Identification of a Novel Human Kinase Supporter of Ras (hKSR-2) That Functions as a Negative Regulator of Cot (Tpl2) Signaling*

Kinase suppressor of Ras (KSR) is an integral and conserved component of the Ras signaling pathway. Although KSR is a positive regulator of the Ras/mitogen-activated protein (MAP) kinase pathway, the role of KSR in Cot-mediated MAPK activation has not been identified. The serine/threonine kinase Cot (also known as Tpl2) is a member of the MAP kinase kinase kinase (MAP3K) family that is known to regulate oncogenic and inflammatory pathways; however, the mechanism(s) of its regulation are not precisely known. In this report, we identify an 830-amino acid novel human KSR, designated hKSR-2, using predictions from genomic data base mining based on the structural profile of the KSR kinase domain. We show that, similar to the known human KSR, hKSR-2 co-immunoprecipitates with many signaling components of the Ras/MAPK pathway, including Ras, Raf, MEK-1, and ERK-1/2. In addition, we demonstrate that hKSR-2 co-immunoprecipitates with Cot and that co-expression of hKSR-2 with Cot significantly reduces Cot-mediated MAPK and NF- (cid:1) B activation. This inhibition is specific to Cot, because Ras-in-duced ERK and I (cid:1) B kinase-induced NF- (cid:1) B activation are not assays (cid:4) g of DNA per 5–7 (cid:4) g/well a 6-well plate. hKSR-2 dose-dependent studies, hKSR-2 plasmid amounts varied to 4 (cid:4) g in a 6-well plate, the Cot plasmid concentration constant at 2 (cid:4) g. For ELISA tests, HeLa cells transfected at FuGENE 6-to-DNA ratio to h post-transfection, the cell from DMEM with 10% fetal bovine serum to DMEM with 0.5% fetal bovine serum and either left untreated or treated with 100 ng/ml recombinant human TNF- (cid:2) (R&D systems, for 5 h. During the treatments, cells were changed from DMEM supplemented with 10% fetal bovine serum to 0.5% fetal bovine serum containing DMEM. The plasmids epitope hKSR-2-FLAG, Cot-Myc, TAK1-FLAG, and TAB1-Myc for various immunoblotting and immunoprecipitation purposes. (pH 7.5), 150 m M NaCl, 1 m M dithiothreitol, 1 m M EDTA, 1 m M EGTA, 20 m M NaF, 20 m M (cid:3) -glycerophosphate, 1 m M NaVO 4 , and protease inhibitors) (Roche Applied Science). Lysates were clarified by centrifugation at 12,000 (cid:2) g for 15 min. Protein concentrations were determined using the BCA protein assay (Pierce). For immunoprecipitation, protein extracts were precleared with (cid:5) -bound Sepharose (Am- ersham Biosciences) beads. To detect hKSR-2-associated proteins, 1 mg of precleared lysate was subjected to anti-FLAG-agarose conjugate im- munoprecipitation. Bound proteins were eluted with a FLAG peptide and separated by two-dimensional gel electrophoresis. Proteins were resolved on the gel in one dimension, and the membranes were probed using a Miniblotter 20SL (Immunetics, Cambridge, MA) with at least 10 different antibodies as mentioned under “Results.” To immunoprecipitate Myc-tagged Cot and TAB1 proteins, 4 (cid:4) g of Anti-myc tag antibody (clone 9E10) (Upstate Cell Signaling Solutions, VA) was added to 250 (cid:4) g of precleared lysates and incubated overnight at 4 °C. 40 (cid:4) l of (cid:5) -bound Sepharose beads were added to each of the incubations for 2 h at 4 °C. The immunoprecipi- tated complexes were washed three times in phosphate-buffered saline and two times in 0.2% Nonidet P-40 containing phosphate-buffered saline, and immunoblotting was performed as described. immunoblot analyses, proteins resolved by electrophoresis and to nitrocellulose membranes (Invitrogen) using standard Immunore- active bands conjugated horseradish peroxidase Immu- noreactive bands visualized using an enhanced chemilumines-cence (ECL) reagent Cot Kinase Assay— Cot kinase immunoprecipitated cell beads M GST-MEK1-His kinase m M M M protease and phosphatase min, the kinase reaction stopped the of LDS loading (Invitrogen) boiled and separated by denaturing SDS-PAGE, and the radiolabeled proteins were visualized using the Molecular Imager FX (Bio-Rad). In Vitro Kinase Assay— Following immunoprecipitation of whole cell lysates with anti-FLAG antibody, the bead bound proteins were eluted off the beads with FLAG peptide. The eluted hKSR-2 was incubated for 30 min °C with recombinant active Cot (expressed in a baculovirus and isolated on nickel-agarose beads) or Raf enzyme (Upstate Biotech-nology) the of 5 (cid:4) of (cid:5) - 33 P]ATP and n M GST-MEK-

KSR 1 constitutes a protein kinase family that is structurally related to the Raf family of kinases (1). However, there are significant functional differences between the KSR and Raf protein families. KSR was initially isolated by selection and complementation of genetic mutations in Drosophila and Caenorhabditis elegans (1)(2)(3). These studies determined that KSR functions downstream of Ras and either upstream of or parallel to Raf kinase. Unlike Raf, KSR does not contain a consensus Ras binding domain (RBD) (3)(4)(5). The amino-terminal regions of Drosophila and mammalian KSR contain four conserved domains, i.e. CA1 to CA4. CA1 is a domain unique to KSR, proline-rich CA2 contains a Src homology 2 (SH2) domain, CA3 is a cysteine-rich domain, and CA4 is serine/threonine-rich domain (3). The carboxyl-terminal region of KSR contains the CA5 domain, which encompasses the 11 conserved kinase subdomains found in all known protein kinases.
