Abstract

A-kinase-anchoring protein 149 (AKAP149) is a membrane protein of the mitochondrial and endoplasmic reticulum/nuclear envelope network. AKAP149 controls the subcellular localization and temporal order of protein phosphorylation by tethering protein kinases and phosphatases to these compartments. AKAP149 also includes an RNA-binding K homology (KH) domain, the loss of function of which has been associated in other proteins with neurodegenerative syndromes. We show here that protein phosphatase 1 (PP1) binding occurs through a conserved RVXF motif found in the KH domain of AKAP149 and that PP1 and RNA binding to this same site is mutually exclusive and controlled through a novel, phosphorylation-dependent mechanism. A collapse of the mitochondrial network is observed upon introduction of RNA-binding deficient mutants of AKAP149, pointing to the importance of RNA tethering to the mitochondrial membrane by AKAP149 for mitochondrial distribution.

INTRODUCTION

Subcellular localization and temporal organization of protein phosphorylation by protein kinase A (PKA) are mediated by proteins collectively termed A-kinase-anchoring proteins (AKAPs) (1). AKAPs bind a PKA regulatory subunit dimer, and a targeting domain specifies subcellular localization. AKAPs also interact with other signalling molecules such as protein kinase C (PKC), phosphodiesterases and protein phosphatases (PPs) including PP1 in a space- and time-regulated fashion (1,2).

AKAP149 (also termed D-AKAP1; AKAP121 in mouse) is a human 149 kDa anchoring protein localized in mitochondria and the endoplasmic reticulum/nuclear envelope (ER–NE) network (3,4). In addition to PKA, AKAP149 co-precipitates PKCα from NE fractions (5) and binds phosphodiesterase PDE4A (6). AKAP121 also binds PTPD1, an Src-associated classical non-receptor protein tyrosine phosphatase, and Src itself (7). These findings suggest that AKAP149 is an important scaffolding molecule for phosphorylation events in the mitochondrial/ER–NE compartments.

PP1, a member of the PPP family of Ser/Thr phosphatases, has been implicated in a range of cellular processes including muscle contraction, cell division and pre-mRNA splicing by dephosphorylation of key proteins such as RNA polymerase II, SR-splicing factors and retinoblastoma protein (8–12). PP1 usually consists of a catalytic and a regulatory subunit (8,13). The latter serves to target the catalytic subunit to the proximity of its substrate. Most PP1 regulators contain an RVXF motif that interacts with the hydrophobic pocket of PP1. AKAP149 harbours two canonical RVXF motifs: the previously characterized 153KGVLF157 (3) and the uncharacterized 627RYVSF631 located in the RNA-binding K homology (KH) domain. We earlier demonstrated that AKAP149 functions as a PP1 regulatory subunit and enhances PP1 phosphatase activity towards B-type lamins upon NE re-formation at the end of mitosis, thus promoting lamin dephosphorylation and polymerization to assemble the nuclear lamina (3,14).

AKAP121, the mouse orthologue of AKAP149, binds the 3′ untranslated region (UTR) of two mitochondrial protein transcripts (MnSOD and F0F) in vitro (15) and of lipoprotein lipase mRNA (16). In both cases, binding of AKAP121 to mRNA is mediated by a single COOH-terminal KH domain. KH domains bind RNA or ssDNA and are found in proteins associated with pre-mRNA splicing, as well as mRNA stabilization, subcellular localization and translational regulation. Interestingly, several diseases such as fragile X mental retardation syndrome and paraneoplastic opsoclonus-myoclonus-ataxia (POMA) neurodegenerative syndrome are associated with the loss of function of a KH domain in specific proteins (17,18). We have recently shown that the KH domain of AKAP149 binds RNA in vivo and is required for the self-association of AKAP149 (19). These findings suggest a functional importance for mRNA binding with the KH domain of AKAP149.

Here, we set out to define the functional importance of RNA and PP1 binding to the KH domain of AKAP149. We show that AKAP149 binds PP1 mostly through the conserved 627RYVSF631 motif found within the RNA-binding KH domain. Binding of PP1 and RNA with the KH domain is mutually exclusive and controlled by an inverse phosphorylation mechanism. Finally, compromising RNA binding by the KH domain yields a mitochondrial aggregation phenotype, suggesting a role for RNA binding to AKAP149 in the maintenance of a normal mitochondrial network.

RESULTS

AKAP149 binds PP1 through the 627RYVSF631 motif in the KH domain

AKAP149 has been shown in vitro to bind PP1 through an NH2-terminal RVXF motif (5); however, AKAP149 contains another RVXF motif, as yet of unknown significance, in the RNA binding core of the KH domain. To gain insight into the regions of AKAP149 interacting with PP1 in vivo, we first transiently expressed AKAP149 fragments NH2-terminally tagged with haemagglutinin (HA) (Fig. 1A) in HeLa cells stably expressing PP1α-enhanced green fluorescent protein (EGFP) or PP1γ-EGFP and performed immunoprecipitations. Both PP1α and PP1γ co-precipitated with HA-AKAP149 (Fig. 1B), indicating that AKAP149 associates with both PP1 isoforms.

Figure 1.

AKAP149 associates with PP1 through the KH domain. (A) AKAP149 constructs used in this study. EGFP or mCherry tag was fused to the COOH terminus, and HA or GST tags were fused to the NH2 terminus, as described earlier. Membrane association (MA), PP1 binding, PKA binding, KH and Tudor domains are shown. KHT RAXA refers to a V629A, F631A double mutation in the RVXF motif within the KH domain; 1–485 RAXA refers to a V155A, F157A double mutation in the NH2-terminus RVXF motif. (B) PP1α-EGFP or PP1γ-EGFP HeLa cell lines transfected with full-length HA-AKAP149 were subjected to immunoprecipitation using anti-GFP or pre-immune control (CTL) antibodies. Proteins were immunoblotted using anti-GFP or anti-HA. (C) PP1α-EGFP or PP1γ-EGFP HeLa cells transfected with HA-AKAP149(1–485) or the corresponding RAXA mutant was immunoprecipitated as in (A). (D) PP1α-EGFP or PP1γ-EGFP HeLa cells transfected with HA-AKAP149 (KHT) or the corresponding RAXA mutant was immunoprecipitated as in (A). (E) GST-tagged AKAP149 (KHT), (KHT, RAXA), (KHT, S630D) or GST alone was mixed with 0.5 µg recombinant PP1α and pulled down. Proteins were immunoblotted using anti-GST or anti-PP1. (F) SPR analysis of the interaction of KHT, KHT RAXA, 1–485 and 1–485 RAXA peptides (0.15 µm) with immobilized PP1α.

