FBXW7 regulates DISC1 stability via the ubiquitin-proteosome system.

Disrupted in schizophrenia 1 (DISC1) is a multi-functional scaffolding protein that has been associated with neuropsychiatric disease. The role of DISC1 is to assemble protein complexes that promote neural development and signaling, hence tight control of the concentration of cellular DISC1 in neurons is vital to brain function. Using structural and biochemical techniques, we show for we believe the first time that not only is DISC1 turnover elicited by the ubiquitin proteasome system (UPS) but that it is orchestrated by the F-Box protein, FBXW7. We present the structure of FBXW7 bound to the DISC1 phosphodegron motif and exploit this information to prove that disruption of the FBXW7-DISC1 complex results in a stabilization of DISC1. This action can counteract DISC1 deficiencies observed in neural progenitor cells derived from induced pluripotent stem cells from schizophrenia patients with a DISC1 frameshift mutation. Thus manipulation of DISC1 levels via the UPS may provide a novel method to explore DISC1 function.

Human phosphodiesterase 4D7 (PDE4D7) expression is increased in TMPRSS2-ERG-positive primary prostate cancer and independently adds to a reduced risk of post-surgical disease progression.

BACKGROUND: There is an acute need to uncover biomarkers that reflect the molecular pathologies, underpinning prostate cancer progression and poor patient outcome. We have previously demonstrated that in prostate cancer cell lines PDE4D7 is downregulated in advanced cases of the disease. To investigate further the prognostic power of PDE4D7 expression during prostate cancer progression and assess how downregulation of this PDE isoform may affect disease outcome, we have examined PDE4D7 expression in physiologically relevant primary human samples. METHODS: About 1405 patient samples across 8 publically available qPCR, Affymetrix Exon 1.0 ST arrays and RNA sequencing data sets were screened for PDE4D7 expression. The TMPRSS2-ERG gene rearrangement status of patient samples was determined by transformation of the exon array and RNA seq expression data to robust z-scores followed by the application of a threshold>3 to define a positive TMPRSS2-ERG gene fusion event in a tumour sample. RESULTS: We demonstrate that PDE4D7 expression positively correlates with primary tumour development. We also show a positive association with the highly prostate cancer-specific gene rearrangement between TMPRSS2 and the ETS transcription factor family member ERG. In addition, we find that in primary TMPRSS2-ERG-positive tumours PDE4D7 expression is significantly positively correlated with low-grade disease and a reduced likelihood of progression after primary treatment. Conversely, PDE4D7 transcript levels become significantly decreased in castration resistant prostate cancer (CRPC). CONCLUSIONS: We further characterise and add physiological relevance to PDE4D7 as a novel marker that is associated with the development and progression of prostate tumours. We propose that the assessment of PDE4D7 levels may provide a novel, independent predictor of post-surgical disease progression.

The cAMP phosphodiesterase-4D7 (PDE4D7) is downregulated in androgen-independent prostate cancer cells and mediates proliferation by compartmentalising cAMP at the plasma membrane of VCaP prostate cancer cells.

BACKGROUND: Isoforms of the PDE4 family of cAMP-specific phosphodiesterases (PDEs) are expressed in a cell type-dependent manner and contribute to underpinning the paradigm of intracellular cAMP signal compartmentalisation. Here we identify the differential regulation of the PDE4D7 isoform during prostate cancer progression and uncover a role in controlling prostate cancer cell proliferation. METHODS: PDE4 transcripts from 19 prostate cancer cell lines and xenografts were quantified by qPCR. PDE4D7 expression was further investigated because of its significant downregulation between androgen-sensitive (AS) and androgen-insensitive (AI) samples. Western blot analysis, PDE activity assay, immunofluorescent staining and cAMP responsive FRET assays were used to investigate the sub-plasma membrane localisation of a population of PDE4D7 in VCaP (AS) and PC3 (AI) cell lines. Disruption of this localisation pattern using dominant-negative protein expression and siRNA knockdown showed that PDE4D7 acts in opposition to proliferative signalling as assessed by electrical impedance-based proliferation assays. RESULTS: Here we identify the differential regulation of the PDE4D7 isoform during prostate cancer progression. PDE4D7 is highly expressed in AS cells and starkly downregulated in AI samples. The significance of this downregulation is underscored by our finding that PDE4D7 contributes a major fraction of cAMP degrading PDE activity tethered at the plasma membrane and that displacement of PDE4D7 from this compartment leads to an increase in the proliferation of prostate cancer cells. PDE4D7 mRNA expression is not, however, directly regulated by the androgen receptor signalling axis despite an overlapping genomic structure with the androgen responsive gene PART1. PDE4D7, which locates to the plasma membrane, acts to supress aberrant non-steroidal growth signals within the prostate or AS metastasis. CONCLUSIONS: PDE4D7 expression is significantly downregulated between AS and AI cell phenotypes. This change in expression potentially provides a novel androgen-independent biomarker and manipulation of its activity or its expression may provide therapeutic possibilities and insights into contributory aspects of the complex molecular pathology of prostate cancer.

