Upregulation of RGS7 may contribute to tumor necrosis factor ... - Nature

© 1999 Nature America Inc. • http://medicine.nature.com

ARTICLES

Upregulation of RGS7 may contribute to tumor necrosis factor-induced changes in central nervous function THOMAS BENZING1, RALF BRANDES2, LORENZ SELLIN1, BERNHARD SCHERMER1, STEWART LECKER3, GERD WALZ1 & EMILY KIM4 1


Department of Medicine, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, Massachusetts 02215, USA 2 Institute for Cardiovascular Physiology, University Hospital of Frankfurt, D-60590 Frankfurt, Germany 3 Department of Cell Biology, Harvard Medical School, 240 Longwood Avenue, Boston, Massachusetts 02115, USA 4 Department of Psychiatry, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, Massachusetts 02215, USA Correspondence should be addressed to E.K.; email: [email protected]

The central nervous dysfunctions of lethargy, fever and anorexia are manifestations of sepsis that seem to be mediated by increased cytokine production. Here we demonstrate that tumor necrosis factor (TNF)-α, an essential mediator of endotoxin-induced sepsis, prevents the proteasome-dependent degradation of RGS7, a regulator of G-protein signaling. The stabilization of RGS7 by TNF-α requires activation of the stress-activated protein kinase p38 and the presence of candidate mitogen-activated protein kinase phosphorylation sites. In vivo, RGS7 is rapidly upregulated in mouse brain after exposure to either endotoxin or TNF-α, a response that is nearly abrogated in mice lacking TNF receptor 1. Our findings indicate that TNF-mediated upregulation of RGS7 may contribute to sepsis-induced changes in central nervous function.

The central nervous system (CNS) is one of the first organ systems affected by sepsis. Sepsis is associated with behavioral changes that include increased sleep, hypomotility, hypophagia and decreased libido. Sepsis-induced CNS dysfunctions have been attributed to changes in mean arterial blood pressure and body temperature, endothelial cell dysfunction, and intravascular coagulopathy. Various cytokines, including TNF-α, interleukin (IL)-1β, IL-6 and IL-8, are released during the onset of endotoxic shock; compelling evidence indicates TNF-α and IL1β are central mediators of this syndrome1–4. Treatment with TNF-α and IL-1β elicits characteristic somnogenic, pyrogenic and anorectic effects in rodents, indicating that these proteins are directly involved in sepsis-induced complications of the central nervous system5–7. The ‘regulator of G protein signaling’ (RGS) proteins stimulate the intrinsic GTPase activity of activated Gα subunits and thereby accelerate G-protein inactivation. By stabilizing activated Gαi and Gαq subunits in a transition state that facilitates the release of phosphate from bound GTP, RGS proteins inhibit signaling downstream of many G protein-coupled receptors8–11. Several RGS proteins, including RGS4, RGS7, RGS8, RGS9 and RGS10, are densely expressed in the mammalian brain12–19. RGS7 is the closest mammalian homolog of EGL-10, a protein that in the nematode Caenorhabditis elegans regulates the frequency of behaviors such as locomotion, foraging and egg laying in a dosedependent manner12. RGS7 is an unstable protein that is rapidly degraded by the ubiquitin–proteasome pathway20. Here, we report that TNF-α augments RGS7 protein levels through the inhibition of proteasomal degradation. This effect is mediated by p38, a stress-activated protein kinase. Furthermore, injection of mice with endotoxin, an inducer of TNF-α release and a sepsis-like synNATURE MEDICINE • VOLUME 5 • NUMBER 8 • AUGUST 1999

drome, rapidly increases RGS7 protein levels in the brain, an upregulation that is substantially curtailed in mice lacking tumor necrosis factor receptor 1 (TNFR1). TNF-α inhibits the degradation of RGS7 RGS7 was poorly expressed compared with RGS9 and EGL-10 after transient transfection in HEK 293T cells (Fig. 1a); however, its expression was increased considerably by treatment with TNF-α, a cytokine involved in many inflammatory and apoptotic pathways induced by sepsis. To demonstrate that this effect was mediated through the stabilization of RGS7 rather than an increase in protein synthesis, we assessed the effect of TNF-α on the half-life of RGS7 while blocking de novo protein synthesis. In the presence of the protein synthesis inhibitor cycloheximide, TNF-α increased the half-life of RGS7, but not that of green fluorescent protein (GFP), used to confirm transfection efficiency and protein loading, at each time point (Fig. 1b). RGS7 is subject to ubiquitin-dependent degradation by the proteasome complex and is ubiquitinated in the presence of the proteasome inhibitor MG132 and HA-tagged ubiquitin. In contrast, TNF-α did not facilitate the accumulation of ubiquitinated RGS7 molecules (Fig. 1c). These findings indicate that TNF-α may induce a post-translational modification of RGS7 that prevents ubiquitination of RGS7. To demonstrate that the effect of TNF-α was specific for RGS7, we assessed the effect of TNF-α on β-catenin, a protein also subject to proteasome degradation21,22. Degradation of β-catenin involves activation of the protein kinase GSK-3β (ref. 21); thus, lithium, a GSK-3β inhibitor, and the proteasome inhibitor MG132 increased steady-state protein levels of β-catenin, whereas TNF-α did not (Fig. 1d). The onset of TNF-α upregulation of RGS7 protein levels was detectable after 2 hours with a progressive decrease after 6 hours (Fig. 1e). Endogenous RGS7 in 913


