PKA‐binding domain of AKAP8 is essential for direct interaction with DPY30 protein

The main function of the A kinase‐anchoring proteins (AKAPs) is to target the cyclic AMP‐dependent protein kinase A (PKA) to its cellular substrates through the interaction with its regulatory subunits. Besides anchoring of PKA, AKAP8 participates in regulating the histone H3 lysine 4 (H3K4) histone methyltransferase (HMT) complexes. It is also involved in DNA replication, apoptosis, transcriptional silencing of rRNA genes, alternative splicing, and chromatin condensation during mitosis. In this study, we focused on the interaction between AKAP8 and the core subunit of all known H3K4 HMT complexes—DPY30 protein. Here, we demonstrate that the PKA‐binding domain of AKAP8 and the C‐terminal domain of DPY30, also called Dpy‐30 motif, are crucial for the interaction between these proteins. We show that a single amino acid substitution in DPY30 L69D affects its dimerization and completely abolishes its interaction with AKAP8 and another DPY30‐binding partner brefeldin A‐inhibited guanine nucleotide‐exchange protein 1 (BIG1), which is also AKAP domain‐containing protein. We further demonstrate that AKAP8 interacts with DPY30 and the RII alpha regulatory subunit of PKA both in the interphase and in mitotic cells, and we show evidences that AKAP8L, a homologue of AKAP8, interacts with core subunits of the H3K4 HMT complexes, which suggests its role as a potential regulator of these complexes. The results presented here reinforce the analogy between AKAP8–RII alpha and AKAP8–DPY30 interactions, postulated before, and improve our understanding of the complexity of the cellular functions of the AKAP8 protein.


Introduction
A kinase-anchoring proteins (AKAPs) play a regulatory role in cyclic AMP (cAMP) signaling by subcellular compartmentalization of cAMP-dependent protein kinase A (PKA) and phosphatases with appropriate substrates within cells. The members of the AKAP family are structurally diverse but functionally related, and the ability to anchor the PKA holoenzyme via the interaction with PKA regulatory subunits is their common feature. In mammalian cells, there are four genes encoding regulatory subunits, such as RIa, RIb, RIIa, and RIIb, and four genes encoding catalytic subunits, such as Ca, Cb, Cc, and PrKX [1][2][3][4]. Upon the elevation of the second messenger cAMP level, cAMP molecules bind to the R subunit which leads to the dissociation of the PKA holoenzyme and release of the C subunits that phosphorylate nearby substrates. Most AKAPs also have scaffolding functions. They interact with other signaling molecules, and by organizing multivalent complexes they provide the integration and dissemination of information within the cell [1][2][3].
The AKAP family comprises over 50 members identified so far, most of which reside in the cytoplasm. They are anchored to defined compartments, such as the plasma membrane, organelles and specific sites in the cytosol where they create focal points for signal transduction [5][6][7][8][9][10]. Up to now, only two AKAPs are found inside the nucleus, that is, AKAP8 and a splicing factor SFRS17A (also known as AKAP17A), binding type II and both type I and type II PKA regulatory subunits, respectively [11][12][13]. Besides, in the nucleus, the homologue of AKAP8, referred to as A-kinase anchoring protein 8 like (AKAP8L) or HA95, is present [14][15][16]. However, AKAP8L, unlike AKAP8, does not bind the R subunit, instead it has been demonstrated to bind PKA catalytic subunit [17]. The cAMP/PKA signaling in the nucleus plays an important role in regulating transcription [18] and RNA splicing [17]. It is accepted that upon activation of the cytosolic PKA holoenzyme induced by cAMP generated by transmembrane adenylyl cyclase (tmAC), a proportion of the C subunit translocates to the nucleus via diffusion. In the nucleus, the catalytic PKA subunit phosphorylates a group of cAMPresponsive nuclear factors to regulate transcription. However, recent studies also demonstrated the presence of a functional pool of PKA holoenzyme in the nucleus involved in the cAMP signaling generated by the soluble nuclear adenylyl cyclase [19,20].
It is currently unknown whether and how nuclear AKAPs participate in the cAMP/PKA signaling. SFRS17A targets PKA to the SFCs to regulate pre-mRNA splicing; however, the precise mechanism of its action remains unknown [11]. AKAP8 is proposed to interact with RIIa subunits at the onset of mitosis, that is, after the breakdown of the nuclear envelope [21]. RIIa subunit is phosphorylated by CDK1 and has been shown to interact with AKAP8 during mitosis; however, considering the controversy about AKAP8 localization on mitotic chromatin, the physiological function of this interaction remains elusive [21]. During interphase, the RIIa subunit is thought to be excluded from the nucleus and to be localized mainly in the centrosome-Golgi area [22,23]. On the contrary, AKAP8 is localized mainly in the nucleus during interphase, however, several lines of evidence suggest the cytoplasmic localization of AKAP8 in a complex with p105 and PKA holoenzyme that is presumably engaged in p105 phosphorylation [24].