Figure 1.

AKAP149 associates with PP1 through the KH domain. (A) AKAP149 constructs used in this study. EGFP or mCherry tag was fused to the COOH terminus, and HA or GST tags were fused to the NH2 terminus, as described earlier. Membrane association (MA), PP1 binding, PKA binding, KH and Tudor domains are shown. KHT RAXA refers to a V629A, F631A double mutation in the RVXF motif within the KH domain; 1–485 RAXA refers to a V155A, F157A double mutation in the NH2-terminus RVXF motif. (B) PP1α-EGFP or PP1γ-EGFP HeLa cell lines transfected with full-length HA-AKAP149 were subjected to immunoprecipitation using anti-GFP or pre-immune control (CTL) antibodies. Proteins were immunoblotted using anti-GFP or anti-HA. (C) PP1α-EGFP or PP1γ-EGFP HeLa cells transfected with HA-AKAP149(1–485) or the corresponding RAXA mutant was immunoprecipitated as in (A). (D) PP1α-EGFP or PP1γ-EGFP HeLa cells transfected with HA-AKAP149 (KHT) or the corresponding RAXA mutant was immunoprecipitated as in (A). (E) GST-tagged AKAP149 (KHT), (KHT, RAXA), (KHT, S630D) or GST alone was mixed with 0.5 µg recombinant PP1α and pulled down. Proteins were immunoblotted using anti-GST or anti-PP1. (F) SPR analysis of the interaction of KHT, KHT RAXA, 1–485 and 1–485 RAXA peptides (0.15 µm) with immobilized PP1α.

Subsequently, PP1 binding to either RVXF motif was examined after expression in PP1α- or PP1γ-EGFP HeLa cells of an HA-AKAP149 (1–485) fragment containing the NH2-terminal RVXF (153KGVLF157) or an AKAP149 fragment comprising the KH-Tudor domain (referred to as KHT here), including the COOH-terminal RVXF (627RYVSF631) (Fig. 1A). Surprisingly, neither PP1α-EGFP nor PP1γ-EGFP co-precipitated with HA-AKAP149 (1–485) (Fig. 1C), with or without an RAXA double mutation known to abolish or reduce PP1 binding (20,21). However, both PP1 isoforms co-precipitated with HA-AKAP149 (KHT) (Fig. 1D), and this co-precipitation was reduced when the RAXA substitutions were introduced into 627RYVSF631 (Fig. 1D). Therefore, AKAP149 interacts with both tagged PP1α and PP1γ preferentially through the COOH-terminal 627RYVSF631 motif in co-immunoprecipitation assays.

The absence of a strong detectable interaction with PP1 mediated by the NH2-terminal RVXF motif of AKAP149 was confirmed by anti-GFP immunoprecipitation of EGFP-tagged full-length AKAP149 constructs harbouring phosphorylation-deficient or constitutively phosphorylated mutations of the serines flanking 153KGVLF157. The phosphorylation state of serines S151 and S159 flanking 153KGVLF157, previously shown to affect binding in vitro (5), did not affect in vivo binding to endogenous PP1 (Supplementary Material, Fig. S1). Whether the 627RYVSF631 motif of AKAP149 directly interacts with PP1 was determined in a GST pulldown experiment using GST-tagged AKAP149 (KHT) and purified PP1α in solution. As shown in Figure 1E (left two lanes), PP1α bound to AKAP149 (KHT) but not to the AKAP149 (KHT, RAXA) mutant, demonstrating a direct interaction of AKAP149 with PP1α through the 627RYVSF631 motif.

Next, we characterized the two RVXF motifs by surface plasmon resonance (SPR) analysis of the interaction of the corresponding KHTpep, KHT RAXApep, 1–485pep and 1–485 RAXApep peptides with immobilized PP1α. Figure 1F shows that the SPR binding response for KHTpep was ∼5-fold higher than that for the KHT RAXApep mutant, whereas none of the 1–485 peptides showed detectable association with immobilized PP1α. Collectively, these results indicate that PP1 binds AKAP149 directly and primarily through the COOH-terminal 627RYVSF631 motif.

Serine 630 phosphorylation in the 627RYVSF631 motif regulates association between PP1 and AKAP149

Binding of PP1 to a regulatory subunit is in several instances regulated by Ser phosphorylation close to the RVXF motif, resulting in the disruption of PP1 binding (22). To determine whether interaction between the 627RYVSF631 motif and PP1 is regulated by phosphorylation, HA-tagged AKAP149 KHT fragments with mutations in the RNA binding groove of the KH domain that mimics a phosphorylation-deficient KHT [HA-AKAP149 (KHT, S630A)] or a constitutively phosphorylated KHT [HA-AKAP149 (KHT, S630D)] were expressed in PP1γ-EGFP HeLa cells. Immunoprecipitation using anti-GFP antibodies co-precipitated HA-AKAP149 (KHT) (Fig. 2A). The S630D pseudo-phosphorylation mutation in 627RYVSF631 resulted in weakened association between HA-AKAP149 (KHT) and PP1γ-EGFP, whereas the S630A phosphorylation-deficient mutant showed reinforced interaction (Fig. 2A). Similar results were obtained when full-length S630A and S630D AKAP149 constructs were used (Supplementary Material, Fig. S2C).

Figure 2.

