Antibiotics are used in medicine by their discriminization between bacterial and eukaryotis cells (e.g. human) meaning that a bacterial infection in humans can be treated without affecting the normal body cells. If mitochondria would be derived from bacteria, we would expect an effect of antibiotics on eukaryotic cells where mitochondria are essential for metabolism. So, that would be more evidence against the use of the antibiotics argument.
Still, I often encounter antibiotics as an argument in favour of the endosymbiotic theory, e.g. see here. For instance erythromycin has been mentioned, but it only affects yeast (here) and can be used in humans without severe side-effects (here). This is not the case with tetracycline, which may has side-effects in some individuals by also inhibiting the mitochondrial 30S ribosomal RNA. but an be used generally in humans (here). It should be realized that drugs work through binding sites in a three-dimensional protein, and related proteins with similar functions (as the 30S RNA) may have similar binding sites. Tetracycline also is not a good example because it still can be used in humans acting as a specific bacterial antibiotic. In some cases, a hereditary mutation can cause hypersensitivity to streptomycin and induce deafness (here).
In all, evidence for the endosymbiotic theory coming from antibiotics is anecdotal and easily explained without an endosymbiotic event.