Inhibitor of Apoptosis (IAP)-like Protein Lacks a Baculovirus IAP Repeat (BIR) Domain and Attenuates Cell Death in Plant and Animal Systems*

A novel Arabidopsis thaliana inhibitor of apoptosis was identified by sequence homology to other known inhibitor of apoptosis (IAP) proteins. Arabidopsis IAP-like protein (AtILP) contained a C-terminal RING finger domain but lacked a baculovirus IAP repeat (BIR) domain, which is essential for anti-apoptotic activity in other IAP family members. The expression of AtILP in HeLa cells conferred resistance against tumor necrosis factor (TNF)-α/ActD-induced apoptosis through the inactivation of caspase activity. In contrast to the C-terminal RING domain of AtILP, which did not inhibit the activity of caspase-3, the N-terminal region, despite displaying no homology to known BIR domains, potently inhibited the activity of caspase-3 in vitro and blocked TNF-α/ActD-induced apoptosis. The anti-apoptotic activity of the AtILP N-terminal domain observed in plants was reproduced in an animal system. Transgenic Arabidopsis lines overexpressing AtILP exhibited anti-apoptotic activity when challenged with the fungal toxin fumonisin B1, an agent that induces apoptosis-like cell death in plants. In AtIPL transgenic plants, suppression of cell death was accompanied by inhibition of caspase activation and DNA fragmentation. Overexpression of AtILP also attenuated effector protein-induced cell death and increased the growth of an avirulent bacterial pathogen. The current results demonstrated the existence of a novel plant IAP-like protein that prevents caspase activation in Arabidopsis and showed that a plant anti-apoptosis gene functions similarly in plant and animal systems.


A novel Arabidopsis thaliana inhibitor of apoptosis was identified by sequence homology to other known inhibitor of apoptosis (IAP) proteins. Arabidopsis IAP-like protein (AtILP) contained a C-terminal RING finger domain but lacked a baculovirus IAP repeat (BIR) domain, which is essential for antiapoptotic activity in other IAP family members. The expression of AtILP in HeLa cells conferred resistance against tumor necrosis factor (TNF)-␣/ActD-induced apoptosis through the inactivation of caspase activity. In contrast to the C-terminal RING domain of AtILP, which did not inhibit the activity of caspase-3, the N-terminal region, despite displaying no homology to known BIR domains, potently inhibited the activity of caspase-3 in vitro and blocked TNF-␣/ActD
All living organisms use a process of cell suicide to achieve and maintain homeostasis during normal development as well as in response to environmental stress or during pathogen challenge (1). This functionally conserved process, known as pro-grammed cell death (PCD) 5 or apoptosis, is genetically regulated and associated with distinct morphological and biochemical characteristics. Extensive study over the past decade has illuminated the biological and molecular mechanisms of the regulation of apoptosis in animal systems (2)(3)(4)(5)(6)(7). Apoptosis is triggered by the sequential activation of cysteine proteases known as caspases, which results in protein cleavage and the breakdown of DNA molecules. This apoptotic cascade is regulated by both initiators and inhibitors and can be activated by diverse stimuli. Caspases are synthesized as zymogens that are activated by proteolytic cleavage at specific aspartic acid residues in the P1 position (8). Compartmentalization of caspases and their cofactors suggests that two major apoptotic pathways exist. One pathway of apoptosis, observed in animal systems, can be induced by the deprivation of serum from tissue culture cells, leading to the release of cytochrome c from mitochondria. Apoptosis activating factor-1 (Apaf1) and cytochrome c form a complex with procaspase-9, which is then activated. Active caspase-9 triggers the common caspase cascade by cleaving procaspase-3 (9 -11). Caspase-3 is responsible either wholly or in part for the proteolytic cleavage of many key proteins, including poly(ADP-ribose) polymerase and lamin A (12)(13)(14). The existence of another apoptosis pathway derives from the observation that caspase-8 is activated when challenged with tissue necrosis factor (TNF-␣) or Fas ligand (15)(16)(17)(18). Loss of caspase activity is observed in cells that express the viral proteins CrmA, from cowpox, and p35, from baculovirus (19 -23). Furthermore, overexpression of these viral caspase inhibitors in insect, nematode, and mammalian cells results in resistance to apoptosis, providing evidence that the components of the apoptotic pathway are highly conserved throughout evolution. This has led to speculation that functional equivalents of these viral proteins may exist in higher organisms.