Studies using KSR knock-out mice and RNA interference (RNAi) technology have confirmed a positive regulatory role of KSR in Ras-mediated MAP kinase activation (2,6). Given that KSR lacks several key properties of a protein kinase, including a conserved lysine in its kinase subdomain II, the role of its kinase activity is still unsettled. KSR may exert its effect through direct phosphorylation of Raf-1 (7)(8)(9) and may also function as a scaffolding protein through its interactions with multiple components of the Ras/MAP kinase pathway such as Raf-1, MEK1/2, and ERK1/2 (10). Thus, it has been proposed that KSR may function as a scaffolding protein or a kinase for the coordination of the Ras/MAP kinase pathway (11)(12)(13).
Cot (Tpl2 in rat), a human proto-oncogene (14,15), is a serine/threonine kinase in the MAP kinase kinase kinase (MAP3K) family (MAP3K8) (16). Cot expression has been shown to induce ERK and JNK activation (17,18). Similar to Raf, Cot is a MEK-1 kinase upstream of ERK pathway. Overexpression of Cot in CD3-activated T cells leads to IL-2 production, suggesting that Cot may play a role in T cell activation (19,20). In addition, Cot has been shown to activate members of the NF-B family, possibly by activating the IB kinase (IKK) complex through the NF-B-inducing kinase (NIK) or by inducing the degradation of the inhibitory protein p105 (21)(22)(23). Recent studies using Cot knock-out mice point to a pivotal role of Cot in the LPS-induced production of TNF-␣ and other pro-inflammatory cytokines (17,24).
Despite these various important cellular functions, the precise molecular mechanism of Cot regulation remains unsolved. Given that activation of both Cot and Raf lead to activation of downstream MAP kinase pathway components and that KSR is known to function as a regulatory scaffold protein in the Raf/ MEK/MAP kinase pathway, we have evaluated the functional interrelationship of Cot and hKSR-2.
Using a structure-based data base mining approach with a * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The conserved kinase domain module, we have identified and cloned a novel KSR family member, hKSR-2. Similar to C. elegans KSR-2, this clone appears to be lacking the CA1 domain of the KSR family and has 66% nucleotide identity and 61% amino acid identity with the known human KSR. hKSR-2 interacts with Cot when ectopically expressed in HEK-293T cells, and hKSR-2 expression attenuates Cot-mediated ERK and NF-B activation. Furthermore, the co-expression of hKSR-2 blocks Cot-mediated IL-8 production in HeLa cells. These data suggest a novel inhibitory role for hKSR-2 in regulating Cot-mediated signal transduction.

MATERIALS AND METHODS
Cloning and Plasmids-Using the predicted hKSR-2 cDNA sequences (see "Results"), two partial cDNA clones (B11 and F7) were generated by Invitrogen using the human testis Marathon-Ready cDNA library. Full-length hKSR-2 was generated by combining two partial cDNA clones via PCR. The PCR product was inserted in to the Gateway entry vector pENTR/SD-TOPO and subsequently cloned into the destination vector pDEST-40. A FLAG tag was incorporated at the carboxyl terminus of each clone (FLAG primer is 5Ј-GACTACAAGGATGA CGACGATAAG-3Ј). The catalytic domain of KSR was generated by designing a PCR primer spanning the lysine residue at position 527 (5Ј-CACCATGGAGCAGCTGGAGATCGGCGAGCTCATT) and the FLAG tag sequence incorporated at the carboxyl terminus (3Ј-CTTCA-GACGTCTCATGATGTTCCTACTGCTGCTATTCATCCGCGGCG). The constructs were confirmed by DNA sequencing.
