Proximal and distal control for ligand binding in neuroglobin: role of the CD loop and evidence for His64 gating

Abstract

Neuroglobin (Ngb) is predominantly expressed in neurons of the central and peripheral nervous systems and it clearly seems to be involved in neuroprotection. Engineering Ngb to observe structural and dynamic alterations associated with perturbation in ligand binding might reveal important structural determinants, and could shed light on key features related to its mechanism of action. Our results highlight the relevance of the CD loop and of Phe106 as distal and proximal controls involved in ligand binding in murine neuroglobin. We observed the effects of individual and combined mutations of the CD loop and Phe106 that conferred to Ngb higher CO binding velocities, which we correlate with the following structural observations: the mutant F106A shows, upon CO binding, a reduced heme sliding hindrance, with the heme present in a peculiar double conformation, whereas in the CD loop mutant "Gly-loop", the original network of interactions between the loop and the heme was abolished, enhancing binding via facilitated gating out of the distal His64. Finally, the double mutant, combining both mutations, showed a synergistic effect on CO binding rates. Resonance Raman spectroscopy and MD simulations support our findings on structural dynamics and heme interactions in wild type and mutated Ngbs.

Figure 1

CO binding to ferrous neuroglobin mutants at 25 °C. CO binding rate constants k_obs were extracted from fitting experimental kinetic traces as single or double exponentials (A) and plotted as a function of CO concentration (B).

Figure 2

Back-bonding correlation line of the ν(Fe-C) and ν(C-O) stretching frequencies of various Ngbs^–,. The corresponding data of sperm whale Mb are also reported^,. The dotted lines indicate the approximate delineation between the frequency zones of the A₀, A₁, and A₃ forms. The human Ngb and swMb show also a third weak H-bonded conformer (A₁) at 505/1956 and 508/1946 cm⁻¹, respectively, not reported in the Table S3.

Figure 3

Comparison of wild type and F106A neuroglobin structures. (A) Unbound (green) and carbo-monoxy (grey) WT Ngb structures^,. (B) Unbound (green) and carbo-monoxy (grey) F106A Ngb structures.

Figure 4

Superposition of Ngb Gly-loop (blue) and Ngb WT (grey) structures. (A) Comparison between the overal fold of the WT and the Gly-loop mutant. The segment missing from electron density maps (Phe49-Ser57) is indicated by a dashed line, the mutated sequence is in red, and the folded D-helix is indicated in dark grey. (B) In WT, the D-helix leans on Phe49, for the Gly-loop mutant, the 2Fo-Fc map is contoured at 1σ for Phe49. (C) Network of interactions between Tyr44, Glu60, Lys67, one of the heme propionates and a water molecule in WT Ngb. In the Gly-loop mutant, this network is abolished due to the absence of Tyr44. (D) The canonical hydrophobic patch in neuroglobin involved in heme support and maintenance of the globin fold is perturbated.

Figure 5

Evidence for His64 swing in Gly-loop structure bound to CO. The electron density indicates the presence of a fraction of swung out His64 (40%, the 2Fo-Fc map is contoured at 1σ). We could reconstruct the CD loop and we observed a re-shaping of the D-helix (from Asp54 to Ser57).

Figure 6

Secondary structure of CD corner at 300 K. The secondary structure was determined by DSSP (hydrogen bond estimation algorithm) on the ensemble of structures, sorted by time in ascending order, at 300 K, for Gly-loop Ngb (A) and WT Ngb (B).