AKAP2-anchored protein phosphatase 1 controls prostatic neuroendocrine carcinoma cell migration and invasion

Prostate cancer (PC) is the second leading cause of cancer-related death in men. The growth of primary prostate cancer cells relies on circulating androgens and thus the standard therapy for the treatment of localized and advanced PC is the androgen deprivation therapy. Prostatic neuroendocrine carcinoma (PNEC) is an aggressive and highly metastatic subtype of prostate cancer, which displays poor prognosis and high lethality. Most of PNECs develop from prostate adenocarcinoma in response to androgen deprivation therapy, however the mechanisms involved in this transition and in the elevated biological aggressiveness of PNECs are poorly defined. Our current findings indicate that AKAP2 expression is dramatically upregulated in PNECs as compared to non-cancerous prostate tissues. Using a PNEC cell model, we could show that AKAP2 is localized both intracellularly and at the cell periphery where it colocalizes with F-actin. AKAP2 and F-actin interact directly through a newly identified actin-binding domain located on AKAP2. RNAi-mediated silencing of AKAP2 promotes the phosphorylation and deactivation of cofilin, a protein involved in actin turnover. This effect correlates with a significant reduction in cell migration and invasion. Co-immunoprecipitation experiments and proximity ligation assays revealed that AKAP2 forms a complex with the catalytic subunit of protein phosphatase 1 (PP1) in PNECs. Importantly, AKAP2-mediated anchoring of PP1 to the actin cytoskeleton regulates cofilin dephosphorylation and activation, which, in turn, enhances F-actin dynamics and favors migration and invasion. In conclusion, this study identified AKAP2 as an anchoring protein overexpressed in PNECs that controls cancer cell invasive properties by regulating cofilin phosphorylation.


Introduction
Prostate cancer (PC) is the second most frequent diagnosed cancer and the 5th most common cause of cancer death in men worldwide [1].Primary PC mainly relies on circulating androgens to sustain its growth and development.In this context, androgen deprivation therapy, which is the primary treatment for patients suffering from advanced PC, can provide satisfactory clinical control over the disease for some years [2].Eventually however, PC relapses in the form of a castration resistant prostate cancer (CRPC), which displays invasive and metastatic properties even in the presence of low concentrations of androgens [2].
Prostatic neuroendocrine carcinoma (PNEC) is an aggressive variant of PC, which is characterized by a poorly differentiated neuroendocrine morphology, the absence of androgen receptor expression and the expression of general neuroendocrine markers such as synaptophysin, chromogranin A, and INSM1 [3].PNEC may arise de novo or in patients previously treated with androgen deprivation therapies for prostate adenocarcinoma.Among newly diagnosed PCs, PNECs are relatively rare and represent about 2 % of total cases.This rate can increase up to 17 % in relapsing PCs [4].The specific molecular determinants driving PNEC cell growth and metastasis are largely unknown with consequent absence of therapies efficiently targeting these processes.
Cancer metastasis is a multi-step process that involves cancer cell detachment from the primary tumor [5], acquisition of a mesenchymal phenotype, migration and invasion into the surrounding tissue and blood vessels, and colonization of distal metastatic sites [5,6].Acquisition of a migratory phenotype is central to the metastatic process and requires a profound reorganization of the actin cytoskeleton of cancer cells [7].Several studies performed on cultured PC cells have shown that reorganization of actin filaments at the leading edge of migrating cells drives the formation of actin-rich protrusions named lamellipodia and filopodia [8,9].Lamellipodia generate the driving force for cell movement, whereas filopodia extend from the cell membrane to sense the environment and promote directional movement [10].Importantly, the kinetics at which these filamentous (F) actin structures polymerize and depolymerize directly influences migration velocity of cancer cells [11,12].
The actin-depolymerizing factor (ADF)/cofilin family members act as central regulators of F-actin dynamics by promoting rapid actin filament severing and disassembly [13].In particular, cofilin-mediated Factin severing generates free-barbed ends that are necessary to the addition of new actin monomers and for actin nucleation by the Arp2/3 complex [13,14].Moreover, cofilin also contribute to the polymerization of new actin filaments by increasing the pool of actin monomers [15].Cofilin activity is directly regulated by the phosphorylation status of Ser3 near its N-terminus.Phosphorylation of Ser3 by LIM kinases 1 and 2 induces cofilin deactivation whereas dephosphorylation by the Slingshot phosphatases or protein phosphatases 1 and 2A (PP1 and PP2A) promotes cofilin activation [16,17].Interestingly, it has been shown that active cofilin accelerates actin dynamics and increases migratory and metastatic properties of prostate cancer cells [18].While these findings highlight a direct link between the cofilin phosphorylation status and PC cell migration, it is currently unknown how cofilin phosphatases and kinases are targeted to the actin cytoskeleton to locally modulate cofilin phosphorylation and activity.
It has become increasingly clear that the subcellular compartmentalization of signaling enzymes controlling the oncogenic properties of cancer cells is regulated by scaffolding and anchoring proteins [6].In particular, A-kinase anchoring proteins (AKAPs) are scaffold proteins that confer spatiotemporal regulation to signaling events in cells [6,19,20].AKAPs can assemble a variety of signaling enzymes including the cyclic adenosine monophosphate (cAMP)-dependent protein kinase (PKA), other kinases, phosphatases, phosphodiesterases, adenylyl cyclases, and GTPases, and they can cluster them in proximity of their downstream effector substrates at precise cellular nanocompartments [21][22][23].
Our current study indicates that AKAP2 is strongly expressed in PNECs as compared to non-cancerous prostate tissues or prostate adenocarcinomas.Using a PNEC cell model and human PNEC tissues, we have demonstrated that AKAP2 forms a signaling complex F-actin and the protein phosphatase 1 (PP1) involved in cofilin dephosphorylation.Accordingly, RNAi-mediated silencing of AKAP2 results in an increase in the phosphorylation status of cofilin and results in a significant decrease in cancer cell migration and invasion.These findings suggest that AKAP2 can control PNEC cell invasive properties by assembling a transduction complex regulating cofilin phosphorylation.

Preparation and immunohistochemical staining of human PNECs
Twelve cases of PNECs, diagnosed on biopsy or trans-urethral specimens at the University Hospital of Varese (ASST Sette Laghi), were included in the study Two cases of non-neoplastic prostatic tissue were included as controls as well.Clinico-pathological information was obtained from patients' clinical records.
The original histopathological slides used for diagnosis were reviewed by two pathologists with expertise in uropathology and endocrine pathology (SLR and SU).Additional slides were cut from significant paraffined inclusions for immunohistochemical and molecular analyses.