Cell Culture and Transfection-HEK-293T, HeLa, and RAW264.7 cell lines were maintained in DMEM supplemented with 10% fetal bovine serum, 2% L-glutamine, and 1% penicillin/streptomycin (Invitrogen). Cells were passaged every 2-3 days as required to maintain log phase growth for all experiments. For immunoblots and kinase assays, HEK-293T cells were transfected either with the full-length or the catalytic domain of hKSR-2 at 10 g per 1000 l of calcium phosphate solution (Profection mammalian transfection systems; Promega, Madison, WI). HeLa and RAW264.7 cells were transfected with Fu-GENE 6 (Roche Applied Science) and LipofectAMINE 2000 (Invitrogen). Luciferase assays were performed with either 2.5-5 g of DNA per well in a 24-well plate or with 5-7 g/well in a 6-well plate. For hKSR-2 dose-dependent studies, the hKSR-2 plasmid amounts were varied from 0.5 to 4 g in a 6-well plate, keeping the Cot plasmid concentration constant at 2 g. For ELISA tests, HeLa cells were transfected at a FuGENE 6-to-DNA ratio of 3 to 2. 48 h post-transfection, the cell culture medium was changed from DMEM with 10% fetal bovine serum to DMEM with 0.5% fetal bovine serum and then either left untreated or treated with 100 ng/ml recombinant human TNF-␣ (R&D systems, Minneapolis, MN) for 5 h. During the treatments, cells were changed from DMEM supplemented with 10% fetal bovine serum to 0.5% fetal bovine serum containing DMEM. The various plasmids with epitope tags used include hKSR-2-FLAG, Cot-Myc, TAK1-FLAG, and TAB1-Myc for various immunoblotting and immunoprecipitation purposes.
Immunoprecipitation and Immunoblotting-Cells were lysed on ice for 20 min in a non-ionic detergent buffer (0.2% Nonidet P-40, 25 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, 1 mM EGTA, 20 mM NaF, 20 mM ␤-glycerophosphate, 1 mM NaVO 4 , and protease inhibitors) (Roche Applied Science). Lysates were clarified by centrifugation at 12,000 ϫ g for 15 min. Protein concentrations were determined using the BCA protein assay (Pierce). For immunoprecipitation, protein extracts were precleared with ␥-bound Sepharose (Amersham Biosciences) beads. To detect hKSR-2-associated proteins, 1 mg of precleared lysate was subjected to anti-FLAG-agarose conjugate immunoprecipitation. Bound proteins were eluted with a FLAG peptide and separated by two-dimensional gel electrophoresis. Proteins were resolved on the gel in one dimension, and the membranes were probed using a Miniblotter 20SL (Immunetics, Cambridge, MA) with at least 10 different antibodies as mentioned under "Results." To immunoprecipitate Myc-tagged Cot and TAB1 proteins, 4 g of Anti-myc tag antibody (clone 9E10) (Upstate Cell Signaling Solutions, Charlottesville, VA) was added to 250 g of precleared lysates and incubated overnight at 4°C. 40 l of ␥-bound Sepharose beads were added to each of the incubations for 2 h at 4°C. The immunoprecipitated complexes were washed three times in phosphate-buffered saline and two times in 0.2% Nonidet P-40 containing phosphate-buffered saline, and immunoblotting was performed as described.
Cot Kinase Assay-Cot kinase was immunoprecipitated from whole cell lysates using anti-Cot antibody (Santa Cruz) bound to GammaBind Plus-Sepharose beads (Amersham Biosciences). The beads were washed in kinase buffer (see below) without ATP and then incubated at 30°C with 5 Ci of [␥-33 P]ATP (PerkinElmer Life Sciences; 3000 Ci/mmol) and 100 nM GST-MEK1-His 6 (Upstate Biotechnology) in kinase buffer containing 20 mM MOPS (pH.7.2), 100 M ATP, 5 mM EGTA, 20 mM MgCl 2 , and protease and phosphatase inhibitors. After 30 min, the kinase reaction was stopped by the addition of LDS loading buffer (Invitrogen) and boiled and separated by denaturing SDS-PAGE, and the radiolabeled proteins were visualized using the Molecular Imager FX (Bio-Rad).
In Vitro Kinase Assay-Following immunoprecipitation of whole cell lysates with anti-FLAG antibody, the bead bound proteins were eluted off the beads with FLAG peptide. The eluted hKSR-2 was incubated for 30 min at 30°C with recombinant active Cot (expressed in a baculovirus and isolated on nickel-agarose beads) or Raf enzyme (Upstate Biotechnology) in the presence of 5 Ci of [␥-33 P]ATP and 100 nM GST-MEK-His 6 . The reactions were stopped, separated, and visualized as above and quantitated on a Bio-Rad densitometer using Quant One software.