AKAP8 interaction with numerous proteins and its scaffolding functions have been revealed in many studies. The interaction with p68 RNA helicase [25] and minichromosome maintenance 2 protein (MCM2) [26] implies its role as a scaffold for assembly of nuclear complexes involved in transcription and replication. AKAP8's possible involvement in modulation of rDNA transcription and alternative splicing of pre-mRNA is exhibited in latest studies [27,28]. It also associates with KMT2b histone methyltransferase (HMT) complex via direct binding to DPY30 protein and regulates the H3K4 methyltransferase activity of this complex [29]. AKAP8 also interacts with D-and E-type cyclins and helps to deliver cyclin D/E to CDK4 [30,31]. Moreover, AKAP8 interacts with AMY-1, a c-Myc-binding protein [32], active caspase 3 CASP3 [33], and connexin 43 [34] to transport and/or retain them in the nucleus. Furthermore, in prometaphase, AKAP8 recruits histone deacetylase 3 (HDAC3) onto chromatin, which leads to deacetylation and subsequent phosphorylation at Ser-10 of histone H3 by Aurora B [35].
The interaction between most AKAPs and PKA regulatory subunits is usually mediated by hydrophobic interface of a conserved amphipathic helix within AKAPs and the N-terminal four-helix bundle in the R subunit dimmer [36][37][38][39]. The PKA-binding domain is located at the C terminus of AKAP8 and interacts with RII alpha subunit in vitro. Other regions of AKAP8 are responsible for its interaction with diverse proteins, DNA, and RNA [25][26][27][28].
In this study, we investigated the interaction of AKAP8 with DPY30 protein, a core subunit of all known H3K4 HMT complexes in yeast and animal cells. The interaction of AKAP8 with DPY30 has been previously shown to enable the association of AKAP8 with MLL2 (KMT2b) HMT complex to regulate histone methylation and gene expression. It is particularly involved in gene induction during ESC-fate transitions [40]. DPY30 is a small (99 aa) protein, important for efficient H3K4 methylation, existing in HMT complexes as a multimer dependent on the complex context [41]. Its dimerization domain is localized at the C-terminus and adopts a canonical four-helix-bundle conformation showing high structural homology to the dimerization/docking (D/D) domain of PKA RIIa subunit [39,42]. Here, we demonstrate that AKAP8 binds to DPY30 through PKA binding domain and that impairment of DPY30 dimerization abolishes its binding to AKAP8. We have also established that the interaction of AKAP8 with RII alpha subunit and DPY30 occurs during both interphase and mitosis. Moreover, we show that AKAP8L, a homologue of AKAP8, associates with all core subunits of H3K4 HMT complexes, which suggests its role as a potential modulator of H3K4 methylation.

AKAP8 and DPY30 interact via their C-terminal domains
Recent data have demonstrated that AKAP8 binds directly to DPY30 protein and, via this interaction, it associates with MLL2 HMT complex and regulates its activity [29]. In order to map domains responsible for interaction between AKAP8 and DPY30 we prepared a series of truncated AKAP8 constructs representing different functional domains of the AKAP8 protein and analyzed their interaction with DPY30 using co-immunoprecipitation (co-IP; Fig. 1A). AKAP8 contains the PKA-binding domain located at the C terminus, while other regions located toward the N-terminus of the AKAP8 protein are responsible for its interaction with diverse proteins, DNA, and RNA. Deletion mutants prepared contained either AKAP8 N-terminal regions of different length devoid of the PKA-binding domain or the C-terminal regions with the PKA-binding domain (Fig. 1A). The truncated AKAP8 constructs, fused EGFP tag, were coexpressed with full length mRFP-DPY30 in human embryonic kidney 293 cell line (HEK293T) cells and purified on GFP-Trap beads. Instead of EGFP-AKAP8, EGFP tag was expressed in control samples. All truncated fragments were expressed with high efficiency and obtained input signals were comparable. As shown in Fig. 1B, all deletion fragments of AKAP8 protein containing the PKA-binding domain (residues 568-692, 477-692, 386-692, and 284-692) as well as the full-length AKAP8 protein co-precipitated with mRFP-DPY30 efficiently. The C-terminal region of AKAP8 encompassing amino acids 568-692 was sufficient to interact with DPY30 in cell. This fragment contained only the domain responsible for anchoring the PKA RII alpha subunit. In contrast, AKAP8 deletion mutants lacking that domain-AKAP8 , AKAP8(1-387), and AKAP8(1-568)-did not copurify with mRFP-DPY30 (Fig. 1B), thus indicating an essential role of the PKAbinding domain of AKAP8 in the interaction between AKAP8 and DPY30.
