Hippocampal long-term depression and depotentiation are defective in mice carrying a targeted disruption of the gene encoding the RI beta subunit of cAMP-dependent protein kinase.

The cAMP-dependent protein kinase (PKA) has been shown to play an important role in long-term potentiation (LTP) in the hippocampus, but little is known about the function of PKA in long-term depression (LTD). We have combined pharmacologic and genetic approaches to demonstrate that PKA activity is required for both homosynaptic LTD and depotentiation and that a specific neuronal isoform of type I regulatory subunit (RI beta) is essential. Mice carrying a null mutation in the gene encoding RI beta were established by use of gene targeting in embryonic stem cells. Hippocampal slices from mutant mice show a severe deficit in LTD and depotentiation at the Schaffer collateral-CA1 synapse. This defect is also evident at the lateral perforant path-dentate granule cell synapse in RI beta mutant mice. Despite a compensatory increase in the related RI alpha protein and a lack of detectable changes in total PKA activity, the hippocampal function in these mice is not rescued, suggesting a unique role for RI beta. Since the late phase of CA1 LTP also requires PKA but is normal in RI beta mutant mice, our data further suggest that different forms of synaptic plasticity are likely to employ different combinations of regulatory and catalytic subunits.

Identification of human cyclin-dependent kinase 8, a putative protein kinase partner for cyclin C.

Metazoan cyclin C was originally isolated by virtue of its ability to rescue Saccharomyces cerevisiae cells deficient in G1 cyclin function. This suggested that cyclin C might play a role in cell cycle control, but progress toward understanding the function of this cyclin has been hampered by the lack of information on a potential kinase partner. Here we report the identification of a human protein kinase, K35 [cyclin-dependent kinase 8 (CDK8)], that is likely to be a physiological partner of cyclin C. A specific interaction between K35 and cyclin C could be demonstrated after translation of CDKs and cyclins in vitro. Furthermore, cyclin C could be detected in K35 immunoprecipitates prepared from HeLa cells, indicating that the two proteins form a complex also in vivo. The K35-cyclin C complex is structurally related to SRB10-SRB11, a CDK-cyclin pair recently shown to be part of the RNA polymerase II holoenzyme of S. cerevisiae. Hence, we propose that human K35(CDK8)-cyclin C might be functionally associated with the mammalian transcription apparatus, perhaps involved in relaying growth-regulatory signals.

Association of mitogen-activated protein kinase with the microtubule cytoskeleton.

Using indirect immunofluorescence microscopy and biochemical techniques, we have determined that approximately one-third of the total mitogen-activated protein kinase (MAPK) is associated with the microtubule cytoskeleton in NIH 3T3 mouse fibroblasts. This population of enzyme can be separated from the soluble form that is found distributed throughout the cytosol and is also present in the nucleus after mitogen stimulation. The microtubule-associated enzyme pool constitutes half of all detectable MAPK activity after mitogenic stimulation. These findings extend the known in vivo associations of MAPK with microtubules to include the entire microtubule cytoskeleton of proliferating cells, and they suggest that a direct association of MAPK with microtubules may be in part responsible for the observed correlations between MAPK activities and cytoskeletal alteration.

A synthetic inhibitor of the mitogen-activated protein kinase cascade.

Treatment of cells with a variety of growth factors triggers a phosphorylation cascade that leads to activation of mitogen-activated protein kinases (MAPKs, also called extracellular signal-regulated kinases, or ERKs). We have identified a synthetic inhibitor of the MAPK pathway. PD 098059 [2-(2'-amino-3'-methoxyphenyl)-oxanaphthalen-4-one] selectively inhibited the MAPK-activating enzyme, MAPK/ERK kinase (MEK), without significant inhibitory activity of MAPK itself. Inhibition of MEK by PD 098059 prevented activation of MAPK and subsequent phosphorylation of MAPK substrates both in vitro and in intact cells. Moreover, PD 098059 inhibited stimulation of cell growth and reversed the phenotype of ras-transformed BALB 3T3 mouse fibroblasts and rat kidney cells. These results indicate that the MAPK pathway is essential for growth and maintenance of the ras-transformed phenotype. Further, PD 098059 is an invaluable tool that will help elucidate the role of the MAPK cascade in a variety of biological settings.

