| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
13Department of Pharmacology, University of Illinois at Chicago, Chicago, Illinois, USA
1Correspondence: Department of Pharmacology (MC 868), University of Illinois, 909 S. Wolcott Ave., Chicago, IL 60612 USA. E-mail: tvy{at}uic.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Here, we report a novel mechanism of G-protein-dependent regulation of ASK1. We demonstrated that G
13 forms a complex with ASK1 in an activation-independent manner. Both N- and C-terminal regulatory domains of ASK1 were essential for the efficient interaction, while its kinase domain was not required. Formation of the G13-ASK1 complex was enhanced by JNK-interacting leucine zipper protein, JLP. Constitutively activated G13Q226L increased ASK1 expression. Short-term activation of a serotonin 5-HT4 receptor that is coupled to G13 also increased ASK1 expression. Importantly, prolonged activation of 5-HT4 receptor in COS-7 cells or prolonged treatment of human umbilical vein endothelial cells with thrombin concomitantly down-regulated both G13 and ASK1. Data showed that G13Q226L reduced the rate of ASK1 degradation, decreased ASK1 ubiquitination, and reduced association of ASK1 with an E3 ubiquitin ligase CHIP, previously shown to mediate ASK1 degradation. Our findings indicate that ASK1 expression levels can be regulated by G13, at least in part via control of ASK1 ubiquitination and degradation.—Kutuzov, M. A., Andreeva, A. V., Voyno-Yasenetskaya, T. A. Regulation of apoptosis signal-regulating kinase 1 degradation by G13.
Key Words: signal transduction • MAP kinases • heterotrimeric G proteins
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
3
4
5
). ASK1 activates MAPKs MKK3/MKK7, and MKK4/MKK6, which, in turn, activate their target MAP kinases JNK and p38, respectively. On the basis of the evidence from ASK1 knockout mice, it was concluded that ASK1 does not mediate transient JNK and p38 activation following stress, but it is essential for their sustained activation, which is required for the induction of apoptosis (6)
.
Although the paradigm of JNK/p38 activation by oxidative stress conditions is now widely accepted, it may not be universally applicable, since in some cellular models, ASK1 does not appear to be activated in response to its known stimulators such as tumor necrosis factor- (TNF-) (7)
, or to be involved in JNK activation (ref. 8
and our unpublished observations). ASK1 may selectively activate JNK rather than p38 (9)
or p38 but not JNK (10)
. An ability of ASK1 to induce caspase-independent cell death in a manner independent of its kinase activity has been reported (11)
.
While in most cases ASK1 activation leads to apoptosis, it may also signal cell survival and differentiation. ASK1 can induce either differentiation or apoptosis in PC12 cells (9
, 12)
and keratinocytes (13
, 14)
. In cardiomyocytes, ASK1 promotes apoptosis (15
, and references therein), nonapoptotic cell death (16)
, but also mediates angiotensin II-induced (17)
and TNF--induced (18)
hypertrophic response. Ca2+/calmodulin-dependent kinase II and a small G protein Rac1 may act upstream of ASK1 in signaling pathways leading to cardiomyocyte hypertrophy (18)
.
ASK1 is activated by autophosphorylation on Thr845, which parallels its kinase activity and can be used to assess the extent of its activation (reviewed in ref. 2
). A number of ASK1 interaction partners that either stimulate or inhibit its activity have been identified, including several scaffold proteins, protein kinases, and phosphatases (2
3
4
5
, 15
, 19
, 20)
. ASK1 exists as a macromolecular complex of 1500–2000 kDa, designated as "ASK1 signalosome" (19)
. Inhibitory phosphorylation of ASK1 occurs at least at three sites in its regulatory N- and C-terminal domains. One of these sites, Ser-967, binds an inhibitory 14–3-3 protein (21)
. Steady laminar flow was found to prevent TNF--induced dissociation of ASK1 from 14–3-3 protein in endothelial cells, thus inhibiting inflammatory action of TNF- (22)
.