In order to further test the importance of AKAP8 PKA-binding domain for the interaction with DPY30, we generated a mutated version of AKAP8 by substituting isoleucine 582 with proline. The I582P substitution has been determined by other authors to abolish AKAP8 binding to PKA RII alpha subunit [43]. We reanalyzed the interactions of the native AKAP8 and the AKAP8 I582P point mutant with DPY30 by co-IP (Fig. 1C). EGFP-AKAP8 or EGFP-AKAP8 I582P was coexpressed with mRFP-DPY30 and their interaction was analyzed by co-IP using GFP-Trap. In control sample, EGFP with mRFP-DPY30 was coexpressed. Despite comparable expression of EGFP-AKAP8 or EGFP-AKAP8 I582P as well as mRFP-DPY30 in analyzed samples, mutation in the PKA-binding domain completely abolished interaction with DPY30 (Fig. 1C). To support the results obtained in co-IP analysis we also performed Far-western blotting (WB) that enables detection of direct interactions between two proteins (Fig. 1D). EGFP-AKAP I582P, EGFP-AKAP8, and, for positive control, EGFPabsent, small, or homeotic 2-like protein (ASH2L) were ectopically expressed in HEK293 cells, separated by PAGE and transferred to membrane. The membrane was incubated with bacterially expressed DPY30 protein fused with His-tag or with His-tag alone followed by immunodetection using Ab recognizing Histag. The results of Far-Western analysis confirm the results obtained in co-IP tests and support the crucial role of the PKA-binding domain in the interaction between AKAP8 and DPY30. Additionally, to further confirm the results of Far-WB, pull-down assay was performed. Glutathione S-transferase (GST)-pull-down of bacterially expressed GST-AKAP8 and GST-AKAP I582P incubated with His-tagged DPY30 showed that AKAP I582P does not interact with DPY30 protein (Fig. 1E).
We next analyzed colocalization of DPY30 with AKAP8 and AKAP8(1-568) deletion mutant devoid of PKA/DPY30-binding domain. EGFP-AKAP8 and EGFP-AKAP8(1-568) deletion mutant were coexpressed with mCherry-DPY30 in HeLa cells. For  control, in order to determine whether overexpression of DPY30 influences AKAP8 localization, the localization of EGFP-AKAP8 and EGFP-AKAP8(1-568) without mCherry-DPY30 overexpression was also analyzed (Fig. 2). Analysis of the subcellular localization of the AKAP8 confirmed its presence in the nucleus, while DPY30 was located in the nucleus and in the cytoplasm. The ectopic expression of DPY30 did not change AKAP8 localization (Fig 2A,C). Contrary with AKAP8, the truncated AKAP8 mutant lacking the PKA/DPY30-binding domain was preferentially localized in the nucleoli, where it did not colocalize with

The role of the Dpy-30 motif in the interaction with AKAP8
In a further analysis, we asked which region of the DPY30 protein is responsible for the interaction with AKAP8. DPY30 protein contains at the C-terminus (residues 49-99) the structural Dpy-30 motif (pfam 05186) important for DPY30 dimerization. Structurally, this motif is very similar to the regulatory subunit of PKA (RIIa; pfam 02197) [39,42]. The N-terminal part of DPY30 does not contain any known functional domain, and the crystallographic analyses have shown that this region is highly disordered [42]. To map the DPY30 interaction domain engaged in AKAP8 binding we generated truncated DPY30 protein comprising either N-terminal (residues 1-49) or C-terminal (residues 49-99) regions fused with tandem affinity purification tag (TAP tag) composed of Protein A and CBP. The DPY30 fragments were coexpressed with full-length EGFP-AKAP8 in HEK293 cells and were tested for binding EGFP-AKAP8 using GFP-Trap co-IP assay. The result of this test showed that the C-terminal region of DPY30 (residues 49-99) encompassing Dpy-30 structural motif is sufficient for binding AKAP8 protein (Fig. 3A). All of the DPY30 protein direct interactions identified so far have been through the Dpy-30 motif. The crystallographic studies have shown that the two protomers of DPY30 form through the Dpy-30 motif a typical X-type four-helix bundle in an antiparallel arrangement [42]. The Dpy-30 motif is largely hydrophobic and comprises four shallow hydrophobic pockets defining the binding site that accommodates amphipathic a helix present in ASH2L and probably in other DPY30-binding proteins, that is, brefeldin Ainhibited guanine nucleotide-exchange protein (BIG1) and BPTF-associated protein of 18 kDa (BAP18) [44]. Crystallographic data and a detailed overlay assay between ASH2L point mutants and DPY30 revealed the presence of the DPY30-binding motif (DBM) in the DPY30 interacting proteins. The DBM consists of , where φ and w represent any hydrophobic and nonhydrophobic residues, respectively, that form the amphipathic helix [44]. To identify the DBM motif in AKAP8 protein, we performed a sequence alignment of the PKA/DPY30-binding domain in AKAP8 with the sequences of the DBM motif present in other proteins directly interacting with DPY30 (Fig. 3B). Our results show that within the domain responsible for the interaction with DPY30, the sequence of AKAP8 is highly similar to the consensus sequence of the DBM in ASH2L, BIG1, and BAP18. The only disagreement is alanine in position 585; however, it apparently does not impair the interaction between AKAP8 and DPY30.
