We have previously generated microphthalmic mice lacking lens fiber cells by targeting the expression of the diphtheria toxin A (DT-A) gene in transgenic mice with regulatory sequences associated with the mouse gamma 2-crystallin gene. Because of the extreme toxicity of DT to animal cells and the potential leakiness of many tissue-specific regulatory regions, we investigated whether there might be an experimental advantage in using a mutant, attenuated form of the DT-A gene (tox-176) fused to gamma 2-crystallin regulatory sequences to ablate fiber cells in the ocular lens. In contrast to the microphthalmia observed in transgenic animals carrying the native DT-A gene, independent lines of mice transgenic for the gamma 2tox176 construct displayed predominantly cataracts or clinical anophthalmia. These contrasting phenotypes were transmitted within each pedigree, although for some lines some phenotypic heterogeneity among offspring was noted. The difference in phenotype between cataractous and clinically anophthalmic transgenic lines could not be ascribed to differences in the transgene copy number. Instead, the results suggest that transgene expression and hence the extent of genetic ablation are modulated by the site of chromosomal integration and, to a lesser extent, by epigenetic events. They also suggest that the attenuated gamma 2tox176 construct can integrate into chromosomal regions that are particularly favorable for expression without compromising embryological development and therefore that the tox-176 gene may be more versatile and effective than the wild-type DT-A gene for achieving genetic ablation with a broad range of cell- or tissue-specific regulatory sequences.