ARTICLES

a

b


c

e

d

f

ECV-304 cells had a similar TNF-mediated upregulation and time course (Fig. 1f). Phosphorylation of candidate MAPK sites stabilizes RGS7 Mutational analysis of RGS7 and the C. elegans homolog EGL-10 showed that upregulation of RGS7 by TNF-α depends on a 79amino-acid region of RGS7 (Q170–T247) between the DEP and Gγ-like domains (Fig. 2a). TNF-α failed to stabilize the truncation RGS7253–469, whereas protein levels of RGS7170–469 and RGS750–247 were promptly increased by TNF-α (Fig. 2b). In contrast, all three truncations were nonspecifically stabilized by the proteasome inhibitor MG132 (Fig. 2b). To confirm that RGS7 Q170–T247 mediated the TNF-α effect, we constructed a series of RGS7–EGL10 chimeras using PstI and NcoI sites conserved in both proteins. Chimeric proteins retaining the Q170–T247 domain of RGS7 were very sensitive to TNF-α upregulation (Fig. 2c). In contrast, chimeras containing the corresponding EGL-10 domain showed minimal augmentation by TNF-α. Alignment of RGS7 and EGL10 showed that their similarity diverges considerably at the 3′ end of the Q170–T247 region (Fig. 2d). In RGS7, this area encodes three candidate phosphorylation sites (S241, T245 and T247) for mitogen-activated protein kinases (MAPKs) such as extracellular signal-regluated kinase (ERK) 1, ERK 2, jun N-terminal kinase (JNK) or p38. To determine whether phosphorylation of these sites could be involved in the TNF-induced stabilization of RGS7, we mutated S241, T245 and T247 of RGS7 to alanines. Mutation of RGS7 at S241, T245 and T247 abolished the TNF-α stabilization of RGS7, whereas this mutant was still upregulated by MG132 (Fig. 2e). 914

Fig. 1 Increased accumulation of RGS7 induced by TNF-α. a, TNF-α (10 ng/ml for 4 h) increases the steady-state levels of RGS7, but not of EGL-10 or RGS9. HEK 293T cells were transiently transfected with Flag-tagged RGS7, EGL-10 and RGS9. Cellular lysates were assessed by western blot analysis with M2 monoclonal antibody against Flag. GFP, control to confirm transfection efficiency and protein loading. b, TNF-α prolongs the half-life of RGS7. Cycloheximide (CHX) was used to block de novo protein synthesis in HEK 293T cells transiently transfected with Flag-tagged RGS7 and GFP. The rapid degradation of RGS7 was substantially slowed by the presence of 10 ng/ml TNF-α. c, Ubiquitination of RGS7 is increased in the presence of a proteasome inhibitor, but not in the presence of TNF-α. HEK 293T cells were transfected with His-tagged RGS7 and multimeric HAtagged ubiquitin; cellular lysates were assessed by western blot analysis with antibody against HA to detect RGS7–ubiquitin conjugates. Left margin, molecular size markers. d, TNF-α does not induce the accumulation of β-catenin, which is stabilized by the proteasome inhibitor MG132 and the GSK-3β inhibitor lithium (Li+), but not by TNF-α. e, TNF-induced upregulation of RGS7 protein levels peak by 6 h, and decrease by 8 h. f, TNF-α induces upregulation of endogenous RGS7 in ECV-304 cells; levels are maximal by 4 h and decrease by 6 h.

