We cloned and sequenced the cDNAs against genomic RNA and mRNA for phosphoprotein (P) of human parainfluenza type 2 virus (PIV-2). cDNA clone from genomic RNA was 1439 nucleotides in length excluding poly(A) and was found to have two small open reading frames encoding proteins of 233 and 249 amino acids. Two different mRNA cDNA clones were obtained; that is, one mRNA contained a smaller reading frame coding 225 amino acids, V protein, and the other mRNA contained a larger reading frame coding 395 amino acids, P protein. Both mRNAs had G cluster in coding frame. The former mRNA contained seven G residues, and two extra G residues were inserted in the latter mRNA. Ten cDNA clones from the genomic RNA were identical and were composed of seven G residues, indicating that genomes analyzed here were a homogeneous population. Therefore, V protein is encoded by faithfully copied mRNA and P protein is translated from mRNA in which two additional G residues are nontemplately inserted immediately after seven genomically encoded G residues. The V and P proteins are amino coterminal proteins and have different C termini. The C terminus of V protein is cysteine-rich and bears some resemblance to metal-binding protein of the zinc finger-type motif. P protein sequence of PIV-2 showed high homologies with SV 5 (40.4%) and mumps virus (35.5%), and a moderate homology with Newcastle disease virus (20.6%). On the other hand, very little homology was found between PIV-2 and other paramyxoviruses including Sendai virus, PIV-3, and measles virus. The cysteine-rich region in V protein was found to be highly conserved in PIV-2, SV 5, and measles virus, suggesting that V protein of paramyxoviruses plays important roles in transcription and/or replication. The predicted cysteine-rich V protein was detected in virus-infected cells using antiserum directed against an oligopeptide specific for the predicted V polypeptide.