Effects of ethidium bromide,tetramethylethidium bromide and betaine B on the ultrastructure of HeLa cell mitochondria in situ

Summary Several investigators have described the ultrastructural changes that occur in the mitochondria of cells in tissue cultures after treatment with the drug ethidium bromide (E). The mitochondria swell and the cristae become greatly altered and finally disappear; in the cristae-free region of the matrix electron-dense granules can be observed. It has been assumed that intercalation of E between the base pairs of the mitochondrial DNA induces the formation of the granular inclusions. To investigate whether intercalation is really the initial step in the generation of dense granules inside the matrix, we performed a comparative incubation study of HeLa-cell mitochondria in situ using three closely related dyes (D), i.e. E, tetramethylethidium bromide (TME) and betaine B (B). They strongly differ with regard to their affinity for DNA and their ability to cross membranes. E was used as a reference dye. TME does not intercalate, but is externally bound to DNA only weakly. The neutral B is not bound at all, but can cross membranes more easily than the cation E. Moreover, in aqueous solutions at pH7.0, B is in equilibrium with its protonated cation BH. BH and E have almost equal affinities for DNA. Therefore B may quickly pass the inner mitochondrial membranes and the cristae, and should then be bound inside the matrix, thus forming a BH-DNA complex. On the assumption that intercalation is necessary for the generation of intramitochondrial electron-dense bodies, we predicted that BH/B should be more efficient than E, while TME should be relatively ineffective. In experiments using HeLa cells, these predictions were found to be inaccurate. E, TME and BH/B produced almost the same mitochondrial alterations, but at different concentrations and after different incubation periods. In contrast to our expectations TME was much more effective than E and BH/B, with the last two behaving rather similarly.Therefore, it seems unlikely that the drugs penetrate the inner mitochondrial membrane system by simple physical diffusion or that intercalation is the preliminary step for the generation of dense granules inside the matrix. Instead, we assume that hydrophobic interaction between the dye cations E, BH and TME and the cristae is the main cause of the mitochondrial changes. The favoured binding partner of the dye cations may be the divalent anion, cardiolipin: this phospholipid is an essential part of the inner membrane system but is absent in other membranes of cells. By distributing the dyes between a lipophilic phase and water, it was shown that TME is more lipophilic than E and BH; this may explain the greater effectiveness of TME. The bound dye cations disturb the organization of the cristae, which become altered and finally disappear. We assume that the electron-dense granules in the matrix are mainly composed of the dyes and former membrane materials such as phospholipids and proteins, as well as perhaps some other hydrophobic matrix materials. This would also explain why it was impossible to digest the dense granules by DNase treatment. The drugs enter the mitochondrial matrix by disordering and finally destroying the cristae.