Translation start site multiplicity of the CCAAT/enhancer binding protein α mRNA is dictated by a small 5′ open reading frame

5540-5547 ©1994 Oxford University Press Nucleic Acids Research, 1994, Vol. 22, No. 25 Translation start site multiplicity of the CCAAT/enhancer binding protein a mRNA is dictated by a small 5' open reading frame Cor F.Calkhoven, Peter R.J.Bouwman, Lenie Snippe and Geert AB* Department of Biochemistry, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands Received October 18, 1994; Revised and Accepted November 14, 1994 EMBL accession no. X66844 The CCAAT/enhancer binding proteins (C/EBP) a and /3 of the bZIP family of transcription factors each occur as multiple forms due to translation Initiation at different in-frame AUG codons from the same messenger RNA. The C/EBPa mRNAs of chicken, rat and Xenopus all contain a small 5' open reading frame (5'ORF) whose size (18 nucleotldes) and distance (seven nucleotides) to the C/EBPa clstron has been conserved in vertebrate evolution. The present studies shows that the small 5'ORF Is crucial to the leaky scanning mechanism of ribosomes causing a fraction of them to ignore the first C/EBPa AUG codon and to start at Internal AUGs. Our data challenge the view that translational start site multiplicity is mainly governed by the sequence context of the potential initiation codons. Western analysis showed that the two major chicken C/EBPa translation products, the full-length cC/EBPa-42 which acts a frans-actlvator in liver and the N-terminally truncated cC/EBPa-29 which lacks transcription activation potential, occur In a fixed ratio which Is similar In different expressing tissues, like liver, lung and small intestine. The presence of a similar, thusfar unnoticed, small ORF 5' to the major initiation codon of C/EBP/3 mRNA suggests that start site multiplicity from this mRNA may be governed by the same mechanism. INTRODUCTION CCAAT/enhancer binding proteins (C/EBPs) belong to the bZIP class of transcription factors (1). The positively charged DNA binding domain and the leucine zipper dimerisation domain are located in the C-terminal section of the C/EBP molecule (2). Several C/EBPs, each encoded by a different member of the family of C/EBP genes, have been identified (3 -17). Of these, the liver-enriched C/EBPa and C/EBP/3 have been most extensively studied (3,11,18,19). They are involved in the transcription of several liver-specific genes, such as the serum albumin gene (20). *To whom correspondence should be addressed C/EBPa and C/EBP/3 share the common feature that, besides the full-length product, N-terminally truncated polypeptides are translated from the same mRNA by use of internal, in-frame AUG codons (9,21). For rat C/EBPa, it has been shown that the smaller translation product lacking the N-terminal 117 amino acids is devoid of its transcription activation potential in liver (21). Studies with the mouse adipocyte cell line 3T3-L1 have shown that the full-length product inhibits cell proliferation, whereas the smaller translation product is not antimitotic (22). The C/EBP/3 gene also issues two proteins, the liver-enriched transcriptional activator (LAP) and inhibitory (LIP) proteins. It is believed that the shorter products have a physiological function as antagonists of their /ra/u-activating counterparts. The formation of multiple polypeptides from the rat C/EBP messenger RNAs is best explained by a ribosome scanning mechanism in which a fraction of ribosomes ignores the first AUG codon and start at internal AUG codons (9,21). We have cloned a member of the C/EBP gene family of chicken (12) to study its role in the liver-specific expression of the estrogen-regulated apoVLDL II gene (23). The encoded protein, called cC/EBP, most strongly resembles C/EBPa of the rodent C/EBP family. The chicken C/EBP/3 gene homologue has been cloned by Katz et al. (17) and encodes the myeloid-specific transcription factor, NF-M. Alignment of the rat and chicken C/EBPa sequences has delineated three conserved regions in the N-terminal moiety of the polypeptide chain in addition to the C-terminal bZIP domain (12). Comparison of the chicken and rat C/EBPa shows that the methionines which are used for the internal starts are located at analogous positions within the variable sequences linking the conserved regions I and II, and between region HI and the bZIP domain. Interestingly, the leader sequences of rat and chicken C/EBPa mRNA were found to contain a small open reading frame at seven nucleotides from the C/EBP coding region (12). The most important finding of the present studies is that this small 5'ORF, which appears to be conserved in vertebrate evolution, is the primary mRNA feature responsible for C/EBPa translation start site multiplicity. Moreover, sequence comparison of C/EBP/3 mRNAs has revealed a, thusfar unnoticed, small ORF in front ABSTRACT Nucleic Acids Research, 1994, Vol. 22, No. 25 5541 of the major C/EBP/3 translation initiation codon. This feature, together with the similar distribution and sequence context of internal start sites, suggests that C/EBPa and /3 may follow the same strategy for generating multiple translation products from the same mRNA. For cC/EBP:mut5'ORF, oligo 5'-CCGGAGCCCTTCAGGATC CCCGGCAGGCTG-3' (mismatches underlined) to mutate the ATG of the small 5'ORF and to introduce a diagnostic BamYQ site; for cC/EBP:KozakATGl, oligo 5'-GCAGGCTGTAGGACCACCATGGAGCAAGCC-3', to create diagnostic AvaO. and Ncol sites; for cC/EBP:mutATG3, oligo 5'-TTCCACGGG.ATC_ CACGGGGCC-3' to mutate Met3 (codon 102) into De and to introduce a diagnostic BamW site; for cCIEBP.dddr, oligo 5'AGGCTGTAGGATCCCG4 7TGGAGC AAGC-3' to create a BamiU just upstream of the first ATG which was used for the excision and cloning of the downstream sequences; for cC/EBP:AdomI, oligo 5'-GGGGATTTCGAATTCCACGGCATG-3' to introduce an EcoKL site just upstream of the third AUG codon (codon 102) which allowed the deletion of upstream sequences. cC/EBP:bZJP was constructed by cloning the SmaVBamHl fragment containing the sequences downstream of codon 220. For the transfection of LMH cells, the reporter plasmid (D)9-tkCAT containing nine copies of the mouse albumin C/EBP binding site 'D' in front of TK minimal promoter and CAT gene was used (5,27). All constructs were checked by sequence or restriction analysis. Transfection COS-1 cells were transfected with 10 ng of pSG5-based expression plasmid as described by Ausubel et al., 1994, and cultured in DMEM/F12/5% fetal calf serum (Gibco). LMH cells (28) were maintained in Waymouth's MB 752/1 medium supplemented with 10% fetal calf serum, 50/tg/ml streptomycin and 50 U/ml penicillin (Gibco). The cells were cultured in a 10% CC^/air mixture at 37°C. DNA transfections were carried out with the calcium phosphate method (29). The day before transfection, the cells were plated at 40% confluence on 6 cm <p dishes. The DNA precipitates contained 2 /tg (DVtkCAT reporter plasmid (5) and 2.5 or 5.0 ng pSG5-based expression Nuclear proteins Nuclei were isolated from liver, lung and small (...truncated)