Interactome of the Autoimmune Risk Protein ANKRD55

The ankyrin repeat domain-55 (ANKRD55) gene contains intronic single nucleotide polymorphisms (SNPs) associated with risk to contract multiple sclerosis, rheumatoid arthritis or other autoimmune disorders. Risk alleles of these SNPs are associated with higher levels of ANKRD55 in CD4+ T cells. The biological function of ANKRD55 is unknown, but given that ankyrin repeat domains constitute one of the most common protein-protein interaction platforms in nature, it is likely to function in complex with other proteins. Thus, identification of its protein interactomes may provide clues. We identified ANKRD55 interactomes via recombinant overexpression in HEK293 or HeLa cells and mass spectrometry. One hundred forty-eight specifically interacting proteins were found in total protein extracts and 22 in extracts of sucrose gradient-purified nuclei. Bioinformatic analysis suggested that the ANKRD55-protein partners from total protein extracts were related to nucleotide and ATP binding, enriched in nuclear transport terms and associated with cell cycle and RNA, lipid and amino acid metabolism. The enrichment analysis of the ANKRD55-protein partners from nuclear extracts is related to sumoylation, RNA binding, processes associated with cell cycle, RNA transport, nucleotide and ATP binding. The interaction between overexpressed ANKRD55 isoform 001 and endogenous RPS3, the cohesins SMC1A and SMC3, CLTC, PRKDC, VIM, β-tubulin isoforms, and 14-3-3 isoforms were validated by western blot, reverse immunoprecipitaton and/or confocal microscopy. We also identified three phosphorylation sites in ANKRD55, with S436 exhibiting the highest score as likely 14-3-3 binding phosphosite. Our study suggests that ANKRD55 may exert function(s) in the formation or architecture of multiple protein complexes, and is regulated by (de)phosphorylation reactions. Based on interactome and subcellular localization analysis, ANKRD55 is likely transported into the nucleus by the classical nuclear import pathway and is involved in mitosis, probably via effects associated with mitotic spindle dynamics.


INTRODUCTION
An intronic single nucleotide polymorphism (SNP, rs6859219) in the ankyrin repeat domain protein-55 (ANKRD55) gene emerged originally as risk factor for rheumatoid arthritis (RA) from a meta-analysis of genome-wide association studies (1)(2)(3). We demonstrated that this SNP constitutes, as well, a genetic risk factor for multiple sclerosis (MS) (4), and subsequently we validated this association with genome-wide significance (5). A proxy of rs6859219 (r 2 = 0.9, D' = 1), rs71624119, was also associated with MS in a comprehensive fine-mapping of autoimmune disease related genomic regions (6). Further studies have underscored pleiotropy of ANKRD55 by linking SNPs in this locus to Crohn's Disease (7,8), type 1 diabetes (9), juvenile idiopathic arthritis (10), celiac disease (11) and inflammatory myopathies (polymyositis and dermatomyositis) (12,13), as well as post-traumatic stress disorder (14), cognitive decline in Alzheimer's disease (15), and type 2 diabetes (16,17). Rs6859219 is significantly associated with expression levels of three ANKRD55 transcripts, i.e., Ensembl (GRCh37.p13 Human Genome Assembly) full-length protein isoform 001, shorter protein isoform 005 and non-coding transcript 007 in PBMCs, and 001 and 005 in CD4 + T cells (18), and acts thus as a cisexpression quantitative trait locus (cis-eQTL). Rs71624119 (19) as wells as rs10065637 (20), a perfect proxy of rs6859219, were independently identified as cis-eQTLs for ANKRD55 expression, the latter so by means of multiple tagging probes. The risk alleles of associated eQTL SNPs such as rs6859219, rs71624119 and rs10065637 are shared among autoimmune diseases and are associated with higher expression of ANKRD55 in CD4 + T lymphocytes (18)(19)(20). Thus, ANKRD55 appears as the nearest annotated gene affected by the risk eQTL SNP rs6859219 or its proxies that emerged from GWAS on various autoimmune diseases. Notwithstanding these strong indications for a role of ANKRD55 gene products in autoimmune pathogenesis, the biological function of ANKRD55 remains unclarified. Bioinformatic inspection of ANKRD55 does not provide clues as to its function beyond the presence of nine ankyrin repeats (ARs) in isoform 001, of which only the last is shared with isoform 005. ARs are 30-34 residue sequence motifs that represent one of the most common, modular, protein-protein interaction motifs in nature (21). They are found in proteins with diverse functions involved in processes such as transcription, cell cycle regulation, cytoskeletal integrity, inflammatory response, development, cellcell signaling, and various transport phenomena (22,23). The involvement of ankyrin repeat proteins in targeting, correct orientation and mechanical stabilization of membrane proteins to specific compartments within the endoplasmic reticulum (ER) and plasma membrane is well-documented (24). Various biologically important proteins including the family of INK4 tumor suppressors, p15, p16, p18, and p19, the signaling protein Notch and the NF-κB inhibitor IκB, as well as 53BP2, a regulator of the tumor suppressor p53, all contain ankyrin repeats (22). Ankyrin repeat motifs comprise typically a β-turn and two antiparallel α-helices. In elucidated complexes, the β-turn and the first α-helix mediate the interactions with the target, and varying numbers of neighboring repeats partake in binding (25)(26)(27).
These observations support a role for the ankyrin repeat as a versatile scaffold for protein-protein interactions (PPIs) (21), with reported binding affinities of natural ankyrin repeat proteins for their partners occurring in the low nanomolecular range (28). Based on its ankyrin repeats, ANKRD55 is thus likely to take part in interactions with diverse proteins. In this study we used a proteomics-based interactome approach in order to identify partners of ANKRD55 in purified complexes. Using this approach, protein partners whose individual roles in the cell have already been documented, were identified, and these thus shed light on the pathways and processes in which ANKRD55 is active.