Binding of the KH-Tudor domain of AKAP149 with PP1 is modulated by S630 phosphorylation in the RVXF motif. (A) PP1γ-EGFP HeLa cells transfected with HA-AKAP149 (KHT), (KHT, S630D) or (KHT, S630A) were subjected to immunoprecipitation, as shown in Figure 1. Graph shows a densitometric analysis of the ratio of anti-HA/anti-GFP signals in the blot. (B) SPR analysis of the interaction of 0.15 µm AKAP149 (KHT), AKAP149 (KHT, S630*pep) and AKAP149 (KHT, RAXA) peptides with immobilized PP1α. One of five experiments is shown. (C) Ligand-binding curves for PP1α with 0.15–5.0 µM of indicated peptides.

Figure 2.

Binding of the KH-Tudor domain of AKAP149 with PP1 is modulated by S630 phosphorylation in the RVXF motif. (A) PP1γ-EGFP HeLa cells transfected with HA-AKAP149 (KHT), (KHT, S630D) or (KHT, S630A) were subjected to immunoprecipitation, as shown in Figure 1. Graph shows a densitometric analysis of the ratio of anti-HA/anti-GFP signals in the blot. (B) SPR analysis of the interaction of 0.15 µm AKAP149 (KHT), AKAP149 (KHT, S630*pep) and AKAP149 (KHT, RAXA) peptides with immobilized PP1α. One of five experiments is shown. (C) Ligand-binding curves for PP1α with 0.15–5.0 µM of indicated peptides.

The negative effect of phosphorylation of 627RYVSF631 on interaction with PP1 was also demonstrated by the absence of interaction of GST-AKAP149 (KHT, S630D) with purified recombinant PP1α in a GST pulldown assay (Fig. 1E, third lane). Furthermore, the SPR analysis demonstrated that phosphorylation of S630 (S630*pep) reduced binding to immobilized PP1α 2-fold (Fig. 2B and C), indicating that S630 phosphorylation of the AKAP149 627RYVSF631 motif reduces the affinity for PP1 (Supplementary Material, Table S1). Therefore, binding of the 627RYVSF631 motif decreases upon S630 phosphorylation within the RVXF (and KH) motif.

AKAP149 binds directly to RNA in a Ser phosphorylation-dependent manner

To confirm previous assessments that the KH domain of AKAP149 associates with RNA (15,19) and extend the observations to an in vivo context, we performed an RNA immunoprecipitation experiment following UV cross-linking (19). 293T cells were UV cross-linked, RNA partially digested with RNase T1 and immunoprecipitation was carried out using anti-AKAP149 antibodies. Immunoprecipitated RNA was γ-32P end-labelled before sodium dodecyl sulphate (SDS)–polyacrylamide gel electrophoresis (PAGE) and autoradiography. Notably, AKAP149-RNA complexes were detected after UV cross-linking, but not under native conditions (Fig. 3A), thus AKAP149 directly associates with RNA in living cells.

Figure 3.

The KH-Tudor domain of AKAP149 binds directly to RNA in an S630 phosphorylation-dependent manner. (A) Autoradiography of RNA binding to immunoprecipitated AKAP149 (A149) after UV-cross-linking compared with native conditions. (B) Autoradiography and western blotting of indicated EGFP-tagged AKAP149 peptide–RNA complexes immunoprecipitated from 293T cells. (C) Urea gel separation and autoradiography of RNA co-precipitated with indicated EGFP-tagged AKAP149 fragments and purified. (D) PP1γ-EGFP HeLa cells transfected with HA-KHT were subjected to immunoprecipitation, as shown in Figure 1, with lysates RNAse A-treated where indicated. (E) PP1α-, PP1δ- or PP1γ-EGFP co-expressed in HeLa cells with indicated AKAP149 fragments were immunoprecipitated, as shown in Figure 1. Densitometric analysis of anti-HA/anti-GFP signals is shown.

Figure 3.

The KH-Tudor domain of AKAP149 binds directly to RNA in an S630 phosphorylation-dependent manner. (A) Autoradiography of RNA binding to immunoprecipitated AKAP149 (A149) after UV-cross-linking compared with native conditions. (B) Autoradiography and western blotting of indicated EGFP-tagged AKAP149 peptide–RNA complexes immunoprecipitated from 293T cells. (C) Urea gel separation and autoradiography of RNA co-precipitated with indicated EGFP-tagged AKAP149 fragments and purified. (D) PP1γ-EGFP HeLa cells transfected with HA-KHT were subjected to immunoprecipitation, as shown in Figure 1, with lysates RNAse A-treated where indicated. (E) PP1α-, PP1δ- or PP1γ-EGFP co-expressed in HeLa cells with indicated AKAP149 fragments were immunoprecipitated, as shown in Figure 1. Densitometric analysis of anti-HA/anti-GFP signals is shown.

To characterize the RNA binding properties of AKAP149, we expressed EGFP-tagged wild-type or various mutant KH-Tudor domain constructs in 293T cells (Fig. 3B). AKAP149–RNA complexes were immunoprecipitated using anti-GFP antibodies and immune complexes subjected to autoradiography. Western blotting with anti-GFP antibodies (αGFP) revealed comparable amounts of EGFP-tagged proteins immunoprecipitated in all conditions (Fig. 3B). Autoradiography showed RNA/protein complexes precipitated by wild-type AKAP149 (KHT)-EGFP and with the pseudo-phosphorylated AKAP149 (KHT, S630D)-EGFP fragment. However, very little if any RNA was co-precipitated with the AKAP149 (KHT, S630A)-EGFP phosphorylation-deficient protein nor with AKAP149 (KHT, V629E)-EGFP, which was substituted in the hydrophobic groove of the KH domain (19) (Fig. 3B). As expected, the PP1-binding deficient AKAP149 (KHT, RAXA)-EGFP mutant did not show any RNA association. Collectively, these results suggest that the KH domain of AKAP149 associates with RNA when S630 is phosphorylated, whereas dephosphorylation of the same residue prevents RNA binding.