RNA isolation and quantitative real-time PCR
To determine AKAP2 expression in human PNECs, total RNA was extracted from three paraffinized histological sections after manual microdissection by using Maxwell® RNA FFPE Kit and Maxwell 16 system (Promega, Madison, USA) according to manufacturer's protocol.RNA was quantified using Qubit™ RNA XR Assay Kit (Invitrogen -Thermo Fisher Scientific, Whaltam, USA).Reverse transcription was carried out on 1000 ng of total RNA using High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Whaltam, USA) using random primers.
q-PCR reactions were prepared using the reverse-transcribed RNA, AKAP2 TaqMan Gene Expression Assays (Thermofisher scientific, Hs02338795_m1) and TaqMan universal PCR Mastermix (Thermofisher scientific, 4324018).The housekeeping gene Beta-2 microglobulin was used as internal control.q-PCR analysis was performed by using a Quantstudio 6 real-time PCR system (Applied Biosystems).
To determine the expression of AKAP2, synaptophysin and chromogranin A mRNAs expression in LnCAP, PC3 and DU145 PC cells, RNA was extracted from approximately 1 million cells using the RNeasy Mini Kit (QIAGEN, 74106).Single-strand cDNA was synthesized from 500 ng of total RNA by using PrimeScript™ RT Master Mix (Takara, RR036A).q-PCR reactions were prepared using 100 ng of the reverse-transcribed RNA, selected TaqMan Gene Expression Assays (Thermofisher scientific) and TaqMan universal PCR Mastermix (Thermofisher scientific, 4324018).The housekeeping gene GAPDH was used as internal control.q-PCR analysis was performed by using a Quantstudio 6 real-time PCR system (Applied Biosystems).

Plasmids and constructs
AKAP2-003 and AKAP2-003Δ3 splice forms were PCR-amplified from retro-transcribed DU145 cells RNA using the following primers: 5′ TTATAAGCGGCCGCAATGGCGTGGCCCCAGCCCGGG 3′ and 5′ GCGCGCGTCGACTTATTCGTTGTCTTCTTCCTCCTGGTT 3′ and then cloned into the NotI and SalI sites in a pEGFP-C1 vector.To generate an AKAP2-003 mutant resistant to siRNA1-mediated silencing, we introduced a silent point mutation in the AKAP2 sequence recognized by the siRNA by site-directed mutagenesis using the following primer: 5′ ACA CGAGTTAACCGCAGGAAG 3′.This AKAP2 silencing resistant construct underwent a second round of site-directed mutagenesis to generate Factin and PP1 binding deficient mutants (AKAP2-ΔAct and AKAP2-ΔPP1, respectively).For the generation of the AKAP2-ΔAct mutant the following primers were used: 5′ GAT ACT GGG CTA TCC CGA GAA GCG CGA TCA CAT GGT CCT 3′, 5′ TTG CCT GGA ATT CTC CAT CAG CTG AGC TTG TTT AGC AGC AGC AGA AGC ATC TAT 3′.For the generation of AKAP2-ΔPP1 we used the following primers: 5′ GAT ACT GGG CTA TCC AAC GCG CAA GCA TGA TTG ACA AAG 3′, 5′ CCT GCC AAG CTT GCG TCC GCG AAA CAG CTG TTT CGC GGA CGA AAG CTT GGC AGA 3′.The human cofilin-pcDNA3.1 and human cofilin S3A-pcDNA3.1 constructs were purchased from Addgene (n • 50853 and n • 50854).The plasmid encoding the GST fusion of the PP1 binding motif of AKAP2 (residues 716 to 727) was purchased from VectorBuilder Inc.A mutant construct was generated by replacing residues 719, 722 and 724 with alanines.

Cell invasion assays
Invasion assays have been performed using a modified Boyden Chamber technique measuring the ability of DU145 cells to migrate across an 8 μm pore-size polycarbonate membrane (Corning, 3422) coated with Matrigel basement membrane matrix.DU145 cells were starved in RPMI1640 (Gibco™, 61870-010) containing 0.5 % FBS for 24 h.After trypsinization, 2.5 × 10 4 quiescent DU145 cells were loaded into the upper chamber.To create a serum gradient across the Matrigel layer, the lower chamber was filled with RPMI1640 supplemented with 10 % FBS.After a 24 h incubation the membrane was fixed in 70 % methanol at − 20 • C for 20 min.DU145 that did not migrate were removed and cells migrated on the underside of the membrane were stained using 2 % crystal violet and counted using a light microscope.

Proliferation assays
1.5 × 10 5 DU145 cells seeded on 25 mm coverslips and cultured in RPMI1640 supplemented with 5 % non-essential amino acids and 10 % FBS, were transfected with 30 pmol of control or AKAP2 siRNAs.24 h after transfection, cells were incubated for an additional 24 h in medium with 10 % FBS followed by 1 h incubation with 10 μM BrdU.The concentration of DMSO in the medium was 0.1 % for all experimental conditions.Cells were then washed twice with PBS, fixed for 10 min in PBS/2 % paraformaldehyde, and permeabilized for 20 min with 0.25 % (w/v) Triton X-100 in PBS.Coverslips were incubated for 10 min with ice-cold 1 N HCl, for 10 min with 2 N HCl at room temperature, and for an additional 20 min with 2 N HCl at 37 • C. HCl was then neutralized by two 10 min washes with 0.1 M boric acid (pH 7.5).Coverslips were blocked for 1 h in buffer A (10 mM Tris-HCl [pH 7.5], 155 mM NaCl, mM EGTA, 2 mM MgCl 2 ) supplemented with 1 % BSA.The incorporation of BrdU was assessed by incubating cells for 2 h with a 1:200 dilution of the rat anti-BrdU antibody (Abcam, ab6326) followed by incubation for 1 h with a 1:250 dilution of Alexa Fluor 488-conjugated anti-rat secondary antibodies (Thermofisher, A-21208) and 10 min incubation with DAPI (1 μg/ml).Immunofluorescent staining was visualized using an Advanced Microscopy Group EVOS fluorescence microscope.

Time-lapse microscopy assay
3 × 10 4 DU145 cells were seeded in 12 well plates (Corning) in the presence of RPMI1640 containing 10 % FBS.Cells were then transfected either with control or AKAP2 specific siRNAs.Random migration was then visualized 48 h after transfection using a Spinning-disk microscope (Nikon, Eclipse Ti2) over a period of 8 h.Images were taken every min.Cell trajectories and velocities (μm/min) were calculated using the Gradientech Tracking Tool ™ (Gradientech).

Wound healing assay
3 × 10 5 DU145 cells were seeded in 6 well plates in the presence of RPMI1640 containing 10 % FBS and transfected either with control or AKAP2 specific siRNAs.48 h after transfection, confluent monolayers were wounded using a 200 μl tip, washed once with PBS, and cultured for another 36 h in RPMI1640 medium without phenol red.Recolonization of the wounded area was imaged every 6-12 h with an Advanced Microscopy Group EVOS fluorescence microscope.The wounded area was measured using the ImageJ software.

