NO] = 200 μM. Enlarged view of the low frequency range in ESI, Fig. S15.

Fig. 8. Transient resonance Raman in low-frequency range compared to steady-state spectra of Ct H-NOX: (a) stationary Ct H-NOX deoxy species with CW excitation at 442 nm; (b) stationary Ct H-NOX deoxy species with sub-ps Raman excitation at 435 nm; (c) transient nitrosylated Ct H-NOX + NO with sub-ps pumping at 565 nm and probing at 435 nm, with time delay Δt = +2 ps between pump and probe pulses; (d) stationary nitrosylated Ct H-NOX with excitation at 442 nm. T = 20 °C.

Fig. 9. (A) Coordination transitions of the heme in Cb and Np H-NOX sensors. The green values in parenthesis refer to the maximum of the Soret absorption band for Np H-NOX. The time constants are: τ_His, rebinding of the proximal His to Fe²⁺; τ_G1 and τ_G2, geminate rebinding of NO to Fe²⁺. Importantly, the temperature range ΔT is different for Cb and Np species. (B) Allosteric equilibrium model incorporating heme distortion in the Ct H-NOX sensor and the time constants in Table 2. k_on is estimated to be diffusion limited (∼10⁸ M⁻¹ s⁻¹).