To confirm this finding, extracts from 293T cells transiently expressing EGFP-tagged AKAP149 (KHT), (KHT, V629E), (KHT, RAXA), (KHT, S630D) and (KHT, S630A) were treated with T1 RNase to partially digest RNA, the protein fragments immunoprecipitated using anti-GFP antibodies and immune complexes were γ-32P end-labelled before RNA elution and visualization by autoradiography. Wild-type KH-Tudor shows association with a population of RNA molecules (Fig. 3C, lane 1). This association was maintained with the S630D substitution in the KHT fragment (lane 7), arguing that in contrast to PP1 binding, S630 phosphorylation does not affect RNA binding. However, RNA co-precipitation was strongly reduced with S630A substitution (lane 9), suggesting a role of S630 phosphorylation in maintaining KHT interaction with RNA. Similar results were obtained when full-length S630A and S630D AKAP149 constructs were used (Supplementary Material, Fig. S2A and B). Furthermore, V629E or RAXA substitution of AKAP149 protein fragments also strongly reduced association with RNA (lanes 4 and 6), indicating that an intact hydrophobic groove within the KH domain is necessary to maintain interaction with RNA.

The V629E mutation weakens interaction between AKAP149 and PP1

The previous experiments show that the KH domain associates with RNA or PP1 through an inverse, phosphorylation-dependent process. We next addressed whether the RNA binding-deficient V629E mutation in 627RYVSF631 showed a similar binding exclusivity for PP1 or RNA. We first verified that PP1 binding to the KH domain was unaltered after RNase treatment in co-immunoprecipitation studies from lysates of PP1γ-EGFP-expressing HeLa cells transfected with HA-AKAP149 (KHT) (Fig. 3D), indicating that RNA is not required for PP1 binding. We next immunoprecipitated EGFP-tagged PP1α, PP1δ or PP1γ from HeLa cells co-expressing HA-AKAP149 (KHT) or AKAP149 (KHT, V629E). The V629E mutant showed reduced association between KHT and all three PP1 isoforms (Fig. 3E). Thus, the V629E mutation strongly diminishes the binding of both PP1 and RNA, unlike the S630A or D mutations. We concluded that binding of PP1 and RNA to the KH domain is mutually exclusive and regulated by an inverse phosphorylation mechanism on S630.

RNA-binding deficient AKAP149 causes a collapse of the mitochondrial network

To gain insights into the functional significance of the KH domain of AKAP149, we examined the effect of disrupted binding to PP1 and RNA in HeLa cells. Full-length AKAP149-EGFP, AKAP149 (V629E)-EGFP, AKAP149 (1–485)-EGFP, AKAP149 (S630A)-mCherry and -(S630D) -mCherry were transiently expressed in HeLa cells. The AKAP149 sequence used for all constructs was the orthologue of the mouse AKAP121-(N0) splice variant, which targets the mitochondrial membrane (23). Accordingly, AKAP149-EGFP localizes to mitochondria in cells showing an apparently normal mitochondrial network (Fig. 4A). A similar phenotype was observed for AKAP149 (S630D) with intact RNA binding and reduced PP1-binding (Fig. 4B and C). Strikingly, however, the expression of both RNA binding-deficient AKAP149 (V629E) and AKAP149 (S630A) resulted in pronounced aggregation of mitochondria in 30–50% of the cells (Fig. 4A–C). A similar effect was observed with AKAP149 (1–485), lacking the KH-Tudor domain (Fig. 4A–C). The aggregates did consist of mitochondria, as judged by co-labelling of ΔMDDX28-EGFP, a construct targeting to the inner mitochondrial membrane (24), with wild-type or mutant AKAP149-mCherry (Fig. 4C). Note that all AKAP149-mCherry constructs targeted mitochondria with a more peripheral location than ΔMDDX28-EGFP, probably corresponding to the outer mitochondrial membrane (Fig. 4C). The aggregation of the mitochondrial network could result from an alteration of the microtubules network, involved in mitochondrial positioning (25) and which has been reported to interact with AKAP149 (26). However, α-tubulin immunolabelling showed that the microtubule network was apparently normal in cells expressing AKAP149 (1–485)-EGFP with aggregated mitochondria (Supplementary Material, Fig. S3). Collectively, these results indicate that the expression of mitochondria-targeted, RNA binding-deficient AKAP149 mutants results in a collapse of the mitochondrial network.

Figure 4.

AKAP149 with defective RNA binding induces mitochondrial collapse. (A) Distribution of indicated EGFP-tagged AKAP149 proteins in HeLa cells. Bar: 10 µm. (B) Quantification of the mitochondrial distribution shown (mean±SD of percent cells showing indicated phenotypes; four experiments) after 24 h. (C) Co-distribution of ΔMDDX28-EGFP, a mitochondrial marker, and indicated AKAP149-mCherry proteins in HeLa cells. Bar: 10 µm.

Figure 4.

AKAP149 with defective RNA binding induces mitochondrial collapse. (A) Distribution of indicated EGFP-tagged AKAP149 proteins in HeLa cells. Bar: 10 µm. (B) Quantification of the mitochondrial distribution shown (mean±SD of percent cells showing indicated phenotypes; four experiments) after 24 h. (C) Co-distribution of ΔMDDX28-EGFP, a mitochondrial marker, and indicated AKAP149-mCherry proteins in HeLa cells. Bar: 10 µm.

DISCUSSION

We demonstrate that AKAP149 contains a functional COOH-terminal 627RYVSF631 motif within the conserved KH domain, in addition to the previously characterized 153KGVLF157 motif situated in the NH2 terminus (3). The existence of two PP1-binding RVXF motifs is not exclusive for AKAP149 and has been reported for two other regulatory subunits of PP1, Sipp1 and Aurora A kinase (20,27). For Sipp1, only disruption of both motifs completely abolishes PP1 binding. Our results indicate that AKAP149 binds PP1 mostly through the evolutionarily conserved 627RYVSF631 motif. However, although only in vitro data are available (5), we cannot rule out the possibility of PP1 binding to the NH2-terminal RVXF motif in vivo. A remaining intriguing question is the different biological responses in which the two PP1-binding RVXF domains of AKAP149 are involved.

