Autoantibodies to citrullinated proteins (anti-CCP) can be detected in up to 80% of rheumatoid arthritis (RA) patients with very high specificity (98%). Citrulline residues, the target of the anti-CCP antibodies, are formed by post-translational modification of arginine residues, catalyzed by peptidylarginine deiminase enzymes (PAD; EC 3.5.3.15).

Our aim was to investigate the full complexity of the family of PAD enzymes and to locate putatively important conserved residues or domains.

We performed a thorough query for PAD sequences using EST and genomic databases. With these data we constructed a complete PAD alignment (ClustalW) and a phylogenetic tree (Treeview), and mapped the different genes.

Next to the previously described PADs, we found several (partial) novel sequences, both mammalian and nonmammalian. From the alignment (available online at http://www.interscience.wiley.com/jpages/0265-9247/suppmat/2003/25/v25.1106.html), it is clear that the carboxyl-terminal half is more conserved than the amino-terminal part. Some of the fully conserved residues (His475 and Cys655) have been suggested to be important for catalytic activity of murine PAD2. These residues would fit in the 'catalytic pocket', as has been described for other arginine converting enzymes (arginine deiminases EC 3.5.3.6. and amidinotransferases EC 2.1.4.1.). The negative charge (on average -14) and low pI (about 5.8) of the enzyme are important for its interactions with arginine substrates and with Ca²⁺ ions, which are essential for PAD activity. A phylogenetic tree confirmed the human PAD5 to be the orthologue of the murine PAD4. By comparing mRNA and genomic sequences, the individual exons of all murine and human PADs could be mapped in a tight conserved gene cluster (human chromosome 1p36.1; mouse #4E1; rat #5q36). Structural characterization of PADs will yield valuable clues regarding the aetiology of RA and for the development of PAD inhibiting drugs.