Inhibitor of Apoptosis Proteins are substrates for the Mitochondrial Serine Protease Omi/HtrA2*

The mature serine protease Omi/HtrA2 is released from the mitochondria into the cytosol during apoptosis. Suppression of Omi/HtrA2 by RNA interference in human cell lines reduces cell death in response to TRAIL and etoposide. In contrast, ectopic expression of mature wildtype Omi/HtrA2, but not an active site mutant, induces potent caspase activation and apoptosis. In vitro assays demonstrated that Omi/HtrA2 could degrade inhibitors of apoptosis proteins (IAPs). Consistent with this observation, increased expression of Omi/HtrA2 in cells increases degradation of XIAP while suppression of Omi/HtrA2 by RNA interference has an opposite effect. Combined, our data demonstrate that IAPs are substrates for Omi/HtrA2, and their degradation could be a mechanism by which the mitochondrially-released Omi/HtrA2 activates caspases during apoptosis.


INTRODUCTION
The genes for inhibitors of apoptosis proteins (IAP) were originally identified in the genome of baculoviruses based on the ability of their gene products to protect infected host cells from virus-induced apoptosis (1). Cellular homologues of the viral IAPs have been identified in insects, nematodes, yeast and mammals (2,3). All IAPs contain one or more conserved domains, referred to as baculovirus IAP repeats (BIRs), that are essential for inhibition of apoptosis (reviewed in (4,5)). The BIR domains and the linker regions between them bind directly to caspases and inhibit their activity. Some IAPs, such as human XIAP, c-IAP1 and c-IAP2, and Drosophila DIAP1 and DIAP2 also contain C-terminal RING domains. The RING domain is important for ubiquitination and proteosome degradation of IAPs and IAP-associated proteins (reviewed in (5,6)).
The anti-apoptotic activity of IAPs is regulated by a group of proteins that bind to the BIR domains of IAPs via a N-terminal conserved 4-residue IAP-binding motif (IBM) (reviewed in (5,7)). In Drosophila melanogaster, five IBM-containing proteins known as Reaper, Hid, Grim, Sickle and Jafrac2 have been identified as direct IAP-binding proteins (5,8). These proteins promote caspase activation by disrupting caspase-IAP complexes and/or inducing autoubiquitination and degradation of IAPs, thus preventing IAPs from inhibiting caspases (4)(5)(6)(7).
In mammals two functional homologues of the Drosophila proteins, known as Smac/Diablo (9,10) and Omi/HtrA2 (11)(12)(13)(14)(15)  In vitro protease assay-The protease activity of Omi was assayed with in vitro translated 35 S-labeled XIAP, c-IAP1, or c-IAP2 as substrates. Bacterially expressed Omi/HtrA2 was purified on Talon affinity resins and then incubated with 35 S-labeled proteins in buffer A (20 mM HEPES pH 7.4, 10 mM KCl, 1.5 mM MgCl2, 1mM EDTA and 1mM EGTA with protease inhibitors in a total 20 µl reaction. The cleavage products were analyzed by SDS-PAGE and visualized by autoradiography.
The S100 extracts were incubated with Omi polyclonal antibody (made in the lab) for 1 h. After incubation, the Omi-antibody complexes were bound to protein G Sepharose and washed several times. The protein G Sepharose-bound complexes were incubated with 35 S-labeled XIAP in buffer A for 1 h at 37° C in a total 20 µl reaction. The cleavage products were analyzed as outlined above.
Omi/HtrA2-induced cell death assay-The ability of Omi/HtrA2 variants to induce cell death in transfected cells was assayed as described (11). 293T cells were seeded in 12 well plates at a density of 1x10 6

Mature Omi/HtrA2 can promote caspase activation and cell death-Recent results
suggest that the serine protease activity of Omi/HtrA2 could be responsible for its potent killing activity (18). To determine whether Omi/HtrA2 protease activity can activate the caspase pathway, we measured caspase activity in 293 cells transfected with constructs encoding mature wildtype or active site mutant Omi/HtrA2 proteins. Expression of mature wildtype, but not the active site mutant Omi/HtrA2 resulted in significant caspase activation ( Fig. 2A). This enhanced caspase activation was associated with increased cell death (Fig. 2B). Similar results were observed in MCF-7 and HeLa cell lines (ref (11) and data not shown). These results suggest that the ability of Omi/HtrA2 to promote caspase activation and cell death in transfected cells is largely dependent on its serine protease activity.
Omi/HtrA2 can degrade IAPs in vitro-Western blot analysis revealed that expression of mature wildtype Omi/HtrA2, but not an active site mutant protein results in significant reduction in the amount of XIAP in the transfected cells (Fig. 2C). These results suggest that degradation 8 of XIAP by Omi/HtrA2 could be responsible for this effect. To test this hypothesis we first determined the physiological concentration of Omi/HtrA2 in 293 and HeLa cells by immunoblot analysis with Omi/HtrA2 antibodies. The concentration of Omi/HtrA2 in 293 was found to be 1 50 nM and in HeLa cells ~ 60 nM. Next we assayed the activity of physiological amounts of recombinant Omi/HtrA2 with 35 S-labeled XIAP, cIAP-1 or cIAP-2 as substrates. As shown in Fig. 3A, physiological amounts of Omi/HtrA2 were able to degrade these IAPs in a dosedependent manner. Consistent with these results, physiological amounts of endogenous Omi/HtrA2 protein isolated by immunoprecipitation from 293 cells were also able to degrade XIAP in a dose dependent manner (Fig. 3B). These results indicate that IAPs are physiological substrates for Omi/HtrA2 protein.
Next we tested the effect of Omi/HtrA2 on the ability of IAPs to inhibit caspase activation in S100 extracts stimulated with cytochrome c and dATP (Fig. 3C). Incubation of XIAP with mature wildtype Omi resulted in a large dose dependent reduction in the inhibitory activity of XIAP. Similar results were obtained with the active site mutant Omi/HtrA2-S306A.
In contrast, incubation of XIAP with protease active Omi/HtrA2 without its IAP binding motif (∆AVPS) did not cause significant reduction in the inhibitory activity of XIAP.
Western blot analysis showed that XIAP is degraded in the reaction mixtures containing wildtype Omi and Omi-∆AVPS, but not Omi/HtrA2-S306A. The IAP binding motif of Omi seems to enhance its ability to cleave XIAP, because the Omi-∆AVPS was less efficient in cleaving XIAP. In addition, a distinct pattern of cleavage was observed with the wildtype Omi/HtrA2 compared to Omi-∆AVPS. Nevertheless, the cleavage products generated by Omi-∆AVPS were able to inhibit caspases indicating that cleavage alone might not be sufficient for inactivation of XIAP. These interesting results suggest that the presence of the IAP binding 9 motif enhances the ability of Omi/HtrA2 to cleave and disrupt the association of the cleavage products of XIAP with caspases. In contrast, in the absence of the IAP binding motif, Omi/HtrA2 cleaves XIAP less efficiently at multiple sites and is not able to disrupt the association of the cleavage products with caspases.