A feature of AKAP149 is the presence of a conserved PP1-binding RVXF motif within the KH RNA-binding domain. This is also the case for the RNA splicing protein tra2-beta1, which shares the presence of an evolutionary conserved PP1-docking motif in the RNA recognition motif with eight other proteins, including SF2/ASF, SRp30c and polypyrimidine tract-binding protein (11). The emerging number of RNA-binding proteins with PP1-binding motifs in their RNA-binding domain suggests a complex phosphorylation-dependent regulation of RNA transport, processing and splicing. An additional level of complexity is provided for AKAP149 by the fact that it also anchors kinases such as PKA and PKC (3,5) and that 627RYVS630 is a consensus PKA phosphorylation site.

PP1 binding to RVXF motifs is frequently regulated by phosphorylation of residues in close proximity to the RVXF motif (22). We show here by mutational analysis that PP1 binding to 627RYVSF631 occurs when S630 is dephosphorylated. One may tentatively speculate that PP1 itself, before binding to the 627RYVSF631 motif, is responsible for S630 dephosphorylation. Our SPR experiments show a slight increase in interaction of the phosphorylated peptide KHT S630*pep at steady state (Fig. 2C). This could correspond to dephosphorylation by PP1, followed by binding of the newly dephosphorylated peptide. Similarly, we show that RNA binding to the KH domain is also affected by the phosphorylation status of the 627RYVSF631 motif, but in the opposite way, in that S630 phosphorylation allows RNA binding. It has already been shown that binding of the 3′-UTR of several mRNAs to AKAP149/AKAP121 is stimulated by PKA phosphorylation (15,16). However, a regulation by phosphorylation of the binding of RNA and PP1 has, to our knowledge, not been previously reported for the same site of a single protein domain. This mutually exclusive phosphorylation-regulated binding of either RNA or PP1 represents a novel and original switching mechanism to control binding of a PP and RNA molecule to the same domain of a signalling complex protein.

Substitutions disrupting the RNA-binding features of the KH domain of AKAP149/AKAP121 lead to a mitochondrial aggregation phenotype. This is similar to the aberrant accumulation of mitochondria observed with a GTP-binding deficient mutant of the mitochondrial AKAP Rab32, a member of the Ras superfamily of small G-proteins (28). The authors propose a role for Rab32 in the regulation of mitochondrial fission. We have noted for the RNA-binding-deficient mutants of AKAP149 that the collapsed mitochondria appear rounded rather than an elongated rod-like shape, suggesting fragmentation of the mitochondrial network. However, the mechanism of this mitochondrial collapse is likely to be different as Rab32 has no characterized RNA-binding domain. AKAP121 has been proposed to target important mitochondrial mRNAs such as Mn-superoxide dismutase or the ATP synthase subunit Fo-f to mitochondria through the KH domain (15). Thus, speculatively, AKAP149 may be involved in the regulation of mitochondrial fusion/fission mechanisms (29), possibly by localizing one or more of the required mRNAs to the mitochondrial membrane.

More generally, correct targeting of mRNAs to the mitochondrial membrane by AKAP149 could play an important role in the regulation of post-transcriptional gene expression. It is becoming increasingly clear that mRNA localization is involved in the sorting of several proteins destined to organelles, especially in mitochondrial biogenesis, which is a complex process requiring the concerted expression of both nuclear and mitochondrial genomes (30). In Saccharomyces cerevisiae, half of the transcripts coding for mitochondrial proteins preferentially localize to the organelle surface (31), and this is also observed in human cells (30).

mRNA mistargeting to the mitochondrial membrane may also result in mitochondrial dysfunction, which is often involved in neurodegenerative diseases (32). It has been proposed that mRNA sorting, by providing a mechanism for non-uniform protein distribution, is critical in the spatial architecture of polarized cells such as neurons (31). An important aspect of gene expression in neurons involves the delivery of mRNAs to particular subcellular domains, where translation of the mRNAs is locally controlled. This local synthesis of protein within dendrites plays a key role in activity-dependent synaptic modifications such as long-term potentiation and long-term depression. In growing axons, local synthesis in the growth cone is important for extension and guidance (33). Perturbations in mRNA transport and local translation could thus result in several human neurodegenerative diseases. As a matter of fact, spinal muscular atrophy is caused by mutation of the SMN1 gene that results in a truncated and unstable version of the SMN protein, which normally promotes an assembly of specific RNA-binding proteins with their target sequences (34). Even more interestingly, fragile X mental retardation syndrome and POMA are neurodegenerative diseases linked to impaired RNA binding, resulting from the loss of function of a specific KH domain (17). It is therefore tempting to speculate that mRNA sorting perturbation arising from impaired RNA anchoring at the KH domain of mitochondrial AKAP149 might be in a similar fashion at the origin of some neurodegenerative disorders.

We report for the first time a phosphorylation-dependent mechanism to regulate binding of RNA and PP1 in a mutually exclusive manner to the single KH domain on the AKAP149 signalling complex. We also show implications for mitochondrial distribution of a defective RNA binding to the same mitochondria-targeted AKAP149 signalling scaffold. Further efforts need to focus on identifying the specific RNA molecule(s) binding to AKAP149 and the nature and functional significance of the phosphorylation/dephosphorylation pathways regulating it.

MATERIALS AND METHODS

Cloning and expression of AKAP149 proteins

Fragments of AKAP149 cDNA (full-length; nucleotides 1–2712; KH-Tudor; nucleotides 1736–2712; 1–485; nucleotides 1–1436) were amplified by PCR, using primers that appended a 5′ EcoRI site and a 3′ BamHI site (EGFP constructs), a 5′ EcoRI site and a 3′ NotI site (GST constructs), a 5′ EcoRI site and a 3′ NotI site (HA constructs) or a 5′ HindIII and a 3′ SacII site (mCherry constructs). Amplified cDNA was digested with appropriate restriction enzymes and cloned into pEGFP-N3 or pmCherry-N1 (Clontech), pGEX-4T-1 (Pharmacia) or pET-BOS-HA (19). For mutagenesis, we used the Quick change Site-Directed Mutagenesis Kit (Stratagene).

