Activation of the Ste20-like Oxidant Stress Response Kinase-1 during the Initial Stages of Chemical Anoxia-induced Necrotic Cell Death

Signal transduction mechanisms activated during the early stages of necrotic cell death are poorly characterized. We have recently identified the Sterile 20 (Ste20)-like oxidant stress response kinase-1, SOK-1, which is a member of the Ste20 kinase family. We report that SOK-1 is markedly activated as early as 20 min after chemical anoxia induced by exposure of Madin-Darby canine kidney or LLC-PK1 renal tubular epithelial cells to 2-deoxyglucose (2-DG) and any one of three inhibitors of the electron transport chain, cyanide (CN), rotenone, or antimycin A. Since oxidant stress activates SOK-1, we postulated that reactive oxygen species (ROS), which are produced by the electron transport chain during chemical anoxia, might be responsible for SOK-1 activation. The time course of CN/2-DG-induced SOK-1 activation and of production of ROS, measured in cells loaded with dichlorofluorescein, were compatible with a role for ROS in SOK-1 activation. Furthermore, preincubation of LLC-PK1 cells with three unrelated scavengers of ROS, pyrrolidine dithiocarbamate, pyruvate, or nordihydroguaiaretic acid, reduced both cellular oxidant stress and activation of SOK-1 by CN/2-DG. An increase in cytosolic free [Ca2+] ([Ca2+] i ) was necessary but not sufficient for CN/2-DG-induced activation of SOK-1. Preincubation of cells with BAPTA-AM prevented activation of SOK-1. Incubation of cells with thapsigargin or the calcium ionophore, A23187, had no effect on SOK-1 activity, but preincubation of cells with either of these agents markedly enhanced CN/2-DG-induced activation of SOK-1 (20-foldversus 7-fold). In summary, chemical anoxia activates SOK-1 via an oxidant stress-dependent mechanism that is both critically dependent upon and markedly amplified by an increase in [Ca2+] i . This requirement for dual inputs of oxidant stress and an increase in [Ca2+] i may prevent inappropriate activation of the kinase by milder degrees of oxidant stress, which are insufficient to generate an increase in [Ca2+] i . The activation of SOK-1 may be one of the cell’s earliest responses to inducers of necrotic cell death.

Ischemia, caused by the interruption of blood flow to a tissue, is the cause of myocardial infarction and stroke, and is a major cause of acute renal failure. Ischemia is characterized by three major components: ATP depletion, acidosis, and metabolite accumulation (1). If blood flow is restored in whole or in part, a fourth component, production of reactive oxygen species, is introduced (2). In the kidney, the cells most adversely affected by ischemia are the tubular epithelial cells. Those cells that survive the ischemic insult may undergo dedifferentiation and, after a variable period of time, re-enter the cell cycle to replace cells which suffered necrotic or apoptotic cell death (1). The signal transduction pathways modulating these responses to ischemic injury are poorly understood.
In tubular epithelial cells in culture, chemical anoxia, induced by inhibition of the electron transport chain combined with inhibition of glycolysis, recapitulates two key features of ischemia: ATP depletion and, if chemical anoxia occurs in a normoxic environment, oxidant stress due to enhanced mitochondrial generation of reactive oxygen species (3)(4)(5). Chemical anoxia, if prolonged, induces necrotic cell death, characterized by what appears to be intact chromatin, by mitochondrial swelling with loss of crista structure, and loss of plasma membrane integrity (6).
To identify signal transduction pathways that might modulate the cell's response to impending necrotic cell death, we have examined activation of various MAP 1 kinase cascades by renal ischemia and chemical anoxia. We have found that the stress-activated protein kinases (SAPKs, or c-Jun N-terminal kinases), while not activated by periods of renal ischemia in vivo or chemical anoxia in vitro, are markedly activated after as little as 5 min of reperfusion of ischemic kidney or 5 min of re-exposure of anoxic cells to glucose. The SAPKs are the predominant c-Jun transactivation domain kinases and the predominant ATF-2 transactivation domain and DNA binding domain kinases in the post-ischemic kidney (7,8).
We recently reported the cloning and characterization of SOK-1 (9), a human serine/threonine kinase from the germinal center kinase group of Ste20-like kinases which also includes hematopoietic progenitor kinase (10,11). SOK-1 appears to be an environmental stress response kinase since it is not activated by ligand-receptor interactions but is activated by oxidant stress (9). Herein we report that chemical anoxia markedly activates SOK-1 during the early and potentially reversible stages of necrotic cell death, and describe the mechanisms responsible for activation.