The inhibitor of apoptosis (IAP) family of proteins plays a central role in apoptotic and inflammatory processes, conferring protection against cell death. IAP family members inter-fere with the transmission of intracellular death signals by inhibiting caspase-dependent apoptotic pathways. The IAP proteins were initially identified in baculovirus as factors that prevented host cell apoptosis, allowing time for the virus to replicate (24,25). Since then, eight mammalian IAPs (XIAP,  HIAP1, HIAP2, ILP2, MLIAP, NAIP, BRUCE, and survivin) and  three Drosophila IAP homologs (DIAP1, DIAP2, and Deterin) have been identified (26 -35). IAP proteins exhibit a modular structure characterized by the presence of one or more baculovirus IAP repeat (BIR) domains. The BIR domain is a zincbinding fold of ϳ70 amino acid residues that is essential for the anti-apoptotic properties of IAP proteins. The fact that all known IAP members have a BIR domain suggests that this domain plays a pivotal role in mediating cellular protection. In addition, with the exception of NAIP, all known IAP family members also contain a RING domain in their C terminus, defined by seven cysteine residues and one histidine residue that together coordinate two zinc atoms (36,37). The RING domain confers E3 ubiquitin ligase activity and has been suggested to play a role in apoptosis regulation by directing the ubiquitination of target proteins for degradation by the proteasome (38 -40). The RING domain is not essential for apoptosis inhibition by human IAP family members, which suggests that the BIR domain is sufficient to protect cells from apoptosis (41)(42)(43).
The genes that control PCD are functionally conserved across wide evolutionary distances (44 -46). For example, homologues of the mammalian Bax-inducible cell death inhibitor BI-1 have been identified in several plants, including Arabidopsis, rice, tobacco, and barley (47)(48)(49)(50). In addition, animal apoptotic regulators, such as human Bcl-2 and Bcl-xl as well as nematode CED-9, can either induce or suppress cell death in transgenic plants (51)(52)(53). In plants, PCD occurs during developmental processes, such as flower development, embryogenesis, seed germination, and vessel and trachea formation. Of note, PCD is crucial for a plant defense response termed hypersensitive response (HR), which serves to restrict the spread of pathogens through the process of PCD (54,55). Studies in plant systems have shown that the biochemical and morphological hallmarks of apoptosis, such as cytoplasmic shrinkage, nuclear condensation, and DNA laddering, are similar in animal and plant cells (56 -58). The cytosolic caspase-mediated apoptotic pathway is well defined in animal cells but has yet to be demonstrated in plant cells. However, evidence from recent studies has suggested that there are some similarities between plant apoptosis and caspase-mediated apoptosis in animal cells, with the exception of the presence of IAP-like proteins. For example, in tobacco cells, caspase-1-like proteases participate in HR, and the presence and subcellular localization of caspase-3-like proteases in barley has been reported (59 -62).
In the current study, we identified and characterized a novel Arabidopsis gene, AtILP (for Arabidopsis thaliana IAP-like protein), which encodes a RING finger protein with homology to mammal IAPs. The expression of AtILP efficiently suppressed apoptosis induced by TNF-␣/ActD and the fungal toxin fumonisin B1 (FB1) by blocking the activation of caspases in HeLa cells. Interestingly, despite lacking a BIR domain, an N-terminal fragment of AtILP conferred anti-apoptotic activity in Arabidopsis. Overexpression of the N-terminal domain of AtILP resulted in the suppression of FB1-induced cell death and attenuated cell death caused by the bacterial effector AvrRpt2. These results suggested that AtILP may act as a negative regulator of PCD in Arabidopsis.
Cell Culture and Cell Viability Assay-Human cervical epitheloid carcinoma (HeLa) cells were purchased from American Type Culture Collection (ATCC). HeLa cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Invitrogen), 2 mM L-glutamine, 100 units/ml penicillin, and 100 units/ml streptomycin in a humidified CO 2 incubator. Cells were transfected with the indicated expression vectors using Lipofectamine (63). Stable transfectants were selected in the presence of G418 (800 g/ml).