Cell Culture
HEK293 and HeLa cell lines were maintained in Dulbecco's Modified Eagle's medium (DMEM) supplemented with 10% inactivated fetal bovine serum (FBSi) and 2 mM L-glutamine (all from Sigma-Aldrich).

DNA Constructs and Cloning Procedures
pCMV6 containing TrueORF Gold Expression validated cDNA clone including the C-terminal DDK-myc tagged ORF of human ANKRD55 transcript 001 (NM_024669; Cat. No. RC221211) was purchased from OriGene. Human ANKRD55 transcript 005 coding sequence was amplified by PCR from pCR-BluntII-TOPO clone IRCMp5012E115D (Source Bioscience) with specific primers including restriction sites from Integrated DNA Technologies (IDT) using Pfu DNA Polymerase (Agilent Technologies) following manufacturer's protocol. Primers used: ATTCGCGATCGCCATGGACAGCAACCTG (005 SgfI Fw), ACCACGCGTATTTTCATCACTGGTGGGGTTGGCAGA (005 MluI Rv). The cycling conditions were set to initial denaturation at 94 • C 45 s, followed by 30 cycles of denaturation at 94 • C 45 s, annealing at 50 • C 45 s and extension at 72 • C 1 min 30 s and a final extension at 72 • C for 10 min. The PCR product size was verified by agarose gel electrophoresis, and the products were purified using GeneJET Gel Extraction and DNA Cleanup Micro Kit (Thermo Fisher Scientific) following manufacturer's protocol. Amplified product was digested using SgfI and MluI restriction enzymes (New England Biolabs) and verified by agarose gel electrophoresis. Purified insert was ligated with pCMV6 vector digested with the same enzymes using T4 ligase (New England Biolabs). Individual colonies were screened by colony PCR and double digestion to identify those that contain the insert and confirmed by Sanger sequencing.

ANKRD55 Expression Analysis and Immunoprecipitation
HEK293 and HeLa cells were cultured in T75 flasks (Sigma-Aldrich) and transfected with FLAG/myc ANKRD55 isoforms 001 or 005 using MACsfectin Reagent (Miltenyi Biotec) according to manufacturer's protocol and incubated for 48 h. Wherever applicable, cells were cross-linked using the membrane-permeable thiol-cleavable homobifunctional crosslinker dithiobis[succinimidyl propionate] (DSP;Thermo Fisher) dissolved in dimethyl sulfoxide. The cells were incubated in PBS supplemented with 100 µg/ml of DSP for 30 min, with gentle vortexing every 5 min. Reactions were quenched by adding 1 M Tris-HCl pH 7.5 to final concentration of 50 mM 15 min at room temperature (RT). For analysis of recombinant ANKRD55 isoform expression, HeLa and HEK293 were fractionated into different subcellular compartments following the protocol from Horton and Holden (29). FLAG-ANKRD55 immunoprecipitation (IP) was carried out from nuclear extracts obtained by sucrose gradient centrifugation (30); membranous organelles from subcellular fractionation and total protein extracts using lysis buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, 1% triton X-100 and 1% cOmplete EDTA-free protease inhibitor cocktail pH adjusted to 8.0) or RIPA buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate and 0.1% SDS, 1 U/ml Benzonase and 1% cOmplete EDTA-free protease inhibitor cocktail). Part of the lysates were kept as input controls. For full interactome analysis, ANKRD55 IP was done from nuclear extracts lysed with RIPA buffer and total protein extracts using lysis buffer. Both buffers included phosphatase inhibitors (1 mM sodium pervanadate, 5 mM beta-glycerophosphate and 5 mM NaF). Part of the lysates were kept as input controls. Protein concentration was estimated using BCA protein assay kit (Pierce). For nuclear enrichment by sucrose gradient centrifugation, cells were washed three times with PBS and resuspended in ice-cold sucrose buffer I (0.32 M sucrose, 3 mM CaCl 2 , 2 mM magnesium acetate, 0.1 mM EDTA, 10 mM Tris-HCl pH 8.0, 1 mM DTT, 0.5% NP-40) in a 50-ml tube. A small aliquot of cells was examined with a phase-contrast microscope to ensure that they were uniformly lysed. Sucrose buffer II (2 M sucrose, 5 mM magnesium acetate, 0.1 mM EDTA, 10 mM Tris-HCl pH 8.0, 1 mM DTT) was added to 50 ml polypropylene centrifuge tubes for JA-20 rotor (sucrose cushion). Nuclei mixture were carefully layered onto the sucrose cushion, the gradient was completed with sucrose buffer I to have the same volume and centrifuged for 45 min at 30,000× g and 4 • C in Beckman Coulter J2-MC High Speed Centrifuge. The supernatant was carefully removed, and nuclei were lysed with RIPA buffer. For the IP of ANKRD55 isoforms together with their binding partners, ANKRD55 and control samples were incubated with Anti-DDK G1 mouse monoclonal antibody coupled to the resin (cat.no. L00432; Genscript) at 4 • C overnight (o/n), on rotating wheel. After incubation, resin was washed six times with washing buffer containing 50 mM Tris-HCl and 150 mM NaCl, pH adjusted to 7.4 (phosphatase inhibitors were included for interactome analysis) and ANKRD55 001 complexes were eluted using acidic elution buffer (0.1 M glycine, 0.15 M NaCl and 0,5% SDS pH 2).