The psychiatric disease risk factors DISC1 and TNIK interact to regulate synapse composition and function.

Disrupted in schizophrenia 1 (DISC1), a genetic risk factor for multiple serious psychiatric diseases including schizophrenia, bipolar disorder and autism, is a key regulator of multiple neuronal functions linked to both normal development and disease processes. As these diseases are thought to share a common deficit in synaptic function and architecture, we have analyzed the role of DISC1 using an approach that focuses on understanding the protein-protein interactions of DISC1 specifically at synapses. We identify the Traf2 and Nck-interacting kinase (TNIK), an emerging risk factor itself for disease, as a key synaptic partner for DISC1, and provide evidence that the DISC1-TNIK interaction regulates synaptic composition and activity by stabilizing the levels of key postsynaptic density proteins. Understanding the novel DISC1-TNIK interaction is likely to provide insights into the etiology and underlying synaptic deficits found in major psychiatric diseases.

Disruption of the cyclic AMP phosphodiesterase-4 (PDE4)-HSP20 complex attenuates the ß-agonist induced hypertrophic response in cardiac myocytes.

The small heat shock protein HSP20 is known to be cardioprotective during times of stress and the mechanism underlying its protective abilities depends on its phosphorylation on Ser16 by PKA (protein kinase A). Although the external stimuli that trigger Ser16 phosphorylation have been well studied, the events that modulate spatial and temporal control of this modification remain to be clarified. Here, we report that inhibition of cAMP phosphodiesterase-4 (PDE4) induces the phosphorylation of HSP20 in resting cardiac myocytes and augments its phosphorylation by PKA following ß-adrenergic stimulation. Moreover, using peptide array technology, in vitro binding studies, co-immunoprecipitation techniques and immunocytochemistry, we show that HSP20 binds directly to PDE4 within a region of the conserved catalytic domain. We also show that FRET-based, genetically-encoded cAMP reporters anchored to HSP20 exhibit a larger response to PDE4 inhibition compared to free cytosolic cAMP reporters, suggesting that the interaction with PDE4 is crucial in modulating the highly localised pool of cAMP to which HSP20 is exposed. Using information gleaned from peptide array analyses, we developed a cell-permeable peptide that serves to inhibit the interaction of PDE4 with HSP20. Disruption of the HSP20-PDE4 complex, using this peptide, suffices to induce phosphorylation of HSP20 by PKA and to protect against the hypertrophic response measured in neonatal cardiac myocytes following chronic ß-adrenergic stimulation.

Tailoring cAMP-signalling responses through isoform multiplicity.

Multiple forms of cAMP phosphodiesterases (PDE), adenylate cyclase and protein kinase A (PKA) allow cells to tailor the responsiveness of the cAMP-signalling system and to allow for its dynamic adjustment. Multiple forms of these enzymes confer spatial and temporal characteristics on cAMP signalling so as to affect compartmentalised responses within a single cell type. The ability to breach the PKA activation threshold can depend upon either or both the activation of adenylate cyclase and inhibition of specific PDE isoforms.

Phosphorylation of PDE4A5 by MAPKAPK2 attenuates fibrin degradation via p75 signalling.