The studies of Tremblay and colleagues have shown that Leu69 in DPY30 is a central residue shaping all hydrophobic pockets involved in the interaction of DPY30 with amphipathic helix of ASH2L and a single L69D substitution completely abolished interaction between these two proteins [44]. Based on this information, we performed pull-down and co-IP tests to verify whether DPY30 L69D point mutation influences the interaction between AKAP8 and DPY30 ( Fig. 3C-D). GST-pull-down of bacterially expressed GSTtagged AKAP8 and ASH2L incubated with His-tagged DPY30 and DPY30 L69D showed that AKAP8, likewise ASH2L, interacted only with wild-type DPY30 (Fig. 3C). The L69D point mutation of DPY30 completely abolished DPY30 interaction both with ASH2L and AKAP8. The in vitro results were confirmed in co-IP experiment that also aimed at the evaluation of the influence of the L69D point mutation on the dimerization ability of the DPY30 protein. mCherry-DPY30 L69D used as a bait protein coexpressed with GFP-DPY30 L69D in HEK293T cells did not copurify with endogenous AKAP8 and BIG1. BIG1 is a DPY30binding proteins that similar to the AKAP8 protein, contains AKAP sequence. In contrast, endogenous AKAP8 and BIG1 were copurified when mCherry-DPY30 L69D was coexpressed with wild-type GFP-DPY30 or when both wild-type DPY30 proteins fused with mCherry and EGFP were coexpressed (Fig. 3D). The obtained results demonstrate similarity of the interaction of DPY30 protein with AKAP8 and BIG1. Moreover, the results suggest that L69D substitution influences dimer formation by DPY30. When this mutation is present in both protomers it significantly decreases the dimerization ability of DPY30 (Fig. 3D).

Interaction of AKAP8 with the DPY30 protein and PKA RIIa subunit during cell cycle
To determine whether the interaction of AKAP8 with DPY30 is cell cycle specific we purified proteins interacting with ectopically expressed EGFP-AKAP8 in HEK293T cells unsynchronized and synchronized in mitosis. For control, HEK293T cell line with ectopic EGFP expression was used in this experiment. Cell synchronization in mitosis was determined by the level of H3S10 phosphorylation. Immunoprecipitation on GFP-Trap beads followed by WB analysis and immunodetection of the endogenous DPY30 showed that AKAP8 interacts with DPY30 in unsynchronized cells and cells synchronized in mitosis (Fig. 4A). Further experiments were done to establish whether the formation of AKAP8 complexes with RII alpha subunit depends on the cell cycle. HEK293T cells with ectopic expression of EGFP-AKAP8 or EGFP were synchronized in G1 phase, S phase, and mitosis. Cell synchronization was monitored by flow cytometry analysis of cellular DNA stained with propidium iodide (PI). In addition, the level of H3S10 phosphorylation was determined by WB. Immunoprecipitation on GFP-Trap beads followed by WB analysis and immunodetection of the endogenous RII alpha (PRKAR2a) showed that RII alpha subunit interacts with EGFP-AKAP8 in G1 phase, S phase and at mitosis (Fig. 4B,C).
To date, the interaction between AKAP8 and DPY30 has been shown to occur in the interphase cells and to be involved in regulation of H3K4 HMT complexes. However, the co-IP tests clearly indicate that AKAP8 exists in the complex with DPY30 protein also in mitotic cells. Microscopy analysis of AKAP8 and DPY30 localization in mitotic cells we performed revealed that during prometaphase, both proteins are enriched and colocalize in the area around condensing chromosomes. At metaphase, both AKAP8 and DPY30 are enriched around the metaphase plate in the area overlapping with spindle MTs, and in anaphase they show the localization enrichment at the spindle midzone (Fig. 5).