Proximity ligation assay
DU145 cells seeded on glass coverslips were fixed using 4 % paraformaldehyde for 20 min at 37 • C. Cells were then permeabilized 10 min with 0.1 % Triton-X-100 at 4 • C and coverslips blocked for 1 in PBS supplemented with 1 % bovine serum albumin.Proximity ligation was performed using the Duolink® Proximity Ligation Assay kit (Merck, DUO92008) according to manufacturer's protocol.Primary antibodies against β-actin (Sigma Aldrich, A3853), cofilin (Thermofisher, GT567) and PP1 (Invitrogen, PA5-102349) were used at 1:500 dilution.Cytoskeleton was stained using Actin Green Alexa 488 Ready Probes (Ther-moFisher, R37110) and coverslips mounted on microscope slides using DAPI-containing mounting medium.Proximity ligation was also performed on human NEPC sections according to the protocol described by Gomes et al. [25].In this case, primary antibodies were used at 1:100 dilution.

Peptide arrays
Peptide arrays were synthetized on cellulose membrane using Mul-tiPep spotter at the Peptide array Facility of the University of Oslo.The membranes were then activated in methanol for 30 s.The arrays were blocked 1 h at room temperature in PBS containing 5 % non-fat dried milk followed by incubation with or without 1 μM filamentous actin (Cytoskeleton, APHL99-A).After 3 washes in TBS-Tween, membranes were incubated with horseradish peroxidase-conjugated anti-actin antibodies (Cell Signaling Technology, 12620).

Expression and purification of glutathione-S-transferase (GST) fusion proteins
GST fusions of the WT and mutated PP1 binding motif of AKAP2 were expressed using the bacterial expression vector pGEX4T1 in the Bl21DE3 strain of Escherichia coli and purified.To induce the expression of fusion proteins, exponentially growing bacterial cultures were incubated 16 h at 16 • C with 1 mM isopropyl 1-thio-β-d-galactopyranoside (IPTG), and subsequently subjected to centrifugation at 4,000 ×g at 4 • C. Pelleted bacteria were lysed in buffer D (20 mM Tris pH 7.4, 150 mM NaCl, 1 % Triton X-100), and protease inhibitors (Thermo Fisher Scientific, catalog n.A32963), sonicated, and centrifuged at 10,000 ×g for 30 min at 4 • C.After incubating the supernatants with glutathione Sepharose beads (Cytiva, catalog n • 17075601) for 2 h at 4 • C, the resin was washed three times with 10 volumes of buffer C (20 mM Tris pH 7.4, 300 mM NaCl, 1 % Triton X-100, and protease inhibitors).The protein content of the beads was assessed by Coomassie Blue staining of SDSpolyacrylamide gels.Beads were stored at − 20 • C.

GST pulldown experiments
5 μg of GST fusions of the WT and mutated PP1 binding motif of AKAP2 were incubated with 100 nM of purified PP1CB (MyBioSource, MBS965220) in 250 μl of buffer B for 4 h at 4 • C. The beads were then washed twice for 10 min with buffer C and twice with buffer B. Proteins were eluted in SDS-PAGE sample buffer, boiled at 95 • C for 5 min and analyzed by Western blotting.

Statistical analysis
Values are presented as mean ± standard deviation (SD).Statistical differences between two groups were assessed using unpaired 2-tailed Student's t-test. 1 or 2-way ANOVA followed by Tuckey tests was used for multiple groups analysis.Values of p < 0.05 were considered as statistically significant.

AKAP2 is overexpressed in prostatic neuroendocrine carcinoma
The AKAP2 gene has been reported to be amplified in up to 27 % of PNECs (www.cbioportal.org;Fig. 1A) [26][27][28] raising the hypothesis that AKAP2 could play a role in coordinating pro-oncogenic responses in this rare but highly metastatic cancer.To address this question, we initially analyzed AKAP2 expression in biopsies obtained from 12 PNEC patients, of whom 8 were affected by de novo PNEC and 4 by androgen deprivation therapy-induced PNEC.The mean age of patients was 69.7 years.The histopathological review confirmed the diagnosis of PNEC in all cases, based on the poorly differentiated neuroendocrine morphology and the expression of general neuroendocrine markers (synaptophysin, chromogranin, and INSM1).All cases showed a small cell morphology, being composed of small to medium-sized (2-4 times the size of a small lymphocyte) round to oval cells with scant cytoplasm and hyperchromatic nuclei without nucleoli.Mitotic figures were extremely frequent (>20 mitotes/2mm 2 ), as well as apoptotic bodies.Ki67 proliferation index was always higher than 60 %.In three cases a residual component of acinar cell carcinoma was also observed.
As shown in Fig. 1B, eleven out of twelve PNEC biopsies displayed increased AKAP2 mRNA levels as compared to non-cancerous prostate tissues.Upregulation was variable and ranged from 2.5 to 520-fold (Fig. 1B).Immunohistochemical analysis of control non-cancerous prostate tissues revealed AKAP2 expression in basal cells of prostatic acini and in vascular pericytes and weak AKAP2 levels in stromal and acinar cells (Fig. 1C, left panels).AKAP2 stainings of biopsies from patients with small-cell PNECs confirmed a significant increase in AKAP2 expression in cancer cells (Fig. 1D, upper panel) whereas no AKAP2 overexpression was detected in prostate adenocarcinomas (Fig. 1D, lower panel).These results suggest that AKAP2 upregulation mainly occurs in PNECs.
To identify human PC cell lines that could represent an in vitro model for PNECs, we took advantage of the gene expression atlas provided by NIH Genomic Data Commons (https://gdc.cancer.gov/).This allowed us to select DU145 and PC3 cells as potential candidates based on the observation that they express typical neuroendocrine markers such as synaptophysin and chromogranin A but not androgen receptors.The androgen-dependent PC cells LNCaP, derived from prostate adenocarcinoma cells, were selected as a non-neuroendocrine prostate cancer control cell line.qPCR analysis of synaptophysin and chromogranin A mRNA expression confirmed that DU145 and PC3 cells display 3-to-5fold increased expression of neuroendocrine markers compared to LNCaP cells (Fig. 1E).In line with our observation that NEPCs but not prostate adenocarcinomas overexpress AKAP2, analysis of AKAP mRNA and protein expression in the three PC cell lines revealed that AKAP2 is selectively expressed in DU145 and PC3 cells but not in LNCaP cells (Fig. 1F-H).As it can be visualized in Fig. 1G, DU145 and PC3 cells express two main AKAP2 forms.The two forms were PCR-amplified from total DU145 cell RNA and identified as the AKAP2 variant 003 and a variant of AKAP2 003 lacking exon 3 (Fig. S1).Flag-tagged fusion of these AKAP2 proteins displays apparent molecular weights comparable to those of the endogenous AKAP2 forms (Fig. S1).

