Abstract
Keratins (Ks) consist of central α-helical rod domains that are flanked by non-α-helical head and tail domains. The cellular abundance of keratins, coupled with their selective cell expression patterns, suggests that they diversified to fulfill tissue-specific functions although the primary structure differences between them have not been comprehensively compared. We analyzed keratin sequences from many species: K1, K2, K5, K9, K10, K14 were studied as representatives of epidermal keratins, and compared with K7, K8, K18, K19, K20 and K31, K35, K81, K85, K86, which represent simple-type (single-layered or glandular) epithelial and hair keratins, respectively. We show that keratin domains have striking differences in their amino acids. There are many cysteines in hair keratins but only a small number in epidermal keratins and rare or none in simple-type keratins. The heads and/or tails of epidermal keratins are glycine and phenylalanine rich but alanine poor, whereas parallel domains of hair keratins are abundant in prolines, and those of simple-type epithelial keratins are enriched in acidic and/or basic residues. The observed differences between simple-type, epidermal and hair keratins are highly conserved throughout evolution. Cysteines and histidines, which are infrequent keratin amino acids, are involved in de novo mutations that are markedly overrepresented in keratins. Hence, keratins have evolutionarily conserved and domain-selectively enriched amino acids including glycine and phenylalanine (epidermal), cysteine and proline (hair), and basic and acidic (simple-type epithelial), which reflect unique functions related to structural flexibility, rigidity and solubility, respectively. Our findings also support the importance of human keratin 'mutation hotspot' residues and their wild-type counterparts.