TNF-mediated stabilization of RGS7 requires p38 activation To ascertain whether a known MAP kinase mediates the TNF-induced effects on RGS7, we assessed the ability of dominant-negative versions of protein kinases upstream of the stress-activated MAPK p38, the JNKs and the MAPKs ERK1 and ERK2 to block the upregulation of RGS7. A dominant-negative MEKK1 and the combination of dominant-negative MKK3 and MKK6 inhibited the TNF-α induced augmentation of RGS7 protein levels; dominantnegative versions of kinases upstream of JNK, such as MKK4 and the small G proteins RhoA, Rac1 and Cdc42, did not inhibit the upregulation of RGS7 (Fig. 3a). As MEKK1 and MKK3/MKK6 are sequential upstream activators of p38, these data indicate that activation of p38 mediates the TNF-α effect on RGS7. Verifying this result, co-expression of RGS7 with wild-type MEKK1 substantially increased RGS7 levels in response to TNF-α, whereas dominantnegative MEKK1 inhibited the TNF-α mediated upregulation of RGS7 (Fig. 3b). Furthermore, dominant-negative MKK3/6 (Fig. 3c) or SB203580, a selective inhibitor of p38 activity (Fig. 3 d), reduced the stabilization of RGS7 induced by TNF-α in a dose-dependent manner. In contrast, MEK1, a kinase upstream of ERK1/2, and dominant-negative mutants of ERK1, ERK2 and several protein kinase C isozymes did not block the effect of TNF-α on RGS7 (Fig. 4a–c). Confirming that RGS7 is a suitable target for p38, recombinant RGS7170–469, which contains the candidate MAPK phosphorylation sites required for TNF-α induced stabilization, was phosphorylated by p38 in vitro (Fig. 3e). In addition, two agents that activate p38, sorbitol and anisomycin, induced the accumulation of RGS7 protein levels (Fig. 3f). These data indicate that p38 is central in the stabilization of RGS7 protein induced by TNF-α. Endotoxin (LPS) upregulates RGS7 in vivo TNF-α is a central mediator of endotoxin-induced behavioral changes. Treatment with TNF-α elicits characteristic somnogenic, pyrogenic and anorectic effects in rodents5–7. In mouse models of sepsis, lipopolysaccharide (LPS) induces a rapid release of TNF-α that peaks in concentration within 60 minutes after injection23. To study endogenous RGS7 protein levels in a mouse model of sepsis, we used an affinity-purified RGS7 antiserum that specifically recognizes and immunoprecipitates human and mouse RGS7. RGS7 was rapidly upregulated in the brain of mice 4–6 hours after intraperitoneal injection of LPS (1 mg) (Fig. 5a). After determining total protein levels of brain lysates, we sepaNATURE MEDICINE • VOLUME 5 • NUMBER 8 • AUGUST 1999


ARTICLES

a

c

b


d

Fig. 2 Stabilization of RGS7 induced by TNF-α depends on candidate MAPK sites in Q170 and P248. a, Domain structure of RGS7. Three putative MAPK sites between S241 and P248 (bold letters) precede two PEST sequences contained within the Gγ-like domain. Filled bar, RGS. DEP, dishevelled/Egl10/pleckstrin. b, Truncations RGS7170–469 and RGS750–247 are stabilized in the presence (+) of TNF-α, whereas RGS7253–469 is not; all truncations accumulate in the presence (+) of the proteasome inhibitor MG132. c, Chimeric proteins were generated using PstI and NcoI sites conserved in RGS7 (R) and EGL-10 (E). Chimeras stabilized by TNF-α contained the PstI–NcoI fragment of RGS7. No other domain of RGS7 confers this ability, indicating that the TNF-responsive element is located between Q157 and T247. d, The similarity (upper-case letters) between EGL-10 and RGS7 between Q157 and T250 decreases con-

rated equal amounts of proteins by electrophoresis to compare cerebral RGS7 in LPS-treated and control mice. As an additional control, we also assessed levels of protein kinase C α (PKCα) to ensure equal amounts of brain lysate protein were used (Fig. 5a). Densitometric analysis of several independent experiments showed an average increase in RGS7 of approximately 250% in in brains of LPS-treated mice compared with those of control mice (Fig. 5c). Total brain homogenates were used for this study; therefore, the measured increase may substantially underestimate the accumulation of RGS7 in distinct brain areas. Intravenous injection of TNF-α resulted in a similar accumulation of cerebral RGS7 (Fig. 5b). Furthermore, LPS-mediated upregulation of RGS7 was nearly abrogated in mice lacking TNFR1 (Fig. 5d), indicating that LPS-induced upregulation of RGS7 is mainly mediated through the binding of TNF-α to TNFR1. Discussion RGS proteins constitute a family of more than 20 proteins that regulate G-protein signaling. RGS proteins stimulate the GTPase activity of activated Gα subunits in vitro and inhibit signaling by ligand-activated G protein-coupled receptors in cellular model systems8–11. Elucidation of the tissue-specific functions of RGS proteins has been complicated by their fairly nonspecific GTPase activity for all Gαi subfamilies and for Gαq. Whereas some RGS proteins are ubiquitously expressed, others are very tissue-specific, and their putative function has been predicted from localization with other signaling molecules. For example, RGS9 is a candidate GTPase accelerator for visual transduction, as it is highly expressed in retinal photoreceptors and accelerates in vitro GTP hydrolysis by the visual G protein transducin24. At least five RGS members, RGS4, RGS7, RGS8, RGS9 and RGS10 mRNAs, are densely expressed in the brain13. RGS7 is present in cortical NATURE MEDICINE • VOLUME 5 • NUMBER 8 • AUGUST 1999