AKAP2 mediates DU145 cell migration and invasion
Based on the above results, we initially investigated the possibility that the increased expression of AKAP2 in PNEC cells might confer them increased migratory and invasive potential.To address this question, DU-145 cells were used as a model system for PNECs since they display the highest expression of AKAP2 and neuroendocrine makers (Fig. 1E-H).We initially assessed the impact of AKAP2 silencing on the migratory capacity of DU145 cells using wound-healing assays and single-cell tracking.Results show that knockdown of AKAP2 in DU145 cells using two independent siRNAs (Fig. 2A-B) reduced wound-healing by nearly 50 % (Fig. 2C-D).In line with these findings, migration velocity calculated from single-cell trajectories was also reduced by approximately 40 % (Fig. 2E-H).The involvement of AKAP2 in migration let us to investigate whether a reduction of AKAP2 expression could affect the invasive properties of DU145 cells, assessed by their ability to migrate through a Matrigel layer.As shown in Fig. 2I-L, AKAP2 silencing inhibited DU145 invasion capacity by 50 to 62 %, suggesting that AKAP2 is a key determinant of DU145 cell migration and invasion.Finally, we could show that AKAP2 knockdown does not significantly impact DU145 cell proliferation as determined by monitoring bromodeoxyuridine (BrdU) incorporation (Fig. S2).

AKAP2 directly interacts with F-actin through a newly identified binding motif
Cell migration is controlled by the reorganization and turnover of the actin cytoskeleton, which provides the driving force for directional movement.Based on our current results linking AKAP2 to DU145 cells migration (Fig. 2), we investigated whether AKAP2 might associate with the actin cytoskeleton.In line with this hypothesis, immunolabeling experiments revealed that AKAP2 is localized both intracellularly and at the periphery of DU145 cells where it exhibits a significant degree of colocalization with F-actin (Fig. 3A).Control experiments indicate that AKAP2 immunostaining is highly specific since it is not detected following RNAi-mediated knockdown of AKAP2 (Fig. 3A).To determine whether AKAP2 and F-actin interact directly and to identify the potential actin binding motif on AKAP2 we used purified F-actin as probe to screen a peptide array of overlapping 20-residue peptides (offset every 3 amino acids) spanning the entire AKAP2 sequence (Fig. 3B).Binding of F-actin was detected using HRP-coupled actin specific antibodies.A major site of interaction was detected between residues 319 and 341 of AKAP2 (Fig. 3B, lower array).In order to delineate specific F-actin binding determinants within this sequence, we generated an array of peptide derivatives where each residue in the binding domain (indicated above the array by single-letter code) was replaced by all 20 amino acids (indicated on the left side of the array) (Fig. 3C).This analysis revealed that the most critical binding determinants within the actin binding site of AKAP2 are represented by arginines at positions 9 and 19 and phenylalanines at positions 5 and 12 (Fig. 3C).In this respect, positions 9 and 10 cannot tolerate residues other than arginines, whereas positions 5 and 12 can accommodate any aromatic amino acid.To determine whether these residues represent the main interacting surface for F-actin on AKAP2, we generated an AKAP2 mutant in which the key arginines and phenylalanines involved in actin binding were substituted by alanines (AKAP2-ΔAct).As shown in Fig. 3D, the ability of immunoprecipitated GFP-tagged AKAP2-ΔAct to bind F-actin in an overlay assay was completely abolished.Altogether, these results suggest that AKAP2 directly interact with F-actin through a newly identified binding motif that requires arginines and phenylalanines.

AKAP2 regulates cofilin phosphorylation
Our current results indicate that AKAP2 regulates DU145 cell motility and that it directly binds to F-actin (Figs. 2 and 3).Knowing that remodeling of the actin cytoskeleton is main driver of cell migration we investigated the possibility whether AKAP2 could influence the function of known regulators of actin dynamics.To this end, we focused our interest on cofilin since it directly favors actin severing and turnover.
We initially determined whether AKAP2 can affect cofilin phosphorylation state, based on the evidence that the actin-severing activity of cofilin is directly controlled by cofilin phosphorylation at serine 3. Results indicate that AKAP2 silencing in DU145 cells results in a significant 2.3-fold increase in phospho-cofilin levels as assessed by Western blotting using antibodies recognizing phosphoserine 3 (Fig. 4A  and B).Conversely, individual overexpression the GFP-tagged AKAP2 003 and AKAP2 003 Δex3 splice variants resulted in a 55 to 83 % decrease in cofilin phosphorylation (Fig. 4C and D).In line with these findings, PLA revealed that AKAP2 and cofilin can form complexes in DU145 cells (Fig. 4E, left panel).Collectively, our findings suggest that AKAP2 interacts in situ with cofilin and favors its dephosphorylation.

AKAP2 anchors PP1 in PNECs
Based on the above results, a possible hypothesis is that AKAP2medited cofilin dephosphorylation could result from the capacity of this anchoring protein to recruit a protein phosphatase to the actin cytoskeleton in proximity of cofilin.A plausible candidate could be represented by the Slingshot protein phosphatase 1 (SSH1), which has been described as the main regulator of cofilin dephosphorylation [16].However, co-immunoprecipitation experiments failed to detect an interaction between AKAP2 and SSH1 (Fig. S3A).Similarly, AKAP2 silencing in DU145 cells did not affect the extent of SSH1 phosphorylation on Ser 937 and 978, which is a prerequisite for subsequent SSH1 activation (Fig. S3B and C).This suggests that AKAP2 could regulate cofilin phosphorylation through an SSH1-independent pathway.Accumulating evidence suggests that, in addition to SSH1, protein phosphatase 1 (PP1) can facilitate cofilin dephosphorylation in cancer cells [17,29].Interestingly, an interaction between AKAP2 and PP1 was identified in DU145 cells when endogenous PP1 was detected in AKAP2 immunoprecipitates (Fig. 5A, upper panel).The ability of AKAP2 to form complexes with PP1 was further confirmed using PLA both in DU145 cells and in sections of primary NEPCs isolated from patients (Fig. 5B and C).As it can be visualized in Fig. 5B and C, complexes could be detected only in samples incubated with both anti-PP1 and anti-AKAP2 antibodies.Interactions between AKAP2 and PP1 were not detected in section from non-cancerous prostates, suggesting that AKAP2-PP1 complexes are mainly formed in NEPC cells (Fig. 5C, lower panels).Interestingly, co-immunoprecipitation experiments performed from PC3 cell lysates did not detect an interaction between AKAP2 and PP1, suggesting that the AKAP2/PP1 complex might not be formed in this PC cell type (Fig. S4).
Analysis of the AKAP2 primary sequence revealed the presence of a potential PP1 interaction site located between residues 709 and 714 and corresponding to the following binding sequence: F-x-x-R/K-x-R/K, where x represents any amino acids (Fig. 5D).To determine whether this site represents the main PP1 binding sequence on AKAP2, we generated an AKAP2 mutant in which the phenylalanine and the two arginines in the putative interacting motif were substituted by alanines (AKAP2 ΔPP1).As shown in Fig. 5E, GFP-AKAP2 but not GFP-AKAP2 ΔPP1 was able to co-immunoprecipitate with the Flag-tagged PP1 catalytic subunit PP1CB, suggesting that AKAP2 interact with PP1 through the predicted motif (Fig. 5E, upper panel).Co-immunoprecipitation experiments failed to detect an interaction between AKAP2 PP1CA and PP1CC suggesting that AKAP2 might preferentially interact with the catalytic subunit PP1CB (Fig. S5).Finally, in vitro pulldown experiments performed using a GST-fusion of the PP1CB binding site of AKAP2 and purified PP1CB were able to detect an interaction between the two proteins (Fig. S6).In contrast, no binding was detected when pulldown experiments were performed using a mutated PP1binding site (Fig. S6).This suggests that the identified PP1CB-interacting motif directly interacts with the phosphatase.