Phosphodiesterases (PDEs) shape local cAMP gradients to underpin the specificity of receptor function. Key to this process is the highly defined nature of the intra-cellular location of PDEs in the cell. PDE4A5 is a PDE isoform that specifically degrades cAMP and is known to associate with the p75 neurotrophin receptor (p75NTR) where it modulates cAMP signalling cascades that regulate extracellular matrix remodelling in the lungs. Here we map and validate novel protein-protein interaction sites that are important for formation of the PDE4A5-p75NTR complex and show, for the first time, that phosphorylation of PDE4A5 by MAPKAPK2 enhances PDE4A5 interaction with p75NTR and that this, in turn, serves to attenuate fibrin degradation.

PDE4 associates with different scaffolding proteins: modulating interactions as treatment for certain diseases.

cAMP is an ubiquitous second messenger that is crucial to many cellular processes. The sole means of terminating the cAMP signal is degradation by cAMP phosphodiesterases (PDEs). The PDE4 family is of particular interest because PDE4 inhibitors have therapeutic potential for the treatment of various inflammatory and auto-immune diseases and also have anti-depressant and memory-enhancing effects. The subcellular targeting of PDE4 isoforms is fundamental to the compartmentalization of cAMP signaling pathways and is largely achieved via proteinprotein interactions. Increased knowledge of these protein-protein interactions and their regulatory properties could aid in the design of novel isoform-specific inhibitors with improved efficacy and fewer prohibitive side effects.

Identification of the pertussis and cholera toxin substrates in normal and N-ras transformed NIH3T3 fibroblasts and an assessment of their involvement in bombesin-stimulation of inositol phospholipid metabolism.

Cholera and pertussis toxin-sensitive G-proteins were examined using specific immunological probes in wild type NIH3T3 cells and in clones of these cells containing the N-ras gene attached to a promotor where expression either was (T15+) or was not (T15-) induced. The major pertussis toxin sensitive-polypeptide had the immunological characteristics of Gi2. Two distinct forms of Gs alpha (45 and 42 kDa) were identified. Long term over-expression of p21N-ras (T15+ cells) did not alter the levels of Gi2 alpha or of Gs alpha. Pretreatment of NIH3T3 or T15 cells with either pertussis toxin or cholera toxin led to the complete in situ ADP-ribosylation of the respective G-proteins. Modification of Gi2 by pertussis toxin, however, had no inhibitory effect on the ability of bombesin to stimulate the production of inositol phosphates in any of these cells lines. Treatment of these cells with cholera toxin elicited a potent inhibition of the bombesin-stimulated production of inositol phosphates. This could be mimicked, however, by other agents which increase intracellular cyclic AMP concentrations. Cholera toxin treatment did not produce a significant alteration in the number of bombesin receptors on the cell surface. These results suggest that, in the T15 cell line, enhanced coupling of bombesin receptors to a phospholipase C-mediated hydrolysis of inositol phospholipids is either produced directly by p21N-ras or that overexpression of this gene product leads to the enhanced expression or function of a cholera and pertussis toxin-insensitive G-protein which then mediates the effect.

Mutant but not normal p21 ras elevates inositol phospholipid breakdown in two different cell systems.

Expression of oncogenic mutant p21 ras either in stably transformed NIH3T3 fibroblasts or transiently in COS-1 cells elevates the basal rate of inositol phosphate production. Additional mutations in the effector region or the carboxy terminal region which abolish transforming capacity on NIH3T3 cells block the effect of oncogenic mutant p21 ras on basal rates. Overexpression of normal (Gly12) p21 ras has no such effect on this second messenger signalling system. These differences between overexpressed normal p21 ras and oncogenic mutant p21 ras strongly suggest that the increased basal rate of inositol phosphate production is a direct consequence of constitutively activated p21 ras activity.

Determination of G-protein levels, ADP-ribosylation by cholera and pertussis toxins and the regulation of adenylyl cyclase activity in liver plasma membranes from lean and genetically diabetic (db/db) mice.