Both AKAP8 and its homologue AKAP8L associate with core subunits of the H3K4 HMT complexes We isolated DPY30-interacting proteins using tandem affinity purification method from HEK293 cells stably transfected with NTAP-DPY30, in order to better understand the cellular functions of DPY30 (data not shown). Consistent with the results obtained by other authors, we identified the presence of AKAP8 among DPY30-interacting proteins [29]. We also found that not only AKAP8, but also its homologue AKAP8L devoid of the domain responsible for binding PKA regulatory subunits copurified with DPY30. These results were confirmed by co-IP experiment in which EGFP-AKAP8L was coexpressed with mRFP-DPY30. In parallel, EGFP-AKAP8 was used as a positive control of interaction with DPY30 (Fig. 6A). The signal observed for DPY30 was much weaker in AKAP8L co-IP despite high levels of expression of AKAP8 and AKAP8L. Next, we examined whether AKAP8L like AKAP8 directly interacts with DPY30. We performed the GST-pull-down assay in which we used bacterially overexpressed GST-DPY30 protein and His-tagged AKAP8 and AKAP8L proteins. The pull-down test revealed that in contrast with AKAP8, AKAP8L does not interact directly with DPY30 (Fig. 6B). In support of the pull-down interaction data, the interaction between DPY30 and AKAP8L was also not detected in Far-WB experiment (Fig. 1D). Jiang and colleagues demonstrated that AKAP8 via interaction with DPY30 associates with MLL2 complex and regulates its H3K4 HMT activity [29]. Given that AKAP8L and AKAP8 have been shown to coexist in diverse complexes [35], we found it reasonable to examine whether AKAP8L can like AKAP8, be associated with H3K4 HMT complexes. To test this, we performed GFP co-IP of ectopically expressed EGFP-AKAP8 and EGFP-AKAP8L in HEK293 cells. In positive control sample DPY30 was overexpressed. In the precipitates, we analyzed the presence of endogenous WD repeat domain 5 (WDR5), retinoblastoma-binding protein 5 (RbBP5), ASH2L, and DPY30 that form WRAD subcomplex being the core module of all H3K4 HMT complexes. When EGFP-AKAP8 and EGFP-DPY30 were used as bait, all WRAD subunits copurified efficiently (Fig. 6A). RbBP5, WDR5, and ASH2L were detected less efficiently in AKAP8L co-IP as compared to EGFP-AKAP8 co-IP (Fig. 6C). Moreover, we did not detect endogenous DPY30 in the Co-IP with EGFP-AKAP8L. The results obtained indicate that AKAP8L copurifies with the core subunits of the H3K4 HMT complexes, however, with lower efficiency than AKAP8.
We also performed co-IP of EGFP-AKAP8 I582P mutant that does not interact with DPY30 (Fig. 1C) to determine its ability to associate with the WRAD module. We found that AKAP8 I582P mutant despite its incapability to interact with DPY30 still preserves the ability to bind other subunits of the WRAD complex (Fig. 6C), which indicates that AKAP8 associates with H3K4 HMT complexes not only via direct binding to DPY30, but also via interaction with another subunit.
To determine which WRAD subunits interact directly with AKAP8 and AKAP8L we performed a series of GST pull-down assays using bacterially expressed and purified GST-fused WDR5, RbBP5, ASH2L, and DPY30 as bait proteins (Fig. 7A). In negative control samples GST protein was used. His-AKAP8 and His-AKAP8L were used as prey proteins. His-AKAP8 was bacterially expressed and purified. Due to the low stability an toxicity of AKAP8L in bacterial expression systems, this protein was produced in an in vitro transcription/translation system. The results obtained in pull-down assays confirmed AKAP8 direct interaction with DPY30. Moreover, the signal with much lower intensity was detected in the pull-down assay of the interaction of AKAP8 with WDR5, which suggests a weak direct interaction between these proteins. Contrary to AKAP8, AKAP8L does not interact with DPY30. Instead, the low intensity signals were detected in pull-down assays of the interaction of AKAP8L with WDR5 and ASH2L (Fig. 7A). To further confirm the results obtained additional GST pull-down assays were performed, in which GFP-AKAP8 and GFP-AKAP8L overexpressed in HEK293T cells, were used as prey proteins (Fig. 7B). AKAP8 and AKAP8L were isolated in stringent conditions to destroy protein complexes and incubated with bacterially expressed GST-fused WDR5, RbBP5, ASH2L, and DPY30 proteins immobilized on glutathione-agarose beads. Similar to the aforementioned GST pull-down assay, the results obtained in this experiment demonstrate that only AKAP8 and not AKAP8L interact with DPY30. The signal corresponding to the AKAP8 protein bound to WDR5 was also present, although with lower intensity compared to the signal indicating AKAP8 interaction with DPY30. In GST pull-down assays demonstrating AKAP8L interactions, the signals corresponding to AKAP8L bound to WDR5, RbBP5, and ASH2L were detected, however, their intensity was very low and may suggest nonspecific binding or very weak interactions (Fig. 7B).

Discussion
The specificity in cAMP signal transduction is provided by A Kinase Anchoring Proteins (AKAPs) that enable the assembly of signaling complexes at distinct intracellular sites and placing PKA close to specific substrates and effectors. Thus far, only two AKAPs are found in the nucleus, that is, AKAP8 and a splicing factor SFRS17A (AKAP17A) [11]. AKAP8 was shown to bind the PKA RII alpha regulatory subunit in vitro, and it serves as a scaffold by interaction with many proteins enabling an assembly of nuclear complexes involved in transcription [25], alternative splicing of pre-mRNA [28], and replication [26]. Recent data indicate that AKAP8 associates with KMT2b (MLL2) HMT complex to regulate its H3K4 methyltransferase activity and gene expression in ESC differentiation. The AKAP8 protein interaction with MLL2 complex is shown to occur via direct binding to DPY30 protein. The N-terminal part of AKAP8 has Due to the unspecific recognition of GST-RbBP5 and GST-ASH2L bait proteins by anti-His antibody, the proteins purified in GST pull-down assays were separated using 6% PAGE to obtain best separation of bait and prey proteins. The nonspecific signal corresponding to GST-ASH2L bait protein recognized by anti-His antibody has been marked with an asterisk. (B) Bacterially expressed GST-fused WRAD subunits were purified on Glutathione Sepharose 4 Fast Flow beads and incubated with His-AKAP8 and His-AKAP8L overexpressed in HEK293T cells and isolated in stringent conditions to destroy protein complexes. GST protein was used in the negative control sample. The experiments were repeated at least two times.

