Are sequences of plasmid DNA used in gene therapy erroneous?

To the editor:

Plasmid DNA offers an attractive way to deliver therapeutic genes for gene therapy¹ and genetic immunization². Nucleotide sequences of plasmids are usually considered as composed of only four bases. However, because of its bacterial origin, plasmid DNA contains the two modified bases 6-methyladenine and 5-methylcytosine.

These methylated bases are the consequence of the existence of bacterial DNA modification systems. Dam and Dcm DNA methylases are two of these DNA methyltranferases³. They are found in almost all of the laboratory strains of E. coli and they respectively methylate the adenine residues of all the GATC sequences and the internal cytosine in all of the CC(A/T)GG motifs. Thus, although four bases are used to synthesize DNA, the nucleotide sequence of any plasmid DNA is in fact composed of six bases as a consequence of this epigenetic phenomenon. The omission of these modifications in the reported sequences of plasmid DNA is probably related to the fact that they do not affect the message of the genetic code. However, consequences of the changes introduced by bacterial methylation is far beyond a simple problem of chemical formula.

Methylation of GATC sequences in plasmid DNA has for instance been reported to introduce artificial hormone responsive elements in plasmid DNA⁴ and to destabilize the double helix⁵. On the other hand, spontaneous base substitution hotspots are known to occur at 5-methylcytosine in E. coli and are associated with a CCAGG→CTAGG change⁶. The presence of such foreseeable nonsense or missense mutations in the coding sequence of even a very small percentage of injected plasmids could have a dramatic effect if they give a dominant gain of aberrant function to the corresponding immunogenic or therapeutic protein. In addition, the fact that difference in the CpG dinucleotide methylation status between eucaryotic and procaryotic DNA is involved in immunostimulatory reactions when plasmid DNA is injected in mammals⁷ raises the intriguing possibility that methylated GATC and CC(A/T)GG sequences may also trigger some yet uncharacterized biological reactions in a dose dependant manner.

Thus, because switching plasmid DNA from a tool for molecular biologists to a drug for human therapy requires exact information of the the chemical formula of injected DNA and because omission of the methylated status of nucleotides induces loss of chemical, biophysical, biochemical and potentially biological and legal information, we suggest that all methylated mucleotides are systematically mentioned in any plasmid DNA relevant to human therapy.