Induction of Chemical Anoxia-Madin-Darby canine kidney (MDCK) or LLC-PK 1 renal tubular epithelial cells were grown to confluence in Eagle's minimum essential medium supplemented with glutamine (2 mM) and 10% fetal calf serum. Chemical anoxia was induced as described (7,8,12). Briefly, cell monolayers (in 10-cm dishes) were incubated in a Krebs-Henseleit buffer (115 mM NaCl, 3.6 mM KCl, 1.3 mM KH 2 PO 4 , 25 mM NaHCO 3 , 1 mM CaCl 2 , 1 mM MgCl 2 , pH 7.4) for 30 min, followed by the addition of either 5 mM CN, 0.5 g/ml rotenone, or 0.5 M antimycin A, and 5 mM 2-DG in the absence of dextrose. All incubations were performed at 37°C in a 95% air, 5% CO 2 incubator. With this protocol, during the CN/2-DG treatment, ATP content of cells was reduced to less than 5% of control within 10 min. Cells were released from chemical anoxia by washing once with phosphate-buffered saline, and then incubating them in Krebs-Henseleit buffer containing 5 mM glucose (7,8).
Determination of ATP Concentration-ATP concentration was determined using a luciferin-luciferase assay (12). Briefly, confluent LLC-PK 1 cells on 10-cm plastic dishes were snap-frozen in a slurry of dry ice and ethanol, and ATP was extracted into 1 ml of 4% perchloric acid in 200 mM Tris. After 30 min on ice, the extract was titrated to pH 8.2 with 7 M KOH in 100 mM Tris, and then centrifuged (10,000 ϫ g, 10 min). Fifty microliters of supernatant were diluted 1:4 in luciferase buffer (50 mM Hepes, 1 mM MgCl 2 , 1 mM EGTA, pH 8.2) containing 500 M luciferin (final concentration), and then luciferase (5 g) was added. Luminescence was measured in triplicate with a Los Alamos Diagnostics model 535 luminometer, and ATP content was determined relative to a standard curve.
Determination of Cellular Oxidant Stress-Production of ROS was measured using the fluorescent probe 5-(and-6)-2Ј,7Ј-dichlorodihydrofluorescein diacetate (DCFH-DA). This probe is nonpolar, and is converted to the membrane-impermeant polar derivative, DCFH, by esterases when it is taken up by the cell. DCFH is minimally fluorescent, but is rapidly oxidized to the highly fluorescent 2Ј,7Ј-dichlorofluo-rescein (DCF) by intracellular H 2 O 2 or hydroxyl radicals (13,14) .
LLC-PK 1 cells on glass coverslips were loaded with 7.5 M DCFH-DA in KH buffer for 30 min. After washing with KH buffer, the coverslips were placed in a microscope tissue chamber which maintained the buffer at 37°C. The chamber was placed on the stage of a Zeiss IM-35 inverted microsope, and DCF fluorescence was excited at 450 and 505 nm using a xenon arc lamp and grating monochromators, alternated by a rotating chopper (Photon Technologies International (PTI), Inc.) as described (15,16). DCF fluorescence was monitored at Ͼ515 nm with a photon-counting photomultiplier tube, and data analysis was performed with PTI software (Felix). Data were expressed as the ratio of fluorescence intensity at excitation 505 nm over the intensity at excitation 450 nm.

RESULTS
To define putative physiological roles of SOK-1 in the response of the cell to environmental stress, we determined whether SOK-1 was activated by the extreme cellular stress of chemical anoxia. SOK-1 was activated in MDCK and LLC-PK 1 cells after exposure to CN (5 mM) and 2-DG (5 mM) in glucosefree medium for as little as 20 min, and activation peaked at 40 min (7-fold; Fig. 1A). Thereafter, SOK-1 activity declined and at 60 min of chemical anoxia, activity was similar to that of control (1.3-fold).
We also examined the effect of allowing ATP stores to recover on the activity of SOK-1. Within 20 min of re-exposing cells, which had been treated previously with CN/2-DG, to glucose (5 mM), SOK-1 activity was identical to control (Fig. 1B).