Luciferase Assay-Transfections were carried out as mentioned above. Cells were lysed for 48 h post transfection in 1 ϫ passive lysis buffer from the Promega Dual luciferase reporter assay kit. Cells were co-transfected with NF-B luciferase and pRL-TK, a renilla luciferase reporter plasmid, to normalize the experimental results. Briefly, 20 l of cell lysate was transferred to a 96-well Cliniplate (Thermo Lab Systems, Franklin, MA) designed specifically for a luciferase assay. The Fluoroscan Ascent FL luminometer (Lab systems) was programmed to perform a 2-s pre measurement followed by a 10-s reading of luciferase activity.
ELISA-Culture media were harvested 48 h following transfection of HeLa cells and frozen at Ϫ80°C until further use. IL-8 concentrations in these samples were determined using the anti-human IL-8 ELISA kit from BIOSOURCE. This assay is specific for IL-8 with a minimum detectable concentration of 15.6 pg/ml. Assays were run in duplicate, and samples were read at 450 nm in an ELISA plate reader (Wallac 1420 Multilable Counter, PerkinElmer Life Sciences). IL-8 concentrations were determined by comparison to a standard with a two-parameter curve fit analysis, and the dilution factor was taken into account.

RESULTS
Cloning of hKSR-2-The hKSR-2 cDNA was derived from genomic data base mining using the structural profiles of the catalytic domains of various kinases. To achieve this, all available x-ray crystal structures of the catalytic domains of serine/ threonine and tyrosine kinases were collected from the SCOP data base (http://scop.berkeley.edu). Structural alignments were performed using the ProCeryon package (www.proceryon. com). The structural alignments were presented in the form of multiple sequence alignments, which were used to search the Celera Human Genome Genscan predictions (release R25h) using the NCBI psiblast package. A predicted gene, which was homologous to the known human KSR, was identified. This predicted sequence was supported by an expressed sequence tag (EST) clone (BF948353). The expressed sequence tag-supported region was selected as a gene capture oligo for cloning, and two isolates (B11 and F7) were obtained from a human testis cDNA library (Fig. 1A). The presence of an in-frame stop codon upstream of the predicted initiator methionine argues that this cDNA contains the complete open reading frame. Based on the multiple sequence alignments of all known KSR family members, the amino terminus of the B11 and carboxyl terminus of F7 were combined by PCR to obtain the hKSR-2 cDNA, which has a 2490-bp long open reading frame that encodes an 829-aa protein kinase. hKSR-2 contains proline-rich CA2, cysteine-rich CA3, the serine/threonine repeats, the consensus MAPK phosphorylation site of CA4, and the conserved kinase catalytic domain CA5 (Fig. 1B). However, hKSR-2 lacks subdomain CA1, which is similar to the known hKSR-1 and is a domain that is unique to the KSR family (3). A murine homolog of hKSR-2 (mKSR-2) was predicted using Genewise software, and a cDNA clone was obtained from a mouse kidney cDNA library. The 5893-bp mouse cDNA has 88.9% identity to hKSR-2 cDNA, and the predicted 830-aa protein has 96.9% identity at the amino acid level with hKSR-2. Northern blot revealed that hKSR-2 is mainly expressed in brain and kidney (data not shown).
Interaction of hKSR-2 with the Members of MAPK Family-To examine whether hKSR-2 interacts with components of the MAP The putative kinase domain has an arginine instead of a lysine residue, which is shown in magenta. B, schematic illustration of the KSR homologues. Drosophila (Dm), C. elegans (Ce), mouse (m), and human (h) KSR members are depicted. Five conserved areas (CA1 to CA5) are shown. hKSR-2 and the known hKSR lack the CA1 domain (pink) similar to CeKSR2 (not shown). The respective positions for the CA2 (gray), CA3 (red), and CA4 (teal) in the amino terminus of the KSR are also shown. The kinase domain (blue), located at the carboxyl terminus, is ϳ300 amino acids long and is conserved across the KSR homologs, and this corresponds to CA5.