The product of the ataxia-telangiectasia group D complementing gene, ATDC, interacts with a protein kinase C substrate and inhibitor.

Ataxia-telangiectasia (AT) is an autosomal recessive human genetic disease characterized by immunological, neurological, and developmental defects and an increased risk of cancer. Cells from individuals with AT show sensitivity to ionizing radiation, elevated recombination, cell cycle abnormalities, and aberrant cytoskeletal organization. The molecular basis of the defect is unknown. A candidate AT gene (ATDC) was isolated on the basis of its ability to complement the ionizing radiation sensitivity of AT group D fibroblasts. Whether ATDC is mutated in any AT patients is not known. We have found that the ATDC protein physically interacts with the intermediate-filament protein vimentin, which is a protein kinase C substrate and colocalizing protein, and with an inhibitor of protein kinase C, hPKCI-1. Indirect immunofluorescence analysis of cultured cells transfected with a plasmid encoding an epitope-tagged ATDC protein localizes the protein to vimentin filaments. We suggest that the ATDC and hPKCI-1 proteins may be components of a signal transduction pathway that is induced by ionizing radiation and mediated by protein kinase C.

Cell-free activation of neutrophil NADPH oxidase by a phosphatidic acid-regulated protein kinase.

The phosphorylation-dependent mechanisms regulating activation of the human neutrophil respiratory-burst enzyme, NADPH oxidase, have not been elucidated. We have shown that phosphatidic acid (PA) and diacylglycerol (DG), products of phospholipase activation, synergize to activate NADPH oxidase in a cell-free system. We now report that activation by PA plus DG involves protein kinase activity, unlike other cell-free system activators. NADPH oxidase activation by PA plus DG is reduced approximately 70% by several protein kinase inhibitors [1-(5-isoquinolinesulfonyl)piperazine, staurosporine, GF-109203X]. Similarly, depletion of ATP by dialysis reduces PA plus DG-mediated NADPH oxidase activation by approximately 70%. Addition of ATP, but not a nonhydrolyzable ATP analog, to the dialyzed system restores activation levels to normal. In contrast, these treatments have little effect on NADPH oxidase activation by arachidonic acid or SDS plus DG. PA plus DG induces the phosphorylation of a number of endogenous proteins. Phosphorylation is largely mediated by PA, not DG. A predominant substrate is p47-phox, a phosphoprotein component of NADPH oxidase. Phosphorylation of p47-phox precedes activation of NADPH oxidase and is markedly reduced by the protein kinase inhibitors. In contrast, arachidonic acid alone or SDS plus DG is a poor activator of protein phosphorylation in the cell-free system. Thus, PA induces activation of one or more protein kinases that regulate NADPH oxidase activation in a cell-free system. This cell-free system will be useful for identifying a functionally important PA-activated protein kinase(s) and for dissecting the phosphorylation-dependent mechanisms responsible for NADPH oxidase activation.

Gene for the catalytic subunit of the human DNA-activated protein kinase maps to the site of the XRCC7 gene on chromosome 8.

The DNA-activated serine/threonine protein kinase (DNA-PK) is composed of a large (approximately 460 kDa) catalytic polypeptide (DNA-PKcs) and Ku, a heterodimeric DNA-binding component (p70/p80) that targets DNA-PKcs to DNA. A 41-kbp segment of the DNA-PKcs gene was isolated, and a 7902-bp segment was sequenced. The sequence contains a polymorphic Pvu II restriction enzyme site, and comparing the sequence with that of the cDNA revealed the positions of nine exons. The DNA-PKcs gene was mapped to band q11 of chromosome 8 by in situ hybridization. This location is coincident with that of XRCC7, the gene that complements the DNA double-strand break repair and V(D)J recombination defects (where V is variable, D is diversity, and J is joining) of hamster V3 and murine severe combined immunodeficient (scid) cells.