Liver plasma membranes prepared from genetically diabetic (db/db) mice expressed levels of Gi alpha-2, Gi alpha-3 and G-protein beta-subunits that were reduced by some 75, 63 and 73% compared with levels seen in membranes from lean animals. In contrast, there were no significant differences in the expression of the 42 and 45 kDa forms of Gs alpha-subunits. Pertussis toxin-catalysed ADP-ribosylation of membranes from lean animals identified a single 41 kDa band whose labelling was reduced by some 86% in membranes from diabetic animals. Cholera toxin-catalysed ADP-ribosylation identified two forms of Gs alpha-subunits whose labelling was about 4-fold greater in membranes from diabetic animals compared with those from lean animals. Maximal stimulations of adenylyl cyclase activity by forskolin (100 microM), GTP (100 microM), p[NH]ppG (100 microM), NaF (10 mM) and glucagon (10 microM) were similar in membranes from lean and diabetic animals, whereas stimulation by isoprenaline (100 microM) was lower by about 22%. Lower concentrations (EC50-60 nM) of p[NH]ppG were needed to activate adenylyl cyclase in membranes from diabetic animals compared to those from lean animals (EC50-158 nM). As well as causing activation, p[NH]ppG was capable of eliciting a pertussis toxin-sensitive inhibitory effect upon forskolin-stimulated adenylyl cyclase activity in membranes from both lean and diabetic animals. However, maximal inhibition of adenylyl cyclase activity in membranes from diabetic animals was reduced to around 60% of that found using membranes from lean animals. Pertussis toxin-treatment in vivo enhanced maximal stimulation of adenylyl cyclase by glucagon, isoprenaline and p[NH]ppG through a process suggested to be mediated by the abolition of functional Gi activity. The lower levels of expression of G-protein beta-subunits, in membranes from diabetic compared with lean animals, is suggested to perturb the equilibria between holomeric and dissociated G-protein subunits. We suggest that this may explain both the enhanced sensitivity of adenylyl cyclase to stimulation by p[NH]ppG in membranes from diabetic animals and the altered ability of pertussis and cholera toxins to catalyse the ADP-ribosylation of G-proteins in membranes from these two animals.

Quinidine and melittin both decrease the fluidity of liver plasma membranes and both inhibit hormone-stimulated adenylate cyclase activity.

Increasing concentrations of either quinidine or melittin gave a dose-dependent inhibition of both the glucagon- and fluoride-stimulated activities of adenylate cyclase in the liver plasma membranes. At similar concentrations these agents increased the order of liver plasma membranes as detected by a fatty acid ESR probe, doxyl stearic acid. This increase in bilayer order (decrease in 'fluidity') is suggested to explain the inhibitory action of quinidine on adenylate cyclase activity but only in part contributes to the inhibitory action of melittin on adenylate cyclase. Arrhenius plots of fluoride-stimulated activity became non-linear in the presence of either quinidine or melittin, with a single well-defined break occurring at around 12 degrees C in each instance. Arrhenius plots of the glucagon-stimulated activity also exhibited such a novel break at around 12 degrees C when either quinidine or melittin were present as well as exhibiting a break at around 28 degrees C, as was seen in the absence of these ligands. The fatty acid spin probe inserted into liver plasma membranes detected a novel lipid phase separation occurring at around 12 degrees C when either quinidine or melittin was present and showed that the lipid phase separation occurring at around 28 degrees C in native membranes was apparently unaffected by these ligands.

Genetically acquired diabetes: adipocyte guanine nucleotide regulatory protein expression and adenylate cyclase regulation.

Adipocyte membranes from diabetic (db/db) animals showed marked elevations in the levels of alpha-subunits for Gi-1 which were almost twice those found in membranes from their normal, lean littermates. In contrast, no apparent differences were noted for levels of the alpha-subunits of Gi-2 and Gi-3, the 42 and 45 kDa forms of Gs and for G-protein beta-subunits. Adenylate cyclase specific activity was similar in membranes from both normal and diabetic animals under basal conditions and also when stimulated by optimal concentrations of either NaF or forskolin. In contrast, the ability of isoprenaline, glucagon and secretin to stimulate adenylate cyclase activity was greater in membranes from normal animals compared with membranes from diabetic animals. Receptor-mediated inhibition of adenylate cyclase, as assessed using PGE1 and nicotinate, was similar using membranes from both sources, but PIA (phenylisopropyladenosine) was a slightly more effective inhibitor in membranes from diabetic animals. A doubling in the expression of Gi-1 thus appears to have little discernible effect upon the inhibitory regulation of adenylate cyclase.

