Characterization of a Novel and Specific Inhibitor for the Pro-apoptotic Protease Omi/HtrA2*

Omi/HtrA2 is a mammalian serine protease with high homology to bacterial HtrA chaperones. Omi/HtrA2 is localized in mitochondria and is released to the cytoplasm in response to apoptotic stimuli. Omi/HtrA2 induces cell death in a caspase-dependent manner by interacting with the inhibitor of apoptosis protein as well as in a caspase-independent manner that relies on its protease activity. We describe the identification and characterization of a novel compound as a specific inhibitor of the proteolytic activity of Omi/HtrA2. This compound (ucf-101) was isolated in a high throughput screening of a combinatorial library using bacterially made Omi-(134–458) protease and fluorescein-casein as a generic substrate. ucf-101 showed specific activity against Omi/HtrA2 and very little activity against various other serine proteases. This compound has a natural fluorescence that was used to monitor its ability to enter mammalian cells. ucf-101, when tested in caspase-9 (−/−) null fibroblasts, was found to inhibit Omi/HtrA2-induced cell death.

Omi/HtrA2 is a human serine protease that has extensive homology to bacterial high temperature requirement A (HtrA) 1 proteins (1,2). Bacterial HtrAs have a dual function, acting as chaperones at normal temperatures and as proteases at elevated temperatures and removing damaged or denatured proteins, allowing the recovery and survival of bacteria following stress (3,4). Mammalian Omi/HtrA is a ubiquitous protein, although tissue-specific alternatively spliced forms have been reported (2,5). Recent studies (6 -10) described Omi as a mitochondrial protein that upon induction of apoptosis is released to the cytoplasm where it binds XIAP (X chromosome-linked inhibitor of apoptosis protein) resulting in caspase-9 activation. In this regard, Omi resembles Smac/DIABLO, which also binds IAPs and as a result activates caspase-9 (11,12). Omi protein is synthesized as a precursor that is processed in the mitochondria to produce the mature protein. This processing exposes an internal tetrapeptide motif, AVPS, at the amino terminus. A similar motif is found in all IAP-binding proteins, including the Drosophila Grim, Reaper, Hid, and Sickle (13)(14)(15)(16). Processing of the precursor Omi polypeptide is an intramolecular reaction and requires an intact protease domain. 2 Omi can also induce apoptosis in a caspase-independent pathway that relies entirely on its ability to function as a serine protease (6,7,9). Neither the mechanism nor the significance of this function of Omi is clearly understood. In order to investigate the caspaseindependent mechanism of apoptosis initiated by Omi, we isolated specific inhibitors of its proteolytic activity. One such family of synthetic compounds was identified and is described in this paper. These heterocyclic compounds showed significant and specific activity against the protease in an in vitro assay. A compound from this family, which exhibited the highest activity against Omi, was selected for more detailed studies. This compound (ucf-101) naturally fluoresces and easily enters mammalian cells, allowing its use in in vivo experiments. Our results clearly showed ucf-101 had a profound effect on the activity of Omi and could substantially inhibit its ability to induce caspase-independent apoptosis in caspase-9 (Ϫ/Ϫ) null fibroblasts (17).
The ucf-101 inhibitor can be used as a tool to dissect the two different activities (caspase-dependent versus caspase-independent) of Omi and their respective contribution to apoptosis in various biological settings. Furthermore, ucf-101, or a similar compound, may be useful as an anti-apoptotic drug that would specifically target caspase-independent cell death under clinical conditions.

EXPERIMENTAL PROCEDURES
Preparation of FITC-labeled, Dephosphorylated Casein-Dephosphorylated-casein (Sigma, 3 ml of 2 mg/ml stock solution) was used in a dialysis cassette (Pierce) placed in labeling buffer (0.03 mg/ml FITC, 50 mM sodium borate, pH 9.3, 40 mM NaCl) and kept for 48 h at 4°C in the dark with constant stirring. After labeling, the FITC-dephosphorylated casein (Dcasein-FITC) was further dialyzed against 50 mM Tris-HCl, pH 7.5, and 50 mM NaCl to remove any residual unlabeled FITC. The Dcasein-FITC conjugate was recovered from the dialysis cassette and stored at 4°C in the dark.