Several lines of evidence suggest that ASK1 may be regulated not only at the level of its kinase activity but also via regulation of its expression levels. Thioredoxin, a well-known inhibitor of ASK1 kinase activity, also promotes ASK1 ubiquitination and degradation (23)
. On the other hand, TNF- and TRAF2 block ubiquitination and stabilize ASK1 (23)
. Laminar flow reduces ASK1 dissociation from thioredoxin in endothelial cells (22)
, at least in part by reducing the expression of thioredoxin-interacting protein, which competes with ASK1 for thioredoxin binding (24)
. In contrast to the stabilizing effect of TNF- on ASK1 observed by Liu and Min (23)
, TNF- signaling via TNF receptor 2 has recently been reported to induce ubiquitination and degradation of ASK1 (as well as TRAF2) in a manner dependent on an E3 ubiquitin ligase c-IAP1 (25)
. Ubiquitination and degradation of ASK1 are also promoted by another E3 ubiquitin ligase and a cochaperone of Hsp70 CHIP, which interacts with ASK1 via its TPR domain (26)
. ASK1 degradation is also stimulated by suppressor of cytokine signaling-1 (SOCS1) in a manner dependent on phosphorylation state of ASK1 Tyr718 (27)
. The latter mechanism has been suggested to attenuate TNF--induced inflammation in endothelial cells (27)
. ASK1 regulation at transcriptional level by a transcription factor E2F1 has also been reported (28)
.
Heterotrimeric G proteins G12 and G13 have been found to stimulate ASK1 activity (1
, 29)
, although molecular details of this stimulation remained undefined. G12 and G13 form one of the four major families of subunits of heterotrimeric G proteins. A number of functions and interaction partners (reviewed in refs. 30
and 31
) have been described for G12 and G13, some of them overlapping and some other specific for either G12 or G13. The most conspicuous difference between the two G proteins is that the absence of G13 is embryonically lethal due to a defect in angiogenesis (32)
, while G12 knockout mice have no apparent phenotype (33)
. Depending on the cell lines, activated G12/G13 may induce either neoplastic transformation or apoptosis (see ref. 29
, and references therein).
In this study, we identified G12 and G13 as novel ASK1 interaction partners, with G13 binding to ASK1 with higher potency as compared to G12. We showed that activated G13 is able to increase ASK1 expression levels by affecting ASK1 ubiquitination and degradation, presumably due to a reduced ASK1 interaction with CHIP. These observations, together with our recent findings that G12 can reduce degradation of eNOS (34)
, suggest that these G proteins may regulate their effectors by affecting their expression levels.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
13 constructs (wild type and G13Q226L) and ASK1 constructs were described previously (1
, 35)
. Constructs for internally EE-tagged G13 and G13Q226L were purchased from the Guthrie Research Institute (Sayre, PA, USA).
Constructs for HA-tagged G12 and G13 (in pcDNA3) were a generous gift from Silvio Gutkind (National Institute of Dental Research, National Institutes of Health, Bethesda, MD, USA). G13 constructs (wild type and G13Q226L) with HA and EE tags were created as follows. Fragments encoding the EE tag and Q226 or L226 were excised from respective EE-tagged constructs using an internal XhoI site and an XbaI site in pcDNA3.1 polylinker. These fragments were ligated into pcDNA3-HA-G13 digested with XhoI and XbaI. The constructs were verified by sequencing.
Constructs for Myc/His6-tagged CHIP, ASK1 deletion mutants, S-tagged JLP and YFP-tagged 5-HT4 receptor were kindly provided by Cam Patterson (University of North Carolina, Chapel Hill, NC, USA), Jacques Landry (Université Laval, Québec, Canada), Danny Dhanasekaran (Fels Institute for Cancer Research and Molecular Biology, Philadelphia, PA, USA) and Evgeni Ponimaskin (University of Göttingen, Göttingen, Germany), respectively.
Antibodies
Polyclonal G12, G13, Gß, S-tag, ß-actin and ß-tubulin antibodies, monoclonal HA, Myc and ASK1-specific antibodies and Protein A/G agarose were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Monoclonal ubiquitin-specific and polyclonal JNK-specific antibodies were from Cell Signaling Technology (Beverly, MA, USA). Monoclonal Hsp90-specific antibody was from BD Biosciences (San Jose, CA, USA). Monoclonal -tubulin antibody was from Sigma.