To confirm that the CN/2-DG-induced activation of SOK-1 was due to effects on the electron transport chain, we employed ATP depletion is likely to be important for CN/2-DG-induced activation of SOK-1 since inclusion of 10 mM glucose in the media during the incubation of the cells with CN/2-DG completely abrogated SOK-1 activation (Table I). To determine whether SOK-1 activation was due to ATP depletion, we exposed LLC-PK 1 cells to either CN/2-DG or the uncoupling agent, FCCP, plus 2-DG, and determined SOK-1 kinase activity. Although FCCP/2-DG reduced cellular content of ATP to levels comparable to that induced by CN/2-DG (Ͻ5% of control), SOK-1 was not activated by FCCP/2-DG (1.0 Ϯ 0.2-fold versus control) but was by CN/2-DG (5.6 Ϯ 0.4-fold versus control). These data indicate that ATP depletion is not sufficient for activation of SOK-1 by chemical anoxia, and suggest that mechanisms in addition to ATP depletion were critical.
Our previous studies indicated that SOK-1 is activated by oxidant stress (9). Therefore, we explored whether ongoing production of reactive oxygen species by the electron transport chain during the period of chemical anoxia might be the mechanism by which SOK-1 was activated. We incubated LLC-PK 1 cells, which had been loaded with 7.5 M DCFH-DA, with CN/2-DG and followed production of ROS for 45 min. In contrast to vehicle treatment (Fig. 3A), beginning at approximately 20 min after CN/2-DG, there was a discernible rise in DCF fluorescence compatible with production of ROS (Fig. 3B). Fluorescence continued to increase compared with control throughout the remainder of the experiment. Thus, CN/2-DG induced the production of ROS in LLC-PK 1 cells, and the time course of ROS production paralleled the time course of SOK-1 activation.
We next determined whether several chemically unrelated agents that react with and thus inactivate ROS inhibited SOK-1 activation by chemical anoxia. We incubated cells with PDTC (100 M) for 120 min before exposure to CN/2-DG. PDTC is a dithiocarbamate, which functions as an antioxidant by delivering thiol groups directly to the cell (17). PDTC abolished the CN-induced activation of SOK-1 and inhibited the CN/2-DG-induced activation by 70% (Fig. 4). An unrelated agent, sodium pyruvate, like other ␣-ketoacids, reduces H 2 O 2 to H 2 O while undergoing nonenzymatic decarboxylation at the 1-carbon position (18,19). Pyruvate markedly inhibited SOK-1 activation by CN/2-DG (Fig. 4).
To confirm that PDTC and sodium pyruvate were functioning as expected to reduce cellular oxidant stress, we followed DCF fluorescence for 45 min after cells, previously treated with PDTC or pyruvate, were exposed to CN/2-DG. Both PDTC and pyruvate markedly reduced CN/2-DG-induced cellular oxidant stress as measured by DCF fluorescence (Fig. 3, C and D).
LLC-PK 1 cells were also incubated in media containing catalase (300 units/ml), or heat-inactivated catalase (55°for 5 min) for 3 h before CN/2-DG (13,23). No cellular toxicity is observed at this concentration and duration of treatment (24). 2 Catalase inhibited activation by 35%, but the heat-inactivated enzyme was ineffective (Table II). These data, taken together, strongly suggest that production of ROS is necessary for the activation of SOK-1 in response to chemical anoxia.
Oxidant stress and chemical anoxia produce an increase in cytosolic free [Ca 2ϩ ] ([Ca 2ϩ ] i ) in part by releasing Ca 2ϩ from endoplasmic reticulum (ER) storage pools, by increasing Ca 2ϩ influx, and by inhibiting cell membrane Ca 2ϩ -ATPase (24 -28). The resulting increase in [Ca 2ϩ ] i is believed to exacerbate cellular injury. Therefore, we examined the role of [Ca 2ϩ ] i in the activation of SOK-1 by chemical anoxia. When [Ca 2ϩ ] i was buffered by incubation of cells with BAPTA-AM (75 M) for 60 min before the addition of CN/2-DG for 40 min, activation of SOK-1 was completely inhibited (Fig. 5A, top). Incubation of cells with BAPTA-AM alone had no effect on SOK-1 activity. Preincubation of cells with EGTA (2 mM) for 10 min also markedly inhibited activation of SOK-1 by CN/2-DG (Fig. 5A, bottom). This suggests extracellular Ca 2ϩ is important for SOK-1 activation. However, the prolonged (50 min) exposure to extracellular EGTA likely also resulted in some depletion of intracellular stores (29,30). In addition, compromise of membrane integrity by chemical anoxia may have allowed extracellular EGTA to buffer [Ca 2ϩ ] i and thus to function similarly to BAPTA.