e

siderably between D213 and T250; only RGS7 contains candidate MAPK phosphorylation sites (P) between S241 and T247 (highlighted). e, Site-directed mutagenesis indicates that candidate MAPK phosphorylation sites S241, T245 and T247 are essential for the TNF-induced stabilization of RGS7.

layers I–VI, the hippocampus and several diencephalic and brain stem nuclei25 as well as muscle; however, its tissue-specific function remains unclear. RGS7 is the closest mammalian homolog to EGL-10, a C. elegans protein that has a very similar pattern of expression in the ventral and dorsal nerve cord, as well as body wall muscle cells of the adult nematode12. EGL-10 antagonizes effects of GOA-1, a C. elegans protein closely related to mammalian Gαo that may be involved in serotonin-mediated signaling, and modulates the frequency of behaviors such as locomotion, foraging, and egg laying in a dose-dependent manner12,26. RGS7 catalyzes the inactivation of Gαi, Gαq and Gαo subunits in vitro17; thus, RGS7 is likely to control signaling pathways mediated by neurotransmitters that use Gαi/o-coupled receptors in vivo. RGS7 accelerates the activation and deactivation kinetics of the G-protein-coupled inward rectifier K channel, indicating that RGS7 can modify neuronal excitability27. Functional specificity as well as temporal activity of RGS proteins seems to derive from patterns of expression, transcriptional regulation and subcellular localization9,10. Transient expression of RGS proteins during cellular activation, sequestration of RGS proteins through protein–protein interaction, and post-translational modification of their Gα target provide additional controls. Our results demonstrate that RGS7 protein levels are regulated through rapid degradation by the ubiquitin-dependent proteasome pathway, and that the halflife of RGS7 can be extended by post-translational modification induced by TNF-α. A short half-life is a common feature of many regulatory proteins that are rapidly degraded by the proteasome–ubiquitin pathway28,29. Post-translational modification, most often serine or threonine phosphorylation, seems to facilitate or prevent the degradation of many regulatory proteins by the proteasome. For example, activation of NFκB by 915



ARTICLES

a

b

c

d

e

f

Fig. 3 Activation of mitogen-activated protein kinase p38 is required for TNF-induced upregulation of RGS7. a, Dominantnegative (DN) MEKK1 and a combination of dominant-negative MKK3 and MKK6 block the TNF-induced increase of RGS7 protein levels. In contrast, neither dominant-negative MKK4, nor a combination of dominant-negative Cdc42, Rac1 and RhoA (G-proteins DN) substantially affect RGS7 stabilization, indicating that JNK does not mediate the TNF-α effect. b, The dominant-negative (DN) MEKK1 blocks the TNF-α response, whereas wild-type (WT) MEKK1 augments the TNF-induced upregulation of RGS7 protein levels. c, Dominant-negative (DN) MKK3 and MKK6, p38 MAPK kinases, reduce the TNFmediated upregulation of RGS7 in a dose-dependent manner. d, The p38 inhibitor SB203580 reduces RGS7 levels induced by TNF. e, Recombinant His-tagged RGS7170–469 us readily phosphorylated by activated p38 in vitro. Phosphorylation of RGS7 was compared with that of PHAS-1, a known substrate for p38. f, Activation of p38 increases RGS7 protein levels. HEK 293T cells were transiently transfected with RGS7 and incubated with TNF-α (10 ng/ml for 4 h), sorbitol (100 mM for 3 h) or anisomycin (100 nM for 3 h). All three activators of p38 induce an increase of RGS7 protein levels.