IAP degradation during apoptosis is dependent on Omi/HtrA2 expression levels-To
provide additional support for the role of Omi/HtrA2 in IAP-degradation we decreased the expression levels of Omi/HtrA2 in the mitochondria by transfection of MCF-7 cells with Omispecific or non-specific siRNAs and then treated the cells with etoposide for different periods of time. Less IAP-degradation was observed in cells transfected with the Omi-specific siRNA compared to the cells transfected with the non-specific siRNA (Fig. 4). Taken together, these results indicate that Omi/HtrA2 plays a direct role in the degradation of IAPs during cell death.

DISCUSSION
IAP degradation is an important mechanism for caspase activation during developmental and perhaps p53-induded cell death in Drosophila (5,6,19). Drosophila DIAP-1 destruction is stimulated by binding of IAP-binding proteins such as Reaper, which is induced during development or DNA damage, to the BIR2 domain of DIAP1. This binding induces autoubiquitination of DIAP1, which targets it for degradation by the proteosomal pathways. So far, a similar mechanism of IAP destruction has not been identified in mammalian cells during cell death. However, recent results demonstrated that c-IAP1 is cleaved in human HeLa cells during p53-dependent apoptosis by a serine protease (20). Based on results showing by guest on March 24, 2020 http://www.jbc.org/ Downloaded from upregulation of Omi/HtrA2 mRNA by p53, it was suggested that Omi/HtrA2 could be responsible for the observed degradation of cIAP-1 during p53-dependent apoptosis (20).
Our biochemical and cellular data indicate that XIAP is also degraded in human cancer cell lines during cell death. Omi/HtrA2 appears to be directly responsible for degradation of XIAP, since increased level of Omi/HtrA2 induces more degradation (Fig. 2C) while decreased level has an opposite effect (Fig. 4). In addition, purified Omi/HtrA2 can directly degrade XIAP and other human IAPs in vitro (Fig. 3). The new results also indicate that the IAP binding motif of Omi/HtrA2 is important for efficient cleavage of IAPs at distinct sites, and for disruption of the association between caspases and the IAP fragments. In the absence of the IAP-binding motif, Omi/HtrA2 can still cleave IAPs with less efficiency but is unable to disrupt the association between caspases and the IAP fragments. Thus, the presence of the IAP-binding motif enhances the ability of Omi/HtrA2 to find and destroy IAPs after its release from the mitochondria into the cytosol during mammalian cell death.
Although the level of XIAP was reduced by the transfected wildtype Omi/HtrA2, there were no XIAP fragments detectable in the cellular extracts of the transfected cells (Fig. 2B). In contrast multiple fragments were detectable in vitro after incubation of Omi/HtrA2 with IAPs ( Fig. 3). These interesting results suggest that in cells, the Omi/HtrA2-generated IAP fragments could be targeted for further degradation by the N-end rule degradation pathway. This is consistent with recent results, which demonstrated that caspase-cleavage of the Drosophila DIAP1 makes it unstable and targets it for degradation by the N-end rule degradation pathway (21). Thus Omi/HtrA2 could promote apoptosis in mammalian cells by two mechanisms. One mechanism relies on its IAP-binding motif to bind to IAPs and disrupts association of active caspases with IAPs and the other mechanism relies on its protease activity to cleave bound IAPs 11 and targets them for further degradation by the proteosomal pathways. However, both mechanisms might be necessary for efficient killing by Omi/HtrA2. This could explain why the active site mutant Omi/HtrA2-S306A can disrupts IAP-caspase association and promote caspase activation in vitro (Fig. 3C) while it is not able to induce efficient cell death or caspase activation in transfected cells (Fig. 2).
Ironically, the similarity between Omi/HtrA2 and the bacterial survival chaperoneprotease DegP/HtrA suggests that Omi/HtrA2 could have an important survival function in the mitochondria (16,17). Omi/HtrA2 may protect the mitochondria against cellular stresses by recognizing unfolded or misfolded proteins in the inter-membrane space and helps refold or degrade them. If this is the case, then loss of this protein could lead to accumulation of damaged proteins in the inter-membrane space of the mitochondria over a long period of time. This is likely to be deleterious to mitochondrial function and increase sensitivity of cells to stress-