Expression and Purification of MBP-Omi and MBP-L56 Proteases-PCR was used to amplify DNA sequences corresponding to Omi/HtrA2 (amino acids 134 -458) and L56/HtrA1 (amino acids 140 -480). The PCR products were digested with MfeI and XbaI and cloned in the * This work was supported by National Institutes of Health Grants R01 DK55734-01 (to A. S. Z.) and R01 AG13487 (to E. S. A.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
corresponding EcoRI and XbaI restriction sites of pMAL-p2X vector (New England Biolabs). MBP-Omi-(134 -458) and MBP-L56-(140 -480) proteases were expressed in TB1 Escherichia coli (New England Biolabs) and purified on an amylose binding affinity column as described by the manufacturer (New England Biolabs). The concentration of purified proteases was determined using the Bradford assay.
Expression and Purification of His-tagged Omi-(134 -458)-PCR was used to amplify the DNA sequence corresponding to Omi/HtrA2 (Omi-(134 -458)). The PCR product was cloned in-frame in the bacterial expression vector pET-28 (Novagen). For bacterial expression, BL21 (DE3) (Novagen) bacteria were transformed with pET-Omi-(134 -458). Single colonies were grown overnight in LB medium containing kanamycin. The overnight culture (1 ml) was used to inoculate 1 liter of LB medium, and growth was continued at 37°C until the A 600 was ϳ0.8. At this time, 1 mM isopropyl-1-thio-␤-D-galactopyranoside was added, and the culture was placed in a shaking incubator overnight at 20°C. Bacteria were harvested by centrifugation and lysed in a buffer containing 20 mM Tris-HCl, pH 7.5, 500 mM NaCl, 100 g/ml lysozyme, and a protease inhibitor mixture (Sigma). The bacterial suspension was then sonicated, and the soluble fraction was purified using nickelnitrilotriacetic acid His-Bind resin (Novagen). The purity of the His-Omi-(134 -458) protein was monitored by SDS-PAGE and Coomassie Blue staining of the resulting gel.
Proteolytic Activity of Omi Using Dcasein-FITC as Substrate-Opaque microtiter plates (Corning Glass) were used in order to minimize background absorbance. A typical assay included 10 g of Dcasein-FITC substrate, 2 g of MBP-Omi-(134 -458) in assay buffer (20 mM Na 2 HPO 4 /NaH 2 PO 4 , pH 7.4, 200 mM NaCl, 5% glycerol) in a final volume of 100 l. Dcasein-FITC conjugate solution (50 l of a 0.2 g/l stock solution) was added in each well using a Multidrop 384 multiple dispenser (Labsystems) and incubated in the Wallac 1420 Victor 2 Multilabel Counter at 37°C for 15 min. After this time, 50 l of MBP-Omi-(134 -458) protease (40 ng/l) was added to each well. The fluorescence (535 nm) of the reactions was recorded every 5 min for a 30-min period.
Combinatorial Library Screening-A Pharma Library Collection (Nanosyn) was screened for compounds, which may inhibit the proteolytic activity of MBP-Omi-(134 -458) in an in vitro assay. Each compound, 10 M final concentration, was preincubated with 50 l of MBP-Omi-(134 -458) (2 g) for 15 min at 37°C. After this time, Dcasein-FITC solution (10 g) was added, and the change in fluorescence was read every 5 min during a 30-min period using a 535-nm bandpass filter. The proteolytic activity of MBP-Omi-(134 -458) was expressed as the percentage inhibition, where 100% refers to the activity in the absence of inhibitor (Me 2 SO replaced the inhibitor in the assay).
