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 (cid:2) protein interaction sites that are important for formation of the PDE4A5 (cid:2) 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.

Cyclic adenosine monophosphate (cAMP), a ubiquitous second messenger molecule, is employed widely by Gs-coupled receptors to orchestrate a variety of receptorspecific responses.PDEs (phosphodiesterases) are the only known enzyme family that can breakdown cAMP, making them crucial for the fine control of receptor signalling cascades.Of the large PDE superfamily, PDE4 is the most versatile as its 25 isoforms, encoded by four genes, locate differentially within cells and tissues (4).Using a combination of approaches, including siRNAmediated knockdown, peptide displacement and dominant-negative isoforms, these enzymes have been shown to have unique, non-redundant roles, which span from inflammatory regulation to cardiac contractility (5).Generally, the modular structures of PDE4 isoforms contain an N-terminal targeting domain, which directs them to specific subcellular locations, UCR1 and UCR2 (upstream-conserved regions 1 and 2), which act as regulatory regions, and a C-terminal catalytic domain (see Fig. 1), which is highly conserved between isoforms (4).
The specific, subcellular targeting of individual PDE4 enzymes underpins the functionality of each isoform and is vital to ensure receptor-specific responses via the three dimensional shaping of cAMP gradients that result from receptor activation (4,6).
Compartmentalized degradation of cAMP in cells is made possible by the ability of PDE4s to integrate into macromolecular complexes, or signalosomes, via interactions with different cellular scaffold proteins, such as A-kinase anchoring proteins (AKAPs) and other signalosome components (4,7).This paradigm also provides a cellular desensitization mechanism, whereby compartmentalized increases in cAMP activate PKA (protein kinase A) pools localized in the vicinity of the PDE4 in order to phosphorylate and activate long PDE4 isoforms (8).MK2 is also known to phosphorylate the PDE4 isoform PDE4A5 (9), although this phosphorylation does not alter the PDE's enzymatic activity on its own.Instead, phosphorylation by MK2 attenuates the activation triggered by PKA phosphorylation, disrupting cellular desensitization to cAMP.Additionally, the MK2 phosphorylated PDE4A undergoes a conformational change, which reduces its binding to interacting proteins which associate with the UCR2 domain (9).Recently, MK2 was shown to interact with PDE4A5 with high fidelity via two separate docking sites, with such a process allowing for the efficient phosphorylation of PDE4A5 by this kinase (10).
The p75 neurotrophin receptor (p75NTR), although most commonly known for its role in the nervous system, also plays a role in respiratory inflammation (11).This low affinity receptor is unusual for a neurotrophin receptor, as it does not have any catalytic activity of its own.Instead, it functions by sequestering other signalling molecules that mediate downstream signalling actions.One example of this type of signalling is p75NTR's regulation of fibrin degradation.Fibrinolysis is an integral part of the matrix remodelling process, which contributes to tissue repair.Studies using mice deficient in p75NTR provided unequivocal evidence that the receptor functions to block fibrin degradation (12).Vital to this important function is the receptor's interaction with the cAMP/PKA pathway.Specifically, p75NTR's direct interaction with PDE4A5 increases local cAMP degradation, reducing the activation of PKA (13,14) resulting in the simultaneous downregulation of tissue plasminogen activator (tPA) and up-regulation of plasminogen activator inhibitor-1 (PAI-1).These actions promote scar formation and inhibit extracellular matrix remodelling.Previous attempts to map the sites of PDE4p75NTR interaction pinpointed domains in the C-terminal, LR1 (linker region 1) and the catalytic regions, of the PDE4 as being important, though little has been done to define these sites and provide a functional assessment (13).
Here we revisited the mapping of the sites on PDE4A that bind to the p75NTR intracellular domain and report that sites in both the UCR2 and catalytic domains of PDE4A5 are, functionally, the most important.Importantly, we show, for the first time, that phosphorylation of PDE4A by the p38MAPK downstream effector, MK2 enhances the interaction of PDE4 with p75NTR restricting local cAMP in order to reduce the degradation of fibrin.

