Reported Translational Bypass in a trpR 0 -lacZ 0 Fusion is Accounted for by Unusual Initiation and 1 Frameshifting Norma M. Wills 1 , Jennifer A. Ingram 1 , Raymond F. Gesteland 1,2 and John F. Atkins 2 * 1 Howard Hughes Medical Institute and 2 Department of Human Genetics, University of Utah Salt Lake City, UT 84112, USA I. Benhar and H. Engelberg-Kulka reported that a 55 nucleotide transla- tional bypass occurs in decoding a fusion of the Escherichia coli trypto- phan repressor, trpR, and lacZ genes. The start of the bypass occurred in the trpR gene and coding resumed in the lacZ gene. It was considered that bypassing likely occurred in expression of trpR itself to produce an additional 10 kDa product which may be biologically important. We report here that bypass is undetectable in the same and related trpR 0 -lacZ 0 fusions. The b-galactosidase activity derived from the fusions is accounted for by unusual internal initiation and 1 frameshifting, both of which occur in the lacZ part of the fusion. The 10 kDa product report- edly encoded by the trpR gene was not detectable to a level of 1% of the full-length 12 kDa tryptophan repressor product, at least when expressed from a T7 promoter. # 1997 Academic Press Limited Keywords: translational bypass; frameshifting; trpR; trp repressor; initiation *Corresponding author Introduction Of several reported cases of translational bypass- ing, two have been extensively investigated: one in decoding T4 gene 60 (Huang et al., 1988) and the other in a trpR 0 -lacZ 0 fusion (Benhar & Engelberg- Kulka, 1993). In T4 gene 60, at least four elements are important for the 50 nt bypass; (1) matched take-off and landing codons, (2) a zero-frame ter- mination codon immediately 3 0 to the take-off site, (3) a stem-loop structure within the coding gap, and (4) a speci®c sequence of 16 amino acids in the nascent polypeptide (Weiss et al., 1990). Translation of the gap is precluded by termination codons in all three reading frames, but recent evidence suggests ribosomal scanning of the gap occurs (F. M. Adamski, B. Moore, J.F.A. & R.F.G., unpub- lished). Although the size of the coding gap in the trpR 0 -lacZ 0 fusion (55 nt) is similar to that of T4 gene 60 (50 nt), there is no reason to assume a com- mon mechanism. None of the features necessary for T4 gene 60 bypass appears to apply in trpR 0 - lacZ 0 bypass, however, there is a requirement for at least ten translatable codons preceding the gap which itself must be translated for bypass to occur (Benhar & Engelberg-Kulka, 1993). The ef®ciency of bypass in the two systems is markedly different; 40 to 70% in T4 gene 60 (Huang et al., 1988; J.F.A. & R.F.G., unpublished) and 5% in trpR 0 -lacZ 0 (Benhar & Engelberg-Kulka, 1993). Bypass in trpR 0 - lacZ 0 has been taken as evidence that bypass occurs in trpR expression (Benhar & Engelberg-Kulka, 1993; Engelberg-Kulka et al., 1993; Engelberg- Kulka & Schoulaker-Schwarz, 1994; Groisman & Engelberg-Kulka, 1995; Engelberg-Kulka & Schoulaker-Schwarz, 1996), but this has not been directly demonstrated. The initial study utilized a chimeric construct with lacZ 0 in the 1 reading frame relative to up- stream trpR 0 (Benhar & Engelberg-Kulka, 1993). The trpR 0 -lacZ 0 fusion contained trpR codons 1 to 78 (nt 387 to 622, see Figure 1b). The transition in reading frame occurred at least 65 amino acids from the amino terminus, a position inaccessible by N-terminal sequencing methods. In order to se- quence the region of interest, a speci®c protease cleavage site for Factor Xa was introduced immedi- ately 5 0 to AUG 65 (Benhar & Engelberg-Kulka, 1991). Protein sequencing of the Factor Xa-cleaved trpR 0 -lacZ 0 fusion protein led to the conclusion that bypass occurred between trpR codon 65, AUG, and lacZ codon 12, UUA (Benhar & Engelberg-Kulka, 1993 and see Figure 1a). The trpR sequences re- Abbreviation used: GST, glutathione-S-transferase. J. Mol. Biol. (1997) 271, 491±498 0022±2836/97/340491±08 $25.00/0/mb971187 # 1997 Academic Press Limited