The cDNAs for two glycoproteins, the 158-kDa submandibular glycoprotein (SGP158) and the 200-kDa parotid glycoprotein (PGP200), have been cloned from rat submandibular and parotid glands, respectively. Both cDNAs encode for identical proteins with repeating peptides -Asp-Gln-Gly-(Asn)-Gln-Thr-Gln-Pro-Arg-Pro-Pro-His-Pro-. A full-length cDNA encoding SGP158 was obtained using the strategy of anchor-PCR, and a full-length cDNA of PGP200 was prepared using RNA-PCR. Sequence analysis of the cDNAs revealed that SGP158 and PGP200 are identical proteins with 23 repeating peptides. Twenty-one peptides contain potential N-glycosylation sites and these two glycoproteins differ only in their glycosylation patterns. Southern-blot analysis showed that a single-copy gene encodes both mRNAs. PGP200 is constitutively expressed, but the synthesis of SGP158 is totally dependent upon treatment of animals with the beta-agonist isoproterenol. The first 106-nucleotide sequence of cDNAs for PGP200 and SGP158, which corresponds to the 5'-untranslated region and sequence encoding the signal peptide, is highly conserved when compared with proline-rich protein and glutamine-rich protein gene sequences. Based on the nucleotide sequences of exon I, a phylogenetic tree was constructed for 35 members of these multigene families. The tree fits with the generally recognized phylogeny of mammalian orders. We propose that exon I sequences of the proline-rich protein and glutamine-rich protein multigene families are relatively new and are possibly generated through exon shuffling during evolution.