AKAP2-anchored PP1 mediates cofilin dephosphorylation
To determine whether anchoring of PP1 to the actin cytoskeleton by AKAP2 controls cofilin dephosphorylation in DU145 cells, we performed experiments that combined knockdown of the endogenous AKAP2 and Below the array, the proposed AKAP2 actin-binding motif, where X represents any amino acids, is shown.D) HEK-293 cells were transfected with either GFP-AKAP2 or the GFP-AKAP2-ΔACT mutant in which phenylalanine and arginine resides contributing to F-actin binding were substituted by alanines.48 h after transfection the AKAP2 constructs were immunoprecipitated from cell lysates using rabbit polyclonal anti-GFP antibodies, separated by SDS-PAGE, and transferred onto nitrocellulose membranes.Membranes were then incubated with F-actin and HRP-conjugated anti-actin antibodies to visualize AKAP2-actin interactions.rescue with GFP-tagged wild type and mutant AKAP2 forms that are refractory to siRNA-mediated silencing.Results indicate that rescue with GFP-AKAP2 reduces cofilin phosphorylation back to the levels observed in cells expressing control siRNAs (Fig. 6A, upper panel, Fig. 6B).In contrast, rescue with GFP-AKAP2 ΔAct or GFP-AKAP2 ΔPP1 did not significantly decrease cofilin hyperphosphorylation in silenced cells (Fig. 6A, upper panel, Fig. 6B).AKAP2 silencing and expression AKAP2 constructs was confirmed by Western blot (Fig. 6A).These findings suggest that AKAP2-mediated cofilin dephosphorylation requires the presence of intact actin and PP1 binding domains on the anchoring protein and are consistent with the idea that AKAP2 anchors PP1 to Factin in proximity of cofilin to facilitate its dephosphorylation.

AKAP2-anchored PP1 mediates DU145 cell migration
Based on the evidence that cofilin dephosphorylation increases its actin-severing activity and consequently actin turn-over, we investigated whether the AKAP2-PP1 complex regulates DU145 cell migration.To address this question, we evaluated the ability of control, AKAP2 silenced, and rescued cells to migrate using a transwell migration assay (Fig. 6C and D).Results indicate that cell migration is reduced by 55 % following AKAP2 silencing, confirming results shown in Fig. 2. Interestingly, while re-expression of wild type AKAP2 in silenced cells completely rescued their ability to migrate, rescue with the F-actin-or PP1-binding deficient mutants of AKAP2 did not (Fig. 6C and D), suggesting that interactions of AKAP2 with F-actin and PP1 are crucial for Fig. 4. AKAP2 regulates cofilin phosphorylation state.A) DU-145 cells were transfected with a control (Ctrl) siRNA or AKAP2 specific siRNA1 and siRNA2.72 h after transfection, cell lysates were loaded on SDS-PAGE gels and analyzed by Western blot.Phosphorylated cofilin was detected using antibodies recognizing phosphoserine 3 of cofilin.The amounts of total cofilin were detected using specific antibodies, as indicated.Silencing of AKAP2 was confirmed using an AKAP2 specific antibody.B) Quantitative analysis of phosphorylated cofilin was obtained by densitometry.The amount of phospho-cofilin was normalized to the total amount of cofilin.C) DU-145 cells were transfected with plasmids encoding either GFP, or two splice forms of GFP-AKAP2 003 that include or miss exon 3 (GFP-AKAP2 003 and GFP-AKAP2 003 Δex3, respectively).Cell lysates were subjected to immunoprecipitation with anti-FLAG antibodies.Western blots of the immunoprecipitates and cell lysates were revealed using rabbit anti-GFP polyclonal antibodies to detect GFP, GFP-AKAP2 or GFP-AKAP2 ΔPP1.Fig. 6.AKAP2-mediated anchoring of PP1CB to the actin cytoskeleton favors cofilin dephosphorylation and DU145 cell migration.A) DU-145 cells were transfected with control (Ctrl) or AKAP2 specific siRNAs (siRNA1) in combination with GFP, AKAP2-GFP or AKAP2-GFP mutants missing the actin-biding domain (AKAP2-ΔAct GFP) or the PP1CB-binsing site (AKAP2-ΔPP1 GFP).72 h after transfection, cell lysates were loaded on SDS-PAGE gels and analyzed by Western blot.Phospho-cofilin was detected using antibodies recognizing phospho-serine 3 of cofilin.The amounts of cofilin, AKAP2 GFP constructs, and actin were detected using specific antibodies, as indicated.Silencing of AKAP2 was confirmed using an AKAP2 specific antibody.B) Quantitative analysis of phosphorylated cofilin was obtained by densitometry.The amount of phospho-cofilin was normalized to the total amount of cofilin.Results are expressed as the mean ± SD of three independent experiments.* p < 0.05; ** p < 0.01.C) DU145 cells transfected as indicated in A) were seeded on transwell membranes with 8 μm pores and cultured for 24 h in RPMI1640 supplemented with 10 % FBS.Membranes were subsequently fixed, and cells migrated on the underside of the membrane were stained using 2 % crystal violet.Representative fields for each condition are shown.D) The percentage of migrated cells was determined on three transwell membranes per condition.Results are expressed as the mean ± SD of three independent experiments.* p < 0.05; ** p < 0.01.E) DU-145 cells were transfected with control (Ctrl) or AKAP2 specific siRNAs (siRNA1) in combination with GFP, cofilin-GFP or a cofilin-GFP phosphorylation-deficient mutant in which serine 3 was mutated into alanine (cofilin S3A-GFP).72 h after transfection, DU145 cells were seeded on transwell membranes and migration assessed as indicated in C. F) The percentage of migrated cells was determined on three transwell membranes per condition.Results are expressed as the mean ± SD of three independent experiments.* p < 0.05; ** p < 0.01 G) Expression of AKAP2, cofilin-GFP constructs and actin in the lysates was assessed by immunoblot blot using specific antibodies, as indicated.H) Model illustrating the role of the AKAP2/ PP1CB signaling complex in mediating migration in PNEC cells.Increased AKAP2 expression favors PP1CB tethering to the actin cytoskeleton, through a direct interaction with F-actin.Anchored PP1CB dephosphorylates Ser3 of cofilin, leading to cofilin activation.This enhances actin turnover, which promotes cell migration and invasion.
To provide evidence that the effects of AKAP2 on cell migration are mediated through cofilin, we assessed the impact of expressing wild type or a constitutively active mutant form of cofilin in which serine 3 is substituted by an alanine (Cofilin S3A) in AKAP2 silenced cells [18].If cofilin acts downstream of AKAP2 to regulate cell migration, then the expression of the cofilin S3A mutant in DU145 cells devoid of AKAP2 should restore their ability to migrate.Our results indicate that while rescue with wild type GFP-tagged cofilin did not significantly enhance migration of AKAP2 silenced cells, re-expression of cofilin S3A-GFP restored DU145 cell migratory capacity (Fig. 6E and F).AKAP2 silencing and expression cofilin constructs was confirmed by Western blot (Fig. 6G).Collectively these findings suggest that the AKAP2/PP1 complex regulates DU145 cell migration through the regulation of cofilin.

