Contrasting expression of KAL in cell-free systems: 5' UTR and coding region structural effects on translation.

We investigated the expression of two different X-linked Kallmann (KAL) gene cDNAs in two different cell-free systems using rabbit reticulocyte lysate: (system A) transcription/translation coupled and (system B) noncoupled. System A yielded a single band of 76 kDa corresponding to anosmin-1, the expected full-length gene product, and upon addition of canine microsomal membranes produced a 85-kDa glycosylated form. System B did not produce any detectable protein band despite the expression of a beta-galactosidase-positive control gene. The first 179 bases of the coding sequence are 74% GC-rich and showed the potential to form imperfect hairpin structures, which in part may explain the translation inhibition of KAL in system B. This has further led us to speculate that coupling transcription to translation may either be preventing translating-inhibiting hairpin formation or be compensating for the lack of certain tissue-specific proteins in reticulocyte lysate that are essential in overcoming inhibitory hairpins during translation. Substitution of the 5'-UTR with an encephalomyocarditis virus internal ribosomal entry site (EMCV IRES) sequence resulted paradoxically in a lower yield of anosmin-1, suggesting that elements in the 5'UTR may be necessary for maintaining a "normal" level of expression. The use of KAL and luciferase reporters (containing different 5'UTRs) demonstrated that the native KAL 5' UTR is not involved in translational efficiency. However, this sequence may influence faithful translation initiation. Theoretical RNA conformation data imply that effective EMCV IRES usage with KAL may require favorable pairing between the IRES and unidentified sequences within the 5' coding region of the gene. This work provides a foundation both for the investigation of KAL regulation and for the characterization of its function.