958
The been shown to be responsible for DPY30 binding [29].
The results presented in this paper do not confirm the data published by Jiang and colleagues [29]. The detailed mapping of interaction domains clearly demonstrates that AKAP8 interacts with DPY30 via the C-terminal part comprising the PKA-binding domain. We have shown that the AKAP8 I582P point mutation abolishing its binding to PKA RII alpha also abolishes DPY30 binding. We have mapped the interaction domain in DPY30 to the Dpy30 motif. This motif shows structural homology to the RII alpha dimerization motif and belongs to the same clan (CL0068). Thus, we conclude that the interaction between AKAP8 and DPY30 shows analogy to the complex composed of AKAP and PKA regulatory subunit. The DPY30 crystallography studies have demonstrated that the Dpy30 motif, similar to dimerization motif of the PKA R subunit, is responsible for dimerization and forms X-type, four-helix bundles via hydrophobic interactions [42]. The dimer interface, in both DPY30 and the PKA R subunits, is largely hydrophobic and enables interaction with hydrophobic interface of amphipathic alpha helices present in proteins binding to DPY30 and PKA regulatory subunits, respectively. Thus far the crystal structure of DPY30 in the complex with ASH2L is resolved. These studies demonstrated that amphipathic alpha helix located at the C-terminus of ASH2L binds to the Dpy-30 motif, and subsequent overlay assays revealed that the Dpy-30 motif can accommodate different amphipathic alpha helices [44]. Based on the detailed overlay assays, the authors defined a consensus sequence of the DBM in DPY30-binding proteins. The sequence alignment of the PKA/DPY30-binding domain in AKAP8 with the consensus sequence of the DBMs in the known DPY30-interacting proteins showed high similarity of DPY30-binding sequence present in AKAP8 to the DBMs in ASH2L, BIG1, and BAP18 identified by other authors. Tremblay and colleagues revealed that in the interaction between DPY30 and ASH2L, the DPY30 Leu69 plays a key role since it shapes all four pockets defining the ASH2L-binding site in DPY30. The current study shows that Leu69 is also crucial in DPY30 binding to AKAP8, because a single L69D substitution completely abolished the DPY30 interaction with AKAP8. In a similar manner, this substitution also abolished the DPY30 interaction with BIG1 (also known as ARFGEF1) that contains AKAP sequence and binds protein kinase A [45]. Nevertheless, this experiment shows that DPY30 L69D mutation also influences the dimerization of DPY30, which has been neglected in Tremblay et al.'s studies [44]. The impaired binding of DPY30 to AKAP8, ASH2L and BIG1 may, therefore, also result from the DPY30 L69D dimerization defects. Current data reveal that the interaction between AKAP8 and DPY30 enables AKAP8 to associate with MLL2 complex to regulate its activity [29]. However, our results show that AKAP8 I582P, unable to interact with DPY30, still associates with the core subunits of the H3K4 HMT complexes. It demonstrates that DPY30 is not the only anchor for AKAP8 in MLL2 complex and that another subunit must also participate in AKAP8 binding with this complex. The results of pull-down assays, aimed at the identification of WRAD subunits that interact with AKAP8, suggest that WDR5 protein can serve as an additional anchor for AKAP8 association with H3K4 HMT-containing complexes. Moreover, we have shown that not only AKAP8, but its homologue AKAP8L also associates with the core subunits of H3K4 HMT complexes. AKAP8L copurifies with the WRAD module with lower efficiency than AKAP8 and, although it copurifies with DPY30 from the cell extracts, it does not interact with DPY30 in in vitro bacterial assays that indicates that it does not bind DPY30 directly. AKAP8L is also shown to associate with the AKAP8 protein [28,35], but it does not bind to AKAP8 directly [14]. Therefore, we suspect that AKAP8L apparently associates with the WRAD module via interacting with proteins other than AKAP8 and DPY30. In pull-down assays, AKAP8L associates very weakly with WRAD subunits except for DPY30. However, due to the low intensity of the signals obtained, the results can suggest the unspecific binding or very weak interaction. Therefore, we suspect that AKAP8L associates with H3K4 HMT-containing complexes via subunits other than WRAD components. Human cells contain at least six complexes with H3K4 methyltransferase activity that are distinguished by six different methyltransferases (Set1a, Set1b, MLL1, MLL2, MLL3, and MLL4) [46]. Moreover, certain H3K4 HMT complexes in human cells contain unique components, such as e.g. menin, ASC2 and HCFs that are responsible for the recruitment of the complexes to target sites in the genome [47]. Thus, in order to identify the subunit of HMT complexes involved in AKAP8L protein binding, further experiments should be performed where the interaction of AKAP8L with other subunits of these complexes should be investigated. AKAP8 and AKAP8L have been shown to copurify from both interphase and mitotic cells [28,35]. At the onset of mitosis, AKAP8L and AKAP8 coexist in a multiprotein complex composed also of HDAC3 and N-CoR among others, which is responsible for histone deacetylation and regulates mitosis by modulating Aurora B kinase activity [35]. On the contrary, the complexes formed by AKAP8 and AKAP8L during interphase have not been detected. Both proteins are engaged in pre-mRNA splicing and in the initiation of replication; however, it is not recognized whether they coexist in the same complexes. The association of AKAP8 and AKAP8L with H3K4 HMT complexes implies their role in regulating histone methylation and gene expression; however, further studies must be carried out to determine whether they coexist in the same H3K4 HMT complexes. A substantial work is also needed to establish the role of AKAP8L in H3K4 HMT complexes.