Discussion
PNEC aggressiveness is linked to its elevated metastatic potential.In this context, understanding how migratory and invasive processes are regulated in PNEC cells would potentially allow the identification of protein targets specifically controlling PNEC metastasis.
In this context, our current findings indicate that AKAP2 expression is strongly increased in PNECs as compared to control prostate tissues (Fig. 1).AKAP2 upregulation appears to be specific to PNECs since it is not observed in prostate adenocarcinomas (Fig. 1).Molecular analysis of the role of AKAP2 in DU145 cells, a prostate cancer cell line expressing neuroendocrine markers, revealed that the anchoring protein assembles a signaling complex that confers increased migratory capacity (Fig. 6H).In particular, AKAP2 recruits PP1CB to the peripheral actin cytoskeleton, where it promotes cofilin dephosphorylation and activation (Fig. 6H).This directly regulates actin turnover, thus favoring cancer cell migration.In line with these findings, AKAP2 silencing in DU145 cells increases cofilin phosphorylation and inhibits migration.Overall, these findings suggest that AKAP2 coordinates a PP1CB based signaling complex coordinating pro-migratory signals at the actin cytoskeleton of NEPCs.
The fact that AKAP2 mRNA and protein expression is strongly upregulated in PNECs but not in prostate adenocarcinomas suggest that transcription of the AKAP2 gene is specifically enhanced during the process of neuroendocrine differentiation.Similarly, AKAP2 expression in PC cell lines highly correlates with the expression of neuroendocrine markers (Fig. 1E-H).Similarly, AKAP2 mRNA expression profiles reported by the NIH Genomic Data Commons (https://gdc.cancer.gov)for various PC cell lines indicate that AKAP2 expression is mainly detected in prostate cancer cell lines expressing neuroendocrine markers (e.g.DU145, PC3 and VCAP cells) [30].Interestingly, analysis of the AKAP2 cistrome using the online platform provided by the Signaling Pathways Project (SPP) (http://www.signalingpathways.org)[31,32], revealed that the promoter region of AKAP2 contains binding sites for transcription factors including N-Myc, Sox2, and Ascl1, and FOXA2 which have been shown to participate in neuroendocrine differentiation of prostate cancers [33][34][35][36][37][38].Based on this evidence, one could speculate that upregulation and activation of these transcriptional regulators might increase AKAP2 expression in PNEC cells and consequently enhance their migratory and invasive properties.
In line with our current findings showing that AKAP2 enhances PC motility and invasiveness, recent studies indicate that AKAPs can control migratory and invasive properties in several cancer cell types.In this respect, it has been shown that AKAP13 (AKAP-Lbc) controls the formation of PKA activity gradients that ultimately control migration colon carcinoma cells and promotes RhoA-dependent migration and invasion of PC3 prostate cancer cells [11,39].Similarly, AKAP220-dependent regulation of GSK3 and IQGAP1 at the leading edge of fibrosarcoma cells influences the polymerization of the actin and microtubule cytoskeleton and, consequently, their migratory properties [40].On the other hand, recent studies highlight the role for AKAP1 (dAKAP1) in modulating migration and invasiveness of breast cancer cells.In particular, depletion of AKAP1-PKA complexes from the mitochondrial membrane, impaired mitochondrial function, and increases the glycolytic potential and invasiveness [41].This suggests that AKAPs, by assembling distinct signaling complexes at focal points in cancer cells, can ultimately modulate cytoskeleton dynamics to regulate cancer cell migration and/or invasion.
Previous findings highlighted an association between AKAP2 expression and prostate cancer development.In this regard, silencing of AKAP2 or disruption of the interaction between AKAP2 and PKA has been shown to decrease PC3 prostate cancer cell proliferation and invasiveness.This suggested that AKAP2 regulates PC3 cell oncogenic properties through its interaction with PKA [42].Surprisingly, this conclusion diverges from our current results, which indicate that AKAP2 silencing does not consistently impact DU145 cell proliferation (Fig. S2).Moreover, co-immunoprecipitation studies performed from DU145 failed to detect regulatory or catalytic subunits of PKA in AKAP2 immunoprecipitates suggesting that AKAP2 does not recruit PKA in this PC cell line (Fig. S7).While the reason of these apparent discrepancies is currently not known, one could speculate that the interaction between AKAP2 and PKA might be regulated by signaling cascades that are differentially activated in DU145 and PC3 cells.
The fact that the AKAP2/PP1 complex is not formed in PC3 cells (Fig. S4) further suggests that AKAP2 assembles profoundly different signaling complexes in DU145 and PC3 cells.In this respect, while DU145 cells rely on the AKAP2/PP1CB signaling complex to transduce promigratory signals, PC3 cells might regulate proliferation and migration through the formation of AKAP2-PKA complexes.The reason why these two cell types use distinct mechanism to regulate cell movement is not clear, but it might be related to the fact PC3 cells might not display the full characteristics of PNECs as shown by the fact that they express lower amounts of chromogranin A compared to DU145 cells (Fig. 1).
Our current findings indicate that the AKAP2/PP1CB signaling complex favors DU145 cell migration by promoting cofilin dephosphorylation.In line with our current results, cofilin activation through dephosphorylation has been shown to promote PC3 prostate cancer cell migration and metastatic spread by enhancing actin severing [18].Similar observations have been made in breast cancer and colon cancer cells, where cofilin activation generates actin filament barbed ends required for the extension of lamellipodia and migration [43,44].Interestingly, PC3 cells, similarly to DU145 cells, express AKAP2 and neuroendocrine markers (Fig. 1).This raises the possibility that AKAP2 might selectively favor cofilin activation and increased migratory function in prostate cancer cells undergoing neuroendocrine differentiation.
SSH1 has been shown to act as the main phosphatase involved cofilin dephosphorylation in several cell types [16].Our present results, however, suggest that AKAP2 favors cofilin dephosphorylation by anchoring PP1CB and not SSH1 (Figs. 5, 6, and S3).Moreover, AKAP2 does not participate in the regulation of SSH1 (Fig. S3).These findings imply that PP1CB functions as a cofilin-phosphatase controlling migratory processes in NEPC cells.In line with this observation, growing evidence suggests that PP1 catalytic subunits play important roles in tumorigenesis by regulating multiple pro-oncogenic pathways promoting cell proliferation and tumor growth, epithelial-mesenchymal transition (EMT) as well as migration and metastasis [45].In this respect, the PP1CA gene has been found to be amplified in up to 17 % of castration resistant PCs [46], whereas PP1CB has been shown to play promigratory and invasive functions by influencing the turnover of focal adhesions [47,48].
Several AKAPs have been shown to anchor PP1 to specific cellular compartments to locally regulate cellular functions.This is the case of AKAP220, which favors PP1-mediated regulation of renal water reabsorption [49] and renal ciliary homeostasis [50]; Yotiao, which targets PP1 to the N-methyl-D-aspartate (NMDA) receptor to regulate channel activity [51]; AKAP149 (D-AKAP1), which regulates nuclear envelope integrity, and AKAP79, which directs PP1 activity towards phospho-PSD-95 [52].In this context, our current findings that AKAP2-directed targeting of PP1CB enhances PC cell migratory properties further highlight the key role AKAP-mediated PP1 subcellular compartmentalization in specifying the cellular function of this widely expressed phosphatase.
The PP1CB anchoring domain identified on AKAP2 (FKLRSR) conformed to a consensus PP1 binding motif (FxxR/KxR/K) also found in AKAP79 but not to other well characterized PP1 binding sequences including the RVxF, MyPhone, or SILK motifs [52].Interestingly, in line with the previous observation that the PP1 interacting motif of AKAP79 enhances PP1 activity towards substrates [52], our current findings indicate that deletion of this consensus sequence from AKAP2 increases cofilin phosphorylation.This raises the possibility that anchoring of PP1 through FxxR/KxR/K sequences directs an active pool of the phosphatase in proximity of its downstream targets.
Because PP1 catalytic subunits are abundantly expressed in most tissues and organs, inhibiting their activity does not represent a viable strategy to inhibit PNEC growth and development because of potential widespread toxic effects [53].Interfering with PP1 subcellular localization by altering the interaction between PP1 anchoring proteins and target sites might offer a more focused approach.However, since the FxxR/KxR/K PP1 binding motif identified on AKAP2 is conserved in other intracellular proteins, one could expect that competitor peptides or small molecules designed to disrupt the interaction between AKAP2 and PP1CB might potentially also interfere with the targeting of the phosphatase to other signaling complexes, and therefore display low specificity.On the other hand, the newly identified actin binding site of AKAP2 is not conserved in other proteins and could therefore serve as basis for the design of cell-permeable competitor peptides that would disrupt the interaction of AKAP2 with F-actin in a more specific manner and consequently inhibit the targeting of PP1CB in proximity of cofilin.In this respect, the F-actin binding sequence of AKAP2 could be fused to cell penetrating cationic peptides to ensure their uptake by PNEC cells [54].Nevertheless, the in vivo success of such an approach might be hampered by an excessive degradation of the peptide by serum proteases.Under these circumstances, one would need to modify the peptide to render it resistant to proteolytic degradation [23].
An alternative strategy could rely on the development of cellpermeable small molecules inhibitors of the AKAP2-F-actin interaction.In this case, the AKAP2-F-actin interaction surface could be targeted by virtual screening of compound libraries.Positive hits would next be tested for their ability to selectively block the interaction [11].Future experiments will determine whether these strategies can efficiently inhibit PNEC cell migration and invasiveness in vitro and in vivo.