Interactome Analysis
In-gel Tryptic Digestion Immunoprecipitated samples were separated by SDS-PAGE on 6% or 10% gels and stained with Pierce Silver Staining kit or SYPRO Ruby Protein Gel Stain (both from Thermo Fisher) according to manufacturer's protocol. Either selected protein bands from silver stained gels were excised, or entire lanes from SYPRO Ruby-stained gels were cut into ten contiguous pieces (done for both the ANKRD55 001 and control lanes). These gel slabs were then cut into small pieces and washed in Milli-Q water. Reduction and alkylation was achieved by incubation with dithiothreitol (DTT, 10 mM in 50 mM ammonium bicarbonate) at 56 • C for 20 min, followed by incubation in iodoacetamide (IA, 50 mM in 50 mM ammonium bicarbonate) for another 20 min, in the dark. Gel pieces were dried and incubated with trypsin (12.5 µg/mL, in 50 mM ammonium bicarbonate) for 20 min on ice. After rehydration, the trypsin supernatant was discarded; gel bands were covered with 50 mM ammonium bicarbonate and incubated overnight at 37 • C. After digestion, supernatant with digested peptides was recovered. Acidic peptides were further extracted from the gel with TFA 0.1% and pooled with the first supernatant. Digested peptides were dried in a RVC2 25 SpeedVac concentrator (Christ). Peptides were resuspended in 0.1% FA and sonicated for 5 min prior to their nano LC MS/MS analysis.

NanoLC-MS/MS and Data Analysis
Peptide mixtures obtained from trypsin digestion were separated by online nanoLC and analyzed by electrospray tandem mass spectrometry. Peptide separation was performed on a nanoAcquity UPLC system (Waters) connected to an LTQ Orbitrap XL mass spectrometer (Thermo Electron). Approximately 40% of each sample was loaded onto a Symmetry 300 C18 UPLC Trap column, 180 µm × 20 mm, 5 µm (Waters). The precolumn was connected to a BEH130 C18 column, 75 µm × 200 mm, 1.7 µm (Waters) equilibrated in 3% acetonitrile and 0.1% FA, and peptides were eluted at 300 nl/min using a 60 min linear gradient of 3-50% acetonitrile directly onto the nanoelectrospray ion source (Proxeon Biosystems). The mass spectrometer automatically switched between MS and MS/MS acquisition in DDA mode. Survey full scan MS spectra (m/z 400-2,000) were acquired in the orbitrap with a resolution of 30,000 at m/z 400. The 6 most intense ions were sequentially subjected to collision-induced dissociation (CID) fragmentation in the linear ion trap. Precursors with charge states of 2 and 3 were specifically selected for CID. Collision-energy applied to each peptide was automatically normalized as a function of the m/z and charge state. Analyzed peptides were excluded for further analysis during 30 s using dynamic exclusion lists. Searches were performed using Mascot Search engine (Matrix Science) on Proteome Discoverer 1.2. software (Thermo Electron). Carbamidomethylation of cysteines as fixed modification, oxidation of methionines as variable modification, 10 ppm of peptide mass tolerance, and 0.5 Da fragment mass tolerance were adopted as search parameters. Two missed cleavages were allowed. Spectra were searched against a UNIPROT/Swissprot database restricted to Homo sapiens. A false discovery rate estimation procedure was applied for peptide identification (FDR <1%). Proteins identified with at least two peptides passing that cutoff were considered for further discussion. For interactome analysis, three replicates were run for each sample type, and a spectral counting approach was carried out in order to select specifically enriched proteins. This approach relies on the counting of the MS/MS spectra matching to a certain protein (31). The comparison of this value for a certain protein in two different samples provides an idea of its putative differential abundance. In our case, proteins with an ANKRD55 IP/Negative Ctrl IP spectral count ratio >5 in 3/3 replicates were selected for further analysis and discussion. In order to determine the relative abundance of each protein in each of the samples, a Normalized Spectral Abundance Factor (NSAF) approach was carried out (32). The spectral counts for a protein were normalized to its length and further expressed as a % of the total of normalized spectral counts present in that particular sample.

