We have exploited differences in the structures of S2' subsites of proteinase 3 (Pr3) and human neutrophil elastase (HNE) to prepare new fluorogenic substrates specific for each of these proteases. The positively charged residue at position 143 in Pr3 prevents it from accommodating an arginyl residue at S2' and improves the binding of P2' aspartyl-containing substrates, as judged by the decreased K(m). As a result, the k(cat)/K(m) for Abz-VADCADQ-EDDnp is over 500 times greater for Pr3 than for HNE, and that for Abz-APEEIMRRQ-EDDnp is over 500 times greater for HNE than for Pr3. This allows each protease activity to be measured in the presence of a large excess of the other, as might occur in vivo. Placing a prolyl residue in position P2' greatly impaired substrate binding to both HNE and Pr3, which further emphasizes the importance of S' subsites in these proteases. HNE and Pr3 activities were measured with these substrates at the surface of fixed polymorphonuclear leukocytes (PMNs) before and after activation. This demonstrated that their active site remains accessible when they are exposed to the cell surface. Both membrane-bound proteases were strongly inhibited by low M(r) serine protease inhibitors, but only partially by inhibitors of larger M(r) such as alpha1-protease inhibitor, the main physiologic inhibitor in lung secretions. Most of membrane-bound HNE and Pr3 can be released from the membrane surface of fixed cells by a buffer containing detergent, suggesting that hydrophobic interactions are involved in membrane binding.