The preparation of high quality plasmid DNA is a necessary requirement for most molecular biology applications. We compared four different large plasmid preparation protocols, which were based on either a liquid-phase approach (Triton lysis) or purification of alkaline lysis bacterial extracts followed by supercoiled plasmid purification on affinity columns. Two host Escherichia coli strains, JM 109 and INValphaF', were used to grow the test plasmids for comparison of product plasmid DNA produced from the four different plasmid isolation methods. While the DNA grown in E. coli strain JM109, prepared by liquid-phase Triton lysis was appropriately restricted by 12 restriction enzymes, this was not the case for any of the JM109-grown DNA purified by any of the affinity column solid-phase approaches. In contrast to this, when the plasmid DNA was grown in E. coli strain INValphaF', most restriction enzymes cut DNA appropriately, irregardless of the plasmid preparation protocol used. It seems that an impurity commonly eluted with the DNA from all three of the solid-phase DNA columns had an equal effect on the above enzymes using the common host strain JM109, but not strain INValphaF'.