hKSR-2 Is a Negative Regulator of Cot/Tpl2
kinase pathway, we performed co-immunoprecipitation experiments using lysates of hKSR-2-transfected HEK-293T cells. Similar to the known KSR, over-expressed hKSR-2 also interacts with endogenous Raf, MEK, and ERK ( Fig. 2A, right panel) (25). Control immunoprecipitates using vector-transfected lysates did not precipitate any of these proteins ( Fig. 2A, left panel). Cot is an important signaling kinase directly upstream of MEK-1 and has been shown to be a strong activator of ERK when overexpressed in cells. Therefore, we performed co-immunoprecipitation experiments to evaluate the potential interaction of hKSR-2 with Cot. FIG. 2. A, hKSR-2 interacts with MAPK family members. HEK-293T cells were transfected with 5 g of plasmid DNA encoding either FLAG-tagged hKSR-2 (right panel) or vector alone (Control; left panel). 48 h post transfection cells were lysed, and 250 g of precleared protein lysates were incubated overnight with FLAG antibody. FLAG tag immunoprecipitates from both vector and FLAGtagged hKSR-2 transfected cells were eluted with a FLAG peptide. FLAG peptideeluted fractions were boiled in Laemmli sample buffer resolved by SDS-PAGE and immunoblotted with phosphorylated ERK1/2 (P.ERK), MEK1/2, phosphorylated MEK1/2 (P.MEK1), and phosphorylated Raf (P.Raf) antibodies using a mini-blotter. B and C, hKSR-2 and Cot interact with each other. HEK-293T cells were transfected with 5 g of plasmid DNA encoding either FLAGtagged hKSR-2 or Myc-tagged Cot or both (hKSR-2 ϩ Cot) as mentioned above. 48 h post transfection cells were lysed, and protein lysates were incubated overnight with FLAG-(B) or Myc-specific (C) antibodies to immunoprecipitate FLAG-tagged hKSR-2 or Myc-tagged Cot, respectively. Immunoprecipitates (IP) were boiled and resolved by SDS-PAGE as mentioned above and immunoblotted (IB) by the antibodies as indicated (B and C). D, hKSR-2 interacts selectively with Cot. HEK-293T cells were transfected with 5 g of plasmid DNA encoding either FLAG-tagged TAK1, Myc-tagged TAB1, or FLAG-tagged hKSR-2. The cell lysates were either immunoprecipitated with anti-Myc (upper panel) or anti-FLAG (lower panel) antibodies and subsequently immunoblotted with antiFLAG antibodies. FLAG-tagged hKSR-2 is not detected in the immune complex directed against co-transfected TAK1/TAB1, suggesting that hKSR-2 selectively associates with Cot and not with overexpressed TAK1/TAB1. Anti-FLAG immunoprecipitation followed by immunoblot analysis with anti-FLAG antibody confirmed the presence of hKSR-2 in these samples (lower panel). hKSR-2 Is a Negative Regulator of Cot/Tpl2 Fig. 2 shows that Cot can be co-immunoprecipitated with epitopetagged hKSR-2 when the two proteins are co-expressed in HEK-293T cells. Both Cot and hKSR-2 were detected in immunoblots of immunoprecipitates performed with the cognate antibody or the reciprocal antibody (Fig. 2, B and C).
To evaluate the specificity of the association of hKSR-2 and Cot, we performed co-immunoprecipitation experiments with hKSR-2 and another MAP3K family member, TAK1. TAK1 is known to form an active complex with TAB1 (26). FLAG-tagged TAK1 is readily detected in anti-Myc immunoprecipitates directed against co-transfected Myc-TAB1 (Fig. 2D). However, FLAG-tagged hKSR-2 is not detected in the immune complex directed against over-expressed TAK1/TAB1 (Fig. 2D, upper  panel), suggesting that hKSR-2 selectively associates with Cot and not with over-expressed TAK1/TAB1. Anti-FLAG immunoprecipitation followed by immunoblot analysis with anti-FLAG antibody confirmed the presence of hKSR-2 in these samples (Fig. 2D, lower panel).  3. A and B, hKSR-2 down-regulates Cot-induced ERK activation in a dose-dependent fashion. HEK-293T cells were transfected with 5 g of indicated plasmids. Cells were lysed 48 h post-transfection in the buffer containing protease and phosphatase inhibitors. A, immunoblots of whole cell lysates were probed with (left of panel) anti-phosphorylated ERK (P.ERK), anti-total ERK, anti-Cot, and anti-FLAG to detect the amounts of activated ERK, total ERK, Cot, and hKSR-2, respectively. Total ERK levels were shown for equal loading of protein. Cot-KD, Cot-kinase dead. B, control immunoblot was performed using C-Raf-transfected cell lysates to show that hKSR-2 inhibition is specific to Cot-induced ERK1 and ERK2 activation. C, increasing concentrations of hKSR-2 progressively decreased the protein levels of phosphorylated ERK, whereas the levels of total ERK and Cot remained constant. The phosphorylated ERK (P.ERK) levels were scanned, normalized based on total ERK loading in each lane, and represented with respect to Cot alone. The amounts of hKSR-2 plasmid transfected into HEK-293T cells were also represented.