Phosphopeptide Enrichment
To enrich for phosphorylated peptides, titanium dioxide chromatography was performed essentially as described by Rigbolt et al. (33) with some modifications. Briefly, an equal volume of Titansphere material (GL Sciences) was mixed with a buffer containing 80% ACN / 1% TFA / 0.6 M glycolic acid. The slurry was incubated for 20 min at RT with end-over-end rotation. Sample buffer was adjusted to a final concentration of 60% ACN / 1% TFA before incubation with the beads. Titansphere beads were added to the sample in a 0.6 mg TiO 2 / 100 peptide µg proportion and mixed 20 min at RTwith endover-end rotation. Finally, Titansphere material with bound phosphopeptides was washed three times with 150 µl 60% ACN / 1% TFA and transferred on top of a C8 disc (Empore) placed in 200 µL pipette-tip. Phopshopeptide samples were sequentially eluted using a buffer containing 5% NH 4 OH and a buffer containing 25% ACN / 10% NH 4 OH, dried in a RVC2 25 SpeedVac concentrator (Christ), and resuspended in 2% ACN / 0.3% TFA before nano LC MS/MS analysis.

Functional and Pathway Analysis
The DAVID tool from the NIH (https://david.ncifcrf.gov/) (34) was used for Gene Ontology (GO) enrichment analysis. Results from the functional annotation clustering were used to select the interesting GO terms, and only results with p < 0.05 for the modified Fisher's test were considered. Ingenuity Pathway Analysis was used for more detailed characterization of the molecular events lying behind the differential protein patterns under analysis. The calculated p-values determine the probability that the association between proteins in the dataset and a given canonical pathway or upstream regulator is explained by chance alone, based on a Fisher's exact test (p-value < 0.05 considered significant).

Validation by Immunocytochemistry
For colocalization studies of endogenous and recombinant ANKRD55 with interacting partners, HEK293 cells were plated on coverslips coated with poly-D-lysine (Sigma-Aldrich) and transfected with ANKRD55 001 using MACsfectin Reagent (Miltenyi Biotec) according to manufacturer's instructions. 24 and 48 h later, cells were fixed with 4% paraformaldehyde in PBS for 20 min at RT. Those coverslips destined to nuclear colocalization were subjected to antigen retrieval treatment by incubating the cells with R-Universal buffer (Cat. No. AP0501-500; Aptum). Next cells were permeabilized with 0.2% Triton No. 20543-1-AP; 1:200), all from Proteintech for 1 h at RT, followed by staining with Alexa Fluor 647 conjugated antirabbit (ab150075; 1:500) and DyLight 550 conjugated anti-mouse (ab98713; 1:500) from Abcam for 1 h at RT and counter-staining with DAPI (1:500; Sigma-Aldrich). Images were obtained using a Leica TCS CW SP8 STED Super-Resolution microscope with a 63x immersion objective and excitation wavelengths of 405, 551, and 650 nm.

Quantification of Colocalization
For the cytoplasmatic colocalization, quantification was carried out using LAS AF (Leica Application Suite Advanced Fluorescence) software and Pearson's correlation coefficient. First, the background signal was extracted in all the studied images. Then, the Pearson's correlation coefficient values from 10 regions of interest (ROIs) were annotated, corresponding each ROI to a unique cell (n = 10 cellular ROIs/condition). In transfected cells the ROIs were selected from transfected cells. Analysis of the colocalization quantification was performed with GraphPad v.6 (GraphPad Software, San Diego, CA, USA). The data is represented as mean ± SEM and the Mann-Whitney test was applied. The level of significance was set at p ≤ 0.05, with p ≤ 0.0001 ( * * * * ) as extremely significant, p ≤ 0.001 ( * * * ) highly significant, p ≤ 0.01 ( * * ) very significant, p ≤ 0.05 ( * ) significant, and p > 0.05 (ns) not significant.