downstream of p38 may phosphorylate RGS7. RGS7 contains multiple potential phosphorylation sites for various protein kinases and is phosphorylated in HEK 293T cells in the absence of stimuli; TNF-α did not substantially increase the degree of overall phosphorylation (data not shown). Phosphoamino-acid analysis to determine the precise location of the residues phosphorylated in response to TNF-α has been complicated by low RGS7 protein levels even after induction with TNF. However, our mutational analysis supports the idea that the MAPK sites S241, T245 and T247 are essential for the TNF effect on RGS7 stability. The mechanism by which TNF-α prevents the recognition and degradation of RGS7 by the ubiquitin-dependent proteasome remains unclear. Although stabilization of RGS7 induced by TNF-α requires an area between Q170 and T247, this domain is not required for the proteasome-dependent degradation of RGS7. Truncations lacking this domain are poorly expressed and fail to respond to TNF-α, yet rapidly accumulate after proteasome inhibition. This observation indicates that the ubiquitin-ligating enzymes recognize peptide sequences distinct from the phosphorylation sites affected by TNF-α. It is possible that TNFα-induced generation of phosphoserines and phosphothreonines facilitates interaction between RGS7 and other

TNF-α requires phosphorylation and subsequent degradation of IκB, whereas phosphorylation of the protooncogene c-jun by JNK results in activation and suppresses ubiquitination and degradation of this transcription factor30,31. Similarly, posttranslational modification of RGS7 may increase its activity by impeding degradation. Site-directed mutagenesis of RGS7 showed that TNF-α-induced stabilization of RGS7 depends on candidate phosphorylation sites for MAPKs at S241, T245 and T247. As TNF-α triggers the activation of many MAPKs (ref. 32), we used dominant-negative mutants of protein kinases upstream of MAPKs to identify the p38 signaling cascade as essential for the TNF-induced upregulation of RGS7. Inhibition of kinases upstream of p38, a dominant-negative MEKK1 and a combination of dominant-negative MKK3 and MKK6 resulted in a substantial loss of the TNF-mediated RGS7 stabilization. In contrast, dominant-negative mutants of ERK1 and ERK2, or dominant-negative mutants of small G proteins, essential for the TNFmediated activation of JNK 1 in HEK 293T cells, had no effect on RGS7 protein levels. Although Fig. 4 Dominant-negative mutants of MEK1, ERK1/2, and PKC isozymes fail to suppress the TNF-mediated upthese results identify p38 as an regulation of RGS7. a, MEKK1, the p38 MAPKK kinase, consistently inhibits the TNF-mediated effects, whereas essential component of the TNF- neither dominant-negative (DN) nor wild-type (WT) MEK1, the ERK1/2 MAPK kinase, affects TNF-mediated mediated upregulation of RGS7, RGS7 protein levels. b, A combination of dominant-negative (DN) ERK1 and ERK2 does not inhibit the TNF-meour findings do not preclude the diated increase of RGS7 protein levels. c, Dominant-negative (DN) mutants of protein kinase C α, β and ε do not possibility that protein kinases inhibit the TNF-mediated increase of RGS7 protein levels.

a

916

b

c

NATURE MEDICINE • VOLUME 5 • NUMBER 8 • AUGUST 1999



ARTICLES

a

b

c

d

203580, from Calbiochem (Cambridge, Massachusetts). Purified activated p38 was purchased from Upstate Biotechnology (Lake Placid, New York). The cDNAs encoding RGS7 (E.K. et al., manuscript submitted), EGL-10 (provided by M.R. Koelle and H.R. Horvitz) and RGS9 (a gift from W. He and T.G. Wensel) were subcloned into CDM8 to obtain FLAG-His-tagged constructs (F.RGS7, F.EGL-10, F.RGS9) with the tag sequence MLDYKDDDDKHHHHHHHHH. FLAG–His-tagged green fluorescent protein (GFP) in CDM8 was used to monitor transfection efficiency. RGS7 truncations were generated as described20. Tagged chimeric proteins were obtained by subcloning MluI–PstI (M1–L156), PstI–NcoI (Q157–N304) and NcoI–NotI (P305–Y469) fragments of RGS7 into F.EGL-10, replacing the corresponding EGL-10 fragments. The use of Cdc42, Rac-1, RhoA and HA-p38 and PKC constructs has been described33. Dominant-negative mutants of MKK3 and MKK6 were provided by R. J. Davis; dominant-negative MKK4, by A. R. Morrison; and MEKK1 constructs, by D. Templeton. The multimeric HAubiquitin was provided by D. Bohmann. The M2 antibody against FLAG was from Sigma, and antibodies against HA, β-catenin, PKC α and 14-3-3 were from Santa Cruz Biotechnology (Santa Cruz, California). RGS7-reactive antiserum was generated as described (T.B. et al., manuscript submitted). Site-directed mutagenesis. Mutagenesis of human RGS7 used a PCR-directed approach with wild-type human RGS7 as template and respective primers. PCR fragments were subcloned into unique preexisting or engineered restriction sites of RGS7. All sequences were verified by automated DNA sequencing.

