Functional domains and sequence alignments of the C/EBP-ϵ isoforms. Functional domains and predicted transcriptional activities of the 4 C/EBP-ϵ isoforms (32, 30, 27, and 14 kDa), based on mutagenesis and transactivation studies of the human and murine full-length isoforms^16,23,49,50  and shorter human isoforms,^2,22  are shown schematically in panel A. Alignments of their amino acid sequences and locations of the various transactivation, repression, and bZIP domains are shown in panel B. All 4 isoforms are identical at their carboxyl terminus, which encodes the RDII repressor, basic DNA binding, and bZIP dimerization domains. The activator isoform C/EBP-ϵ³², a 281-amino acid protein, contains 2 transcriptional activation domains (TADI, TADII), 2 repressor domains (RDI, RDII), and the DNA-binding domain composed of the basic region (BR) and leucine zipper (LZ). C/EBP-ϵ³⁰, a 250-amino acid protein, is derived from an alternative translation start site 100 bp (33 amino acids) downstream of the start site for C/EBP-ϵ³². C/EBP-ϵ²⁷, a 253-amino acid protein derived through alternative RNA splicing, contains a unique 68-amino acid N-terminal repression domain we have designated as RD27.²²  The shortest isoform, C/EBP-ϵ¹⁴, a 130-amino acid protein, consists mainly of the basic DNA binding and leucine zipper domains, the RDII repressor domain, with no transactivation domain. The RDI domain contains a highly conserved “VKEEP” sumoylation consensus sequence (boxed in panel B), through which sumoylation increases the transcriptional activity of the murine activator isoforms.^23,49,50