DPPI is a structurally unique member of the C1 family of cysteine proteases because it exists as a tetramer and retains a portion of the pro-peptide in its active form (Dolenc et al., 1995; Wolters et al., 1998; Turk et al., 2001). DPPI activation requires proteolytic cleavage at three sites: two in the pro-peptide and a third within the mature enzyme. In humans, these processing events generate a 163-aa heavy chain and a 68-aa light chain. Cleavage of the 206-aa pro-peptide releases an 87-aa fragment, leaving behind 119 amino acids of the N-terminal portion of pro-peptide termed the exclusion domain (Turk et al., 2001) which partially blocks the active site, rendering the enzyme a dipeptidase (Turk et al., 2001).

The mechanisms of activation of many lysosomal cysteine proteases have yet to be defined in vivo. In vitro, several cathepsins (K, L, and S) are capable of autoactivation upon exposure to low pH (McQueney et al., 1997; Menard et al., 1998; Vasiljeva et al., 2005). In contrast, studies using recombinant human pro-DPPI suggest that pro-DPPI cannot activate itself (Dahl et al., 2001). However, pro-DPPI can be activated in vitro by cathepsin L or S (Dahl et al., 2001). To establish if these in vitro findings correspond to in vivo functions, we studied DPPI activation in C57BL/6 mice deficient in cathepsins L and S generated by crossing cathepsin L^-/- (Potts et al., 2004) with cathepsin S^-/- (Shi et al., 1999) mice.

We used mast cells to study DPPI activation because they are an abundant source of DPPI (Wolters et al., 1998) and are convenient to derive in cell culture. First, we sought to determine if cathepsin L and cathepsin S are expressed in mast cells. RT-PCR of bone marrow mast cell (BMMC) RNA using specific primer pairs for cathepsin L and cathepsin S generated single-band amplimers of the predicted sizes (Figure 1A). Primer specificity was demonstrated by the failure to generate amplimers on BMMC RNA obtained from cathepsin L/S^-/- mice. However, RNA from these cells was intact, as primers specific for G3PDH produced an amplimer of the predicted size (data not shown).

To confirm cathepsin L and cathepsin S were translated into active proteins, we labeled the active site of lysosomal cysteine proteases in mast cells with the [¹²⁵I]-JPM-ethyl ester (Shi et al., 1992; Liu et al., 2006) and found that mouse BMMCs contain several active cathepsins, including cathepsins B, C, H, L and S (Figure 1B). The presence of each cathepsin was established by their characteristic size (Liu et al., 2006) on the autoradiogram and the absence of cathepsin L and cathepsin S bands in cathepsin L/S^-/- BMMC lysates.

Next, we compared DPPI and chymase activities in lysates of BMMCs from cathepsin L/S^+/+ and cathepsin L/S^-/- mice. Previous results from our laboratory demonstrated that DPPI is essential for activating mast-cell chymase (Wolters et al., 2001) and that BMMCs lacking DPPI have no active DPPI or chymases (Table 1). Thus, if cathepsin L or S is required for DPPI activation, we predicted that BMMCs from cathepsin L/S^-/- mice would have no active DPPI or chymase. Instead, we found that cathepsin L/S^-/- BMMCs contain levels of DPPI and chymase activity indistinguishable from those of cathepsin L/S^+/+ controls (Table 1), indicating that cathepsin L and cathepsin S are not required for activation of DPPI in cultured BMMCs.

BMMCs are differentiated in vitro and do not fully reflect the phenotype of various mast-cell subtypes in vivo. Thus, it is possible that cathepsin L or S is required for DPPI activation in mast cell subtypes in vivo. DPPI is required for chymase activation (McCoy et al., 1988; Wolters et al., 2001). Therefore, if cathepsin L and cathepsin S are essential for activating DPPI, cathepsin L/S^-/- mice should recapitulate the DPPI^-/- phenotype and lack chymase activity using the enzyme histochemical chloroacetate esterase stain (Wolters et al., 2001). To test this possibility, we examined sections of tongue, trachea, lung, spleen, stomach, and skin from cathepsin L/S^-/- mice for chloroacetate esterase activity and found positive cells in all of these tissues, but not tissues obtained from DPPI^-/- mice (Figure 2, and data not shown). Esterase-positive cells had a distribution, number and morphology typical for mast cells identified in wild-type mice (Wolters et al., 2001). The presence of esterase activity in these tissues demonstrates that active chymase and, by extension, active DPPI are present in tissue mast cells of cathepsin L/S^-/- mice.

Finally, we investigated whether, despite maintaining activity, DPPI was processed at different sites in cathepsin L/S^-/- mice. Normally, three proteolytic processing events, yielding a heavy chain, a light chain and an exclusion domain, are required for DPPI activation (Dahl et al., 2001; Turk et al., 2001). If cathepsin L and cathepsin S were involved in processing, we predicted we would detect the absence or a change in the size of some or all of these domains in cathepsin L/S^-/- BMMCs. To investigate these possibilities, we probed immunoblots of wild-type and cathepsin L/S^-/- BMMC lysates with rabbit anti-bovine DPPI. Immunoblots showed that the antibody recognized both heavy and light chains of mouse DPPI (Figure 3). Furthermore, the size of the heavy and light chains were the same in wild-type and cathepsin L/S^-/- BMMC lysates, indicating they were processed to their expected size in the absence of cathepsins L and S.

Studies using recombinantly expressed cysteine proteases demonstrate that both cathepsins L and S activate DPPI in vitro (Dahl et al., 2001). In contrast, our data show that cathepsins L and S are not required for DPPI activation in vivo. Furthermore, given the evidence that DPPI is processed into appropriately sized heavy and light chains in the absence of cathepsins L and S, it is also unlikely that they secondarily process DPPI following activation by another protease. While these findings do not eliminate the possibility that cathepsin L or S participates in the activation of DPPI in vivo, they do suggest that, at minimum, other proteases activate DPPI and that there may be redundancy in the DPPI activation system. This redundancy would serve to ensure activation of DPPI and the proteases it activates to guarantee they maintain their activity and are able to participate in important biological processes (Pham and Ley, 1999; Wolters et al., 2001; Adkison et al., 2002).

The identity of the other protease(s) responsible for DPPI activation in vivo remains unknown. However, although we cannot rule out the possibility that BMMCs contain cathepsins at levels below the detection limit for the JPM label, in mast cells, the protease is unlikely to be a lysosomal cysteine protease because the only active cathepsins we detected in these cells are cathepsins B, C (DPPI), H, L, and S. Previous studies using recombinant DPPI demonstrated that DPPI cannot autoactivate itself in vitro (Dahl et al., 2001). As discussed, cathepsins L and S are not required to activate DPPI. Cathepsin B will not likely activate DPPI because it is preferentially a carboxypeptidase (Turk et al., 1997) and does not activate pro-DPPI in vitro (Dahl et al., 2001). Finally, cathepsin H is an aminopeptidase and therefore is not likely to be capable of generating the endoproteolytic cleavages required for DPPI activation.