Fig. 5 LPS and TNF-α increase RGS7 protein levels in mouse brain. a, Western blot analysis of brain lysates obtained from mice 4-6 h after injection of LPS (1 mg) or solvent (Control). b, Western blot analysis of brain lysates obtained from mice 4 h after injection of TNF-α (20 µg) or solvent (Control). c, Quantitative assessment of the LPS- and TNF-α mediated upregulation of RGS7 shows a 200–250% increase. d, TNFR1-deficient mice fail to upregulate RGS7 in response to LPS. Western blot analysis of brain lysates obtained from mice 4 h after injection of LPS (1 mg) or solvent (control). Detection of PKCα confirms equal protein loading.

Western blot analysis. HEK 293T cells seeded in 6-well plates were transiently transfected with 1–2 µg of DNA per well using the calcium phosphate method. Flag-tagged GFP (F.GFP) was co-transfected to confirm transfection efficiency. Proteins were separated by 10% SDS–PAGE, transferred to polyvinylidene difluoride (PVDF) membranes, and were visualized using specific antibodies in combination with enhanced chemiluminescence (Pierce, Rockford, Illinois). Equal loading was confirmed by staining the PVDF membranes with amido black. In some experiments, proteins were quantified by densitometry in the linear range of film exposure by video-scanning densitometry.

cytoplasmic proteins that impede the ubiquitin-dependent degradation of RGS7. Alternatively, TNF-α-mediated phosphorylation could induce conformational changes of RGS7 that directly impede ubiquitin-dependent degradation. To study the regulation of RGS7 levels in vivo, we determined whether a condition associated with high systemic TNF-α levels induced the accumulation of RGS7 in the brain. Endotoxin (LPS), which elicits a rapid release of TNF-α during the course of endotoxin-induced sepsis syndrome, upregulated RGS7 protein levels in the brain. Furthermore, endotoxin failed to induce the accumulation of RGS7 in TNFR1-deficient mice, indicating that the RGS7 response to endotoxin is mainly mediated through the elaboration of TNF-α and its binding to TNFR1. TNF-α is considered an essential mediator of sepsis-induced behavioral changes and temperature regulation5-7. As many neurotransmitters act through GαI- and Gαq-coupled receptors, a TNF-mediated increase of RGS7 protein levels could alter certain behavioral responses during sepsis. RGS7 is highly expressed in the hypothalamus, diencephalon and brain stem, areas of the central nervous system that regulate neurovegetative responses of arousal, appetite and libido, as well as temperature. Defining the function of RGS7 in the brain and the possible consequences of its upregulation during sepsis will require further studies to identify the neuronal subtypes that express RGS7 and to determine the neurotransmitter pathways regulated by RGS7.

Ubiquitination assay. HEK 293T cells seeded in 10-cm dishes were transiently transfected with constructs (10 µg of total DNA) using the calcium phosphate method. After 24 h, cells were incubated with TNF-α, MG132 or media alone, lysed in 6 M guanidinium-HCl, 0.1 M Na2HPO4/NaH2PO4, pH 8.0, and 5 mM imidazole, and then sonicated for 1 min. After centrifugation for 15 min at 4 ºC, the His-tagged RGS7 conjugates were precipitated from the cleared lysate with Ni2+-nitrilotriacetic acid agarose (Qiagen) for 2 h at room temperature. Complexes were washed with 8 M urea, 0.1 M Na2HPO4/NaH2PO4, pH 6.3, and 0.01 M Tris, pH 8.0, and then resuspended in sample buffer. Proteins were separated by SDS–PAGE, electroblotted to a PVDF membrane, and labeled with rabbit polyclonal antiserum against HA and M2 antibody against FLAG, followed by incubation with horseradish peroxidase-coupled immunoglobulin (Dako, Carpinteria, California). Immobilized antibodies were detected by chemiluminescence (Pierce, Rockford, Illinois).

Methods Reagents, plasmids and antibodies. TNF-α was obtained from PeproTech (Rocky Hill, New Jersey); Escherichia coli O127:B8 LPS, from Sigma; and SB NATURE MEDICINE • VOLUME 5 • NUMBER 8 • AUGUST 1999