Cell viability was determined by the crystal violet staining method. Briefly, HeLa cells plated in a 12-well dish were exposed to TNF-␣ (100 ng/ml)/ActD (100 ng/ml). Cells were stained with a solution of 0.5% crystal violet in 30% ethanol and 3% formaldehyde for 10 min at room temperature, after which the plates were washed three times with tap water. After drying, cells were lysed in 1% SDS, and dye uptake was measured at 550 nm using a 96-well plate reader. Cell viability was calculated as dye intensity relative to untreated samples.
DEVDase Activity Assay-Cell pellets were washed with icecold PBS and then resuspended in 100 mM HEPES buffer (pH 7.4) containing protease inhibitors (5 mg/ml aprotinin and pepstatin, 10 mg/ml leupeptin, and 0.5 mM phenylmethylsulfonyl fluoride). The cell suspension was lysed by three freeze-thaw cycles, and then the cytosolic fraction was obtained by centrifugation at 100,000 ϫ g for 1 h at 4°C. DEVDase activity was evaluated by measuring proteolytic cleavage of the chromogenic substrate Ac-DEVD-pNA, which serves as a substrate for caspase-3-like proteases. Briefly, cell lysate (40 g of protein) was mixed with 150 l of reaction buffer containing Ac-DEVD-pNA (240 M) in a 96-well plate. The reaction mixture was incubated at 37°C for 90 min. The increase in enzymatically released pNA was measured every 15 min by absorbance at 405 nm; DEVDase activity was calculated from initial velocity.
For measuring DEVDase activity assay in plants, leaves were ground and homogenized in caspase extraction buffer (50 mM HEPES (pH 7.5), 1 mM EDTA, 1 mM DTT, 1% BSA, 1 mM PMSF, 20% glycerol). Samples were mixed with 50 l of caspase assay buffer (caspase extraction buffer containing 150 M Ac-DEVD-pNA) and then incubated at 37°C for 1 h. The increase in enzymatically released pNA was measured every 15 min by absorbance at 405 nm; DEVDase activity was calculated from the initial velocity.
Plants-A. thaliana seedlings were germinated on MS medium containing 2% sucrose and 0.6% Phytagel and maintained in a temperature-and light-controlled growth chamber. Arabidopsis seedlings were grown for 14 days before being transferred to fresh MS plates or to fresh MS plates supplemented with FB1.
For the DNA fragmentation assay, 10 g of genomic DNA was separated by electrophoresis on a 0.8% agarose, 0.6% Meta-Phor-agarose gel and then transferred to a Hybond membrane. As a probe, 50 ng of total genomic Arabidopsis DNA was labeled using a commercially available random labeling kit. Following hybridization, the membrane was washed with 0.1ϫ SSC, 0.1% SDS at 65°C for 2 h.
Bacteria-Bacterial strains were grown at 28°C on KB medium containing the appropriate antibiotics for selection. For assessing ion leakage and to score HR phenotype, plants were infiltrated with 10 7 cfu/ml (A 600 ϭ 0.2) of Pseudomonas syringae pv. phaseolicola (Pph) strain NPS3121 using a needleless 1-ml syringe (see Table 1 and Fig. 6). Pph strain NPS3121 harboring AvrRpt2 was used for the ion leakage and cell death assay. For ion leakage measurements, eight leaf discs (8 mm in diameter) were removed immediately following infiltration (t ϭ 0) and allowed to float in 40 ml of water. After 30 min, the wash water was replaced with 10 ml of fresh water, and then conductance over time was measured using a Fisher brand conductivity meter.
For growth experiments using P. syringae pv. maculicola (Pma) strain M6C⌬E (64) harboring empty vector (pVSP61) or its derivative encoding AvrRpt2 (Fig. 7), the leaves of 5-weekold plants were inoculated with bacterial suspensions in 10 mM MgCl 2 using a needleless 1-ml syringe. After the indicated periods of time, three leaf discs for each sample were ground in 10 mM MgCl 2 and then serially diluted and plated to determine bacterial number.
Subcellular Localization of AtILP Fusion Protein-PCR was used to generate a cDNA fragment encoding full-length AtILP. The cDNA fragment was digested with XbaI and BamHI and then ligated in-frame with soluble modified green fluorescent protein (smGFP) to create AtILP::smGFP. The AtILP::smGFP fusion construct was introduced into Arabidopsis protoplasts using polyethylene glycol-mediated transformation. The expression of red fluorescent protein fused to a nuclear localization signal (RFP::NLS) was used as a positive control for nuclear localization. Transformed protoplasts were incubated at 22°C in the dark. Expression of fusion protein was observed 2 days after transformation by fluorescence microscopy (Olympus AX70) using standard FITC and rhodamine filters.