Overexpressed ANKRD55 Isoforms Are Located in Cytosol, Membranous Organelles and Nucleus
Human ANKRD55 001 and 005 isoforms were transiently overexpressed in HEK293 and HeLa cells, and cells were fractionated into three subcellular compartments; cytosol (C), nucleus (N) and membranous organelles (MO). In transfected HEK293 cells, similar levels of isoform 005 were detected in Cells were fractionated at 48 h after transfection. ANKRD55 was detected with specific Ab (Sigma-Aldrich) by WB. Enrichment for subcellular fractions was verified using histone H3 as nuclear marker (N), GRP94 as marker of membranous organelles (MO) and GAPDH as cytosolic marker (C). Specific bands corresponding to overexpressed isoforms 001 and 005 are indicated by black and white arrowheads, respectively. Black cross marks endogenous ANKRD55 isoform 001. Black diamonds, and black and white triangles show three different types of truncated immunoreactive products.
the three fractions. Isoform 001 (72 kDa) showed a more abundant signal than 005 (43 kDa), and was as well present in the three compartments though slightly less abundant in the cytosol than in the other two fractions (Figure 1). In HeLa cells, isoform 001-enriched fraction from the cytosol showed a stronger signal than that of the other two fractions. Isoform 005 detection was slightly weaker in nuclei than in the rest of the fractions. Moreover, truncated immunoreactive products were present. These included (i) a 55-60-kDa band in HEK293 and HeLa cells in isoform 001-transfected cells present in all 3 compartments though at levels differing from the full-length proteins, (ii) a 20-kDa band in MO of HEK293 cells, and (iii) a series of bands with Mr<34 kDa especially enriched in the cytosol of isoform-005 transfected HEK293 and HeLa cells. Additionally, the results showed the endogenous isoform 001 in both cell lines demonstrating the intrinsic capacity of HEK293 and Hela cells to produce ANKRD55 and therefore to engage its biological pathway(s). Thus, western blot analysis of biochemically fractionated cells revealed presence of the isoforms 001 and 005 of ANKRD55 in all three analyzed subcellular compartments of both cell lines (Figure 1). Further experiments showed that it was possible to efficiently purify ANKRD55 isoform 001, more so than isoform 005, using anti-FLAG agarose from MO fraction, nuclear (NE) and total (TPE) protein extracts in the presence or absence of cross-linker DSP (Supplementary Figure 1).  Figure 2). Use of the crosslinker DSP had generally little effect on overall intensities of the bands, suggesting that the interaction with ANKRD55 was unaffected by lysis and purification conditions. Bands corresponding to the Mr of recombinant FLAG-ANKRD55 isoforms and those not present in the control samples were cut, trypsin digested and identified by nLC-MS/MS analysis (Supplementary Table 1). Proteins belonging to the 14-3-3 family (YWHAE, YWHAZ, YWHAH, YWHAB, YWHAQ, YWHAG) were repeatedly identified in IPs of ANKRD55 001 or 005 from TPE but not those from NE or MO extracts. Heat shock proteins (HSPA9, HSPA5, HSPA1A, HSP90B1, HSP90AB1, HSP90AA1, HSPA8, HSPD1) were also abundant among ANKRD55 interactors especially so in the MO fraction. HSP90AB1, HSP90AA1 and the unique HSP90 ER homolog GRP94 (HSP90B1) were identified in ANKRD55 IPs from MO fractions of both HEK293 and HeLa cells. Members of the tubulin superfamily were found in ANKRD55 001 IPs from TPE and NE but not MO (TUBB, TUBA1B, TUBB4B, TUBB4A). Proteins participating in initiation and elongation phases of eukaryotic translation (EEF1A1, EIF4A1, TUFM, EIF4A3, EEF1G, EEF2) were identified in all extracts. Members of the mitochondrial transporter family SLC25 (SLC25A1, SLC25A5, SLC25A6) were identified in TPE and MO extracts. YWHAE was identified in presence or absence of cross-linker as well as in two independent experiments from HEK293 TPE. The protein kinase PRKDC was found in three independent experiments in HEK293 NE. Thus, this first analysis revealed that ANKRD55 interacts (in)directly with a combination of functionally distinct proteins in MO, NE and TPE. Of these, 14-3-3 proteins appear to present the Frontiers in Immunology | www.frontiersin.org best-resolved interactors in silver-stained gels (various examples in Supplementary Figure 2).

The ANKRD55 Interactome
We compared and analyzed in more detail ANKRD55 and control IPs in order to search for and identify additional interacting partners as well as to confirm those identified above. In essence, rather than focusing on single visualized gel bands, we analyzed the complete gel lanes corresponding to ANKRD55 001 or mock (not transfected cells) FLAG eluates. Low abundant ANKRD55-interacting proteins may be present that are not resolved as single bands, and therefore could be missed by the first method. Together such bands tend to produce a darker stainable protein smear in ANKRD55 compared to mock lanes (several such examples can be seen in Supplementary Figure 2). By subtracting the "mock proteome" from the ANKRD55 proteome, a specific and comprehensive ANKRD55 interactome can be deduced. We focused this proteomic analysis on FLAGtagged ANKRD55 isoform 001, which was more efficiently IPed than isoform 005, and showed highest number of potential interacting partners. FLAG-ANKRD55 isoform 001 was overexpressed in HEK293 cells and IPed from TPE and sucrose gradient-purified NE in triplicates (three independent experiments each that were performed on distinct days from distinct batches of cells). WB with FLAG antibody was performed to verify the quality of ANKRD55 overexpression and IP (Figure 2A). ANKRD55 complexes were then separated by SDS-PAGE, and stained using SYPRO Ruby. The full gel lanes starting from the boundary of the stacking gel with the resolving gel up to the gel front were cut into ten pieces (one contained ANKRD55 001 or its control), trypsin digested, and identified by nLC-MS/MS ( Figure 2B). To build the ANKRD55 interactome, all potential ANKRD55 interacting partners were filtered according to their spectral count enrichment. Spectral counting accounts for the total number of MS/MS spectra matching to peptides that belong to a certain protein, and provides a valuable parameter for relative protein abundance comparison (see section Materials and Methods for more details). We identified enrichment of 148 proteins in TPE and 22 proteins in NE over the three replicates, based on stringent selection criteria including both at least two unique peptides and a 5-fold increase in ANKRD55 spectral counts in the ANKRD55 IP compared to control IP in all three replicates. Several of these proteins were selected for further functional analysis and validation ( Figure 2C, Table 1, Supplementary Table 2). The average NSAF of these proteins was calculated in order to estimate their relative abundance within the IPed samples. Although not considered for further analyses, an additional, less robust, dataset of putative ANKRD55 interactors was substracted in addition to the set of reliably enriched proteins. For this purpose, proteins enriched with a more relaxed criteria of ANKRD55 IP/negative control IP spectral count ratio of at least 5 in only 2 out of 3 replicates were selected, providing additional 144 putative-ANKRD55 partners in total protein extracts and 34 in nuclear extracts (Supplementary Table 3).