hKSR-2 Is a Negative Regulator of Cot/Tpl2
The association between hKSR-2 and Cot is observed when either the ϳ105-kDa full-length hKSR-2 or the 40-kDa catalytic domain of hKSR-2 is used. Similarly, a 69-aa carboxyl terminal truncation of Cot did not effect its interaction with hKSR-2, as we could co-immunoprecipitate both truncated Cot and hKSR-2 in this system (data not shown). These data suggest that the carboxy terminus of hKSR-2 is sufficient for this association. Similarly, the carboxyl terminal 69-aa region of Cot is not required for its interaction with KSR.
hKSR-2 Attenuates Cot-induced ERK Activation in a Dosedependent Fashion-To determine the significance of the association of hKSR-2 and Cot in the downstream MAP kinase signaling pathway, we probed HEK-293T cell lysates in which both hKSR-2 and Cot were over-expressed with antibodies to phosphorylated ERK1/2 and total ERK1/2 (Fig. 3A, top and middle panels). As expected, over-expression of Cot activates ERK as demonstrated with anti-phospho ERK1/2 immunoblots. Co-transfection of hKSR-2 reduces Cot mediated ERK activation (Fig. 3A) by ϳ83% (as judged by densitometry). The expression levels of total ERK, Cot, and hKSR-2 were monitored with the respective antibodies, and similar levels of ex-pression of these proteins were detected (Fig. 3A, middle and  bottom panels). The Raf-mediated ERK activation is not significantly (20% by densitometry) attenuated by the co-expression of hKSR-2 (Fig. 3B, top panel), which indicates that hKSR-2 selectively inhibits Cot-mediated ERK activation.
hKSR-1 has been reported to have a positive and a negative regulator of Ras-dependent signal transduction based on its expression levels (27). To rule out the possibility that the inhibitory effect on Cot by hKSR-2 is not due to its over-expression, we have titrated down hKSR-2 concentrations while keeping Cot levels constant. We found a dramatic decrease in the levels of Cot-induced ERK activation by hKSR expression in a dose-dependent manner (Fig. 3C, top panel). The expression levels of total ERK, Cot, and hKSR-2 were monitored with the respective antibodies, and similar levels of expression of these proteins were detected (Fig. 3C, middle and bottom panels). The phosphorylated ERK (P.ERK in Fig. 3) levels were scanned, normalized based on the total ERK loading in each lane, and reported with respect to the Cot alone sample. We conclude that the hKSR-2 inhibition of Cot is not an artifact of transient transfections. hKSR-2 Is a Negative Regulator of Cot/Tpl2 hKSR-2 Attenuates Cot-induced NF-B Activation-Cot activity regulates NF-B activation by targeting the IKK complex (possibly through the NF-B-inducing kinase, which is an intermediate kinase) (28) or through a mechanism that enhances the proteolysis of the NF-B inhibitory protein p105 (21,28,29). In addition, Cot can associate with and promote the degradation of p105 and thereby release Rel family members (29). Therefore, by regulating NF-B activation, Cot plays a significant role in the pathophysiology of inflammation. To examine whether hKSR-2 expression has an effect on the Cot-induced NF-B activation, we used a dual luciferase reporter gene assay system to determine NF-B activation. As shown in Fig.  4A, over-expression of Cot-induces a 6-fold increase in expression of the luciferase reporter gene compared with control. Over-expression of hKSR-2 alone did not activate NF-B in this system. However, Cot-induced NF-B activation is inhibited by the co-expression of hKSR-2 by ϳ90%. The comparable expression levels of various proteins in the experiment were detected by immunoblot (Fig. 4A). Similar to anti-phosphorylated ERK levels, in a co-transfection experiment hKSR-2 reduced Cotmediated NF-B levels in dose-dependent manner (Fig. 4B), again suggesting that hKSR-2 inhibition of Cot is not an artifact of transient transfections.
hKSR-2 Does Not Inhibit IKK␤-mediated NF-B Activation-We examined whether the inhibition of Cot-induced NF-B activation by hKSR-2 is mediated directly through Cot or through other signaling components of the pathway. To address this question, we co-transfected IKK␤ and hKSR-2 and determined the effect of hKSR-2 on IKK␤-induced NF-B activation. In contrast to our observation with Cot, co-expression of hKSR-2 did not have any effect on the IKK␤-induced NF-B activation (Fig. 5). These results demonstrate the selectivity of the observed Cot inhibition by hKSR-2 and show that the inhibition is not mediated through IKK␤.
hKSR-2 Regulates Kinase Activity of Cot in an in Vitro Kinase Assay-Given that the kinase activity of Cot is necessary for the activation of ERK and NF-B, we performed in vitro kinase assays to determine whether hKSR-2 is effective at inhibiting Cot kinase activity. We utilized the recombinant Cot in the presence or absence of immunoprecipitated hKSR-2 using a kinase-inactive MEK1 as a substrate. hKSR-2 reduced Cot activity by ϳ60% at a 2-g concentration of hKSR-2 (Fig. 6,  lanes 2-5), whereas Raf-mediated MEK phosphorylation was up-regulated by 70% at a 2-g concentration of hKSR-2 (Fig. 6,  lanes 6 -8). This indicates that hKSR-2 selectively blocks Cotmediated but not Raf-mediated MEK phosphorylation in this system. The percent control of MEK phosphorylation by either Cot or Raf-1 was indicated at the respective concentrations of hKSR-2 shown at the bottom of Fig. 6.