Materials and Methods
Peptide array 25-mer peptides were immobilized on cellulose membranes using an Autospot Robot ASS222 (Intavis Õ ) as previously described (15).Arrays were activated in 100% ethanol, washed in TBS-T and blocked with 5% milk powder solution (Marvel Õ ), before incubation in a solution of recombinant protein diluted in 1% milk powder solution.Arrays were thoroughly washed in TBS-T and incubated in a solution of the appropriate primary antibody diluted in 1% milk powder solution.Again, the arrays were washed and incubated in a solution of the appropriate secondary antibody-HRP conjugate diluted in 1% milk powder solution.The arrays were subjected to final washes and visualized using chemiluminescence.

PDE activity assay
Cyclic nucleotide phosphodiesterase activity assays were carried out in a two-step radioassay, as previously described (16).Samples were first incubated with [8-3 H] 3', 5'-cyclic adenosine monophosphate substrate.The labelled cAMP was then hydrolyzed to 5'-AMP by PDEs in the sample.This was then incubated with snake venom, which dephosphorylated the 5'-AMP to adenosine.Dowex ion exchange resin (1 : 1 : 1 ratio of Dowex : dH 2 O : EtOH) was then used to separate the negatively charged cAMP from the uncharged adenosine.Scintillation counting was used to measure the amount of [8-3 H] adenosine, which allowed the determination of the rate of cAMP hydrolysis.

Fibrin breakdown assay
Fibrin matrices (0.1 U/ml thrombin, 2.5 U factor XIII, 2 mg fibrinogen, 2 mg Na-citrate, 0.8 mg NaCl, 3 ml plasminogen per 1 ml DMEM) were added to cell culture plates and left to clot at room temperature.The matrices were soaked with DMEM supplemented with 10% NCS and 0.1% penicillin/streptomycin.Cells were seeded upon the matrices at high density and maintained in medium (DMEM supplemented with 10% NCS, 2 mmol/l L-glutamine and 0.1% penicillin/streptomycin) for 812 days with medium and replaced every 2 days.Where indicated, treatments were added in fresh medium.Phase contrast microscopy (Zeiss) was used to analyze cell invasion into the 3 D fibrin matrix.Fibrin degradation was quantified by weighing the remaining gel and determining the decrease in gel weight, which corresponds to fibrin degradation.

DNA manipulation
A QuikChange Õ Site-Directed Mutagenesis Kit (Stratagene Õ ) was used to carry out site-directed mutagenesis as per the manufacturer's instructions.The concentration was measured using a spectrophotometer (Nanodrop Õ ).Mini-and Maxi-prep Kits (Qiagen Õ ) were used to produce purified plasmid DNA, which was eluted in sterile water and stored at À20 C.

Preparation of whole cell lysates
Whole cell lysates were prepared by washing cells in sterile PBS before the addition of 3T3 lysis buffer (25 mM HEPES, pH 7.5, 2.5 mM EDTA, 50 mM NaCl, 50 mM NaF, 30 mM sodium pyrophosphate, 10% (w/v) glycerol and 1% Triton X-100, supplemented with Complete Ultra TM protease inhibitors (Roche Õ ; 1 tablet per 10 mls buffer).Cells were scraped and mixed at 4 C before clearing using centrifugation.Cell lysates were stored at À80 C. Protein concentration was determined using BSA as a standard with Bradford reagent (Bio-Rad Õ ) and measured using a spectrophotometer.

Immunoprecipitation and immunoblotting
Whole cell lysates were pre-cleared by incubation with G-Sepharose beads for at least 30 min.Equalized, pre-cleared lysates were incubated with the indicated antibody for a minimum of 3 h to allow the formation of immunocomplexes.The immunocomplexes were captured by incubation with fresh G-Sepharose beads, washed and eluted using Laemmli buffer and heat.