Bioinformatic Analysis of ANKRD55 Protein Partners
To interpret the potential functional role of ANKRD55-interacting proteins, ANKRD55 interactome FIGURE 3 | Functional enrichment analysis of ANKRD55 interactome from total protein extracts. The algorithm DAVID was used to analyze the ANKRD55 interactome using the GO terms Cellular Component (blue), Biological Process (green), Molecular Function (orange) and Up Keywords (yellow). The GO and other terms with FDR<5% are shown. Plotted as -log (p-value) significance.
Frontiers in Immunology | www.frontiersin.org results were subjected to enrichment analysis using DAVID bioinformatics resource tool. We used functional annotation clustering to select the interesting Gene Ontology (GO) and other terms. The results from GO biological process, GO cellular compartment, GO molecular function and Uniprot (UP) keywords with a FDR<5% are shown in Figures 3, 4. A more detailed summary of the functional annotation clustering with p < 10 −3 is provided in Supplementary Table 4. The functional analysis suggested that ANKRD55 protein partners from total protein extracts are tightly related to nucleotide and ATP binding (p < 1.18 × 10 −13 ). Bioinformatics analysis predicted the binding ability of ANKRD55 to DNA, ATP, ADP and GTP nucleotides, but not to RNA (Supplementary Table 5). GO categories were enriched in nuclear transport terms including Ran GTPase binding, intracellular protein transport, protein import into nucleus and nuclear pore (p < 2.00 × 10 −03 ). Total protein extracts dataset was also linked to processes that are associated with cell cycle and RNA, lipid and amino acid metabolism. Additionally, bioinformatics identified protein biosynthesis-related terms such as translation, structural constituent of ribosome, SRP-dependent cotranslational protein targeting to membrane, tRNA aminoacylation for protein translation and regulation of translational fidelity. GO cellular component annotated focal adhesion, ribosome, nuclear pore, and nuclear chromosome, among others, as the localization of the involved proteins in the total protein interactome. In the enrichment analysis of the ANKRD55-protein partners from nuclear extracts, ∼20 and 30% of proteins are related to sumoylation and RNA binding, respectively. GO categories were involved in processes that are associated with cell cycle including cell division, mitotic nuclear envelope disassembly, mitosis, DNA repair and damage. Nuclear extracts interactome was also linked to processes including RNA transport, nucleotide and ATP binding, among others. GO cellular component analysis indicated nuclear pore as the localization of the involved proteins in the nuclear interactome. To further examine the ANKRD55 interactome, we performed an Ingenuity Pathway Analysis (IPA, QIAGEN) to identify major canonical pathways associated with ANKRD55-interacting proteins. Sixty out of one hundred ninety-six pathways were significantly enriched in total protein extracts (p < 0.05; -log (p-value) > 1.30). The top ten pathways, ranked by significance included tRNA Charging, Signaling. In nuclear extract interactome analysis, SMC1A and SMC3 proteins were related to Mitotic Roles of Polo-Like Kinase and ATM Signaling; PRKDC and PHB2 to Estrogen Receptor Signaling, VIM and TUBA1C to 14-3-3-mediated Signaling and PRKDC and TUBA1C to Sirtuin Signaling Pathway. All information on the IPA canonical pathways analysis of both interactomes is included in Supplementary Table 6.