Our results indicate that AKAP8 interacts with DPY30 not only during the interphase but also during mitosis. While AKAP8's role in regulating the activity of H3K4 HMT complexes is well documented, the role of the interaction between the AKAP8 and DPY30 proteins in mitosis remains unknown. Recent studies have shown that the mixed lineage leukemia (MLL) complex subunits play an important and chromatin-independent function during mitosis [48,49]. The depletion of DPY30 and other WRAD components by RNAi caused defects in S-and M-phase progression. However, in the further studies, the role of MLL and WDR5 is studied in further more detail, while the DPY30 and ASH2L function is not addressed. These studies demonstrated that MLLl/ WDR5 complex, by interaction with kinesin family member 2A (Kif2A), regulates chromosome alignment and spindle assembly [49]. Given the mitotic role of the MLL subunits, further studies are needed to establish whether AKAP8 associates with WRAD during mitosis to cooperate with this complex in mitotic functions.
Our microscopic analyzes of DPY30 and AKAP8 localization during mitosis have shown that during prometaphase both proteins are enriched in the area around condensing chromosomes. In metaphase and anaphase, they show the localization enrichment around the metaphase plate and at the spindle midzone, respectively. Our data are in line with recently published data obtained by L opez-Soop and colleagues, who demonstrated the interaction of AKAP8 with translocated promoter region protein (TPR) in mitosis and AKAP8-dependent enrichment of TPR in the spindle microtubule area in metaphase and in the spindle midzone area in anaphase [50]. By interacting with TPR, AKAP8 is suggested to be involved in proper spindle assembly checkpoint (SAC) function. The fact that the localization of DPY30 is similar to that of the AKAP8 and TPR proteins may suggest an association of DPY30 with AKAP8 and TPR mitotic functions. However, further experiments are needed to confirm this.
Our analysis of the interaction between AKAP8 and the PKA RII alpha subunit during a cell cycle revealed that AKAP8 formed complexes with the PKA RII alpha subunit during G1 phase, S phase, and mitosis. The interaction of AKAP8 with RII alpha in interphase cells is controversial. During interphase, AKAP8 is mainly localized in the nucleus where it presumably does not function as an AKAP because of the exclusion of PKA RII alpha subunit from the nucleus [22,23]. Collas and colleagues did not identify this interaction in the interphase cells [51], whereas other authors demonstrate such interaction occurring in the cytoplasm [24]. Several lines of evidence suggest the cytoplasmic localization of AKAP8. Wall and colleagues showed the presence of a small fraction of AKAP8 protein in the cytoplasm in RAW 264.7 macrophages and suggested the interaction of AKAP8 with p105 (also known as Nfkb1) being responsible for retaining AKAP8 in the cytoplasm. The cytoplasmic pool of AKAP8 existing in a complex with p105 and PKA holoenzyme is presumably engaged in p105 phosphorylation [24].
The interaction between AKAP8 and the PKA RII alpha regulatory subunit during mitosis has also been reported previously by other authors. This interaction was suggested to regulate the remodeling chromatin during mitosis [21]. However, the localization of AKAP8 on mitotic chromatin has been questioned recently. Our microscopic analyses did not confirm AKAP8 localization on mitotic chromosomes and support the data recently published by L opez-Soop and colleagues [50]. Thus, the physiological function of the interaction between AKAP8 and the PKA RII alpha regulatory subunit during mitosis remains elusive [21,27,50]. Further studies are also needed to determine the role of the interaction between the AKAP8 protein and the PKA RII alpha subunit during the S phase.

