Sequence composition analysis on arsenic-binding proteins in human cells

Summary form only given. Arsenic is shown to participate in many of transduction pathways in cancer cells. However, up to now, the mechanism of protein-arsenic interactions is still remaining unknown. This study aims at investigating whether the sequence composition of arsenic-binding proteins is distinct to that of background distribution. We first collected two sets of potential arsenic-binding proteins in human lung cancer cells and breast cancer, respectively, based on recent studies. These two sets of proteins were identified previously according to different chemical methods and affinity chromatography coupled to mass spectrometry. Eight proteins that are both present in these two lists were deleted from the breast set of binding proteins to avoid redundancy. The frequency of each type of amino acids present in a protein set was calculated. This frequency was divided by the background frequency of that amino acid observed in human proteins. Finally, the logarithms of the ratios were recorded for each list. One hundred sets of 100 randomly selected human proteins were generated to produce a background distribution in order to calculate the z-scores of the derived log ratios. The results show that the compositions calculated based on these two lists of arsenic-binding proteins are quite similar. In addition to the potential arsenic-binding proteins collected from published literatures, we also collected a protein list based on local experiments in Chinese hamster ovary (CHOA) cells. The third set of proteins also concurs that the amino acids with negative (D and E) and positive (K) charges are more frequently observed on arsenic-binding proteins than that in general. This observation deserves further studies, in order to develop computational methods for predicting arsenic-binding proteins efficiently in the future.