Uneven distribution of plasmid-based expression vectors to daughter cells during bacterial cell division results in an increasing proportion of plasmid free cells during growth. This is a major industrial problem leading to reduction of product yields and increased production costs during large-scale cultivation of vector-carrying bacteria. For this reason, a selection must be provided that kills the plasmid free cells. The most conventional method to obtain this desired selection is to insert some gene for antibiotic resistance in the plasmid and then grow the bacteria in the presence of the corresponding antibiotic. We describe here a host/plasmid Escherichia coli system with a totally stable plasmid that can be maintained without the use of antibiotic selection. The plasmid is maintained, since it carries the small essential gene infA (coding for translation initiation factor 1, IF1) in an E. coli strain that has been deleted for its chromosomal infA gene. As a result only plasmid carrying cells can grow, making the strain totally dependent on the maintenance of the plasmid. A selection based on antibiotics is thus not necessary during cultivation, and no antibiotic-resistance genes are present neither in the final strain nor in the final plasmid. Plasmid-free cells do not accumulate even after an extended period of continuous growth. Growth rates of the control and the plasmid harboring strains are indistinguishable from each other in both LB and defined media. The indicated approach can be used to modify existing production strains and plasmids to the described concept. The infA based plasmid stability system should eliminate industrial cultivation problems caused by the loss of expression vector and use of antibiotics in the cultivation medium. Also environmental problems caused by release of antibiotics and antibiotic resistance genes, that potentially can give horizontal gene transfer between bacterial populations, are eliminated.