Bacterial biotherapeutic delivery vehicles have the potential to treat a variety of diseases. This approach obviates the need to purify the recombinant effector molecule, allows delivery of therapeutics via oral or intranasal administration, and protects the effector molecule during gastrointestinal transit. Lactic acid bacteria have been broadly developed as therapeutic delivery vehicles though risks associated with the colonization of a genetically modified microorganism have so-far not been addressed. Here, we present an engineered strain with reduced colonization potential. We applied a dual-recombineering scheme for efficient barcoding and generated mutants in genes encoding five previously characterized and four uncharacterized putative adhesins. Compared with the wild type, none of the mutants were reduced in their ability to survive gastrointestinal transit in mice. CmbA was identified as a key protein in adhesion to HT-29 and enteroid cells. The nonuple mutant, a single strain with all nine genes encoding adhesins inactivated, had reduced capacity to adhere to enteroid monolayers. The nonuple mutant producing murine IFN-β was equally effective as its wild-type counterpart in mitigating radiation toxicity in mice. Thus, this work established a novel therapeutic delivery platform that lays a foundation for its application in other microbial therapeutic delivery candidates and furthers the progress of the delivery system towards human use.IMPORTANCEOne major advantage to leverage gut microbes that have co-evolved with the vertebrate host is that evolution already has taken care of the difficult task to optimize survival within a complex ecosystem. The availability of the ecological niche will support colonization. However, long-term colonization of a recombinant microbe may not be desirable. Therefore, strategies need to be developed to overcome this potential safety concern. In this work, we developed a single strain in which we inactivated the encoding sortase, and eight genes encoding characterized/putative adhesins. Each individual mutant was characterized for growth and adhesion to epithelial cells. On enteroid cells, the nonuple mutant has a reduced adhesion potential compared with the wild-type strain. In a model of total-body irradiation, the nonuple strain engineered to release murine interferon-β performed comparable to a derivative of the wild-type strain that releases interferon-β. This work is an important step toward the application of recombinant in humans.