Basal expression of the cystic fibrosis transmembrane conductance regulator gene is dependent on protein kinase A activity.

The cystic fibrosis transmembrane conductance regulator (CFTR) functions as a Cl- channel that becomes activated after phosphorylation by cAMP-dependent protein kinase (PKA). We demonstrate that PKA also plays a crucial role in maintaining basal expression of the CFTR gene in the human colon carcinoma cell line T84. Inhibition of PKA activity by expression of a dominant-negative regulatory subunit or treatment with the PKA-selective inhibitor N-[2-(p-bromocinnamylamino)ethyl]-5-isoquinolinesulfonamide (H-89) caused a complete suppression of CFTR gene expression without affecting other constitutively active genes. Basal expression of a 2.2-kb region of the CFTR promoter linked to a luciferase reporter gene (CFTR-luc) exhibited the same dependence on PKA. The ability of cAMP to induce CFTR over basal levels is cell-type specific. In T84 cells, both the endogenous CFTR gene and CFTR-luc exhibited only a modest inducibility (approximately 2-fold), whereas in the human choriocarcinoma cell line JEG-3, CFTR-luc could be induced at least 4-fold. A variant cAMP-response element is present at position -48 to -41 in the CFTR promoter, and mutation of this sequence blocks basal expression. We conclude that cAMP, acting through PKA, is an essential regulator of basal CFTR gene expression and may mediate an induction of CFTR in responsive cell types.

MEKK1 phosphorylates MEK1 and MEK2 but does not cause activation of mitogen-activated protein kinase.

A constitutively active fragment of rat MEK kinase 1 (MEKK1) consisting of only its catalytic domain (MEKK-C) expressed in bacteria quantitatively activates recombinant mitogen-activated protein (MAP) kinase/extracellular signal-regulated protein kinase (ERK) kinases 1 and 2 (MEK1 and MEK2) in vitro. Activation of MEK1 by MEKK-C is accompanied by phosphorylation of S218 and S222, which are also phosphorylated by the protein kinases c-Mos and Raf-1. MEKK1 has been implicated in regulation of a parallel but distinct cascade that leads to phosphorylation of N-terminal sites on c-Jun; thus, its role in the MAP kinase pathway has been questioned. However, in addition to its capacity to phosphorylate MEK1 in vitro, MEKK-C interacts with MEK1 in the two-hybrid system, and expression of mouse MEKK1 or MEKK-C in mammalian cells causes constitutive activation of both MEK1 and MEK2. Neither cotransfected nor endogenous ERK2 is highly activated by MEKK1 compared to its stimulation by epidermal growth factor in spite of significant activation of endogenous MEK. Thus, other as yet undefined mechanisms may be involved in determining information flow through the MAP kinase and related pathways.

Convergent and parallel activation of low-conductance potassium channels by calcium and cAMP-dependent protein kinase.

