<jats:p>Resequencing estA3, an allele of the methanol-soluble heat-stable enterotoxin of Escherichia coli showed that the proline triplet 19 is in fact an alanine codon; thus, estA alleles 3 and 4 were shown to be identical. Resequencing has also shown that the carboxy terminus of another allele, estA2, is not the previously inferred histidine triplet but the same tyrosine codon reported for all other estA alleles. The improperly inferred histidine codon was used in constructions to fuse estA2 to the B subunit of the heat-labile enterotoxin gene, and the fused gene products as well as three amino acid insertional mutants containing histidine-72 were not efficiently secreted. We show that the defective secretion is not due to histidine as a carboxy-terminal residue, since site-directed mutagenesis of wild-type tyrosine-72 to histidine did not influence the localization of the activity of the methanol-soluble heat-stable enterotoxin.