Identification of an Apoptosis Inhibitor in Arabidopsis and
Demonstration of Anti-apoptotic Activity in Animal Cells-Some aspects of the signaling mechanisms that control apoptosis, including IAP family members, are functionally conserved across wide evolutionary distances. HIAP1 and HIAP2 are functional anti-apoptotic proteins in Homo sapiens (65)(66)(67). To determine whether higher plants carry HIAP-like proteins, homology searches against the Arabidopsis genome sequence database were performed using the sequences of HIAP1 and -2 as the queries. The searches yielded one gene, At4g19700, encoding a putative protein with significant similarity to other IAPs. In particular, the protein contained a RING domain in its C terminus. This protein was named AtILP, for A. thaliana IAP-like protein. The full-length AtILP cDNA was isolated from an Arabidopsis cDNA library. It consisted of 915 nucleotides encoding a putative open reading frame of 305 amino acids (Fig. 1A). Amino acid sequence alignment of the RING domain of AtILP with human HIAP1, HIAP2, XIAP, and KIAP showed that AtILP encodes a perfect C-terminal C3HC4 signature (Fig. 1B). Aside from the highly conserved RING domain, AtILP did not appear to encode any other known conserved domains. IAP proteins are characterized by the presence of one or more BIR domains, a structurally distinct, zinc finger fold domain composed of ϳ70 amino acid residues. It is widely acknowledged that the BIR domain is essential for the antiapoptotic properties of the IAP proteins in animal systems. To determine whether AtILP possessed anti-apoptotic activity, despite not having a BIR domain, HeLa cells were transfected with expression vectors for AtILP or Gpx or empty vector (pcDNA) as a control using Lipofectamine (63), and the response to TNF-␣/ActD-induced cell death was analyzed. Gpx was used as a positive control for apoptosis inhibition (68,69). As shown in Fig. 2, TNF-␣/ActD-induced cell death was considerably reduced in cells expressing AtILP, even more so than in GPx-expressing cells. The viability of AtILP-expressing cells exceeded 85%, whereas that of GPx-expressing cells was ϳ55%. These results indicated that AtILP is a RING finger protein with structural and possibly functional homology to human IAPs and that a gene involved in apoptosis inhibition in plants functions in a similar manner in an animal system.
The N-terminal Domain of AtILP Blocks TNF-␣/ActD-induced Caspase Activation-To define the molecular determinants of AtILP anti-apoptotic activity, four different AtILP protein fragments were constructed (Fig. 3B). HeLa cells were transfected with expression vectors for full-length AtILP or one of the AtILP fragments (fragment a, b, c, or d), and then antiapoptotic activity was measured. Because AtILP did not have a BIR domain, and computer-based sequence homology searches revealed no other similarities with other IAP proteins, we initially expected that the functional domain would map to the C-terminal RING domain. As shown in Fig. 3A, cells transfected with empty vector or fragments c and d underwent cell death in response to TNF-␣/ActD. In contrast, transfection expression vectors for full-length AtIPL, fragment a, or fragment b significantly reduced TNF-␣/ActD-induced apoptosis. Fragments a and b retained ϳ75% of the inhibitory activity of the full-length protein, whereas the anti-apoptotic activity of fragments c and d, which contained the C-terminal RING domain, were comparable with control conditions (Fig. 3B). The various AtILP fragments are depicted schematically in Fig.   3B. Full-length AtILP and the AtILP fragments were all stably expressed in HeLa cells (Fig. 3C). These results indicated that fragment b, which contained the N-terminal 150 amino acid residues of AtILP, contains the main determinant(s) of antiapoptotic activity.
Because caspases are critical mediators of apoptosis, we next examined whether caspase inactivation played a role in the anti-apoptotic activity of AtILP. DEVDase activity was evaluated by measuring the proteolytic cleavage of a chromogenic substrate, Ac-DEVD-pNA, which serves as a substrate of caspase-3-like proteases. As seen in Fig. 4, the inhibitory effects of full-length AtILP and each of the AtILP fragments on DEV-Dase inactivation correlated with the results of the cell viability assay. These data clearly suggested that the activity of AtILP in inhibiting cell death is mediated by the N-terminal domain through the suppression of caspase activation (Fig. 4).
