The observation that activation domains (AD) of procarboxypeptidases are rather long compared to the pro-regions of other zymogens raises the possibility that they could play additional roles apart from precluding enzymatic activity within the proenzyme and helping in its folding process. In the present work, we compared the overall pro-domain tertiary structure with several proteins belonging to the same fold in the structural classification of proteins (SCOP) database by using structure and sequence comparisons. The best score obtained was between the activation domain of human procarboxypeptidase A4 (ADA4h) and the human U1A protein from the U1 snRNP. Structural alignment revealed the existence of RNP1- and RNP2-related sequences in ADA4h. After modeling ADA4h on U1A, the new structure was used to extract a new sequence pattern characteristic for important residues at key positions. The new sequence pattern allowed scanning protein sequences to predict the RNA-binding function for 32 sequences undetected by PFAM. Unspecific RNA electrophoretic mobility shift assays experimentally supported the prediction that ADA4h binds an RNA motif similar to the U1A binding-motif of stem-loop II of U1 small nuclear RNA. The experiments carried out with ADA4h in the present work suggest the sharing of a common ancestor with other RNA recognition motifs. However, the fact that key residues preventing activity within the proenzyme are also key residues for RNA binding might have induced the activation domains of procarboxypeptidases to evolve from the canonical RNP1 and RNP2 sequences.