Validation of Selected ANKRD55 Interacting Partners by IP and Reverse IP
We confirmed the interaction of a selection of identified proteins with recombinant ANKRD55 isoform 001 by IP followed by WB using antibodies against the selected partners (Supplementary Table 7). From the cytosolic interactome, we selected (i) six of the first ten proteins ranked according to NSAF value (14-3-3 isoforms YWHAE, YWHAB, YWHAH and YWHAG, RPS3, and TUBB6); (ii) CLTC with a lower NSAF value which is mainly localized to coated vesicles and coated pits, in addition to cytosol and mitotic spindle; and (iii) proteins that were found to be present as well in the nuclear interactome (SMC1A and PRKDC). From the nuclear interactome, three proteins (VIM, SMC1A and SMC3) from the top ten in the list were analyzed. Several of these proteins had at first been identified in individual bands cut from silver-stained gels including PRKDC, SMC1A and SMC3 (see Supplementary Table 1). For detection of TUBB6, we used an anti-β-tubulin antibody capable of recognizing different βtubulin isoforms including TUBB3, TUBB2B, TUBB, TUBB4B, TUBB2A, and TUBB4A (three of these were identified in two out of three interactome replicates, and several of these had been identified as well as in individual bands analyzed in the first approach, see Supplementary Table 1). Anti-FLAG IP experiments and WB demonstrated that endogenous SMC1A, TUBB6, RPS3, 14-3-3 isoforms, PRKDC, CLTC in total protein extracts (Figures 5A,B) and SMC1A, SMC3, VIM in nuclear extracts interacted specifically with ANKRD55 isoform 001 ( Figure 5C). Neither ANKRD55 nor interacting partners were detected in the control IP. The IP of FLAG-ANKRD55 with 14-3-3 isoforms and PRKDC was also validated in total protein extracts of HeLa cells by WB (Figure 5D).
We performed also reverse IP using antibodies against 14-3-3 epsilon, the highest scoring partner of ANKRD55 in terms of NSAF value (Table 1), RPS3 and TUBB (Figure 6). Recombinant ANKRD55 was detected in the IPs of the specific antibodies but not in the IgG control IPs confirming that interactions of ANKRD55 with the target proteins were specific. We could not detect ANKRD55 in IPs from untransfected cells, likely due to the low levels of endogenous protein (Figure 1). FIGURE 6 | Detection of ANKRD55 in reverse immunoprecipitates of RPS3, 14-3-3ε and β-tubulin. RPS3 (A), 14-3-3ε (B), and β-tubulin (C) were immunoprecipitated from HEK293 transfected with ANKRD55 (001) vector or left untransfected (CTRL). Mock IPs were performed with control IgG antibodies as described in the Materials and Methods. IPs were resolved on SDS-PAGE, blotted and ANKRD55 was detected by means of anti-ANKRD55 Ab.
We checked freely available protein-protein databases (BioGRID (40), STRING (41) and HuRI 1 ) to confirm some of the identified interactions. The BioPlex 2.0 dataset (42,43), which has been deposited into the BioGRID database, described thirteen ANKRD55-interacting partners; seven were included in at least two out of three replicates from our interactome analysis (YWHAE, YWHAQ, IFT52, IFT74, MIB1, YWHAZ, and YWHAG). Intriguingly, HIF1AN, an ANKRD55-interacting partner described in the BioPlex 2.0 dataset, catalyzes hydroxylation of highly conserved asparaginyl residues within AR of NFKB1, NFKBIA, NOTCH1, ASB4, PPP1R12A and several other AR-containing proteins (44). 1 http://interactome.baderlab.org/ Therefore, this could suggest that hydroxylation of asparaginyl residues from ANKRD55 may take place. However, in our experimental set-up, we did not clearly detect HIF1AN in ANKRD55 interactomes by mass spectrometry, and neither were we able to demonstrate clearly its interaction with ANKRD55 by WB analysis using specific antibodies (Supplementary Figure 3). STRING database also included ANKRD55-partners shared with our interactome, specifically RPS3 and ACLY in Homo sapiens and Rps3 and Hsp90ab1 in Mus Musculus. Human Reference Protein Interactome Project (HuRI) included thirty unpublished ANKRD55interacting partners; none of these were classified in our protein list (Supplementary Table 8). However, some of the HuRI ANKRD55 partners are involved in ciliogenesis and its regulation (CFAP206, LRGUK, and DEUP1), just as are some proteins from our ANKRD55 interactome (MIB1, IFT74, ARF4, CCT4, CCT8, TTC26, FLNA, and IFT52), and ANKRD55 itself was described as a novel intraflagellar transport complex protein (45). Moreover, the BioGRID database also identified other members included in the same process (IFT46, TTC30A, and TTC30B).