Cell lines and culture condition
HeLa and HEK293T cells were grown in Dulbecco minimum essential medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin and were incubated in humidified atmosphere containing 5% CO 2 at 37°C. To synchronize HEK293T cells at G1 phase, a double-thymidine block was performed. Cells were blocked for 16 h with 2.5 mM thymidine, released for 10 h by washing out the thymidine, and then blocked again with

Plasmids
The coding sequence of DPY30 (NM_03257) was PCR amplified using cDNA isolated from HEK293 cells and was cloned into a pcDNA3 vector (Invitrogen, Waltham, MA, USA). A PCR-amplified DPY30 coding sequence was subcloned into pEGFP-C1 (Clontech, Mountain View, CA, USA), pmRFP-N1 (vector based on pEGFP-N1), mCherry-C1 (vector based on the pEYFP-C1 from Clontech), pGEX-6SP1 (GE Healthcare, Chicago, IL, USA), and pRSET-A (Thermo Fisher Scientific, Waltham, MA, USA) vectors. The site-directed mutagenesis was used to prepare the following constructs: pEGFP-DPY30 L69D, mCherry-DPY30 L69D, and pRSET-A-DPY30 L69D. DPY30 truncation mutants were prepared by PCR and inserted into pcDNA5/FRT/TO/NTAP vector (a kind gift from Bernhard L€ uscher). The pcDNA3-Flag-AKAP8 vector was a kind gift from Hiroyoshi Ariga. pEGFP-AKAP8 construct was generated by cloning of PCR amplified AKAP8 coding sequence from a pcDNA3-Flag-AKAP8 vector into pEGFP-C1, pGEX-6SP1, and pRSET-A. AKAP8 truncation mutants were created by PCR in pEGFP-C1 vector. The pEGFP-AKAP8 and pGEX-6SP1-AKAP8 sequence was altered by site-directed mutagenesis to generate pEGFP-AKAP8 I582P and pGEX-6SP1-AKAP8 I582P, respectively. The coding sequence of human AKAP8L (NM_014371) was PCR-amplified from cDNA isolated from Hela cells and cloned into a pEGFP-C3 and pRSET-A vectors, respectively (Clontech), in order to prepare a pEGFP-AKAP8L and His-AKAP8L expression constructs. The coding sequence of human PKA regulatory subunit II alpha (NM_001321982) was PCR-amplified from cDNA isolated from Hela cells and cloned into pEYFP-C1 GST pull-down assay GST-tag and His-tag fusion proteins were expressed in Escherichia coli (E. coli) BL21-CodonPlus-RP strain or Lemo21(DE3; New England Biolabs). His-AKAP8L was synthesized in vitro using Promega's TnT T7 Quick Coupled Transcription/Translation System without radiolabeling. EGFP-fused AKAP8 and AKAP8L proteins used for pull-down assays were overexpressed in HEK293T cells. at 4°C. The beads were washed five times with PBS containing 1% Triton X-100 and complexes recovered from the beads were resolved by SDS/PAGE followed by WB analysis.
Far-western blotting DPY30-interacting proteins overexpressed in HEK293T cells were separated on SDS-polyacrylamide gels and transferred to a PVDF membrane by WB. Proteins were denatured by incubating the membrane in AC buffer (20 mM Tris pH 7.5, 1 mM EDTA, 0.1 M NaCl, 10% glycerol, 0.1% Tween-20, 2% non-fat milk, 1 mM DTT) containing 6 M guanidine-HCl for 30 min at RT. Subsequently, the membrane was washed with AC buffer containing 3 M guanidine-HCl for 30 min at RT. This step was followed by washing with AC buffer containing 0.1 M and no guanidine-HCl AC buffer at 4°C, for 30 min and overnight respectively. The membrane was blocked for 8 h at 4°C in 5% nonfat milk in the PBST buffer and subsequently incubated with 1 lgÁmL À1 His-DPY30 protein or His-tag alone overnight at 4°C. Unbound proteins were washed off with PBST buffer three times, each for 15 min. Standard immunodetection was performed to detect bait proteins bound to the proteins on the membrane.

Confocal microscopy analysis
Cells were grown on coverslips and transfected using Effectene Qiagen (Hilden, Germany) according to the manufacturer's protocols. At 24 h after transfection, cells were fixed with 4% formaldehyde for 15 min at RT and then permeabilized in À20°C methanol for 5 min on ice. The cells were rehydrated stepwise. All washes between each step were performed in PBS. DNA was stained with Hoechst 33342. Cells were mounted in Fluoroshield TM (Sigma-Aldrich) and subjected to immunofluorescence microscopic analysis using a NIKON A1R confocal microscope. For immunofluorescence analysis, cells fixed with formaldehyde and permeabilized with methanol were rehydrated stepwise and blocked with an Image-iT TM FX signal enhancer (#A-31624; Invitrogen) followed by sequential incubation with primary antibody (anti-AKAP8; 1 : 600, #WH0010270M1; Sigma Aldrich) for 1 h and Alexa FluorÒ 594 Goat Anti-Mouse SFX Kit conjugated secondary antibody (#A-31624; Invitrogen).