The N-terminal Domain of AtILP Confers Resistance to FB1induced Apoptosis in Arabidopsis-To evaluate the role of AtILP in apoptosis inhibition in plants, transgenic Arabidopsis   DECEMBER 9, 2011 • VOLUME 286 • NUMBER 49 lines that constitutively expressed full-length AtILP or the N-terminal (amino acids 1-150) or C-terminal (amino acids 151-304) domain of AtILP under the control of the cauliflower mosaic virus (CaMV) 35S promoter were generated. Several transgenic plants that exhibited high levels of expression of fulllength, N-terminal, or C-terminal AtILP were selected for further analysis (supplemental Fig. 1). The N-terminal and C-terminal domains consisted of 150 and 154 amino acids, respectively. A striking example of plant apoptosis, HR is a cell death program triggered in host cells at or around the site of pathogen infection, resulting in cellular collapse and the formation of necrotic lesions (70). Because it is well known that the fungal toxin FB1 induces HR in plants (56,71,72), we examined the effect of the overexpression of full-length AtILP or the N-or C-terminal domain on FB1-induced HR in Arabidopsis. Wildtype Arabidopsis ecotype Col-0 and transgenic Arabidopsis plants were grown for 2 weeks on MS agar medium, transferred to MS medium containing 3 M FB1, and then observed for morphological changes 4 days after transfer. As shown in Fig.  5A, the leaves of wild-type and transgenic plants harboring the C-terminal fragment of AtILP were completely macerated, and death lesions were readily apparent. In contrast, transgenic plants expressing full-length AtILP or the N-terminal domain exhibited some lesions in the upper leaves but overall were highly resistant to FB1-induced cell death compared with wildtype and C-terminal domain transgenic plants.

Attenuation of Cell Death by BIR-absent IAP-like Protein
Caspase-like activity and a role for caspase-like proteases in HR have been reported in plants, and HR can be prevented through the inhibition of caspase-like proteases (59,73). To determine whether the anti-apoptotic activity of AtILP in plants exposed to FB1 was mediated by caspase-like protease inactivation, as was seen in HeLa cells (Fig. 4), protein extracts from wild-type and transgenic plants were prepared. As seen in Fig. 5B, caspase inactivation correlated with the ability of fulllength AtILP and the N-and C-terminal domains to suppress FB1-induced apoptosis. Treatment with FB1 induced the activation of caspase-like proteases in wild-type and C-terminal domain transgenic plants. In contrast, the overexpression of full-length AtILP or the N-terminal domain effectively suppressed caspase-like protease activation (Fig. 5B). These results suggested that the isolated N-terminal domain of AtILP can prevent plant cell death by suppressing caspase-like protease activation.

Effect of AtILP on the Interaction between Arabidopsis and the
Bacterial Pathogen P. syringae-It was next examined whether the expression of AtILP altered effector protein-induced HR and the associated cell death. Gram-negative plant pathogenic bacteria secrete a complex set of effectors, making it difficult to detect changes in HR induced by a single effector protein. To overcome this limitation, a strain of Gram-negative phytopathogenic bacteria, Pph strain NPS3121, was used that expressed the avirulent gene AvrRpt2. Using this strain, it was possible to measure HR and electrolyte leakage in response to an avirulent bacterial pathogen. The leaves of 6-week-old Arabidopsis plants (wild-type and AtILP transgenic lines) were infiltrated with Pph strain NPS3121 expressing AvrRpt2 (Pph (AvrRpt2)) at a dose of 10 7 cfu/ml (see "Experimental Procedures"). Within 16 h, most wild-type and transgenic plants overexpressing the C-terminal domain of AtILP exhibited confluent tissue collapse at the site of pathogen infiltration, which is a characteristic feature of HR-associated cell death. However, most of the leaves of the transgenic plants overexpressing fulllength AtILP or the N-terminal domain did not show serious signs of HR, although a small percentage developed a weak HR at 16 h. This weak HR in full-length AtILP and N-terminal domain transgenic plants was restricted to a small area surrounding the point of infiltration and was not confluent. Confluent tissue collapse was observed in most of the inoculated leaves of the transgenic plants by 20 h postinoculation (Table 1).
