ABSTRACT
A novel human serum protein with a molecular mass of 87,000 daltons was purified to homogeneity and subjected to amino acid sequence analyses. These sequences were used to design oligonucleotide primers and to isolate a full-length cDNA. The amino acid sequence encoded by the cDNA shares strong similarity to albumin family members and shares the characteristic pattern of Cys residues observed in this family. In addition, the gene maps to chromosome 4 as do other members of the albumin gene family. Based upon these observations, we conclude that the 87,000-dalton protein, which we designate afamin (AFM), is the fourth member of the albumin family of proteins. Afamin cDNA was stably transfected into Chinese hamster ovary cells and recombinant protein (rAFM) was purified from conditioned medium. Both rAFM and AFM purified from human serum react with a polyclonal antibody that was raised against a synthetic peptide derived from the deduced amino acid sequence of AFM.
Subject(s)
Albumins/genetics , Blood Proteins/genetics , Carrier Proteins , Glycoproteins , Multigene Family , Serum Albumin/genetics , Vitamin D-Binding Protein/genetics , alpha-Fetoproteins/genetics , Amino Acid Sequence , Animals , Base Sequence , CHO Cells , Chromosome Mapping , Cricetinae , Cricetulus , DNA, Complementary , Humans , Molecular Sequence Data , Protein Conformation , Sequence Homology, Amino Acid , Serum Albumin, HumanABSTRACT
We used molecular cloning and functional analyses to extend the family of Neu differentiation factors (NDFs) and to explore the biochemical activity of different NDF isoforms. Exhaustive cloning revealed the existence of six distinct fibroblastic pro-NDFs, whose basic transmembrane structure includes an immunoglobulin-like motif and an epidermal growth factor (EGF)-like domain. Structural variation is confined to three domains: the C-terminal portion of the EGF-like domain (isoforms alpha and beta), the adjacent juxtamembrane stretch (isoforms 1 to 4), and the variable-length cytoplasmic domain (isoforms a, b, and c). Only certain combinations of the variable domains exist, and they display partial tissue specificity in their expression: pro-NDF-alpha 2 is the predominant form in mesenchymal cells, whereas pro-NDF-beta 1 is the major neuronal isoform. Only the transmembrane isoforms were glycosylated and secreted as biologically active 44-kDa glycoproteins, implying that the transmembrane domain functions as an internal signal peptide. Extensive glycosylation precedes proteolytic cleavage of pro-NDF but has no effect on receptor binding. By contrast, the EGF-like domain fully retains receptor binding activity when expressed separately, but its beta-type C terminus displays higher affinity than alpha-type NDFs. Likewise, structural heterogeneity of the cytoplasmic tails may determine isoform-specific rate of pro-NDF processing. Taken together, these results suggest that different NDF isoforms are generated by alternative splicing and perform distinct tissue-specific functions.