Purification and in vitro phosphorylation of His-RGS7170–469. The RGS7170–469 truncation was subcloned into pET30b and expressed in E. coli BL21-DE3 (both from Novagen, Madison, Wisconsin). His-RGS7170–469 protein was affinity-purified using Ni2+–NTA–agarose resin. To determine whether RGS7 can be phosphorylated in vitro by p38, 10 µg His-RGS7170–469 were incubated at 30 °C in kinase buffer (20 mM HEPES, pH 7.4, 20 mM NaF, 20 mM MgCl2, 1 mM Na3VO4, 1 mM dithiothreitol and 20 µM ATP) containing 10 µCi γ-32P-ATP in the presence or absence of activated p38 kinase (Stratagene, La Jolla, California). PHAS-1 (3 µg), a known substrate for p38, served as control for the kinase reaction. After 30 min of incubation, the phosphorylation reaction was terminated by addition of Laemmli sample buffer. Proteins were separated by 12% SDS–PAGE, followed by staining of gels with Coomassie blue, and then autoradiography. Septic mouse model. Female BALB/c mice (20 g in body weight; Charles River, Wilmington, Massachusetts) were either injected intraperitoneally with 1 mg E. coli O127:B8 LPS or sterile saline in a total volume of 1 ml, or 917


ARTICLES


injected with 20 µg murine TNF-α intravenously in the tail vein. Then, 4–6 h after injection, the whole brain was removed, ‘snap-frozen’ in liquid nitrogen and stored at –80 ºC. TNFR1-deficient mice and C57BL/6J control mice were obtained from The Jackson Laboratories (Bar Harbor, Maine). For preparation of protein extracts, frozen brains were homogenized and resuspended in 2 ml lysis buffer (20 mM Tris, pH 7.5, 50 mM NaCl, 1 mM EDTA, 1% Triton X-100 and 0.2% sodium dodecyl sulfate, containing protease inhibitor mix (Boehringer) and 44 µg/mL PMSF), and then incubated on ice for 20 min. Insoluble material was removed by centrifugation at 15,000g for 15 min at 4 ºC, and the protein concentration of the cleared brain lysate was determined using the BioRad DC protein assay. Equal amounts of protein (35 mg per lane) were separated by 10% SDS–PAGE. Loading of equal amounts of protein was confirmed by amido black staining of membranes and by reprobing against PKCα. To confirm equal protein loading, filters were reprobed against PKCα. Quantitative assessment was made using the NIH Image Program.

Acknowledgments We would like to thank A. Goldberg for advice on this manuscript. T.B. was supported by a fellowship of the Deutsche Forschungsgemeinschaft (DFG). This work was supported by NIH-RO1 DK52897 (G.W.) and PHS grant MH01147 (E.K.).

RECEIVED 17 MARCH; ACCEPTED 2 JUNE 1999 1. Haimovitz-Friedman, A. et al. Lipopolysaccharide induces disseminated endothelial apoptosis requiring ceramide generation. J. Exp. Med. 186, 1831–1841 (1997). 2. Selby, P. et al. Tumour necrosis factor in man: clinical and biological observations. Br. J. Cancer 56, 803–808 (1987). 3. Waage, A., Brandtzaeg, P., Espevik, T. & Halstensen, A. Current understanding of the pathogenesis of gram-negative shock. Infect. Dis. Clin. North Am. 5, 781–791 (1991). 4. Remick, D.G. & Kunkel, S.L. Pathophysiologic alterations induced by tumor necrosis factor. Int. Rev. Exp. Pathol. 34, 7–25 (1993). 5. Kapas, L. et al. Somnogenic, pyrogenic, and anorectic activities of tumor necrosis factor-alpha and TNF-alpha fragments. Am. J. Physiol. 263, R708–715 (1992). 6. Dunn, A.J. & Swiergiel, A.H. The role of cytokines in infection-related behavior. Ann. NY Acad. Sci. 840, 577–85 (1998). 7. Fang, J., Wang, Y. & Krueger, J.M. Mice lacking the TNF 55 kDa receptor fail to sleep more after TNFalpha treatment. J. Neurosci. 17, 5949–5955 (1997). 8. Dohlman, H.G. & Thorner J. RGS proteins and signaling by heterotrimeric G proteins. J. Biol. Chem. 272, 3871–384 (1997). 9. Berman, D.M. & Gilman A.G. Mammalian RGS proteins: barbarians at the gate. J. Biol. Chem. 273, 1269–1272 (1998). 10. Kehrl, J.H. Heterotrimeric G protein signaling: roles in immune function and fine- tuning by RGS proteins. Immunity 8, 1–10 (1998). 11. Koelle, M.R. A new family of G-protein regulators - the RGS proteins. Curr. Opin. Cell Biol. 9, 143–147 (1997). 12. Koelle, M.R., Horvitz HR. EGL-10 regulates G protein signaling in the C. elegans