Determination of IC 50 for ucf-101 and ucf-102-The IC 50 value for each of the two selected inhibitors was obtained using His-Omi-(134 -458) and the EnzCheck assay kit (Molecular Probes) that contains BODIPY FL-casein as a generic substrate. Briefly, 35 l of His-Omi-(134 -458) (500 nM) diluted in 20 mM Na 2 HPO 4 /NaH 2 PO 4 , pH 7.4, 200 mM NaCl, 5% glycerol was incubated with 5 l of various concentrations of the inhibitors (0.1-1000 M) in 100% Me 2 SO (final concentration of Me 2 SO was 10%); 10 l of BODIPY-FL casein (2.5 M final concentration) diluted in buffer was added. The assay was carried out using 384-well microtiter plates. Fluorescence was monitored continuously for 10 min at 37°C on a Tecan SpectraFluorPlus microtiter plate reader (Tecan, Crailsheim, Germany) at an excitation wavelength 485 nm/ emission wavelength 530 nm. IC 50 values were calculated using the GraFit 4 program (Erithacus Software, Middlesex, UK).
Activity of ucf-101 against Various Serine Proteases-The amount of the inhibitor giving a 50% (IC 50 ) decrease of the enzyme activity compared with the control reaction was estimated for various serine proteases. Briefly, 35 l of the indicated enzyme diluted in the corresponding buffer was incubated for 10 min with 5 l of various concentrations of inhibitor (0.1-1000 M) in 50% Me 2 SO, 50% buffer. After this time, 10 l of the corresponding peptidic colorimetric or fluorogenic substrate diluted in buffer was added. In the case of the FVIIa/TF assay, a preincubation of both proteins for 10 min at room temperature was performed prior to addition of the inhibitor to allow complex formation. For this experiment the following materials were used: canine FXa, rat FXa, and rabbit FXa (Enzyme Research Laboratories); trypsin from bovine pancreas and human thrombin (Sigma); human plasmin, human t-PA, human u-PA, and human kallikrein (Calbiochem-Novabiochem);  I Screening for inhibitors of the proteolytic activity of Omi A combinatorial library of 528 selected compounds was used in a high throughout screening for potential inhibitors of the proteolytic activity of Omi. Several compounds had weak activity against Omi, but after two rounds of selection only one chemical showed substantial and reproducible inhibition against Omi in the assay used. Subcellular Localization of ucf-101-HeLa cells were grown on coverslips using F-12 (Ham's) nutrient mixture (Invitrogen) supplemented with 10% fetal calf serum (Sigma), 2 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin (Invitrogen). Different concentrations of ucf-101 were added to the cell media, and after 2 h cells were washed three times with phosphate-buffered saline and fixed with 4% paraformaldehyde. The coverslips were then placed on glass slides using Fluoromount-G solution (Southern Biochemical Association). The subcellular localization of ucf-101 was monitored using a LSM510 confocal laser-scanning microscope (Zeiss).

Activity of Omi Protease against Dcasein-FITC-Because no
physiological substrates for Omi protease are known, ␤-casein was used as a generic substrate in in vitro assays to monitor the protease activity of Omi. For high throughput screening, we developed a new assay that uses a mixture of ␣-casein, ␤-casein, and dephosphorylated casein (Dcasein) coupled to FITC, as a substrate. The use of Dcasein-FITC increases the fluorescence during the reaction with less fluctuation and lower background (data not shown). A typical assay uses 10 g of Dcasein-FITC and 2 g of MBP-Omi-(134 -458) in 100 l of reaction buffer (20 mM sodium phosphate buffer, pH 7.5, 200 mM NaCl, and 5% glycerol) at 37°C for 30 min. There was a rapid increase in fluorescence for the first 20 min followed by a slower  Bovine-trypsin Ͼ500 Ͼ500 Hu-plasmin Ͼ500 400 Hu-thrombin Ͼ500 Ͼ500 Hu-rec.-FVIIa Ͼ500 Ͼ500 Hu-t-PA 390 390 Hu-u-PA 500 500 Hu-APC 500 500 increase for the remaining 10 min of the assay. Various concentrations of substrate, Dcasein-FITC, as well as MBP-Omi-(134 -458) were used for kinetic studies. Combinatorial Library Screening-By using an in vitro high throughput assay system, we screened a combinatorial library of synthetic compounds (Nanosyn). This collection represented compounds that had commonly accepted pharmaceutical hit structures with possible pharmacological properties. The assay was performed in multiple 96-well plates. The inhibition activity of each compound was expressed as the percentage of relative fluorescence change (decrease) compared with the control (no compound). The final concentration of each compound tested was 10 M. Fifty two compounds showed more than a 20% inhibition of MBP-Omi-(134 -458) activity in the initial screening, but only one compound (ucf-101) showed greater than a 50% inhibition at the concentration tested after two rounds of screening and selection ( Table I).