Antibodies

Monoclonal anti-AKAP149 antibodies (mAbs) (Transductions Laboratories) were against residues 66–212 of human AKAP149. Rabbit polyclonal anti-GFP was from BD Biosciences and monoclonal anti-GFP antibodies were from Roche. Anti-HA mAbs were from Nordic Biosite. Anti-α-tubulin and anti-γ-tubulin were from Sigma. Polyclonal anti-PP1 was from Santa Cruz. Cy2-, Cy3- and peroxidase-conjugated secondary antibodies were from Jackson Laboratories.

Cell culture and transfection

HeLa cells were cultured in Dulbecco's modified Eagle's medium (DMEM)/1% glutamine/1% Na-pyruvate/1% non-essential amino acids/10% fetal calf serum. 293T cells were cultured in RPMI1640 supplemented as above. Cells were transfected using Effectene Transfection (Qiagen). Briefly, for immunofluorescence, cells were plated in DMEM on cover slips at 0.8×105 cells/well of a 12-well plate. After 24 h, 200 µl EC Buffer (Qiagen) containing 4–10 µl enhancer, 5–15 µl Effectene and 0.5–1.5 µg DNA was added. Cells were incubated for 24 h and processed for microscopy. For immunoprecipitation, cells were seeded in 75 cm2 culture flasks to 40–70% confluence 24 h before transfection.

Immunological procedures

Western blotting was performed as described earlier (35) using anti-AKAP149 (1:1000), anti-GFP (1:1000) or anti-HA (1:1000) antibodies. For immunoprecipitation, HeLa cells were harvested, washed, suspended in immunoprecipitation buffer (10 mm HEPES, pH 7.5, 10 mM KCl, 2 mM EDTA, 1% Triton X-100) adjusted to 100–250 mM KCl and allowed to swell for 15 min before sonication (1 min and 20 s). The lysate was centrifuged at 15 000g for 5 min. RNAse A-treated lysates were prepared, as described (19). After pre-clearing, immunoprecipitation was carried out with relevant antibodies (1:50; 4°C for 2 h) and incubation with Magnetic Protein G-Sepharose beads (Dynal) at 4°C for 1 h. Immune precipitates were washed three times with immunoprecipitation buffer adjusted to 100–250 mM KCl, and proteins were eluted in SDS sample buffer. Immunofluorescence was performed after paraformaldehyde fixation and permeabilization with 0.1% Triton X-100 or 1 minute methanol fixation, as described earlier (35). Cells were probed with anti-α-tubulin (1:1000) or anti-γ-tubulin (1:1000), followed by secondary antibodies conjugated with Cy3 (1:300). DNA was labelled with 0.25 µg/ml DAPI (Sigma). Samples were examined on an Olympus BX51 microscope using AnalySIS (Soft Imaging Systems).

Synthesis of peptides for SPR

Peptides used for SPR were synthesized by the Biotechnology Centre Peptide Synthesis Facility (University of Oslo, The Biotechnology Centre). Purity was analysed by HPLC and mass spectroscopy. The peptides span 30 amino acids with the potential PP1-binding motif in the centre: KH 614-PKHLVGRLIGKQGRYVSFLKQTSGAKIYIS-643, KH RAXA 614-PKHLVGRLIGKQGRYASALKQTSGAKIYIS-643, KH S630* 614-PKHLVGRLIGKQGRYVSFLKQTSGAKIYIS-643, 1–485 140-AKSIPLECPLSSPKGVLFSSKSA EVCKQDS-169, 1–485 RAXA 140-AKSIPLECPLSSPKGALASSKSAEVCKQDS-169. All peptides were dissolved in 25 mm Tris–HCl, pH 7.4. Exact peptide concentration was determined using a Biochrom 30 aminoacid analyser.

SPR analysis

SPR analysis was performed with a Biacore T100 instrument and analysed with the Biacore T100 evaluation software (Biacore AB). Recombinant PP1α (Sigma) (5 µg) diluted in the coupling buffer [10 mm acetate, pH 5.5, 1 mM DTT, 2 mM MnCl2 and 5 µM okadaic acid (sodium salt; Sigma] was immobilized covalently on a hydrophilic carboxymethylated dextran matrix [Sensor chip CM5 (Biacore AB)] with the amine coupling procedure essentially as described before (36) with 10 µl/min flow for 600 s. The remaining free succinimide ester groups were blocked with 1 M ethanolamine. During coupling and binding experiments, the assay buffer [25 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1 mM DTT, 2 mM MnCl2, 0.05% P20 (Biacore AB), 5% DMSO] was used. For binding studies, ±1600 RU recombinant PP1α was immobilized. The reference channel was treated similarly, but without recombinant PP1α. After coupling, assay buffer containing 0.05% SDS was passed through both surfaces (5 µl/min) overnight to minimize non-specific binding of the analytes to the biosensor surface. The following day plain assay buffer was passed through both channels 5 h before experimental start to remove SDS from the chip area and stabilize the baseline. All biosensor analyses were performed at 25°C with the assay buffer at a standard constant flow of 10 µl/min. The binding of KHT, KHT RAXA, KHT S630*, 1-485 and 1–485 RAXA peptides (analytes) on the immobilized enzyme was studied by injecting (10 µl/min) different concentrations of the analytes for 300 s. All sensograms were processed by first subtracting the SPR response observed for the reference surface to correct for any contribution of non-specific interaction with the dextran matrix.

RNA immunoprecipitation

Procedure was described previously (19). Briefly, transiently transfected 293T cells were UV-crosslinked (200 mJ/cm2, 254 mn, twice). Cells were harvested, washed and suspended in IPB containing 150 mm KCl. Cells were RNase T1 (0.002 U) digested for 8 min before sonication (10×30 s). The lysate was centrifuged at 15 000g for 15 min. After pre-clearing, immunoprecipitation was performed with Magnetic Protein G–Sepharose beads coated with relevant antibody (1:50) at 4°C for 2 h. Immune precipitates were washed four times in alternating fashion with RIPA (1× PBS, 0.05% SDS, 0.1% deoxycholate, 1% NP-40) and RIPA containing 1 M NaCl+1 mM EDTA. Precipitates were dephosphorylated with calf intestinal phosphatase (Roche) for 10 min at 37°C and washed twice with buffer D (50 mM Tris–HCl, pH 7.4, 20 mM EGTA) and twice with buffer C (50 mM Tris–HCl, pH 7.4, 10 mM MgCl2, 0.5% NP-40). Precipitates were incubated overnight at 16°C in a Thermomixer with a 30 s shake at 1000 rpm every 3 min. Precipitates were end-labelled with [γ-32P]ATP and either (i) resolved in a 7.5% urea-gel and visualization by autoradiography, (ii) dissolved in SDS sample buffer separated in a 10% Tris–HCl protein gel and visualized by autoradiography, or (iii) eluted in SDS sample buffer and visualized by SDS–PAGE.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.