918

nervous system and shares a conserved domain with many mammalian proteins. Cell 84, 115–125 (1996). 13. Gold, S.J., Ni, Y.G., Dohlman, H.G. & Nestler, E.J. Regulators of G-protein signaling (RGS) proteins: region-specific expression of nine subtypes in rat brain. J. Neurosci. 17, 8024–8037 (1997). 14. Wang, J. et al. RGSZ1, a Gz-selective RGS protein in brain. Structure, membrane association, regulation by galphaz phosphorylation, and relationship to a gz gtpase-activating protein subfamily. J. Biol. Chem. 273, 26014–26025 (1998). 15. Glick, J.L., Meigs, T.E., Miron, A. & Casey, P.J. RGSZ1, a Gz-selective regulator of G protein signaling whose action is sensitive to the phosphorylation state of gzalpha. J. Biol. Chem. 273, 26008–26013 (1998). 16. Ingi, T. et al. Dynamic regulation of RGS2 suggests a novel mechanism in G-protein signaling and neuronal plasticity. J. Neurosci. 18, 7178–7188 (1998). 17. Shuey, D.J., Betty, M., Jones, P.G., Khawaja, X.Z. & Cockett, M.I. RGS7 attenuates signal transduction through the G(alpha q) family of heterotrimeric G proteins in mammalian cells. J. Neurochem. 70, 1964–1972 (1998). 18. Thomas, E.A., Danielson, P.E. & Sutcliffe, J.G. RGS9: a regulator of G-protein signalling with specific expression in rat and mouse striatum. J. Neurosci. Res. 52, 118–124 (1998). 19. Saugstad, J.A., Marino, M.J., Folk, J.A., Hepler, J.R. & Conn, P.J. RGS4 inhibits signaling by group I metabotropic glutamate receptors. J. Neurosci. 18, 905–913 (1998). 20. Kim, E. et al. Interaction between RGS7 and polycystin. Proc. Natl. Acad. Sci. USA 96, 6371–6376 (1999). 21. Aberle, H., Bauer, A., Stappert, J., Kispert, A. & Kemler, R. beta-catenin is a target for the ubiquitin-proteasome pathway. Embo. J. 16, 3797–3804 (1997). 22. Salomon, D. et al. Regulation of beta-catenin levels and localization by overexpression of plakoglobin and inhibition of the ubiquitin-proteasome system. J. Cell. Biol. 139, 1325–35 (1997). 23. Abu-Amer, Y., Ross, F.P., Edwards, J. & Teitelbaum, S.L. Lipopolysaccharidestimulated osteoclastogenesis is mediated by tumor necrosis factor via its P55 receptor. J. Clin. Invest. 100, 1557–1565 (1997). 24. He,W., Cowan, C.W. & Wensel, T.G. RGS9, a GTPase accelerator for phototransduction. Neuron 20, 95–102 (1998). 25. Khawaja, X.Z. et al. Immunohistochemical distribution of RGS7 protein and cellular selectivity in colocalizing with Galphaq proteins in the adult rat brain. J. Neurochem. 72, 174–184 (1999). 26. Segalat, L., Elkes, D.A. & Kaplan, J.M. Modulation of serotonin-controlled behaviors by Go in Caenorhabditis elegans. Science 267, 1648–1651 (1995). 27. Saitoh, O. et al. RGS7 and RGS8 differentially accelerate G protein-mediated modulation of K+ currents. J. Biol. Chem. 274, 9899–9904 (1999). 28. Hershko, A. & Ciechanover, A. The ubiquitin system. Annu. Rev. Biochem. 67, 425–479 (1998). 29. Rechsteiner, M. PEST sequences are signals for rapid intracellular proteolysis. Semin. Cell Biol. 1, 433–440 (1990). 30. Pulverer, B.J., Kyriakis, J.M., Avruch, J., Nikolakaki, E. & Woodgett, J.R. Phosphorylation of c-jun mediated by MAP kinases. Nature 353, 670–674 (1991). 31. Musti, A.M., Treier, M. & Bohmann, D. Reduced ubiquitin-dependent degradation of c-Jun after phosphorylation by MAP kinases. Science 275, 400–402 (1997). 32. Boone, E., Vandevoorde, V., De Wilde, G. & Haegeman, G. Activation of p42/p44 mitogen-activated protein kinases (MAPK) and p38 MAPK by tumor necrosis factor (TNF) is mediated through the death domain of the 55-kDa TNF receptor. FEBS Lett. 441, 275–280 (1998). 33. Arnould, T. et al. The polycystic kidney disease 1 gene product mediates protein kinase C alpha-dependent and c-Jun N-terminal kinase-dependent activation of the transcription factor AP-1. J. Biol. Chem. 273, 6013–6018 (1998).

NATURE MEDICINE • VOLUME 5 • NUMBER 8 • AUGUST 1999