The redox properties of azurin from pseudomonas aeruginosa as studied by high frequency proton NMR

Abstract

Since their 3-dimensional structures have become available [1–3] the blue copper proteins plastocyanin (Pc) and azurin (Az) have gained increased attention from spectroscopists [4]. In both proteins the Cu atom appears to be coordinated by 2 histidines, a cysteine an a methionine, and also in other respects the two structures exhibit a remarkable similarity. Yet in their electron transfer properties and their conformational behaviour as a function of pH Az and Pc show quite notable differences. In Pc the Cu-coordinating His-87 becomes protonated at low pH (pH < 5.4) [1] and moves away from the metal, thus producing a 3-coordinated copper(I) centre. This leads to stabilization of the Cu(I)- state and loss of redox-activity. Moreover, Pc exhibits a large variation in the rate of electron transfer with other proteins (including itself) [5–7], which has been considered as indicative of specific protein-protein interactions. The existence of charged patches on the protein surface is in accordance with the purported electrostatic nature of these interactions [5]. In the case of Az, on the other hand, it was only known, that the protein undergoes a switch from a redox inactive to a redox active state when the pH drops below about pH = 7 [8]; we therefore started a proton NMR study of the redox properties of Az, the results of which are summarized here. By studying the effect of slight oxidation on the NMR spectrum of Az from Pseudomonas aeruginosa the proton signals of the ligand residues His-46, His-117 and Met-121 were identified and subsequently their pH behaviour was studied [9]. As an example the case of Met-121 is shown here. The signal (labelled M 6) of the ϵ-methyl group of this residue at low pH appears at −0.05 ppm from TSS. The unusual upfield shifted resonance position is the result of the combined ring current effects of Phe-15 and His-46. With increasing pH the signal slowly loses intensity and at pH > 8 has disappeared. This was ascribed ▪ tentatively to a weakening of the CuS (Met-121) bond and increased motional freedom of the Met-121 methyl group leading to a broadening of the NMR-signal [9]. Our recent experiments show that when peak M 6 disappears a new singlet appears (labeled M′ 6) at a 0.11 ppm that overlaps with the C γ1-methyl triplet signal of Ile-7 (labeled R 5). The single nature of M′ 6 is demonstrated in the 2-dimensional J-resolved spectrum reproduced in the figure. It is clear therefor, that a conformational change of the protein occurs when the pH is varied, which affects the position of theϵ-methyl group of Met-121 with respect of Phe-12 and/or His-46. Saturation transfer experiments, to be detailed elsewhere, prove that the signals M 6 and M′ 6 are in slow exchange. Further study of the other ligand signals provides additional evidence that Az may exist in two conformations which interconvert on a time scale of 10–100 ms and that the residues His-35, His-46 and Met-121 are involved in the interconversion. On combining the crystallographic and the NMR data, it transpires that the function of His-35 probably is that of a pH-dependent relay between the Cu coordination shell and the protein. In the course of the oxidation experiments it became apparent that the broadening of the NMR signals induced by the paramagnetism of the Cu(II) species would yield information about the rate of electron self exchange of Az. A thorough analysis of the experimental data showed that, for a few NMR signals, the broadening could be analyzed according to the ‘strong pulse limit’. In this limit the broadening is completely determined by the lifetime of the protein in the diamagnetic state and this leads directly to a value of the rate of electron self exchange, k. Although not very accurate (estimated accuracy ±50%) the value of 2 × 10 6 M −1 s −1 found at 50 °C for k in this way [10] clearly is in agreement with the self exchange rate inferred by Wherland and Pecht from the rate of electron transfer between Az and a variety of other redox proteins [6], though not with the data calculated by Gray an coworkers on the basis of a Marcus treatment of the heterogeneous electron transfer between Az and a series of inorganic transition metal compounds [11]. The high rate of self exchange of Az as well as the relatively fast electron exchange between Az and other redox proteins seems to indicates that nonspecific hydrophobic interactions govern the reaction of Az with its reaction partners. This is consistent with the findings from Cr titration experiments [12] and with conclusions from the crystallographic work [2], that there are no pronounced charged patches on the Az surface.