Processing of a Pestivirus Protein by a Cellular Protease Specific for Light Chain 3 of Microtubule-Associated Proteins

ABSTRACT The genome of the cytopathogenic (cp) bovine viral diarrhea virus (BVDV) JaCP contains a cellular insertion coding for light chain 3 (LC3) of microtubule-associated proteins, the mammalian homologue of yeast Aut7p/Apg8p. The cellular insertion induces cp BVDV-specific processing of the viral polyprotein by a cellular cysteine protease homologous to the known yeast protease Aut2p/Apg4p. Three candidate bovine protease genes were identified on the basis of the sequence similarity of their products with the Saccharomyces cerevisiae enzyme. The search for a system for functional testing of these putative LC3-specific proteases revealed that the components involved in this processing have been highly conserved during evolution, so that the substrate derived from a mammalian virus is processed in cells of mammalian, avian, fish, and insect origin, as well as in rabbit reticulocyte lysate, but not in wheat germ extracts. Moreover, two of these proteases and a homologous protein from chickens were able to rescue the defect of a yeast AUT2 deletion mutant. In coexpression experiments with yeast and wheat germ extracts one of the bovine proteases and the corresponding enzyme from chickens were able to process the viral polyprotein containing LC3. Northern blots showed that bovine viral diarrhea virus infection of cells has no significant influence on the expression of either LC3 or its protease, bAut2B2. However, LC3-specific processing of the viral polyprotein containing the cellular insertion is essential for replication of the virus since mutants with changes in the LC3 insertion significantly affecting processing at the LC3/NS3 site were not viable.

Tissue-specific promoters of the α human folate receptor gene yield transcripts with divergent 5′ leader sequences and different translational efficiencies

The α human folate receptor (αhFR), or KB cell folate receptor, gene contains two major promoters that produce transcripts, KB1 and KB4, varying only in the length and sequence of their 5′ untranslated regions (UTRs). Using RNase protection assays specific for each isoform, we show that the level of expression of these two transcripts is tissue-specific, indicating that promoter usage is regulated, not constitutive. RNA stabilities and translational efficiencies of the KB1 and KB4 transcripts were compared to determine the functional significance of the different 5′ UTRs. Analyses of RNA turnover in vivo with actinomycin D to block new transcription and in vitro with a cytoplasmic extract indicate no discernible differences in the stabilities of the two transcripts. However, the KB4 transcript is 2–3-fold more efficiently translated in wheat germ extracts in vitro and transfected CHO cells in vivo. Also, high ionic strength, which favours the formation of RNA secondary structure, differentially affects the translational efficiencies of the two transcripts. Translation of the longer KB1 mRNA is 2–5-fold more inhibited by hypertonic conditions than translation of the KB4 mRNA. Because the 5′ UTR of KB1 is approximately four times longer than the 5′ UTR of KB4, 149 bp (75%) of the KB1 5′ UTR were deleted to determine whether the long leader sequence inhibited translation. The resulting derivative, dKB1, has a 5′ UTR similar in length, but not sequence, to the 5′ UTR of KB4. dKB1 is translated at a level approaching that of KB4 in wheat germ extracts, indicating that the upstream portion of the 5′ leader sequence contributes to the relative translational inefficiency of KB1. Hence, one consequence of tissue-specific promoter usage is the production of αhFR transcripts with different 5′ non-coding regions that affect translational efficiency.