Although it is not clear whether intracellular stores, extracellular Ca 2ϩ , or both are necessary for the increase in [Ca 2ϩ ] i , the data clearly suggested that the increase in [Ca 2ϩ ] i was required for the activation of SOK-1 by chemical anoxia. Therefore, we determined whether the Ca 2ϩ ionophore, A23187 (2 M), or thapsigargin (1 M), which releases ER Ca 2ϩ stores by selectively inhibiting the ER Ca 2ϩ -ATPase, activated SOK-1. Although neither compound alone activated SOK-1, preincubation of cells with either agent markedly enhanced SOK-1 activation by chemical anoxia (Fig. 5B, top and bottom). If, however, EGTA (2 mM) was included in the preincubation with A23187, thus buffering any increase in [Ca 2ϩ ] i , activation of  Data are -fold activation compared to the (Ϫ) Glucose control. LLC-PK 1 cells were incubated in KH buffer without (Ϫ) or with (ϩ) 10 mM glucose for 15 min prior to addition of vehicle (Control), or 5 mM CN plus 5 mM 2-DG (CN/2-DG) for 40 min. SOK-1 kinase assays were performed as described under "Experimental Procedures." With this protocol, ATP levels remain greater than 50% of control. SOK-1 by CN/2-DG was completely prevented. These data suggest that an increase in [Ca 2ϩ ] i markedly amplifies the activation of SOK-1 by chemical anoxia. The increase in [Ca 2ϩ ] i is necessary but not sufficient for activation of SOK-1. Furthermore, it is the increase in [Ca 2ϩ ] i and not depletion of intracellular Ca 2ϩ pools that is the critical signal for SOK-1 activation, since A23187, which releases intracellular Ca 2ϩ stores, was ineffective when [Ca 2ϩ ] i was buffered.

The Ste20 Family of Protein Kinases
Epistasis analyses suggest that Ste20 is upstream of a MAP kinase in a cascade consisting of a MAP kinase kinase kinase (MAPKKK) (Ste11), a MAPKK (Ste7), and two MAP kinases (Kss1 and Fus3) in the pheromone response pathway of Saccharomyces cerevisiae (31)(32)(33)(34). This pathway is activated by binding of mating factors to their cognate serpentine receptors, which are linked to heterotrimeric G proteins. Ste20 can also be activated by binding to the small G protein, Cdc42. The p21activated kinases (PAKs) are human homologs of Ste20 and are activated by binding to GTP-loaded Rac and Cdc42Hs via a conserved sequence motif in the N-terminal regulatory domain of the kinases (35)(36)(37). Like Ste20, the PAKs are activated by ligands, including thrombin and the chemoattractant, f-Met-Leu-Phe, which bind to receptors linked to heterotrimeric G proteins (38). Downstream targets of the PAKs include the MAP kinases, p38, and the SAPKs (39 -41).
A second group of Ste20-like kinases are not functional homologs of Ste20, have no Rac/Cdc42 binding domain, and, in contrast to Ste20 and the PAKs, have an N-terminal catalytic domain. They do, however, share a high degree of amino acid sequence identity with Ste20 and a Ste20-like kinase, Sps1 (42), indicating they too have been conserved throughout evolution. Three members of this group have been demonstrated to activate MAP kinase cascades, suggesting they are equivalent in this respect to the Ste20 proteins and PAKs. Specifically, germinal center kinase, hematopoietic progenitor kinase (HPK1), and Nck-interacting kinase activate the SAPKs via a Rac/Cdc42-independent mechanism (11, 43, 44). Physiologic activators of this family of Ste20-like kinases are not well The curve shown is the ratio of fluorescence intensity at excitation 505 nm over intensity at excitation 450 nm. Other LLC-PK 1 cells were pretreated with PDTC (100 M, 120 min) (C) or sodium pyruvate (50 mM, 30 min) (D) before CN/2-DG. In C and D, the abrupt rise in the ratio at approximately 42 min is due to the addition of H 2 O 2 (3 mM). This was done to confirm that the cells could respond to oxidant stress with an increase in the 505/450 ratio. described. TNF␣ activates germinal center kinase (43), but no other physiologically relevant activators of this group are known and, thus, their role in the cell remains unclear. In yeast, most of the known MAP kinase cascades are not activated by ligand-receptor interactions, but by environmental stresses such as osmolar stress, heat, and nutritional starvation (31). We have recently reported the cloning and characterization of a Ste20-like kinase from the germinal center kinase group (9). This kinase, SOK-1, does not appear to be activated by ligand-receptor interactions, but is activated by oxidant stress and thus appears to be more similar in function to the environmental stress response kinases (9).
