Several methods have been described in the literature for removal of DNA from protein samples prior to proteome analysis. They in general involve protein precipitation techniques. In other protocols, DNAse treatment is suggested prior to precipitation of proteins in excess acetone. All these methods have been evaluated and found to perform poorly in DNA removal, as illustrated by two-dimensional (2D) maps where horizontal and vertical sample streaking are still substantial. Such removal is in general necessary in tissue lysates and especially when analysing sub-cellular organelles, such as nuclei, where the high DNA levels strongly interfere with proteome analysis. Another method is proposed here for efficient DNA removal: two-phase extraction of DNA in chloroform/phenol/isoamyl alcohol, a procedure commonly used to rid DNA samples of protein contaminants, but rarely applied to protein preparation. This extraction is not very efficient if performed at slightly acidic to neutral pH values, but it performs extremely well at pH values of 9.5 or higher. The 2D maps thus obtained of Escherichia coli lysates as well as extracts from purified nuclei of eukaryotic cells are not only devoid of any vertical or horizontal streaking, but exhibit many more spots, especially in the alkaline region of the 2D gels, suggesting that these basic proteins were in general lost to proteome analysis due to co-precipitation in tenacious protein-DNA complexes. It is hypothesized that the alkaline pH values adopted in the two-phase extraction help to fully disrupt any residual DNA-protein complexes, due to strong Coulombic repulsion.