K+ channels, which have been linked to regulation of electrogenic solute transport as well as Ca2+ influx, represent a locus in hepatocytes for the concerted actions of hormones that employ Ca2+ and cAMP as intracellular messengers. Despite considerable study, the single-channel basis for synergistic effects of Ca2+ and cAMP on hepatocellular K+ conductance is not well understood. To address this question, patch-clamp recording techniques were applied to a model liver cell line, HTC hepatoma cells. Increasing the cytosolic Ca2+ concentration ([Ca2+]i) in HTC cells, either by activation of purinergic receptors with ATP or by inhibition of intracellular Ca2+ sequestration with thapsigargin, activated low-conductance (9-pS) K+ channels. Studies with excised membrane patches suggested that these channels were directly activated by Ca2+. Exposure of HTC cells to a permeant cAMP analog, 8-(4-chlorophenylthio)-cAMP, also activated 9-pS K+ channels but did not change [Ca2+]i. In excised membrane patches, cAMP-dependent protein kinase (the downstream effector of cAMP) activated K+ channels with conductance and selectivity identical to those of channels activated by Ca2+. In addition, cAMP-dependent protein kinase activated a distinct K+ channel type (5 pS). These data represent the differential regulation of low-conductance K+ channels by signaling pathways mediated by Ca2+ and cAMP. Moreover, since low-conductance Ca(2+)-activated K+ channels have been identified in a variety of cell types, these findings suggest that differential regulation of K+ channels by hormones with distinct signaling pathways may provide a mechanism for hormonal control of solute transport and Ca(2+)-dependent cellular functions in the liver as well as other nonexcitable tissues.

Regulated assembly of tight junctions by protein kinase C.

We have previously shown that protein phosphorylation plays an important role in the sorting and assembly of tight junctions. We have now examined in detail the role of protein kinases in intercellular junction biogenesis by using a combination of highly specific and broad-spectrum inhibitors that act by independent mechanisms. Our data indicate that protein kinase C (PKC) is required for the proper assembly of tight junctions. Low concentrations of the specific inhibitor of PKC, calphostin C, markedly inhibited development of transepithelial electrical resistance, a functional measure of tight-junction biogenesis. The effect of PKC inhibitors on the development of tight junctions, as measured by resistance, was paralleled by a delay in the sorting of the tight-junction protein, zona occludens 1 (ZO-1), to the tight junction. The assembly of desmosomes and the adherens junction were not detectably affected, as determined by immunocytochemical analysis. In addition, ZO-1 was phosphorylated subsequent to the initiation of cell-cell contact, and treatment with calphostin C prevented approximately 85% of the phosphorylation increase. Furthermore, in vitro measurements indicate that ZO-1 may be a direct target of PKC. Moreover, membrane-associated PKC activity more than doubled during junction assembly, and immunocytochemical analysis revealed a pool of PKC zeta that appeared to colocalize with ZO-1 at the tight junction. A preformed complex containing ZO-1, ZO-2, p130, as well as 330- and 65-kDa phosphoproteins was detected by coimmunoprecipitation in both the presence and absence of cell-cell contact. Identity of the 330- and 65-kDa phosphoproteins remains to be determined, but the 65-kDa protein may be occludin. The mass of this complex and the incorporation of ZO-1 into the Triton X-100-insoluble cytoskeleton were not PKC dependent.

Phosphorylation of the human leukemia inhibitory factor (LIF) receptor by mitogen-activated protein kinase and the regulation of LIF receptor function by heterologous receptor activation.

We used a bacterially expressed fusion protein containing the entire cytoplasmic domain of the human leukemia inhibitory factor (LIF) receptor to study its phosphorylation in response to LIF stimulation. The dose- and time-dependent relationships for phosphorylation of this construct in extracts of LIF-stimulated 3T3-L1 cells were superimposable with those for the stimulation of mitogen-activated protein kinase (MAPK). Indeed, phosphorylation of the cytoplasmic domain of the low-affinity LIF receptor alpha-subunit (LIFR) in Mono Q-fractionated, LIF-stimulated 3T3-L1 extracts occurred only in those fractions containing activated MAPK; Ser-1044 served as the major phosphorylation site in the human LIFR for MAPK both in agonist-stimulated 3T3-L1 lysates and by recombinant extracellular signal-regulated kinase 2 in vitro. Expression in rat H-35 hepatoma cells of LIFR or chimeric granulocyte-colony-stimulating factor receptor (G-CSFR)-LIFR mutants lacking Ser-1044 failed to affect cytokine-stimulated expression of a reporter gene under the control of the beta-fibrinogen gene promoter but eliminated the insulin-induced attenuation of cytokine-stimulated gene expression. Thus, our results identify the human LIFR as a substrate for MAPK and suggest a mechanism of heterologous receptor regulation of LIFR signaling occurring at Ser-1044.