Electrolyte leakage due to membrane damage as a result of plant-pathogen interaction is a characteristic and quantitative feature of HR-associated cell death (74). To determine whether  the attenuation of HR was related to membrane damage, electrolyte leakage in wild-type and AtILP transgenic plants was measured after Pph (AvrRpt2) infiltration (10 7 cfu/ml). Leaves from wild-type and transgenic plants overexpressing the C-terminal domain of AtILP reached close to maximal conductivity 12-16 h postinoculation. Transgenic plants overexpressing full-length AtILP or the N-terminal domain exhibited a similar pattern, but the magnitude of the response was significantly lower, and maximal conductivity was reached later compared with wild-type or C-terminal domain transgenic plants (Fig. 6). However, a difference in conductivity was not observed when plants were treated with virulent bacterial pathogen Pph (data not shown). These results indicated that the overexpression of the N-terminal domain of AtILP significantly impairs HR-associated cell death elicited by an avirulent bacterial pathogen.
To further explore the role of AtILP in plant defense response, the effect of attenuated cell death on bacterial growth was assessed using the bacterial pathogen Pma strain M6C⌬E. Disease phenotype was assessed following the inoculation of this virulent strain of P. syringae (Pma strain M6C⌬E) into wild-type and AtILP transgenic lines. All of the plants (transgenic and wild type) exhibited visible chlorosis 3-4 days after inoculation, which progressed over time on the infected leaves. The plant lines were indistinguishable in terms of the severity of chlorosis (data not shown). In addition, there were no differences in bacterial titer among wild-type and AtILP transgenic lines (Fig. 7A). These results indicated that the overexpression of AtILP does not alter the defense response to infection with virulent Pma strain M6C⌬E.
To determine whether AtILP-mediated HR attenuation affected the growth of an avirulent strain of Pma, strain M6C⌬E carrying the avirulence gene AvrRpt2 was used as the inoculum. In this case, the overexpression of full-length AtILP or the N-terminal domain resulted in a 30 -40-fold increase in bacterial growth, which indicated that the overexpression of AtILP decreases resistance to avirulent Pma strain M6C⌬E (Fig. 7B).
AtILP Localizes to the Nucleus and Blocks DNA Fragmentation-Genomic DNA fragmentation during the process of PCD occurs as a result of the activation of cell deathspecific endonucleases that cleave nuclear DNA into oligonucleosomal units. Genomic DNA was extracted from wildtype and transgenic plants treated with or without 3 M FB1. As shown in Fig. 8A, in transgenic plants harboring full-length AtILP or the N-terminal domain, FB1-induced DNA fragmentation was inhibited. Given that DNA fragmentation is a hallmark of apoptosis, these results confirmed that AtILP blocks apoptosis in plants and that the N-terminal domain of AtILP is important for this anti-apoptotic activity.
To confirm that AtILP was present in the nucleus to mediate genomic DNA fragmentation, the subcellular localization of AtILP in vivo was analyzed in Arabidopsis. A C-terminal smGFP fusion protein of AtILP was generated and expressed in Arabidopsis protoplasts. Fluorescence microscopy revealed that AtILP localized to the nucleus (Fig. 8B), and there was some overlap with the control protein, RFP::NLS. These results indicated that AtILP is targeted exclusively to the nucleus in plant cells.

DISCUSSION
A number of genes that regulate PCD, both positively and negatively, have been identified; however, the mechanisms that control PCD in plants remain largely unknown. In the current study, a novel Arabidopsis RING finger protein, AtILP, was identified and shown to be a negative regulator of PCD in Arabidopsis. Overexpression of AtILP suppressed effector proteinand FB1-induced cell death. In addition, AtILP blocked TNF-/ ActD-induced cell death via the suppression of caspase activation in HeLa cells, suggesting that the function of AtILP in inhibiting cell death is preserved across species.