FUNDING

This work was supported by the Norwegian Functional Genomics Programme, Research Council of Norway, the Norwegian Cancer Society and the University of Oslo.

ACKNOWLEDGEMENTS

We thank L. Trinkle-Mulcahy for PP1α- and PP1γ-EGFP cell lines, M. Bollen for PP1α-, PP1δ- and PP1γ-EGFP plasmids, J. Solheim for protein purification and L. Stijak and K. Vekterud for technical assistance.

Conflict of Interest statement. The authors report no conflicts of interest.

REFERENCES

1
Beene
D.L.
Scott
J.D.
A-kinase anchoring proteins take shape
Curr. Opin. Cell Biol.
 , 
2007
, vol. 
19
 (pg. 
192
-
198
)
2
Tasken
K.
Aandahl
E.M.
Localized effects of cAMP mediated by distinct routes of protein kinase A
Physiol. Rev.
 , 
2004
, vol. 
84
 (pg. 
137
-
167
)
3
Steen
R.L.
Martins
S.B.
Tasken
K.
Collas
P.
Recruitment of protein phosphatase 1 to the nuclear envelope by A-kinase anchoring protein AKAP149 is a prerequisite for nuclear lamina assembly
J. Cell Biol.
 , 
2000
, vol. 
150
 (pg. 
1251
-
1262
)
4
Trendelenburg
G.
Hummel
M.
Riecken
E.O.
Hanski
C.
Molecular characterization of AKAP149, a novel A kinase anchor protein with a KH domain
Biochem. Biophys. Res. Commun.
 , 
1996
, vol. 
225
 (pg. 
313
-
319
)
5
Kuntziger
T.
Rogne
M.
Folstad
R.L.
Collas
P.
Association of PP1 with its regulatory subunit AKAP149 is regulated by serine phosphorylation flanking the RVXF motif of AKAP149
Biochemistry
 , 
2006
, vol. 
45
 (pg. 
5868
-
5877
)
6
Asirvatham
A.L.
Galligan
S.G.
Schillace
R.V.
Davey
M.P.
Vasta
V.
Beavo
J.A.
Carr
D.W.
A-kinase anchoring proteins interact with phosphodiesterases in T lymphocyte cell lines
J. Immunol.
 , 
2004
, vol. 
173
 (pg. 
4806
-
4814
)
7
Cardone
L.
Carlucci
A.
Affaitati
A.
Livigni
A.
DeCristofaro
T.
Garbi
C.
Varrone
S.
Ullrich
A.
Gottesman
M.E.
Avvedimento
E.V.
Feliciello
A.
Mitochondrial AKAP121 binds and targets protein tyrosine phosphatase D1, a novel positive regulator of src signaling
Mol. Cell. Biol.
 , 
2004
, vol. 
24
 (pg. 
4613
-
4626
)
8
Ceulemans
H.
Bollen
M.
Functional diversity of protein phosphatase-1, a cellular economizer and reset button
Physiol. Rev.
 , 
2004
, vol. 
84
 (pg. 
1
-
39
)
9
Mermoud
J.E.
Cohen
P.
Lamond
A.I.
Ser/Thr-specific protein phosphatases are required for both catalytic steps of pre-mRNA splicing
Nucleic Acids Res.
 , 
1992
, vol. 
20
 (pg. 
5263
-
5269
)
10
Boudrez
A.
Beullens
M.
Groenen
P.
Van
E.A.
Vulsteke
V.
Jagiello
I.
Murray
M.
Krainer
A.R.
Stalmans
W.
Bollen
M.
NIPP1-mediated interaction of protein phosphatase-1 with CDC5L, a regulator of pre-mRNA splicing and mitotic entry
J. Biol. Chem.
 , 
2000
, vol. 
275
 (pg. 
25411
-
25417
)
11
Novoyatleva
T.
Heinrich
B.
Tang
Y.
Benderska
N.
Butchbach
M.E.
Lorson
C.L.
Lorson
M.A.
Ben-Dov
C.
Fehlbaum
P.
Bracco
L.
, et al.  . 
Protein phosphatase 1 binds to the RNA recognition motif of several splicing factors and regulates alternative pre-mRNA processing
Hum. Mol. Genet.
 , 
2008
, vol. 
17
 (pg. 
52
-
70
)
12
Rubin
E.
Mittnacht
S.
Villa-Moruzzi
E.
Ludlow
J.W.
Site-specific and temporally-regulated retinoblastoma protein dephosphorylation by protein phosphatase type 1
Oncogene
 , 
2001
, vol. 
20
 (pg. 
3776
-
3785
)
13
Cohen
P.T.
Protein phosphatase 1—targeted in many directions
J. Cell Sci.
 , 
2002
, vol. 
115
 (pg. 
241
-
256
)
14
Steen
R.L.
Beullens
M.
Landsverk
H.B.
Bollen
M.
Collas
P.
AKAP149 is a novel PP1 specifier required to maintain nuclear envelope integrity in G1 phase
J. Cell Sci.
 , 
2003
, vol. 
116
 (pg. 
2237
-
2246
)
15
Ginsberg
M.D.
Feliciello
A.
Jones
J.K.
Avvedimento
E.V.
Gottesman
M.E.
PKA-dependent binding of mRNA to the mitochondrial AKAP121 protein
J. Mol. Biol.
 , 
2003
, vol. 
327
 (pg. 
885
-
897
)
16
Ranganathan
G.
Phan
D.
Pokrovskaya
I.D.
McEwen
J.E.
Li
C.
Kern
P.A.
The translational regulation of lipoprotein lipase by epinephrine involves an RNA binding complex including the catalytic subunit of protein kinase A
J. Biol. Chem.
 , 
2002
, vol. 
277
 (pg. 
43281
-
43287
)
17
Musunuru
K.
Darnell
R.B.
