Protoglobins : novel features from an ancient globin

Very recently a new subfamily of globins, called protoglobins, have been identified in the genome sequences of the Actinobacterium Thermobifida fusca, the green non-sulfur bacterium Chloroflexus aurantiacus, and the two Archaea Aeropyrum pernix (ApPgb) and Methanosarcina acetivorans (MaPgb). Both ApPgb and MaPgb have been partly characterized and demonstrated to conform to the globin sequence motifs, to contain heme, and to bind CO, NO, and O2 in the ferrous form, being very oxygen-sensitive [1]. Their broad ligand range and their sensitivity to oxygen predisposes protoglobins to functioning in low oxygen environments. Thus they appear to be the molecular fossil that thus far, most closely resembles the haemoglobin that was present in organisms possibly as far back as the Last Universal Common Ancestor, or LUCA. A phylogenetic analysis has suggested that protoglobins are the ancestral protein of the globin-coupled sensors (GCSs) [2]. GCSs were first described as regulators of the aerotactic responses in Bacillus subtilis and Halobacterium salinarum (heme-based aerotaxis transducers: HemATs), but also identified in diverse microorganisms that appear to have roles in regulating gene expression. MaPgb and ApPgb are both exactly 195 amino acids long, the same length of the globin domain in archaeal GCS aerotaxis transducer HemATs. The HemATs globin domain, however, retains bound oxygen for much longer periods of time than the protoglobins [3]. This evidence indicates that the protoglobins are not simple equivalents of the liberated GCS globin domains and that there is a fundamental structural and/or functional difference between the two. Here, we report here the first crystal structure of a protoglobin, from Methanosarcina acetivorans, and we analyse how it correlates to the classical globin fold and to its functionally-related globin-coupled sensors. The protein structure MaPgb was solved by means of multiple wavelength anomalous diffraction (MAD), based on the anomalous scattering of the iron atom. The protein model was refined at 1.3 Å resolution using data collected at the remote wavelength during the MAD experiment performed at the ESRF ID23-1 beamline.