To determine the structural basis for the inhibition of apoptosis by AtILP, the effects of various fragments of AtILP on caspase activity in vitro and on apoptosis suppression in HeLa cells were analyzed. The RING domain of AtILP failed to inhibit the activity of caspase-3, whereas an N-terminal fragment that had no homology to any known BIR domain potently inhibited the activity of caspase-3 in vitro and blocked TNF-␣/ActDinduced apoptosis (Figs. 3 and 4). Amino acid sequence alignment with other IAP proteins indicated that AtILP lacks homology to known BIR domains. The secondary structure of AtILP was investigated, and it was found that AtILP and human IAPs share a common motif consisting of three consecutive ␤ strands, an ␣ helix, a ␤ strand, and an ␣ helix (supplemental Fig. 2). One possibility is that these common structural motifs determine the caspase-inhibitory activity of AtILP.  Based on amino acid sequences analysis, AtILP belongs to the family of RING proteins, members of which have diverse biological functions in plants (75). AtILP contained a well conserved RING domain at its C terminus. Overexpression of AtILP in plants resulted in the reduction of cell death in response to an avirulent bacterial pathogen and to low doses of FB1. Most RING finger proteins have enzymatic activities that catalyze reactions within the ubiquitination/26S proteasome protein degradation system (75,76). Many IAP proteins exhibit E3 ubiquitin ligase activity, and the RING domain is critical for biological activity and regulation of PCD (77)(78)(79)(80). In fact, Arabidopsis RING1, which demonstrates E3 ubiquitin ligase activity in vitro, has been implicated in cell death (76). The biochem-ical activity and putative function of another RING domain protein in Arabidopsis, AtHAL1, remain to be elucidated.
Transgenic Arabidopsis lines that overexpressed AtILP demonstrated anti-apoptotic activity when challenged with the fungal toxin FB1. This suppression of cell death was accompanied by the inhibition of caspase activation and DNA fragmentation. The anti-apoptotic activity of AtILP mapped to the N-terminal domain and correlated with the results of similar experiments in HeLa cells. To investigate the role of AtILP in cell death inhibition, T-DNA insertion mutagenesis was carried out, and several AtILP knock-out plant lines were identified and characterized. Mutation of AtILP did not result in any phenotypic differences in terms of germination, flowering, and growth rate as compared with wild-type plants. In addition, plant responses to FB1 and P. syringae pv. tomato DC3000 expressing AvrRpt2 were indistinguishable from wild-type Arabidopsis, indicating that there may be other as yet unidentified genes in Arabidopsis that can compensate for the loss of the cell death inhibition activity of AtILP (data not shown).
Gram-negative plant pathogenic bacteria secrete a complex set of type III effectors directly into host cells via the type III secretion system. For example, the wild-type Pto strain delivers at least 33, and perhaps as many as 50, type III effectors (81,82). Thus, HR in response to bacterial strain Pto is a cumulative effect of multiple effector proteins. As a result, it is almost impossible to detect HR induced by a single effector protein. In the current study, Gram-negative phytopathogenic bacteria Pph strain NPS3121 expressing AvrRpt2 was used. Pph does not trigger HR, which can obscure other defense responses. Pph is a model pathogen that causes halo blight in bean but not in Arabidopsis (83). Thus, the use of this pathogen enabled us to measure the effect of AvrRpt2 on HR and electrolyte leakage.
In many cases, PCD and disease resistance are intricately linked in higher plants (84). During incompatible interactions between plants and bacterial pathogens, HR-associated cell death often triggers the development of plant disease resistance, resulting in the halting of pathogen growth in plant tissues. Cell death, however, can be uncoupled from the resistance response. For example, the Arabidopsis mutant dnd1 (defense no death) is resistant to Pst without HR-associated cell death (85). In the current study, overexpression of AtILP caused a  A, wild-type and transgenic plants were grown on an MS plate with or without 3 M FB1. Genomic DNA was isolated, separated by electrophoresis, and then visualized by staining with ethidium bromide. B, 10-day-old Arabidopsis protoplasts were cotransformed with 10 g of AtILP::smGFP and RFP::NLS expression constructs. RFP::NLS was used as a control for nuclear localization. Images labeled Bright, GFP, and RFP were obtained by fluorescence microscopy. Co-localization of GFP and RFP (Merge) appears as yellow. Scale bar, 10 m. decrease in the stress response to an avirulent strain of Pph that resulted in reduced HR cell death. In addition, transgenic Arabidopsis lines overexpressing AtILP supported higher levels of bacterial growth compared with wild-type plants after inoculation with Pma M6C⌬E harboring AvrRpt2. These results indicate that AtILP has distinct functions in regulating PCD and disease resistance (i.e. a negative role in AvrRpt2-induced PCD and a positive role in RPS2-mediated resistance). Furthermore, neither the overexpression (Fig. 7A) nor mutation of AtILP (data not shown) affected the response of plants to a virulent strain (Pma M6C⌬E). Therefore, reduced PCD in AtILP plants is likely to be unrelated to the defense response to virulent pathogens.