hKSR-2 Attenuates Cot-mediated IL-8 Production in HeLa Cells-Cot kinase activity is upstream of NF-B, ERK, and JNK activation and is required for the production of LPSinduced inflammatory cytokines such as TNF-␣ and IL-1 (17,24). IL-8, a potent neutrophil chemoattractant and a proinflammatory cytokine, has been shown to be regulated, at least in part, by ERK pathway signaling (30); therefore, we investigated the possibility that Cot kinase activity also regulates IL-8 production. There was a significant increase in IL-8 production upon the over-expression of Cot in the HeLa cell line. This Cot-induced IL-8 production was markedly reduced by the co-expression of hKSR-2 (Fig. 7). Similar to previous reports (31,32), over-expression of TAK1/TAB1 in this system also increased spontaneous IL-8 production. However, co-expression of hKSR-2 with TAK1/TAB1 did not cause a reduction in the levels of TAK1/TAB1-induced IL-8 production. Thus, the ability of hKSR-2 to attenuate Cot-induced IL-8 production in HeLa cells is selective, inasmuch as hKSR-2 had no effect on TAK1/ TAB1-induced IL8 production. These results clearly show a selective inhibitory role of hKSR-2 in Cot-induced IL-8 production. DISCUSSION KSR was originally identified as a positive regulator of the Ras-MAP kinase signaling pathway in Drosophila and C. elegans (33). Previous reports have demonstrated that KSR may exerts its function either as a scaffold protein through interaction with various signaling proteins, including Ras, Raf, MEK, ERK, and 14-3-3 (10,34,35) or through direct phosphorylation of Raf (7)(8)(9). In the current study, we characterize a novel member of KSR family, hKSR-2, which was identified by genomic data base mining using structural profiles of catalytic domains of various serine/threonine and protein-tyrosine kinases. A human testis cDNA library was used to isolate hKSR-2, and it has 66% identity at the nucleotide level and 61% identity at the peptide level with the known human KSR. Similar to the case of the known KSR, we have demonstrated that hKSR-2 is an integral component of the MAP kinase pathway through its physical interaction with the members of MAP kinase pathway, including Raf, MEK, ERK1, and ERK2. We have also observed a physical association of hKSR-2 with Cot, a member of MAP3K family. This association between hKSR-2 and Cot is observed when either the 106-kDa fulllength hKSR-2 or the 40-kDa catalytic domain of hKSR-2 is used. This indicates that the carboxyl terminus of hKSR-2 is sufficient for this association. Similarly, we have observed that the 40-kDa domain spanning the 30 -398-aa region of Cot is sufficient to interact with the carboxyl terminus of hKSR-2.
Cot was originally identified in a screen for transforming genes from a human thyroid carcinoma cell line (15). Although the precise mechanism of Cot regulation is not fully understood, a recent report by Kane et al. indicates that Akt (protein kinase B) up-regulates Cot activity by the phosphorylation of serine 400 at the carboxyl terminus (21). In this report we provide evidence for a negative regulatory mechanism of Cot by hKSR-2. In an in vitro kinase assay, immunoprecipitated hKSR-2 decreased the activity of recombinant Cot when an inactive MEK-1 is used as a substrate. Cot kinase activity is upstream of NF-B, ERK, and JNK activation (18,36,37). As reported previously in several cell systems, Cot activity regu- HEK-293T cells were transfected with the corresponding plasmids as described under "Material and Methods," and luciferase assays were performed as in Fig. 4. The statistical significance of the results was determined by a paired two-sample Student's t test of the means and found out to be not significant (NS). Equal protein loading is shown in the bottom panel.
lates NF-B activation by targeting the IKK complex (possibly through the NF-B-inducing kinase, which is an intermediate kinase) (28) or through a mechanism that enhances the proteolysis of the NF-B inhibitory protein p105 (21,23,29). We demonstrate here that Cot-induced NF-B activity is significantly inhibited by the co-expression of hKSR-2. This inhibition of NF-B by hKSR-2 is specific for Cot-induced NF-B activation and is independent of or upstream of IKK␤, because IKK␤induced NF-B is not affected by the co-expression of hKSR-2. hKSR-2 selectively inhibited the Cot-mediated activation of MEK by ϳ 60%. In contrast, hKSR-2 up-regulated the Rafmediated MEK activation by up to ϳ70%. These data indicate that hKSR-2 specifically blocks Cot kinase activity and, thus, subsequently attenuates downstream signaling in the MAP kinase and NF-B pathways.