Immunoprecipitates and other protein samples were resolved using SDS-PAGE and transferred onto nitrocellulose membranes.Five percent milk powder dissolved in TBS-T was used to block non-specific binding before incubation with the indicated primary antibody diluted in 1% milk powder in TBS-T.The bound antibody was detected using appropriate HRP-conjugated secondary antibodies (Sigma-Aldrich Õ ), and visualized using enhanced chemiluminescence (Pierce Õ ).

PDE4A4 and p75NTR interact directly
Peptide array screening has been successfully utilized by us previously to pinpoint sites of proteinprotein interaction that tether PDE enzymes to a variety of different protein complexes (see e.g.(15,18,19)).
Here we used this technique to identify the amino acids that underpin the interaction between PDE4A4, the human orthologue of rodent PDE4A5 and p75NTR.Initial screening of 25-mer peptides that spanned the entire sequence of PDE4A identified positive interaction sites for p75NTR within UCR2 and the catalytic unit (Fig. 1).Subsequent alanine scanning peptide arrays in which each amino acid was sequentially mutated to an alanine residue highlighted E237, Y251, K406 and K407 as being essential for the interaction (Fig. 1a).Further validation of these amino acids was sought using site-specific mutations in PDE4A4 constructs.Mutants and wild-type PDE4A4 were transfected into HEK-293 cells to produce lysates for immunoprecipitation analysis.It was confirmed by PDE assay that these mutations did not affect the activity of PDE4A4 (Fig. 1e).The level of PDE4A4 present in p75 immunoprecipitates was significantly reduced when cells were transfected with Y251A and KK406/7AA mutants compared to wild type (WT) PDE4A4 (Fig. 1b and c).This data strongly suggests that residues Y251 within UCR2 and KK406/7 within the PDE4 catalytic subunit are essential for the interaction between PDE4A4 and p75NTR.Conversely, interaction sites for PDE4A5 on P75 were discovered within the juxta-membrane region of p75 using peptide array (Fig. 1d).Our data agrees with previous experiments that utilized truncated p75 forms, requiring sequences between 275 and 343 (13).
p75NTR decreases fibrin degradation dependent on PDE4 activity As one of the binding sites for p75NTR was contained within the UCR2 domain (Fig. 1) and previous work has shown that phosphorylation of PDE4A4, by MK2, alters the proteinprotein interactions of the PDE4 in that area (10), we were keen to discover whether this signalling axis could affect fibrin degradation by modulating p75NTRPDE4A4 association.To do this we established a fibrin degradation assay in NIH/3T3 cells stably expressing p75NTR (NIH/3T3-p75NTR cells).NIH/3T3 cells are known to have little endogenous p75NTR expression (20) and this was confirmed under the microscope in our study (Supplementary Fig. 1).NIH/3T3-p75NTR cells exhibited a three-fold reduction in fibrin degradation compared to WT NIH/ 3T3 cells, indicating that in the presence of p75NTR, the process of fibrinolysis is inhibited (Fig. 2a and b).To verify that PDE4 activity associated with p75NTR is involved mechanistically in the fibrinolysis process, we treated the NIH3t3-p75NTR cells with PDE4 selective inhibitor, rolipram.As expected, the three-fold reduction in fibrin degradation promoted by p75NTR was abolished following pharmacological attenuation of PDE4 activity (Fig. 2a and b).To determine whether phosphorylation of PDE4A4 by MK2 influences action of the PDE4A4 pool associated with p75NTR, we treated NIH3t3-p75NTR cells with agents (TNFa or anisomycin) that activate the p38 MAPK cascade and have previously been shown to activate the downstream kinase MK2 in such a way as to promote the phosphorylation of PDE4A4 at Ser147, the MK2 consensus phosphorylation site located within the UCR1 regulatory domain (7).