Conclusions
In conclusion, our current findings have several implications.Firstly, they provide evidence that AKAP2 is strongly upregulated in human PNECs where it forms a complex with PP1CB.Secondly, they highlight that the AKAP2-PP1CB signaling complex is targeted to the actin cytoskeleton through a direct interaction with F-actin.Thirdly, they indicate that AKAP2 confers migratory and invasive properties to prostate cancer cells by favoring PP1CB-mediated dephosphorylation and activation of cofilin.This suggests that AKAP2 functions as a signal organizer that enhances cofilin regulation and PNEC cell oncogenic properties.

Ethical approval and consent to participate
Collection and processing of human prostate cancer tissues used for the present study have been carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans and was approved by the Ethical Committee of the ASST Sette Laghi, Varese, Italy (prot.nr: 0008465).

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig. 1.Neuroendocrine prostate cancers display an increased expression of AKAP2.A) % of neuroendocrine prostate cancers displaying amplification of the AKAP2 gene as reported in cBioPortal (https://www.cbioportal.org).B) qPCR analysis of AKAP2 mRNA expression in biopsies of control prostates and neuroendocrine prostate cancers.C) Anti-AKAP2 (upper panels) and hematoxylin/eosin (lower panels) stainings of sections from human control prostate (left panels) and PNEC (right panels) biopsies.D) Anti-AKAP2 stainings of sections of human PNEC (upper panels) and prostate adenocarcinoma (lower panels).E) qPCR analysis of synaptophysin and chromogranin A mRNA expression in LnCAP, DU-145, and PC3 prostate cancer cells.Data are expressed as a mean ± SD of 3 independent experiments * p < 0.05; ** p < 0.01.F-G) qPCR (F) and Western blot (G) analysis of AKAP2 expression in LnCAP, DU-145, and PC3 prostate cancer cells.Data are expressed as a mean ± SD of 3 independent experiments * p < 0.05; *** p < 0.001.H) Quantitative analysis of AKAP2 expression in DU-145 cell lysates was obtained by densitometry.

