Spectral Deconvolution

To determine if excess DCNh was present in solution before saturation of the active apoHb heme-binding pockets during the apoHb titration assay, the concentration of each chemical species following incremental addition of DCNh was determined using a spectral deconvolution algorithm developed in MATLAB (Mathworks, Natick, MA). The analysis was performed in 10 mm quartz cuvettes with 47.6. μM apoHb in triplicate. This concentration maintained the 500–600 nm absorbance readings between 0.1 to 1 AU so that the spectra of DCNh and rHbCN could be differentiated. The absorbance spectrum for each of the pure species was obtained from 450 to 700 nm. For comparison to the 420 nm absorbance plots, the solutions were further diluted by 7× so the absorbance at 420 nm ranged between 0.1 and 1 AU. With the 420 nm absorbance data, the DCNh titration analysis was performed as previously described in this study. To estimate the molar fraction of each species in each DCNh apoHb solution, the MATLAB algorithm squared the difference between the Fourier transforms of the raw absorbance spectra between 510 nm - 585 nm for each sample with a linear combination of Fourier transforms of the spectra of pure DCNh or HbCN. The linear combination of the pure species’ spectra with the lowest error was used to estimate each species’ concentration in solution. The wavelength range was selected based on two criteria: (1) all absorbance values ranged from 0.1 to 1 AU, and (2) regions where potassium ferricyanide does not absorb⁶⁶.