Immunopurification of p75NTR from cell lysates was undertaken and these probed for PDE4A4 in order to determine whether PDE4A phosphorylation had altered the fidelity of interaction.To our surprise, TNFa (tumour necrosis factor a) treatment, which increased PDE4A4 phosphorylation by MK2 dramatically increased association of stably over-expressed p75NTR with endogenous PDE4A (Fig. 2c).This increase in proteinprotein interaction could be ablated by inhibition of p38MAPK signalling using the p38MAPK inhibitor, SB203580 (Fig. 2c).In order to verify that the MK2 phosphorylation of PDE4A4 underpins the increase in p75NTRPDE4A4 association, we transiently transfected PDE4A5 wild type (Fig. 2d) and PDE4A5 S147A (phospho-null) (Fig. 2e) into NIH3t3-p75NTR cells, and repeated the experiment with another MK2 activator anisomycin.The increased amounts of p75NTRPDE4A complex seen with endogenous PDE4A (Fig. 2c) were also observed when PDE4A wild type was over-expressed (Fig. 2d) but not with the corresponding phospho-null mutant (Fig. 2e).Previous work had shown that MK2 phosphorylation could inhibit certain UCR2-dependent proteinprotein interactions with PDE4A (e.g.DISC1 and AIP) and not affect others (e.g.betaarrestin with PDE4D5) (9).Here we show that this post-translational modification can also enhance protein partner binding with p75NTR.
We next investigated whether increased PDE4/ p75NTR interaction had a functional role in fibrin degradation.Treatment of NIH/3T3-p75NTR with anisomycin caused a profound loss of fibrin degradation from 19 ± 2% in untreated cells to 1.4 ± 1.1% (n = 4 ± SD) in treated cells (Fig. 2a and b).The treatment had no effect on WT NIH/3T3 cells.This robust functional effect can be ascribed to the increased sequestration of PDE4A5 by p75NTR, which would further decrease the cAMP levels, ultimately causing a dramatic reduction in the fibrinolysis pathway.Incubation of NIH/3T3-p75NTR cells with SB203580 before anisomycin treatment resulted in a restoration of fibrinolysis activity comparable to untreated cells, which confirmed that the effects of anisomycin were due to the activation of the p38 MAP kinase pathway, rather than any nonspecific, off target activities of such treatment (Fig. 2a  and b).Finally, when NIH/3T3-p75NTR cells were incubated with rolipram prior to anisomycin treatment.Under such circumstances, the level of fibrin degradation (35 ± 3%) was markedly increased compared to treatment with anisomycin alone (1.4 ± 1.1%), but not quite to the level of rolipram alone (51 ± 9%) (n = 4 ± SD), suggesting that the two treatments had opposite effects (Fig. 2a and b).The effects of p38 MAP kinase pathway activation and the effects of PDE4 inhibition were competing with each other to yield a result midway between the two separate treatments.
Effects of PDE4A5/p75NTR interaction enhanced by p38 MAP kinase phosphorylation of PDE4A5 in primary cells Having discovered a novel role for the p38 MAP kinase pathway in the PDE4/p75NTR-mediated regulation of fibrinolysis in a stable cell line, we next wanted to verify this concept in primary cells (Fig. 3).Mouse embryonic fibroblast cells (MEFs) isolated from PDE4A knockout (PDE4A À/À ) mice showed more than a three-fold increase in fibrin degradation compared to MEFs from wild type (WT) mice (Fig 3a and b).Inhibition of all PDE4 isoforms with rolipram resulted in significantly increased fibrin degradation in WT MEFs (57 ± 3%) and this was comparable with levels in PDE4A À/À MEFs (60 ± 6%), whose fibrinolysis activity was not altered by rolipram treatment (Fig. 3a and b).Taken together, these data support the notion of a non-redundant role of PDE4A in the regulation of fibrinolysis.
The role of PDE4A5 phosphorylation by MK2 was also substantiated in MEF cells.Treatment with anisomycin caused fibrin degradation to be all but abolished in WT MEFs (3 ± 2%) while degradation in PDE4A À/À MEFs remained at basal levels (59 ± 5%) (Fig. 3a and b).When cells were treated with p38 MAPK inhibitor, SB203580, then the effects of the p38MAPK activator, anisomycin on WT cells was ablated, and the amount of fibrin degradation (20 ± 5%) was no different from that observed in untreated cells (Fig. 3a and b).Consistent with results on NIH/3T3 cells, the activation of the p38 MAPK pathway, and subsequent phosphorylation of PDE4A5, resulted in reduced fibrin degradation due to increased association of PDE4A5 with p75NTR.