Fig. 2 .
Fig. 2. AKAP2 controls migration and invasion of DU-145 prostate cancer cells.A) Western blot analysis of AKAP2 expression in DU-145 cells transfected with control (Ctrl siRNA) or AKAP2 specific siRNAs (siRNA1 and shRNA2).B) Quantitative analysis of AKAP2 expression in DU-145 cell lysates was obtained by densitometry.C) Wounded DU-145 cell monolayers transfected with Ctrl siRNA, siRNA1 or shRNA2 were cultured for 36 h.Recolonization of the wounded area was imaged every 6-12 h.Scale bar, 1000 μm.D) Quantitation of the wound-healing process.*p < 0.05, **p < 0.01, ***p < 0.001 compared with the wounded area measured in monolayers transfected with control siRNAs.E-H) DU-145 cells were transfected as indicated in A. Random migration was imaged using time-lapse microscopy for 8 h.Migration trajectories (E-G) and velocity (μm/min) (H) were determined on a total of 90 cells per condition.***p < 0.001 compared with the migration velocity measured in cells transfected with control siRNAs.I-K) 2.5 × 10 4 DU-145 cells transfected as indicated in A, were seeded on Matrigel-coated transwell membranes with 8 μm pores and cultured for 30 h in RPMI1640 supplemented serum.FBS concentration was 0.5 % in the upper chamber and 10 % in the lower chamber.Membranes were fixed and cells that migrated on the underside of the membrane were stained using 2 % crystal violet.Representative fields for each condition are shown.L) The percentage of invading cells was determined on three transwell membranes per condition.*p < 0.05, **p < 0.01 compared with the percentage of invading cells measured in cells transfected with control siRNAs.

Fig. 3 .
Fig. 3. AKAP2 interacts with actin in DU-145 cells through a newly identified motif.A) DU-145 cells were transfected with control (Ctrl siRNA) or AKAP2 specific shRNAs (siRNA1 and shRNA2).72 h post transfection, cells were fixed and stained.AKAP2 was stained in green using a polyclonal anti-AKAP2 antibody in combination with Alexa Fluor 488 goat anti-rabbit secondary antibody.Actin was detected using Texas red Phalloidin.B) Peptide arrays analysis of the F-actin-binding site on AKAP2.Arrays of overlapping 20-amino acid peptides spanning the entire sequence of AKAP2 were incubated with (F-actin array) or without (Control array) 1 μM of F-actin.Solid phase binding was then assessed with HRP-conjugated anti-actin antibodies.The main F-actin binding peptides are numbered and their sequence shown.C) Two-dimensional array of AKAP2 peptides where each residue in the F-actin binding sequence (indicated above the array) was replaced with all 20 amino acids (indicated to the left of the array) are shown.F-actin binding was assessed as indicated in A. Replacement of key phenylalanine and arginine resides within the binding motif perturbs F-actin binding (red boxes).Internal control peptides corresponding to the native AKAP2 sequence are indicated by white circles.Below the array, the proposed AKAP2 actin-binding motif, where X represents any amino acids, is shown.D) HEK-293 cells were transfected with either GFP-AKAP2 or the GFP-AKAP2-ΔACT mutant in which phenylalanine and arginine resides contributing to F-actin binding were substituted by alanines.48 h after transfection the AKAP2 constructs were immunoprecipitated from cell lysates using rabbit polyclonal anti-GFP antibodies, separated by SDS-PAGE, and transferred onto nitrocellulose membranes.Membranes were then incubated with F-actin and HRP-conjugated anti-actin antibodies to visualize AKAP2-actin interactions.
Fig. 4. AKAP2 regulates cofilin phosphorylation state.A) DU-145 cells were transfected with a control (Ctrl) siRNA or AKAP2 specific siRNA1 and siRNA2.72 h after transfection, cell lysates were loaded on SDS-PAGE gels and analyzed by Western blot.Phosphorylated cofilin was detected using antibodies recognizing phosphoserine 3 of cofilin.The amounts of total cofilin were detected using specific antibodies, as indicated.Silencing of AKAP2 was confirmed using an AKAP2 specific antibody.B) Quantitative analysis of phosphorylated cofilin was obtained by densitometry.The amount of phospho-cofilin was normalized to the total amount of cofilin.C) DU-145 cells were transfected with plasmids encoding either GFP, or two splice forms of GFP-AKAP2 003 that include or miss exon 3 (GFP-AKAP2 003 and GFP-AKAP2 003 Δex3, respectively).72 h after transfection, the levels of phospho-cofilin, cofilin and GFP constructs in cell lysates were analyzed as indicate in B. D) Quantitative analysis of phosphorylated cofilin was performed as indicated in C. Results are presented as the mean ± SD of three independent experiments.* p < 0.05; ** p < 0.01.E) Proximity ligation assay for DU-145 cells incubated with anti-AKAP2 and anti-cofilin antibodies alone or in combination.Nuclei were stained with DAPI and shown in blue whereas F-actin was stained with the Actin Green Ready probe.PLA signals representing positive AKAP2/cofilin interactions are shown in red.

Fig. 5 .
Fig. 5. AKAP2 forms a complex with protein phosphatase 1. A) Lysates from DU-145 cells were subjected to immunoprecipitation with either non-immune IgGs or affinity purified anti-AKAP2 antibodies.The presence of PP1 in the lysates and immunoprecipitates was identified by immunoblot using specific antibodies.B) Proximity ligation assay for DU-145 cells incubated with anti-AKAP2 and anti-PP1 antibodies alone or in combination.Nuclei were stained with DAPI and shown in blue whereas F-actin was stained with the Actin Green Ready probe.PLA signals representing positive AKAP2/PP1 interactions are shown in red.C) Proximity ligation assay for sections of human NEPC biopsies incubated with anti-AKAP2 and anti-PP1 antibodies alone or in combination.Nuclei were stained with DAPI and shown in blue whereas PLA signals representing positive AKAP2/PP1 interactions are shown in red.D) Schematic representation and protein domain organization of AKAP2.The actin binding site (residues 319-341), the PKA-binding domain (residues 565-582), and the PP1 interacting motif are shown.Alanine substitutions introduced in the AKAP2 sequence to suppress PP1 binding are indicated in red.E) HEK-293 cells were transfected with the empty pFLAG vector or the vector encoding Flag PP1CB in combination with plasmids encoding either GFP, GFP-AKAP2 or the GFP-tagged PP1-binding deficient mutant of AKAP2 (GFP-AKAP2 ΔPP1).Cell lysates were subjected to immunoprecipitation with anti-FLAG antibodies.Western blots of the immunoprecipitates and cell lysates were revealed using rabbit anti-GFP polyclonal antibodies to detect GFP, GFP-AKAP2 or GFP-AKAP2 ΔPP1.