Cell culture and transfection
Transient transfection of COS-7 cells was performed using lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to manufacturer’s instructions. COS-7 cells and HUVECs were cultured as described previously (refs. 36
and 34
, respectively).
Immunoprecipitation, Western blot analysis, and quantitation of ECL data were performed as described previously (36)
.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
12 and G13 form a protein complex with ASK1 independently of G activation state12 and G13 (1
, 29)
. To test a possibility that G12 or G13 and ASK1 might coexist in the same macromolecular complexes, COS-7 cells were cotransfected with wild-type ASK1 and HA-tagged G12 or G13. Immunoprecipitation with HA antibody showed that both G12 and G13 coimmunoprecipitated with ASK1, although interaction with G13 was considerably stronger (Fig. 1
A). ASK1 was not detectable in HA immunoprecipitates from the cells transfected with ASK1 alone (Fig. 1A
).
|
The interaction could be detected both in the absence and in the presence of AlF4–, an activator of subunits of heterotrimeric G proteins that induces a conformation similar to the transition state for GTP hydrolysis (37)
(Fig. 1A
). Reciprocal immunoprecipitation with ASK1 antibody confirmed activation state-independent interaction of G13 with ASK1 (Fig. 1B
). Since HA tag modifies the N terminus of G13, it was essential to ensure that HA-tagged G13 is competent for a transition into its active conformation under our experimental conditions. G13 interaction with its effector protein p115RhoGEF is known to occur only when G13 is in the activated state (38
, 39)
. Therefore, we performed similar coimmunoprecipitation assays with HA-G13 and p115RhoGEF in the absence or in the presence of AlF4–. HA-G13 interacted with p115RhoGEF only in the presence of AlF4–(Fig. 1C
), which was consistent with previously published data (38
, 39)
.
Experiments similar to that shown in Fig. 1A
were also performed with EE-tagged wild-type G13 and constitutively active EE-G13Q226L. When coexpressed with ASK1 and immunoprecipitated with EE antibody, both G13 forms were found to interact similarly with ASK1 (data not shown), confirming that the interaction with ASK1 is not dependent on the activation state of G. Since interaction of ASK1 with G12 was weaker than with G13 and could not be confirmed by reciprocal immunoprecipitation, we concentrated our further efforts on characterizing ASK1-G13 interaction.
Both N- and C-terminal domains of ASK1 are involved in the interaction with G13
To establish which of the three domains of ASK1 (N-terminal, catalytic or C-terminal) is involved in the interaction with G12/13, we used ASK1 deletion mutants (11)
, lacking either the catalytic and C-terminal domain (ASK1-N), N-terminal domain (ASK1
N), or kinase domain with N- and C-terminal domains spliced together (ASK1K) (Fig. 1D
). ASK1K was able to interact with G13 (Fig. 1D
). Separate experiments with varying amounts of full-length ASK1 and ASK1K showed that ASK1K bound G13 as potently as full-length ASK1 (data not shown). These results demonstrated that kinase domain of ASK1 did not contribute to its interaction with G13. Deletion of the N-terminal domain completely abrogated G13 binding (Fig. 1D
), suggesting that this region may be responsible for G13 binding. However, the N-terminal domain alone was not sufficient for efficient interaction with G13 (Fig. 1D
). These data suggest that both N- and C-terminal domains of ASK1 are necessary for the interaction with G13.
G13–ASK1 complex formation is facilitated by JLP
JNK-interacting leucine zipper protein, JLP, has been reported to act as a scaffold protein for G13 (40)
. JIP4, a closely related isoform, has been found to interact with ASK1 (41)
. Therefore, we examined whether JLP might also be able to interact with ASK1 and thus promote assembly of the G13–ASK1 complexes. Increasing amounts of a construct encoding S-tagged JLP were coexpressed with ASK1 and HA-tagged G13Q226L in COS-7 cells, and the amounts of ASK1 were assessed in HA immunoprecipitates. Coexpression of JLP increased the amounts of ASK1 coprecipitating with G13 in a dose-dependent manner (Fig. 1E
). These results indicate that JLP may act as a scaffolding protein to assemble macromolecular complexes that involve G13 and ASK1.