The effects of p21N-ras expression in NIH-3T3 cells upon cyclic AMP metabolism.

The effects of overexpression of p21N-ras upon cyclic AMP metabolism have been examined in the inducible T15 cell line. In cells overexpressing the N-ras gene product, beta-adrenergic stimulation of cyclic AMP generation was reduced. The reduction was more pronounced the longer the ras gene was expressed and in chronically transformed cells a reduction in forskolin-stimulated cyclic AMP generation was also observed. The transformed cells exhibited a reduction in beta-adrenergic binding sites, but no change in the apparent EC50 for agonist induced cyclic AMP generation. Treatment of the cells with dibutyryl cyclic AMP induced a dose-dependent inhibition of proliferation, with the transformed cells being more sensitive than the control cells.

Dimethylnitrosamine inhibits the glucagon-stimulated adenylate cyclase activity of rat liver plasma membranes and decreases plasma membrane fluidity.

The effect of the hepatocarcinogen dimethylnitrosamine on rat liver plasma membrane adenylate cyclase activity and lipid fluidity was assessed. Glucagon-stimulated adenylate cyclase activity exhibited a complex response to increasing concentrations of dimethylnitrosamine, whereas fluoride-stimulated adenylate cyclase activity was progressively inhibited. Maximal inhibitory effects were observed at a concentration of 15 mM in both cases. The activity of detergent-solubilized adenylate cyclase was unaffected by dimethylnitrosamine. ESR analysis using a fatty acid spin probe showed that dimethylnitrosamine produced a marked, dose-dependent reduction in the fluidity of the plasma membrane with a maximal effect occurring at 20 mM. Dimethylnitrosamine also elevated the temperature at which the lipid phase separation occurred in rat liver plasma membranes, from 28 degrees C to 31 degrees C. The non-carcinogenic but structurally similar compound, dimethylamine hydrochloride neither inhibited adenylate cyclase nor decreased plasma membrane fluidity. It is suggested that the decrease in membrane fluidity, induced by dimethylnitrosamine, via its effects on membrane fluidity, could influence plasma membrane function and cellular regulation.

Perturbations of liver plasma membranes induced by Ca2+ are detected using a fatty acid spin label and adenylate cyclase as membrane probes.

Ca2+ decreased the lipid fluidity of rat liver plasma membranes labeled with 5-nitroxide stearate, I(12,3), as indicated by the order parameter (S). These effects form a reversible, saturable process with an association constant of 1 x 10(3) M-1. Arrhenius-type plots of S indicated that the lipid phase separation, present in the external leaflet of native membranes between 28 and 19 degrees C, is perturbed by mM Ca2+ such that the high temperature onset is elevated to 32-34 degrees C. Fluoride-stimulated adenylate cyclase was similarly inhibited by Ca2+ (ID50 = 1 mM) for the enzyme in membrane-bound or solubilized states. The glucagon-stimulated activity was more sensitive to Ca2+ inhibition with an ID50 of 0.2 mM. These inhibitory effects are due neither to perturbations of glucagon binding to its receptor nor to fluidity changes, but are instead attributed to direct Ca2+-enzyme interactions. Such binding desensitizes the enzyme to fluidity alterations induced by temperature elevation or benzyl alcohol addition. With Ca2+, Arrhenius plots of glucagon-stimulated activity indicated breaks at 32 and 16 degrees C, whereas those of fluoride-stimulated activity showed one break at 17 degrees C. Without Ca2+, Arrhenius plots exhibited one break at 28 degrees C for glucagon-stimulated activity, whereas fluoride-stimulated plots were linear. We propose that Ca2+ achieves these effects through asymmetric perturbations of the membrane lipid structure.

The glucagon receptor of rat liver plasma membrane can couple to adenylate cyclase without activating it.