To better define the role that this kinase may play in the response to pathophysiologic processes, we examined whether it was activated by chemical anoxia. We found that SOK-1 was markedly stimulated by chemical anoxia. In distinct contrast to the SAPKs, we found that SOK-1 was activated during the period of chemical anoxia, but activity rapidly returned to control levels when cells were re-exposed to glucose. Thus, during chemical anoxia and recovery from it, SOK-1 appears to be "on" when the SAPKs are "off" and vice versa (7). These data are consistent with our observations that SOK-1 does not activate the SAPKs when plasmids encoding both kinases are transfected into COS7 cells (43).
It is important to note that at 20 -40 min of chemical anoxia, when SOK-1 activity is greatest, damage to the cells is largely reversible and the vast majority of cells are viable (12). Thus, activation of SOK-1 is not a consequence of impending cell death.

Mechanisms of Activation of SOK-1
Role of Oxidant Stress-Consistent with our prior studies, which suggested that oxidant stress was a relatively specific activator of SOK-1 (9), we found that chemical anoxia-induced activation of SOK-1 appeared to be dependent upon generation of reactive oxygen species. Two unrelated redox-active agents, the dithiocarbamate, PDTC (17), and sodium pyruvate, which scavenges H 2 O 2 (19), markedly reduced CN/2-DG-induced cellular oxidant stress (as determined by DCF fluorescence) and inhibited activation of SOK-1. NDGA also markedly inhibited activation of SOK-1 by chemical anoxia. The iron chelator, deferoxamine, and Me 2 SO, which readily scavenges hydroxyl radical, also inhibited CN/2-DG-induced activation of SOK-1, although the inhibition was less marked than with the other three agents. Exogenously added catalase, which is taken up by many types of cells in culture and can, therefore, scavenge cytosolic H 2 O 2 (13,23), also inhibited SOK-1 activation by chemical anoxia. Inhibition of SOK-1 activation by six chemically unrelated scavengers of ROS strongly suggests SOK-1 activation by chemical anoxia requires generation of ROS and the resultant cellular oxidant stress.
Most redox-active agents that limit cellular oxidant stress, including those used in this study, have additional effects on the cell which could account for the observed reduction in SOK-1 activation by CN/2-DG. This is the principal reason behind our choice of six unrelated agents. For example, while chelation of iron by deferoxamine will have effects on the cell beyond reduction in ROS production, the inhibition by the other agents of SOK-1 activation cannot be ascribed to iron chelation.
We believe the similar time courses of CN/2-DG-induced generation of ROS and activation of SOK-1, and the inhibition of SOK-1 activation (and reduction in the level of cellular oxidant stress) by unrelated ROS "scavengers," especially when viewed in light of our previous work demonstrating direct activation of the kinase by H 2 O 2 and by menadione, strongly suggests production of ROS by the electron transport chain is the primary mechanism of activation of SOK-1 by chemical anoxia. We did consider alternative mechanisms of activation of SOK-1, in addition to generation of ROS, which occur after chemical anoxia. ATP depletion per se is not likely to be the signal that activates SOK-1 since exposure of cells to FCCP plus 2-DG, which produced equivalent ATP depletion to that seen with CN/2-DG, did not activate SOK-1. Although not sufficient, ATP depletion may be necessary for CN/2-DG-induced activation of SOK-1 since activation was blocked if cells were incubated with CN/2-DG in the presence of glucose. These data suggest ATP depletion may be acting via secondary mechanisms to enhance SOK-1 activation by chemical anoxia, and the mechanism may, in part, be related to the inability of ER and cell membrane Ca 2ϩ -ATPases to reduce [Ca 2ϩ ] i (see below).
We also considered whether changes in intracellular pH could play a role in the chemical anoxia-induced activation of SOK-1 (45). To explore this possibility, LLC-PK 1 cells were incubated in media of pH 6.9, 7.4, or 7.9, and SOK-1 activity was assayed before and after CN/2-DG. Acidification or alkalinization of media causes a corresponding change in intracellular pH (46). We found no effect of pH on either basal activity or CN/2-DG-induced activation of SOK-1 (data not shown).