To confirm that the cAMP dynamics in the vicinity of p75NTR are uniquely regulated by PDE4A, similar experiments were undertaken in PDE4B À/À MEFS (Fig. 3g and h).These cells behaved no differently to WT cells following treatment with rolipram and/or p38 MAPK activators and inhibitors.PDE4 inhibition resulted in the same increase in fibrin degradation in both WT and PDE4B À/À MEFs while anisomycin caused the opposite effect, decreasing fibrin degradation.Finally, the addition of anisomycin together with SB203580 ablated the stimulatory effect of anisomycin in both WT and PDE4B À/À knockout cells (Fig. 3g and h).
As PDE4A À/À MEFs provide the perfect background for a 'rescue' experiment we wanted to confirm the role that MK2 phosphorylation plays by transiently expressing PDE4A5 (Fig. 3i).To this end, wild type PDE4A5 and a MK2-phosphorylation null mutant (PDE4A5 S147A) were exogenously delivered to PDE4A À/À MEFS (Fig. 3i).Re-expression of WT PDE4A5 in these cells resulted in a complete rescue of the effects seen in WT MEFs, while the MEFs overexpressing PDE4A5 S147A reacted no differently than the PDE4A À/À cells (Fig. 3e and f).Reflecting the similar enzymatic activities of the PDE4A5 enzymes, the inhibition of PDE4 with rolipram caused the same increase in fibrin degradation in cells ectopically expressing either WT or S147A PDE4A5.However, when the p38 MAP kinase pathway was activated, cells expressing WT PDE4A5 had decreased fibrin degradation comparable to WT MEF cells.In striking contrast, the same treatment caused no significant effect on cells re-expressing PDE4A5 S147A.Such evidence robustly supports the idea that PDE4A5 phosphorylation by MK2 results in a decrease in fibrin degradation (Fig. 3e and f).Finally, to prove that increases in PDE4 activity per se, rather than any other scaffolding functions of PDE4, are responsible for changes in fibrinolysis, experiments were undertaken on PDE4A À/À MEFs -re-expressing either WT PDE4A5 or a dominant negative (D/N), catalytically inactive PDE4A5 mutant (Fig. 3c and d).The D/N PDE4A5 clearly displaced the endogenous active PDE4A5 component from p75NTR as ectopic expression of the catalytically inactive, dominant negative, PDE4A5 mutant markedly increased the fibrin degradation to levels comparable to that observed by the PDE4A À/À MEFs (65 ± 5%) (Fig. 3c and d).Additionally, rolipram and anisomycin had no effects on the PDE4A À/À MEFs these cells, which still displayed high levels of fibrin degradation.

Discussion
Cyclic-AMP signalling via p75NTR can inhibit scar formation and suppress extracellular matrix remodelling and these positive outcomes are opposed by recruitment of PDE4A to the receptor (13,14).We report that that this process is also fine-tuned by phosphorylation of PDE4A5 by MAPKAPK2, the downstream kinase of the p38 MAPK inflammatory signalling cascade.
