The N protein coded by bacteriophage λ plays an essential role in the completion of λ transcription by recognizing the boxB sequence in nascent transcripts and then aggregating with Escherichia coli RNA polymerase and four other E. coli proteins into an unstoppable transcription complex. In order to explore the functionality of N protein and the specific recognition between N and boxB, 14 amino acid positions near the amino-terminal end of N(λ) were mutated extensively. The mutant proteins were scored for N function in vivo by a two-plasmid construct that visualizes readthrough transcription as lacZ expression in colonies of E. coli. Mutation was achieved by single TAG replacements, translated through suppression into 13 different amino acids, or by scrambling at assorted three-codon sets. Of the 14 amino acid positions tested (Tables 5 and 6), six remained functional with a wide variety of substitutions, while substitution was sometimes deleterious at one Ala and two Gln positions. At each of the five Arg positions, however, maintenance of Arg occupancy proved important for N function. Despite effective screening for increased N function at boxBs with defective loops, no N mutant, simple or complex, was found to change the order of preference of wild-type N(λ) for boxBs with defective loops. Thus, although multiple amino-terminal Arg positions are found to be important for N function, mutations in the region spanning the five Arg residues were not found to compensate for defects in boxB loop. © 1993 Academic Press Limited.