Serine proteases constitute the major protein content of the cytoplasmic granules of several hematopoietic cell lineages. These proteases are encoded from four different loci in mammals. One of these loci, the chymase locus, has in rats experienced a massive expansion in the number of functional genes. The human chymase locus encodes 4 proteases, whereas the corresponding locus in rats contains 28 such genes. One of these new genes has changed tissue specificity and has been found to be expressed primarily in vascular smooth muscle cells, and therefore been named rat vascular chymase (RVC). This β-chymase has been claimed to be a potent angiotensin-converting enzyme by cleaving angiotensin (Ang) I into Ang II and thereby having the potential to regulate blood pressure. To further characterize this enzyme, we have used substrate phage display and a panel of recombinant substrates to obtain a detailed quantitative view of its extended cleavage specificity. RVC was found to show a strong preference for Phe and Tyr in the P1 position, but also to accept Leu and Trp in this position. A strong preference for Ser or Arg in the P1' position, just C-terminally of the cleavage site, and a preference for aliphatic amino acids in most other positions surrounding the cleavage site was also seen. Interesting also was a relatively strict preference for Gly in positions P3' and P4'. RVC thereby shares similarity in its specificity to the mouse mucosal mast cell chymase mMCP-1, which efficiently converts Ang I to Ang II. This similarity adds support for the role of β-chymases as potent angiotensin converters in rodents, as their α-chymases, which have the capacity to efficiently convert Ang I into Ang II in other mammalian lineages, have become elastases. However, interestingly we found that RVC cleaved both after Arg2 and Phe8 in Ang I. Furthermore this cleavage was more than two hundred times less efficient than the consensus site obtained from the phage display analysis, indicating that RVC has a very low ability to cleave Ang I, raising serious doubts about its role in Ang I conversion.