G13 increases ASK1 expression levels
When ASK1 and epitope-tagged G13 were coexpressed in the experiments described above, in some cases, the levels of ASK1 appeared to be affected by G13. We examined whether coexpression with untagged G13 would affect ASK1 expression. Coexpression of ASK1 with untagged G13Q229L increased ASK1 levels two- to fourfold (Fig. 2
A).
|
As ASK1 was reported to be stimulated by G13 (1
, 29)
, we examined whether the G13Q226L effect on ASK1 expression requires its kinase activity. Expression of the inactive ASK1K709R mutant was increased by G13Q226L similarly to that of wild-type ASK1 (Fig. 2B
), indicating that kinase activity of ASK1 is not required for the effect of G13.
We next examined the dose-response relationship of ASK1 expression levels vs. those of G13. To be able to estimate relative amounts of ASK1 and G13, we coexpressed the HA-tagged forms of both proteins. The levels of G13Q226L required to saturate its effect on ASK1 expression were approximately equimolar to those of ASK1 (Fig. 2C
). Probing the same blots with G13-specific antibody showed that saturating levels of G13Q226L corresponded to a 1.5- to 2-fold increase in total (i.e., endogenous + HA-tagged) G13 (Fig. 2C
; quantification data not shown). Coexpression with wild-type G13 did not reproducibly affect the ASK1 levels (data not shown).
Regulation of expression levels of endogenous ASK1 in COS-7 cells by G13
To further validate the findings described above, we examined whether G13 would affect expression of endogenous ASK1. Increasing amounts of the G13Q226L construct were transfected into COS-7 cells, endogenous ASK1 was isolated by immunoprecipitation, and its expression levels were assessed by Western blot analysis. Similar to overexpressed ASK1, the levels of endogenous ASK1 were increased considerably in the cells expressing activated G13 as compared to control cells transfected with empty vector (Fig. 3
). These data indicated that activated G13 does up-regulate ASK1 at physiological concentrations of the latter. It should be noted that since G13Q226L expression results in only moderate increase in total G13 content in the cells (see above), the levels of G13 were also close to physiological in these experiments.
|
Coexpression of constitutively active serotonin 5-HT4 receptor increases ASK1 levels
To determine whether activation of endogenous G13 by a G protein-coupled receptor (GPCR) will increase ASK1 expression, we coexpressed ASK1 and YFP-fused 5-HT4 receptor, a GPCR selectively coupled to G13 but not G12 (42)
. 5-HT4 receptor was shown to have high constitutive activity that induces stimulation of the endogenous G13 (42)
. Six hours after cotransfection with 5-HT4 receptor, ASK1 was significantly up-regulated, while the levels of G13 and JNK1, a downstream effector of ASK1, were unaffected (Fig. 4
A).
|
Both G13 and ASK1 are down-regulated by prolonged treatments that activate G13
When 5-HT4 receptor and ASK1 were coexpressed for 24 h, down-regulation of both G13 and ASK1 was observed, while JNK1 was not affected (Fig. 5
A). We have shown previously that prolonged treatment of human umbilical vein endothelial cells (HUVECs) with thrombin concomitantly down-regulates both endogenous G12 (which stabilizes eNOS) and eNOS (34)
. To examine whether thrombin exerts similar effect on G13 and ASK1, we treated HUVECs with thrombin for 14 h and examined the levels of G13 and ASK1, as well as Gß and Hsp90, in treated and in control cells. Neither Gß nor Hsp90 levels were down-regulated by thrombin treatment, while those of G13 and ASK1 were markedly reduced (Fig. 5B
). Reciprocal plot of the relative levels of these two proteins in different samples showed a clear correlation (correlation coefficient 0.82) between the extent of down-regulation of G13 and that of ASK1 (Fig. 5B
). While these data alone cannot rule out independent down-regulation of G13 and ASK1 by a common mechanism, they are consistent with ASK1 expression levels being positively regulated by activated G13.
