DNA-SEQUENCING AND COMPARATIVE SEQUENCE-ANALYSIS REVEAL THAT THE ESCHERICHIA-COLI GENOMIC DNA MAY REPLACE THE TARGET DNA DURING MOLECULAR-CLONING - EVIDENCE FOR THE ERRONEOUS ASSEMBLY OF ESCHERICHIA-COLI DNA INTO DATABASE SEQUENCES

DNA sequencing and similarity search of databases provide experimental evidence that portions of the host Escherichia coli genome may get li gated into the cloning vector, resulting in clones containing nontarge ted inserts. Several lines of evidence suggest that this non-targeted ligation, as observed by us while subcloning troponin I cDNA, is presu mably due to a recombination mediated mechanism by which host DNA repl aces the target DNA in the cloning vector. The E. coli genome mapping to 64-65 min and 92.8-00.1 min, the latter containing insertion sequen ces, appears to be the hotspot regions involved in this process. We ex amined the possibility that some sequences reported in the databases m ay also contain genomic sequences of E. coli. A search of current data bases revealed that a rat hepatic glutathione transporter cDNA contain s a 2.2-kb-long portion of the E. coli genome that has been wrongly as sembled into its 5' untranslated and coding regions. In addition, abou t 30 sequences in databases, including a Yersinia pestis toxin gene, s howed relatively high sequence identity with those portions of the E. coli genome that were present in the nonauthentic clones. (C) 1997 Els evier Science Inc.