Role of Increased [Ca 2ϩ ] i -A prominent early feature of chemical anoxia is a disruption of Ca 2ϩ homeostasis. [Ca 2ϩ ] i begins to increase as early as 5-10 min after the onset of   (47,48). The production of reactive oxygen intermediates may cause a leak of Ca 2ϩ from stores in the ER or other intracellular compartments, and an influx of extracellular Ca 2ϩ (25)(26)(27)49 (25). The activation of SOK-1 by chemical anoxia may be so marked because in addition to oxidant stress, Ca 2ϩ homeostasis is disrupted.
There are many possible mechanisms whereby increases in [Ca 2ϩ ] i may exert their effects. First, [Ca 2ϩ ] i could have a direct effect on the kinase. This is unlikely, however, since neither thapsigargin nor A23187 activated SOK-1 in the absence of CN/2-DG. In addition, performing SOK-1 kinase assays in standard assay buffer (very low [Ca 2ϩ ]) or high [Ca 2ϩ ] buffer (2 mM CaCl 2 ) had no effect on kinase activity (data not shown). Second, the increase in [Ca 2ϩ ] i caused by chemical anoxia likely leads to an increase in [Ca 2ϩ ] in the mitochondria via uptake by the Ca 2ϩ uniporter down an electrochemical gradient. An increase in mitochondrial [Ca 2ϩ ] has been postulated to enhance generation of reactive oxygen species and formation of hydroperoxides (50). Thus, the increase in [Ca 2ϩ ] i could exert its effect on SOK-1 activation by enhancing CN/2-DG-induced production of reactive oxygen species by the electron transport chain, thus worsening oxidant stress. It seems equally, if not more, likely that the increase in [Ca 2ϩ ] i synergizes with reactive oxygen species and ATP depletion to enhance activation of (or damage to) a putative sensor which effects the activation of SOK-1 (Fig. 6). (2 M, 10 min; A23187 ϩ), or CN/2-DG (ϩ) after pretreatment with BAPTA, TG, A23187, or A23187 plus EGTA (2 mM). Bottom, quantitation of three separate experiments with Ca 2ϩ mobilizing agents. LLC-PK 1 cells were incubated, as above, with vehicle only (CONTROL), CN and 2-DG, thapsigargin, A23187, or CN/2-DG after pretreatment with thapsigargin, A23187, or A23187 plus EGTA. Either of the Ca 2ϩ mobilizing agents markedly enhanced the SOK-1 activation observed with CN plus 2-DG (from 7-fold to approximately 20-fold), unless the increase in [Ca 2ϩ ] i after A23187 was buffered by simultaneous addition of EGTA.

Role of SOK-1 in Necrotic Cell Death
Apoptotic cell death requires that the cell be an active participant in the progression to death. Furthermore, it has recently become clear that multiple signal transduction pathways may be activated in the cell which modulate the progression to apoptotic death (51)(52)(53)(54)(55). In contrast, necrotic cell death has often been viewed as an inexorable progression to death unless the inciting stimuli are removed or their effects on the cell are minimized. Only recently has it become apparent that signaling molecules, including cPLA 2 , might hasten, and the anti-apoptotic proteins, Bcl-2 and Bcl-x L , and inhibitors of caspases might inhibit the progression to necrotic cell death (4,6,24). These data not only suggest overlapping mediators of necrotic and apoptotic cell death, but also suggest that various responses of the cell may prolong or shorten the time to irreversible injury and determine whether or not a cell will survive.
The simplest strategies used to determine the biological role of protein kinases include the creation of kinase-inactive variants, which, ideally, function as dominant interfering mutants. With SOK-1 and the related germinal center kinase, however, kinase-inactive mutants and even the non-catalytic regions of the proteins retain some of the biological effects of the wild type kinases and can activate downstream targets (43). 3 Consequently, it is not clear as yet what physiologic role SOK-1 plays in the response of the cell to chemical anoxia. Chemical anoxia induces two pathologic responses, generation of reactive oxygen species and disruption of Ca 2ϩ homeostasis, which have been postulated to trigger many forms of necrotic and apoptotic cell death (4, 6, 25, 49, 56 -61). It is these same two pathologic responses that combine to activate the kinase. The requirement for dual inputs for activation, both of which would be present only during conditions of extreme cellular stress, combined with the rapid and complete inactivation of the kinase when energy depletion is corrected, indicates that the activity of the kinase is very strictly regulated in the cell. Such strict regulation suggests SOK-1 may trigger events that would be deleterious to the cell if activated inappropriately, but which might be protective to the cell in the setting of threatened necrotic cell death.