|
Activated G13 decreases the rate of ASK1 degradation
Observations depicted in Figs. 2
3
4
5
indicate that activated G13 is able to increase steady-state levels of ASK1. Since ASK1 expression levels are affected by proteasome-dependent degradation (23
, 26
, 27)
, we examined whether regulation of ASK1 degradation may be affected by G13. ASK1 was expressed in COS-7 cells in the absence or in the presence of G13Q226L. On addition of cycloheximide to suppress protein synthesis, the decrease in ASK1 content was assessed by Western blot analysis and quantitated. The apparent half-life of ASK1 varied in the range of 2–5 h in different experiments. In the presence of G13Q226L, the rate of ASK1 degradation was consistently reduced as compared to control cells (Fig. 6
), indicating that G13 decreases the rate of ASK1 degradation.
|
Activated G13 reduces ubiquitination of ASK1
Proteasome-dependent degradation of ASK1 is preceded by its ubiquitination, which is controlled by at least three separate mechanisms (25
26
27)
. Therefore, we examined whether activated G13 would affect ASK1 ubiquitination. ASK1 was expressed in COS-7 cells in the absence or in the presence of G13Q226L. Cell lysates were immunoprecipitated with ASK1 antibody and polyubiquitinated aggregated forms of ASK1 detected by Western blot analysis using ubiquitin- or ASK1-specific antibodies. The levels of polyubiquitinated ASK1 were clearly reduced in the cells coexpressing G13Q226L (Fig. 7
A). Similarly, probing the same blots with ASK1 antibody revealed reduced presence of high molecular weight bands in the G13Q226L-expressing cells (Fig. 7A
). These results indicate that activated G13 is likely to regulate ASK1 expression levels by affecting its ubiquitination and proteasome-dependent degradation.
|
Activated G13 interferes with ASK1-CHIP interaction
Recently described SOCS1-dependent ASK1 ubiquitination requires ASK1 phosphorylation on Tyr718 (27)
. Since Tyr718 is absent in the ASK1-N construct and we observed that this fragment is still up-regulated by activated G13 (data not shown), the involvement of SOCS1 in regulation of ASK1 expression by G13 seemed unlikely. In addition, we were unable to detect Tyr phosphorylation of ASK1 immunoprecipitated from COS-7 cells using phosphoTyr-specific antibody (data not shown). Another mechanism of ASK1 ubiquitination is via its interaction with CHIP (26)
. Therefore, we examined whether activated G13 would affect ASK1-CHIP interaction. ASK1 and Myc-tagged CHIP were coexpressed in the absence or in the presence of HA-tagged G13Q226L. Cell lysates were immunoprecipitated with Myc-specific antibody, and the amount of ASK1 in immunoprecipitated material was quantitated by Western blot analysis (Fig. 7B
). In line with published observations (26)
, we were able to detect ASK1 interaction with Myc-tagged CHIP in these assays. In the presence of increasing amounts of G13Q226L, the amount of ASK1 associated with CHIP progressively declined (Fig. 7B
), indicating that G13Q226L interferes with ASK1-CHIP interaction.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, c-IAP1 (25)
or an SH2 domain-containing protein SOCS1 (27)
.
Here we present the evidence that ASK1 and subunits of heterotrimeric G proteins G12/G13 may exist as components of the same macromolecular complexes, and report that steady state levels of ASK1 are increased by activated G13. Furthermore, we demonstrated that activated G13 is able to increase expression levels of ASK1 by reducing the rate of its ubiquitination and degradation, likely by interfering with ASK1 interaction with CHIP. Thus, activated G13 would signal via ASK1 in a dual way: first, by stimulating ASK1 kinase activity as described previously (1
, 29)
, and second, by increasing expression levels of ASK1, as reported here, which would potentiate ASK1 signaling (Fig. 8
).