Much literature exists on the roles that p38MAPK has in the orchestration of signalling events involved in inflammatory responses in a range of cell types.Of particular importance is the ability of the kinase to coordinate downstream signalling in macrophages challenged with LPS, monocytes exposed to IL-18 and neutrophils activated by phorbol 12-myristate 13-acetate (PMA) or TNFa (21,22).Data from p38 null mice also suggest that p38 activity is important for the production of TNFa, Interleukin (IL)-1b and IL-10 via Toll-like receptor 4 signalling (23).Indeed, p38MAPK has been linked to the production of a variety of pro-inflammatory mediators such as IL-12, IL-1b, TNF-a, prostaglandin E2 (PGE2) and IL-12 as well as cyclooxygenase 2 (COX-2), IL-8, IL-6, IL-3, IL-2 and IL-1 (reviewed in (24)).Correspondingly, p38MAPK inhibitors are now being assessed as antiinflammatory agents (25).In the context of the lung, p38MAPK inhibitors have been shown to protect against LPS (lipopolysaccharide)induced acute lung inflammation by retarding LPS triggered production of TNFa (26).Interestingly, increase levels of activated p38 have been observed in the lungs of chronic obstructive pulmonary disease (COPD) patients (27,28), and p38MAPK inhibitors have been thought of as a therapeutic target for the disease for over a decade (29).Recent double-blind, randomized, placebo-controlled trials of p38MAPK inhibitors in COPD patients have shown promising results.The first trial looked at whole-blood bio-markers and found that p38MAPK inhibition significantly reduced the phosphorylation of HSP-27 (a p38 substrate) and LPS-induced TNFa production (30).The second trial assessed changes in lung function and reported an increase in forced expiratory volume in COPD patients assigned to p38MAPK inhibition therapy (31) and the third observed that P38 inhibition could reduce plasma fibrinogen levels of patients (32).Fibrinogen has been characterized as a negative biomarker for COPD i.e. associated with poor clinical outcomes (33).
Indeed p38MAPK signalling (via activation of ASK1 (apoptosis signal-regulating kinase 1)) has also been implicated in the deposition of fibrin following the up-regulation of tissue factor in acute lung injury caused by sepsis (34).Additionally, modulation of the p38MAPK pathway by intracellular glutathione inhibits plasminogen activator inhibitor-1 (PAI-1) expression in a cellular model of pulmonary fibrosis (35).Taken as a whole, the above information robustly indicates that p38MAPK activity exacerbates inflammatory signalling associated with COPD and has the ability to regulate extracellular matrix remodelling in the lungs.The data we present here describes a new mechanism that may underpin both of these functions.The mechanism relies on the activation of MAPKAPK2 by p38MAPK and the subsequent phosphorylation of PDE4 by MAPKAPK2 (Fig. 4).
This action enhances the PDE's interaction with p75NTR causing a paucity of cAMP around the receptor and a concomitant diminution of fibrin breakdown (Fig. 4).This concept is supported by data suggesting that both p38MAPK inhibitors and PDE4 inhibitors are capable of attenuating TNFa release in COPD in alveolar macrophages isolated from COPD lung transplant tissue (36).
The specificity of the interaction between only PDE4A (but not PDE4B, C or D) enzymes and p75NTR identifies this proteinprotein interaction as a new potential target for therapeutic intervention, however, the use of pharmacological approaches to inhibit the active site of PDE4s in this setting has lacked success due to side-effects caused by lack of specificity relating to the highly homologous catalytic domain i.e. almost identical in all PDE4 isoforms.Efforts to counteract this problem have resulted in the development of PDE4B (37) and PDE4D (38,39) selective inhibitors, which have been directed toward disease states such as Alzheimer's disease.In this context, both inhibitor types can play a role with PDE4B inhibition attenuating neuro-inflammation by reducing cytokine production in the brain following beta-amyloid insult, whereas PDE4D compounds act to augment phospho-CREB enhancing cognition (40).To date, however, efforts to develop a pharmacological approach to specifically attenuate PDE4A activity have not been fruitful (41).