Based on the Cot knock-out mouse data (17), it is believed that intact Cot kinase activity is required for the production of LPS-induced production of inflammatory cytokines such as TNF-␣ and IL-1 and other inflammatory mediators, such as prostaglandin E 2 , produced by cyclooxygenase-2 (COX-2) (38). We provide evidence that Cot activity may also be required for production of the proinflammatory cytokine IL-8. Over-expression of Cot in HeLa cells causes an increased production of IL-8, and this increase requires the kinase activity of Cot, because the over-expression of a kinase-dead Cot mutant does not have an effect on IL-8 production. Consistent with the inhibition of MAP kinase and NF-B, Cot-induced IL-8 production can be selectively blocked by hKSR-2. IL-8, a C-X-C chemokine, is a potent neutrophil chemoattractant and a proinflammatory chemokine that is produced by many cell types including monocytes and polymorphonuclear cells in response to inflammatory stimuli. IL-8 is present in the infiltrate of many inflammatory diseases including glomerulonephritis, rheumatoid arthritis, Crohn's disease, and bacterial meningitis (39) and has been implicated in the chronic progression of these inflammatory diseases. Our study suggests that Cot controls IL-8 production, FIG. 6. hKSR-2 inhibits Cot but not Raf activity in an in vitro kinase assay in a dose-dependent manner. Recombinant Cot or Raf was incubated with increasing amounts of a FLAG peptideeluted fraction of hKSR-2 with GST-MEK as a substrate in the presence of 5 Ci of [␥-33 P]ATP. Reactions were stopped after 30 min of incubation with sample buffer at 30°C. Samples were boiled, centrifuged, and separated by SDS-PAGE. The gel was dried onto a Whatman paper and autoradiographed. The phosphorylated MEK (P.MEK) levels were scanned, and the densitometry units are reported at the bottom. The amounts of FLAG peptideeluted hKSR-2 protein levels were indicated as well, showing that hKSR-2 attenuated Cot-mediated MEK phosphorylation in a dose-dependent manner but not Raf-mediated MEK phosphorylation. hKSR-2 Is a Negative Regulator of Cot/Tpl2 perhaps through MAP kinase and NF-B pathways. Our data confirm and extend the results of previous reports that showed that IL-8 is regulated by MAP kinase pathway signaling (30, 40 -43).
TAK1, another member of MAP3K family, is also shown to induce IL-8 production by activating the NF-B pathway when coexpressed together with its coactivator, TAB1 (32). Although we confirm this data in our findings, we also report that hKSR-2 does not have any effect on TAK1/TAB1-induced IL-8 production (Fig. 7). Consistent with this observation, we were not able to detect any physical association between these two proteins in co-immunoprecipitation experiments (Fig. 2C). The selectivity of hKSR-2-mediated attenuation of Cot but not of TAK1/TAB1-induced IL-8 strengthens the argument that the association of hKSR-2 and Cot is a physiologically relevant event. The difference in regulation of the three MAP3Ks (Cot, Raf, and TAK1) by hKSR-2 is quite intriguing, although the precise mechanism remains unclear.
Interestingly, overexpression of Cot in HeLa cells did not lead to an increase of other pro-inflammatory cytokines such as TNF-␣, IL-1, and IL-6 (data not shown). These observations are inconsistent with the physiology of the Cot knock-out mouse in which Cot mediates LPS-induced TNF-␣ and IL-1 production in monocytes and macrophages (17). Further investigations are required to delineate the effect of Cot on IL-8 production and help elucidate its precise role in cytokine production.
Although our data are consistent with a model in which the effect of hKSR-2 is specific to Cot, the precise mechanism of this negative regulation is unknown. Given that both Cot and KSR are known to interact with MEK and that the association of hKSR-2 with endogenous MEK is readily detected in HEK-293T cells ( Fig. 2A), one potential mechanism of KSR-mediated inhibition of Cot may involve the blocking of a Cot binding site on MEK. However, the fact that KSR does not inhibit Ras/Rafinduced ERK activation through MEK argues against the possibility that KSR is binding to and blocking a Cot phosphorylation site on MEK. Thus, a likely model of hKSR-2 inhibition of Cot is that hKSR-2 directly binds to Cot to prevent its activation of MEK.
In conclusion, we have identified and isolated a novel human KSR cDNA called hKSR-2, and our data indicate that hKSR-2 functions as a negative regulator for Cot-mediated signaling. We also provide evidence that Cot controls IL-8 production, possibly through the ERK and NF-B signaling pathways, and that Cot-induced IL-8 production is inhibited dramatically by co-expression of hKSR-2. These findings contrast with the positive regulatory role of KSR in the Ras-MAP kinase-signaling pathway. To our knowledge, this is the first report suggesting that a KSR family member may function as a positive or a negative regulatory protein, depending on its targets.