Determination and augmentation of RNA sequence specificity of the Nova K-homology domains
Nucleic Acids Res.
 , 
2004
, vol. 
32
 (pg. 
4852
-
4861
)
18
Valverde
R.
Edwards
L.
Regan
L.
Structure and function of KH domains
FEBS J.
 , 
2008
, vol. 
275
 (pg. 
2712
-
2726
)
19
Rogne
M.
Landsverk
H.B.
Van
E.A.
Beullens
M.
Bollen
M.
Collas
P.
Kuntziger
T.
The KH-Tudor domain of a-kinase anchoring protein 149 mediates RNA-dependent self-association
Biochemistry
 , 
2006
, vol. 
45
 (pg. 
14980
-
14989
)
20
Llorian
M.
Beullens
M.
Andres
I.
Ortiz
J.M.
Bollen
M.
SIPP1, a novel pre-mRNA splicing factor and interactor of protein phosphatase-1
Biochem. J.
 , 
2004
, vol. 
378
 (pg. 
229
-
238
)
21
Trinkle-Mulcahy
L.
Ajuh
P.
Prescott
A.
Claverie-Martin
F.
Cohen
S.
Lamond
A.I.
Cohen
P.
Nuclear organisation of NIPP1, a regulatory subunit of protein phosphatase 1 that associates with pre-mRNA splicing factors
J. Cell Sci.
 , 
1999
, vol. 
112
 (pg. 
157
-
168
)
22
Kwiek
N.C.
Thacker
D.F.
Datto
M.B.
Megosh
H.B.
Haystead
T.A.
PITK, a PP1 targeting subunit that modulates the phosphorylation of the transcriptional regulator hnRNP K
Cell Signal.
 , 
2006
, vol. 
18
 (pg. 
1769
-
1778
)
23
Huang
L.J.
Wang
L.
Ma
Y.
Durick
K.
Perkins
G.
Deerinck
T.J.
Ellisman
M.H.
Taylor
S.S.
NH2-terminal targeting motifs direct dual specificity A-kinase-anchoring protein 1 (D-AKAP1) to either mitochondria or endoplasmic reticulum
J. Cell Biol.
 , 
1999
, vol. 
145
 (pg. 
951
-
959
)
24
Valgardsdottir
R.
Ottersen
O.P.
Prydz
H.
Regulated compartmentalization of the putative DEAD-box helicase MDDX28 within the mitochondria in COS-1 cells
Exp. Cell Res.
 , 
2004
, vol. 
299
 (pg. 
294
-
302
)
25
Boldogh
I.R.
Pon
L.A.
Interactions of mitochondria with the actin cytoskeleton
Biochim. Biophys. Acta.
 , 
2006
, vol. 
1763
 (pg. 
450
-
462
)
26
Cardone
L.
de
C.T.
Affaitati
A.
Garbi
C.
Ginsberg
M.D.
Saviano
M.
Varrone
S.
Rubin
C.S.
Gottesman
M.E.
Avvedimento
E.V.
Feliciello
A.
A-kinase anchor protein 84/121 are targeted to mitochondria and mitotic spindles by overlapping amino-terminal motifs
J. Mol. Biol.
 , 
2002
, vol. 
320
 (pg. 
663
-
675
)
27
Katayama
H.
Zhou
H.
Li
Q.
Tatsuka
M.
Sen
S.
Interaction and feedback regulation between STK15/BTAK/Aurora-A kinase and protein phosphatase 1 through mitotic cell division cycle
J. Biol. Chem.
 , 
2001
, vol. 
276
 (pg. 
46219
-
46224
)
28
Alto
N.M.
Soderling
J.
Scott
J.D.
Rab32 is an A-kinase anchoring protein and participates in mitochondrial dynamics
J. Cell Biol.
 , 
2002
, vol. 
158
 (pg. 
659
-
668
)
29
Westermann
B.
Molecular machinery of mitochondrial fusion and fission
J. Biol. Chem.
 , 
2008
, vol. 
283
 (pg. 
13501
-
13505
)
30
Sylvestre
J.
Margeot
A.
Jacq
C.
Dujardin
G.
Corral-Debrinski
M.
The role of the 3′ untranslated region in mRNA sorting to the vicinity of mitochondria is conserved from yeast to human cells
Mol. Biol. Cell.
 , 
2003
, vol. 
14
 (pg. 
3848
-
3856
)
31
Corral-Debrinski
M.
mRNA specific subcellular localization represents a crucial step for fine-tuning of gene expression in mammalian cells
Biochim. Biophys. Acta.
 , 
2007
, vol. 
1773
 (pg. 
473
-
475
)
32
DiMauro
S.
Hirano
M.
Mitochondrial encephalomyopathies: an update
Neuromuscul. Disord.
 , 
2005
, vol. 
15
 (pg. 
276
-
286
)
33
Schuman
E.M.
Dynes
J.L.
Steward
O.
Synaptic regulation of translation of dendritic mRNAs
J. Neurosci.
 , 
2006
, vol. 
26
 (pg. 
7143
-
7146
)
34
Bassell
G.J.
Kelic
S.
Binding proteins for mRNA localization and local translation, and their dysfunction in genetic neurological disease
Curr. Opin. Neurobiol.
 , 
2004
, vol. 
14
 (pg. 
574
-
581
)
35
Collas
P.
Le Guellec
K.
Tasken
K.
The A-kinase-anchoring protein AKAP95 is a multivalent protein with a key role in chromatin condensation at mitosis
J. Cell Biol.
 , 
1999
, vol. 
147
 (pg. 
1167
-
1180
)
36
Stenlund
P.
Frostell-Karlsson
A.
Karlsson
O.P.
Studies of small molecule interactions with protein phosphatases using biosensor technology
Anal. Biochem.
 , 
2006
, vol. 
353
 (pg. 
217
-
225
)