Subtraction of atmospheric water contribution in Fourier transform infrared spectroscopy of biological membranes and proteins

Abstract

The contribution of the absorption of H₂O vapor in the IR spectrum of proteins Interferes with the analysis of the shape of amide I and II bands and prevents correct assignments to be performed, in particular after Fourier self-deconvolution of the spectra Mathematical treatments of the spectra have previously been proposed to subtract the water vapor contribution from the sample spectrum. Here we propose to take advantage of the intrinsic bandwidth difference existing between the absorption bands of the water vapor and these of the liquid or solid sample. When a nominal resolution of 8 or 4 cm⁻¹ is chosen, atmospheric water bands are broad and rather featureless. The subtraction coefficient applied may vary by about 50% according to the operator. Conversely, when the spectrum of the same sample in the same conditions is recorded with a nominal resolution of 0.5 cm⁻¹, subtraction coefficients are exactly evaluated by integration and the visual evaluation can not be mistaken by more than 5%. The very sharp features arising from imperfect matching between the atmospheric water band shapes in the reference and in the sample spectrum completely disappear if the difference spectrum is now convoluted to a final resolution of 4 cm⁻¹. Incidence of the subtractions obtained at different resolutions on the evaluation of protein secondary structure is evaluated.