As the functions of individual PDE4 isoforms are dictated by their discrete cellular location (7), displacing localized PDE4 'pools' from their site of action is a therapeutic strategy that has potential in this context (42).Detailed information on the docking sites for both protein partners is required for this approach and here we provide evidence of novel binding sites on PDE4A for p75NTR in the UCR2 and catalytic domains.The latter domain is of interest as recently UCR2 has been recognized as a binding partner for integrin a5 and this interaction serves to target PDE4s to focal adhesions to promote pro-inflammatory actions (43).Interestingly, integrin a5 association suppresses the inhibitory phosphorylation of the PDE4 within the catalytic site by ERK MAP kinase, ensuring that highly active PDE is available in focal adhesions to regulate inflammatory signalling.Other PDE4 binding partners that use UCR2 as a docking domain include AIP (aryl hydrocarbon receptor interacting protein) (44) and DISC1 (45).Both of these are released from PDE4A5 following phosphorylation of the PDE on serine 147 by MK2 (9), whereas the present study provides proof that the interaction of PDE4A5 and P75 is enhanced following this posttranslational modification.Such evidence suggests that the modification of serine 147 on PDE4A5 acts as a relocation 'switch' to tailor the binding partner profile of the enzyme in order to shape cAMP gradients in discrete cellular locations (Fig. 4).In the lung, enhancement of localized PDE4A5 activity in the vicinity of P75 caused by MK2 phosphorylation would lead to deleterious promotion of scar formation and this could be of particular concern to COPD patients where the human form of the enzyme (PDE4A4) is selectively upregulated (46).Hence local inhibition of p38MAPK or downstream kinase MK2 may represent an alternative strategy to the disruption of the PDE4AP75 complex proposed above.

Fig. 1
Fig. 1 PDE4A4 mutants reduced interaction with p75 without altering enzymatic activity.(a) A schematic representation of PDE4A4.Full length PDE4A4, in the form of immobilized peptide 'spots' of overlapping 25-mer peptides, sequentially shifted five positions, was overlaid with purified His-p75-ICD.Detection of p75 by immunoblotting identified positively interacting peptides (dark spots) that were taken forward for alanine scanning.The rows of spots show peptide arrays based on the 25-mer interacting peptides, Glu 223 Leu 242 , Tyr 250 Ser 242 and Leu 404 Iso 450 in which the indicated peptide was mutated to alanine.These arrays were probed with His-p75-ICD and immunoblotting allowed detection of interaction.Peptide spots circled indicate peptides in which alanine substitution eliminated interaction consistently, identifying amino acids crucial for the interaction (n = 4).(b) Wild-type or mutated GFP-PDE4A4 (E237A, Y251A, K406A, K407A, KK406/7AA) were transfected into HEK-293 cells and lysates prepared.The lysates were subjected to PDE-activity assays in the indicated concentrations.The mutations did not significantly affect enzymatic activity (e).(d) Full length p75, in the form of immobilized peptide 'spots' of overlapping 25-mer peptides, sequentially shifted five positions, was overlaid with purified GST-PDE4A5.Detection of GST by immunoblotting identified positively interacting peptides (dark spots).(b and c) Lysates from NIH-3T3 fibroblasts stably overexpressing p75NTR and transfected with GFP-PDE4A4 wt or the indicated mutants were subjected to immunoprecipitation and immunoblotted for GFP and p75.The Y251A mutation and the KK406/ 7AA mutations significantly decreased the amount of PDE4A4 bound p75 (*P50.05)when compared with control.Data shown are typical of experiments performed at least 3 times.

Fig. 3
Fig. 3 Effects of PDE4A5/p75NTR interaction enhanced by p38 MAP kinase phosphorylation of PDE4A5.(a, b) WT MEF cells and PDE4A À/À cells were seeded on 3D fibrin gels, subjected to treatment and subsequently analyzed for fibrin degradation.10Â magnification of cells treated with rolipram (10 mM) or anisomycin (10 mM) with or without pre-treatment with SB203580 (25 mM).A quantitative analysis of three independent experiments (**P50.01)when compared with PDE4A À/À mefs).(c, d) Experiments were repeated as in (a) this time including PDE4A À/ À cells with PDE4A5 dominant negative (DN) reintroduced via transient transfection.A quantitative analysis of three independent experiments (*P50.05)when compared with PDE4A5-D/N.(e, f) Experiments were repeated as in (a) using PDE4A À/À cells with PDE4A5 or PDE4A5 S147A (phsopho-null mutant) reintroduced via transient transfection.A quantitative analysis of three independent experiments (**P50.01)when compared with control (PDE4A5 transfected).(g, h) A quantitative analysis of three independent experiments using WT MEF cells and PDE4B À/À MEF cells subjected to treatments as in (a).Data are means ± SD. (i) Confirmation of reintroduction of PDE4A5 forms via western blotting.