|
Several important issues remain to be elucidated. One is whether physical association of G13 and ASK1 reported here is related to the up-regulation of ASK1 expression. On one hand, saturation of the effect of G13Q226L on ASK1 expression at approximately equimolar ratio of both proteins is compatible with a requirement for a complex formation. On the other hand, ASK1 interacts similarly with activated and nonactivated G13, while only activated G13 is efficient in up-regulating ASK1 expression, which would argue against the requirement of G13–ASK1 complex formation for ASK1 up-regulation. In addition, overexpression of JLP, which we have identified as a potential scaffold protein to assemble G13–ASK1 complexes, had no consistent effects on ASK1 expression levels (our unpublished observations), although it dramatically increased G13–ASK1 association in our experiments. Establishing whether endogenous JLP and/or any of several related proteins (see refs. 40
and 41
for references) mediate the observed G13–ASK1 interaction would require a study involving systematic knockdown of these isoforms. It is also possible that G13–ASK1 complexes are heterogeneous and may involve more than one scaffolding protein or that ASK1 and G13 may also be able to interact directly.
The precise events leading to a decreased association of CHIP with ASK1 in the presence of activated G13 also remain to be identified. A direct competition between G13 and CHIP for ASK1 binding does not seem plausible since both activated and nonactivated G13 interact with ASK1 with similar potency, while only activated G13 is able to stabilize ASK1. An indirect effect via activation of Rho/ROCK, a well-established pathway stimulated by G13, can be excluded since dominant negative RhoAT19N mutant and a ROCK inhibitor Y27632 had no effect on ASK1 expression (our unpublished observations).
The ability of the same signaling pathway or protein to regulate not only ASK1 activity but also its cellular levels is not without a precedent. Thioredoxin, a well-known inhibitor of ASK1 kinase activity, was also found to promote ASK1 ubiquitination and degradation (23)
. Subsequent studies by the same group showed that a proinflammatory cytokine TNF- not only stimulates ASK1 kinase activity but can also induce its deubiquitination and stabilization. These authors have identified SOCS1 as a protein that targets ASK1 for ubiquitination, presumably by coupling to the elongin/cullin-2 ubiquitination machinery (27)
. Another mechanism of ASK1 ubiquitination involves an E3 ubiquitin ligase CHIP (26)
. Our work shows that CHIP ability to ubiquitinate ASK1 and therefore reduce ASK1 expression can be regulated by G protein-coupled signaling pathways. Together with the data mentioned above, our findings outline a new paradigm in ASK1 signaling: the same factors that affect ASK1 kinase activity (thioredoxin, TNF-, G13) can also modulate its expression levels, which would either potentiate or dampen respective signaling pathways.
Together with our previous study (34)
(also see a comment in ref. 43
), the findings reported here also outline a new paradigm in G protein-coupled signaling. We have shown previously that G12 can increase expression levels of eNOS by stabilizing both mRNA and protein (34)
. Intriguingly, eNOS has been reported to interact with CHIP (44)
, although the role of CHIP in eNOS ubiquitination has not been established. Whether regulation of ubiquitination is involved in stabilization of eNOS by G12 remains to be seen. It is worth noting that both endogenous eNOS (34)
and ASK1 (this study) are down-regulated in endothelial cells by prolonged thrombin treatment concomitantly with down-regulation of either G12 or G13. Potential similarity of the mechanisms in both cases (eNOS and ASK1) would indicate that the ability of G12 or G13 to regulate effector protein levels may be a more general property not limited to eNOS and ASK1 and would warrant more extensive studies in this respect of other proteins known to interact with these G subunits and/or proteins known to be ubiquitinated and degraded in a CHIP-dependent manner. The ability of G12 or G13 to inhibit degradation of two unrelated proteins (eNOS and ASK1) indicates that these G proteins may play a previously unappreciated role in the control of protein degradation.
|
ACKNOWLEDGMENTS |
|---|
Received for publication March 27, 2007. Accepted for publication May 17, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B inhibits TNF-induced accumulation of ROS that mediate prolonged MAPK activation and necrotic cell death. EMBO J. 22,3898-3909[CrossRef][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
