The mechanism for electrical charge conduction in DNA has been the subject of much recent interest and debate. Many of the measurements of DNA conductivity have been made in aqueous solution, with an aromatic photooxidant moiety such as anthraquinone or a rhodium(III) complex covalently tethered to the DNA. Such studies, however, have given discrepant results, for instance, regarding the relative ability of AT- and GC-rich sequences to conduct charge and the possibility of thymine cyclobutane dimer repair through the DNA from a distance. A recent paper on conduction in DNA immobile four-way junctions using the rhodium photooxidant reported conduction in all four helical arms, contrary to what is known about the three-dimensional structure and stacking of 4-way junctions. We have reexamined conduction in such junctions using rhodium [Rh(phi)(2)(byp*)Cl-3] as well as the anthraquinone photooxidants, and find that although our rhodium data agree with the previously published work, the anthraquinone data reveal conduction in only two of the four helical arms, consistent with the known tertiary structure of four-way junctions. An electrophoretic investigation revealed the formation of intermolecular aggregates in the rhodium-derivatized junctions, but not in the anthraquinone-labeled junctions. Rhodium-specific aggregation was also observed with simple DNA duplexes under the same experimental conditions. A characteristic property of aggregation was that all participating DNA molecules required the rhodium derivatization, and underivatized molecules did not aggregate with the derivatized ones. It is conceivable that the results reported here will help reconcile the various discrepancies that have been reported from charge conduction experiments carried out on DNA utilizing different photooxidants.