Induction of the wild type cholera toxin operon (ctxAB) from multicopy clones in Escherichia coli inhibited growth and resulted in low yields of cholera toxin (CT). We found that production of wild type CT or its B subunit (CT-B) as a periplasmic protein was toxic for E. coli, but by replacing the native signal sequences of both CT-A and CT-B with the signal sequence from the B subunit of E. coli heat-labile enterotoxin LTIIb we succeeded for the first time in producing CT holotoxin in high yield in E. coli. Based on these findings, we designed and constructed versatile cloning vectors that use the LTIIb-B signal sequence to direct recombinant native proteins with high efficiency to the periplasm of E. coli. We confirmed the usefulness of these vectors by producing two other secreted recombinant proteins. First, using phoA from E. coli, we demonstrated that alkaline phosphatase activity was 17-fold greater when the LTIIb-B signal sequence was used than when the native leader for alkaline phosphatase was used. Second, using the pspA gene that encodes pneumococcal surface protein A from Streptococcus pneumoniae, we produced a 299-residue amino-terminal fragment of PspA in E. coli in large amounts as a soluble periplasmic protein and showed that it was immunoreactive in Western blots with antibodies against native PspA. The vectors described here will be useful for further studies on structure-function relationships and vaccine development with CT and PspA, and they should be valuable as general tools for delivery of other secretion-competent recombinant proteins to the periplasm in E. coli.