1. Activation of adenylate cyclase in rat liver plasma membranes by fluoride or GMP-P (NH)P yielded linear Arrheniun plots. Activation by glucagon alone, or in combination with either fluoride or GMP-P(NH)P resulted in biphasic Arrhenius plots with a well-defined break at 28.5 +/- 1 degrees C. 2. The competitive glucagon antagonist, des-His-glucagon did not activate the adenylate cyclase but produced biphasic Arrhenius plots in combination with fluoride or GMP-P(NH)P. The break temperatures and activation energies were very similar to those observed with glucagon alone, or in combination with either fluoride or GMP-P(NH)P. 3. It is concluded that although des-His-glucagon is a potent antagonist of glucagon, it nevertheless causes a structural coupling between the receptor and the catalytic unit.

The lipid environment of the glucagon receptor regulates adenylate cyclase activity.

1. The lipids composition of rat liver plasma membranes was substantially altered by introducing synthetic phosphatidylcholines into the membrane by the techniques of lipid substitution or lipid fusion. 40-60% of the total lipid pool in the modified membranes consisted of a synthetic phosphatidylcholine. 2. Lipid substitution, using cholate to equilibrate the lipid pools, resulted in the irreversible loss of a major part of the adenylate cyclase activity stimulated by F-, GMP-P(NH)P or glucagon. However, fusion with presonicated vesicles of the synethic phosphatidylcholines causes only small losses in adenylate cyclase activity stimulated by the same ligands. 3. The linear form of the Arrhenius plots of adenylate cyclase activity stimulated by F- or GMP-(NH)P was unaltered in all of the membrane preparations modified by substitution or fusion, with very similar activation energies to those observed with the native membrane. The activity of the enzyme therefore appears to be very insensitive to its lipid environment when stimulated by F- or gmp-p(nh)p. 4. in contrast, the break at 28.5 degrees C in the Arrhenius plot of adenylate cyclase activity stimulated by glucagon in the native membrane, was shifted upwards by dipalmitoyl phosphatidylcholine, downwards by dimyristoyl phosphatidylcholine, and was abolished by dioleoyl phosphatidylcholine. Very similar shifts in the break point were observed for stimulation by glucagon or des-His-glucagon in combination with F- or GMP-P(NH)P. The break temperatures and activation energies for adenylate cyclase activity were the same in complexes prepared with a phosphatidylcholine by fusion or substitution. 5. The breaks in the Arrhenius plots of adenylate cyclase activity are attributed to lipid phase separations which are shifted in the modified membranes according to the transition temperature of the synthetic phosphatidylcholine. Coupling the receptor to the enzyme by glucagon or des-His-glucagon renders the enzyme sensitive to the lipid environment of the receptor. Spin-label experiments support this interpretation and suggest that the lipid phase separation at 28.5 degrees C in the native membrane may only occur in one half of the bilayer.

Altered G-protein expression and adenylate cyclase activity in platelets of non-insulin-dependent diabetic (NIDDM) male subjects.

Adenylate cyclase activity and levels of guanine nucleotide regulatory proteins (G-proteins) were compared in platelets from normal and non-insulin-dependent diabetic (NIDDM) male subjects. Whilst no differences were noted in basal and NaF-stimulated adenylate cyclase activities the degree of stimulation achieved by both forskolin and prostaglandin, E1 was lower by some 34 and 52% respectively, in platelet membranes from diabetic subjects compared with those from normal control subjects. Altered alpha 1-adrenoceptor-mediated inhibition of prostaglandin E1-stimulated adenylate cyclase activity was evident; it being some 34% lower in platelet membranes from diabetic subjects compared to controls. Analysis of G-protein alpha-subunits, using specific anti-peptide antisera, showed that platelets from all subjects exhibited the Gi-2 and Gi-3, but not the Gi-1 forms of the inhibitory G-protein 'Gi' and all expressed the 42 kDa species of alpha-subunit of the stimulatory G-protein Gs. Whilst platelets of diabetic subjects had levels of Gs which were comparable to those found in control subjects their levels of Gi-2 and Gi-3 were some 49 and 75%, respectively, of those found in platelets from control subjects. It is suggested that changes in adenylate cyclase functioning and G-protein expression may contribute to altered platelet functioning in non-insulin-dependent diabetic subjects.