Inhibition of the Proteolytic Activity of MBP-Omi-(134 -458) Using ucf-101 and Various Analogs-We obtained several analogs of ucf-101, and the chemical structures are shown in Fig.  1. The activity of these ucf analogs was tested using MBP-Omi-(134 -458) in the same in vitro assay system. All three analogs inhibited the proteolytic activity of MBP-Omi-(134 -458) but to a lesser extent than ucf-101 (Fig. 2). ucf-104 (20 M) inhibited 58% MBP-Omi-(134 -458) activity, whereas the same concentration of ucf-101 inhibited more than 78% MBP-Omi-(134 -458) activity. When MBP-Omi-(134 -458) was preincubated with various concentrations of ucf-101, -102, -103, and -104 at 37°C for longer periods (40 min or 1 h) inhibition was reduced, indicating the inhibitors did not irreversibly bind the enzyme (results not shown). MBP-Omi-(134 -458) was also incubated with various amounts of Dcasein-FITC in the presence of 10 M ucf-101. The rate of the reaction decreased by 50% when the concentration of the substrate was doubled and even reached the rate of the control reaction at higher concentrations of Dcasein-FITC (results not shown). These results suggest ucf-101 is a competitive inhibitor of Omi protease.
Assay of His-Omi-(134 -458) Activity Using Unlabeled ␤-Casein and SDS-PAGE Analysis-The inhibitory effect of ucf-101 on the activity of His-Omi-(134 -458) was monitored by incubating inhibitor and enzyme together for 10 min at room temperature prior to the addition of ␤-casein as a generic substrate. Fig. 3 shows that bacterially made His-Omi-(134 -458) had substantial activity against ␤-casein, and after 30 min total degradation of the substrate occurred. ucf-101 inhibited HtrA1 are members of the same family of mammalian serine proteases with extensive homology in their catalytic domains (1). We used bacterially made His-Omi-(134 -458) and His-L56-(140 -480) to investigate whether ucf-101 was able to inhibit the activity of His-L56-(140 -480) in a similar manner. Fig. 4 shows ucf-101 is an effective inhibitor of the activity of His-L56-(140 -480) but to a lesser extent than His-Omi-(134 -458).
IC 50 of ucf-101 against Various Serine Proteases-The enzymatic activity of His-Omi and the determination of the IC 50 for ucf-101 and ucf-102 were performed as described under "Experimental Procedures." These values are shown in Table II. ucf-101 had an IC 50 of 9.5 M, and ucf-102 had an IC 50 of 45.9 M.
The specificity of ucf-101 was determined using several unrelated serine proteases. Their susceptibility to inhibition by ucf-101 was tested. The IC 50 values for ucf-101 inhibitor from these experiments are shown in Table III. These results suggest ucf-101 has very high specificity for the Omi protease.
Intracellular Localization of ucf-101-To investigate the potential use of ucf-101 inhibitor in in vivo experiments, we tested its ability to enter mammalian cells. ucf-101 has natural fluorescence at 543 nm that was used to detect the presence of the compound. HeLa cells were treated with different concentrations of ucf-101 and observed by confocal microscopy. Intense red fluorescence, due to the presence of ucf-101, was observed in the cytoplasm of the treated cells (Fig. 5) ucf-101 Can Inhibit Omi-induced Caspase-independent Apoptosis-Mouse embryo caspase-9 (Ϫ/Ϫ) null fibroblasts (17) were transiently transfected with pEGFP-N1 (control) or M-Omi-GFP (7). Cytoplasmic expression of M-Omi-GFP induced apoptosis in ϳ70% of the transfected cells. When cells were treated with increasing concentrations of ucf-101, it was found that increasing the concentration of inhibitor up to 10 M gradually reduced Omi-induced apoptosis up to 40%. ucf-101 had no effect on the control (vector alone) transfected cells. When the concentration of the ucf-101 increased to 25 M, the anti-apoptotic activity of the inhibitor was compromised due to a cytotoxic side effect that caused apoptosis both in the M-Omi-GFP, as well as control transfected cells (Fig. 6).