Isolation of a yeast protein kinase that is activated by the protein encoded by SRP1 (Srp1p) and phosphorylates Srp1p complexed with nuclear localization signal peptides.

Srp1p, the protein encoded by SRP1 of Saccharomyces cerevisiae, is a nuclear-pore-associated protein. Its Xenopus homolog, importin, was recently shown to be an essential component required for nuclear localization signal (NLS)-dependent binding of karyophilic proteins to the nuclear envelope [Gorlich, D., Prehn, S., Laskey, R. A. & Hartman, E. (1994) Cell 79, 767-778]. We have discovered a protein kinase whose activity is stimulated by Srp1p (Srp1p fused to glutathione S-transferase and expressed in Escherichia coli) and is detected by phosphorylation of Srp1p and of a 36-kDa protein, a component of the protein kinase complex. The enzyme, called Srp1p kinase, is a protein-serine kinase and was found in extracts in two related complexes of approximately 180 kDa and 220 kDa. The second complex, when purified, contained four protein components including the 36-kDa protein. We observed that, upon purification of the kinase, phosphorylation of Srp1p became very weak, while activation of phosphorylation of the 36-kDa protein by Srp1p remained unaltered. Significantly, NLS peptides and the nuclear proteins we have tested greatly stimulated phosphorylation of Srp1p, suggesting that Srp1p, complexed with karyophilic proteins carrying an NLS, is the in vivo substrate of this protein kinase.

Evidence for a role of endothelin 1 and protein kinase C in nitroglycerin tolerance.

We sought to examine mechanisms responsible for increased vasoconstriction that occurs during development of nitroglycerin tolerance. Rabbits were treated for 3 days with nitroglycerin patches (0.4 mg/hr), and their aortic segments were studied in organ chambers. This treatment resulted in attenuated in vitro relaxations to nitroglycerin and increased contractile sensitivity to angiotensin II, serotonin, phenylephrine, KCl, and a direct activator of protein kinase C, the phorbol ester phorbol 12,13-dibutyrate. The protein kinase C antagonists calphostin C (100 nM) and staurosporine (10 nM) corrected the hypersensitivity to constrictors in tolerant vessels, yet had minimal effects on constrictions in control vessels. Paradoxically, constrictions caused by endothelin 1 were decreased in nitrate-tolerant vessels. Immunocytochemical analysis revealed intense endothelin 1-like and big endothelin 1-like immunoreactivity in the media of nitroglycerin-tolerant but not of control aortas. The enhanced vasoconstriction to angiotensin II, serotonin, KCl, and phenylephrine could be mimicked in normal vessels by addition of subthreshold concentrations of endothelin 1, and this effect was prevented by calphostin C. We propose that increased autocrine production of endothelin 1 in nitrate tolerance sensitizes vascular smooth muscle to a variety of vasoconstrictors through a protein kinase C-mediated mechanism.

Cdc37 is required for association of the protein kinase Cdc28 with G1 and mitotic cyclins.

Studies of the temperature-sensitive cdc37-1 mutant of Saccharomyces cerevisiae suggest that Cdc37 is required for passage through the G1 phase of the cell cycle, but its precise function is not known. We have investigated the role of Cdc37 in the regulation of the cyclin-dependent protein kinase Cdc28. We find that G1 arrest in the cdc37-1 mutant is accompanied by a decrease in the Cdc28 activity associated with the G1 cyclin Cln2. This defect appears to be caused by a decrease in the binding of Cdc28 and Cln2. cdc37-1 mutants also exhibit a defect in the binding and activation of Cdc28 by the mitotic cyclin Clb2. Thus Cdc37 may be a regulator that is required for the association of Cdc28 with multiple cyclins.