Phylogenetic Analysis of the Apolipoprotein B mRNA-editing Region : EVIDENCE FOR A SECONDARY STRUCTURE BETWEEN THE MOORING SEQUENCE AND THE 3′ EFFICIENCY ELEMENT

An algorithm and a computer program have been prepared for determining RNA secondary structures within any prescribed increment of the computed global minimum free energy. The mathematical problem of determining how well defined a minimum energy folding is can now be solved. All predicted base pairs that can participate in suboptimal structures may be displayed and analyzed graphically. Representative suboptimal foldings are generated by selecting these base pairs one at a time and computing the best foldings that contain them. A distance criterion that ensures that no two structures are "too close" is used to avoid multiple generation of similar structures. Thermodynamic parameters, including free-energy increments for single-base stacking at the ends of helices and for terminal mismatched pairs in interior and hairpin loops, are incorporated into the underlying folding model of the above algorithm.

Mammalian apolipoprotein B (apo B) exists in two forms, each the product of a single gene. The shorter form, apo B48, arises by posttranscriptional RNA editing whereby cytidine deamination produces a UAA termination codon. A full-length complementary DNA clone encoding an apo B messenger RNA editing protein (REPR) was isolated from rat small intestine. The 229-residue protein contains consensus phosphorylation sites and leucine zipper domains. HepG2 cell extracts acquire editing activity when mixed with REPR from oocyte extracts. REPR is essential for apo B messenger RNA editing, and the isolation and characterization of REPR may lead to the identification of other eukaryotic RNA editing proteins.

Evidence suggests that intestinal apo-B48 is colinear with the amino-terminal half of hepatic apo-B100. To investigate the mechanism of apo-B48 production, we examined cDNA clones from human and rabbit small intestine. All clones contained a single C----T base difference from the hepatic sequence, resulting in a translational stop at codon 2153. Amplification by the polymerase chain reaction of cDNA from human and rabbit small intestine, rabbit liver, and the human hepatoma cell line HepG2 showed that the stop codon was only present in intestinal mRNA. Enterocyte genomic DNA did not contain the stop codon. We suggest that a co- or posttranscriptional C----U change may result in the production of apo-B48, which represents the amino-terminal 2152 amino acids of apo-B100. This is the first example of tissue-specific modification of a single mRNA nucleotide resulting in two different proteins from the same primary transcript.