DISCUSSION
Omi/HtrA2 is a mitochondrial serine protease that is released to the cytoplasm upon induction of apoptosis. In the cytoplasm, Omi binds to XIAP and relieves its inhibition of caspase-9 (6,7,9). In this respect, Omi acts in a manner similar to Smac/DIABLO, another mitochondrial protein that also binds to XIAP (11,12). This interaction with XIAP is mediated via an AVPS motif that is exposed at the amino terminus of mature Omi protein after processing (6,7). Omi is also able to induce apoptosis in a caspase-independent manner that exclusively relies on its ability to function as a protease (6 -10). This caspase-independent pathway is not well understood; its contribution to overall cell death is not known, and the role of Omi as a protease in this pathway is not clear. Furthermore, although the catalytic domain of all HtrAs including Omi protein is conserved from E. coli to humans, the AVPS tetrapeptide is replaced at least in bovine Omi by SVLG (18). This suggests the IAP binding activity of human HtrA2 that leads to caspase-dependent apoptosis may not be critical to the overall function of Omi. Because the caspase-independent pathway relies on the ability of Omi to function as a protease, it suggests that proteolytic cleavage of specific proteins is involved. This cleavage might inactivate and remove apoptotic inhibitors, or it may activate precursor proteins whose function might be necessary for caspase-independent cell death. In order to investigate the contribution of the proteolytic activity of Omi to its overall pro-apoptotic function, we decided to screen for specific inhibitor(s) of its activity. One such specific inhibitor was identified and called ucf-101. This heterocyclic compound was able to inhibit the proteolytic activity of Omi in vitro in a very specific and reversible manner. The specificity of ucf-101 was tested against a panel of several unrelated serine proteases, and no significant activity was detected. When L56/HtrA1 was used in the same assay, ucf-101 showed specific inhibition against this protease. L56/HtrA1 belongs to the same family of proteases as Omi, and they both share extensive homology throughout their respective catalytic domains (1). The normal function of L56/ HtrA2 is not known, and unlike Omi that localizes to mitochondria, L56 is secreted (19). The natural fluorescence of ucf-101 was used to monitor its ability to enter mammalian cells. This property is essential for ucf-101 to be useful for in vivo experiments designed to test inhibition of Omi that is intracellular. To test the ability of ucf-101 to inhibit the proteolytic activity of Omi in vivo, we used an assay where transient overexpression of a cytoplasmic form of Omi induces caspase-independent cell death (7). The cells used in this assay were caspase-9 (Ϫ/Ϫ) null mouse embryonic fibroblasts (17) in which overexpression of a cytoplasmic form of Omi can induce the caspase-independent pathway of apoptosis. This form of cell death caused by Omi relies entirely on its ability to function as a protease. Therefore, if ucf-101 blocks the activity of Omi, it will interfere with apoptosis of the caspase-9 (Ϫ/Ϫ) mouse embryonic fibroblasts. When these fibroblasts were transfected with the vector encoding the cytoplasmic Omi and treated with ucf-101, apoptosis was dramatically reduced. ucf-101, like many known protease inhibitors, was cytotoxic to mammalian cells when used at high concentrations. Therefore, the ability of ucf-101 to inhibit caspase-independent apoptosis at low doses was compromised by its cytotoxicity at higher doses. This toxicity was seen only at high concentrations of the chemical and does not exclude its use in in vivo experiments where it exhibits activity against Omi at much lower concentrations. ucf-101 will be a useful tool to elucidate the role of Omi in caspase-independent cell death. Several clinical conditions, such as neurodegenerative diseases, show excessive caspase-independent cell death (20,21). If Omi through its proteolytic activity plays a role in the pathogenesis of any of these clinical conditions, ucf-101 or a similar compound would be useful for therapeutic intervention against untimely or excessive cell death.