Immunofluorescence (IF) Colocalization of ANKRD55 and Partners
Proteins that interact tend to reside within the same or adjacent subcellular compartments (46). Fluorescence colocalization microscopy can thus be used to assess potential links between distinct molecules (47). Subcellular localization of endogenous and recombinant ANKRD55 using anti-ANKRD55 Ab with a selection of interacting partners was analyzed to verify sharing of specific subcellular localizations. Under conditions used for visualization, anti-ANKRD55 antibody detected ANKRD55 in transfected cells but not in untransfected cells (Supplementary Figure 4). However, endogenous ANKRD55 could be visualized in untransfected cells using over-/underexposure Look-up Table scale (Supplementary Figure 5). In the IF Figures, we include the Pearson's correlation coefficient for colocalization of RPS3, TUBB, 14-3-3 and VIM with both endogenous ANKRD55 (Untrans.) and recombinant ANKRD55 (Trans.) at 24 and 48 h after transfection (B panels in Figures 7-9, Supplementary Figure 6). The results showed a highly significant colocalization (p ≤ 0.001) between recombinant ANKRD55 isoform 001 and either RPS3 (Figure 7) or β-tubulin isoforms (Figure 8) in both the cytoplasm and cell membrane. Correlation of colocalization between 14-3-3 and ANKRD55 was not higher with recombinant ANKRD55 (Figure 9) than with the endogenous one, and that with VIM only increased at 48 h (Supplementary Figure 6), suggesting already-high occupancy of these proteins with the endogenous ANKRD55. Recombinant ANKRD55 colocalization with β-tubulin and 14-3-3 isoforms remained constant, but decreased over time after transfection with RPS3 (Figures 7-9) and increased with VIM (Supplementary Figure 6). As shown in Figures 7,  8 ANKRD55 was poorly detectable in the nucleus, despite its presence in biochemically or sucrose-gradient obtained nuclear fractions (Figure 2). The procedure was modified to include an antigen retrieval treatment step to try to increase the fluorescent signal of the nuclei. This modification was successful when used in combination with a highly sensitive anti-FLAG Ab and was applied to study the nuclear colocalization of overexpressed ANKRD55 together with β-tubulin isoforms ( Figure 10) and VIM (Supplementary Figure 7). After antigen retrieval treatment, recombinant ANKRD55 became visible in the nuclei of cells in division, though these cells exhibited a lower fluorescence intensity signal than not-dividing cells. Quantification was not performed due to the small sample size. ANKRD55 positively colocalized with diffuse β-tubulin isoforms in the nucleus and specifically with the mitotic spindle (Figure 10). It also colocalized with diffuse VIM in the nucleus without an associated specific structure (Supplementary Figure 7).

ANKRD55 Is a Phosphoprotein
An important subset of proteins identified in this interactome analysis belong to the 14-3-3 protein family. 14-3-3 proteins are ubiquitously expressed phosphoserine/phosphothreonine (pS/pT)-binding proteins that are members of a large family of isoforms (48). In general, they bind to their ligands through RSXpS/TXP or RXXXpS/TXP sequences (49), but they can also bind their ligands in a phosphorylation-independent manner (50,51). Thus, we determined whether ANKRD55 was phosphorylated. Our primary sources for this step were phosphoproteome repositories, such as PhosphoSitePlus (35) and dbPAF (36) databases. Moreover, 14-3-3-Pred web server (37) was used to predict putative 14-3-3 binding phosphosites on ANKRD55. Two out of ninety predicted phosphorylation sites (S436 and S597) fulfill the cut-off (Supplementary Table 9). We identified specific serine and threonine phosphorylation sites in overexpressed and immunoprecipitated ANKRD55 isoform 001 from HEK293 nuclear and total proteins extracts by TiO 2 -based phosphopeptide enrichment prior to nLC-MS/MS analysis. The phosphorylation sites at T6 and S436 in ANKRD55 were confirmed and the S475 site was found in nuclear and total protein extracts previously identified in mouse ANKRD55 ( Table 2). In conclusion, the data revealed five phosphorylation sites in human ANKRD55 at T6, T11, T189, S436 and S475. 14-3-3-Pred web server predicted with the highest score S436 as 14-3-3 binding phosphosite.

DISCUSSION
ANKRD55 emerged as the nearest annotated gene of which expression is affected by a series of cis-eQTL risk SNPs that were identified through multiple GWAS and other studies in diverse multifactorial-predominantly autoimmune, diseases (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17). However, the biological function(s) of ANKRD55 has (have) not been elucidated. Ankyrin repeats function as versatile scaffolds for protein-protein interactions. ANKRD55 isoform 001, the main focus of this study, contains nine ankyrin repeats and consequently should take part in diverse PPIs as has been reported for other members of the ANK family (21)(22)(23)(24)(25)(26)(27). Similar to naturally produced ANKRD55 (18), overexpressed recombinant ANKRD55 is present in the nuclei and cytosol. We therefore characterized the ANKRD55 protein interactome from both TPE and sucrose-gradient-purified NE. The limitations of using co-immunoprecipitation in overexpressing cell lines to determine protein interactions should be noted, as this approach may lead to non-specific interactions and prevents spatial or temporal analysis of protein interactions (61). A larger group of ANKRD55-interacting partners was present in total protein extracts than in nuclear extracts with 148 vs. 22 proteins consistently identified over three independent replicates. Functional analysis suggested that ANKRD55-protein partners from nuclear (eight and seven proteins) and total protein extracts (fifty-two and forty-three proteins) are related to nucleotide and ATP binding, respectively. 30% of proteins from the nuclear interactome are related to RNA binding.
Experiments were performed to validate top-scoring partners of ANKRD55 by western blot, reverse IP and/or colocalization in IF microscopy. Four distinct 14-3-3 proteins (epsilon, beta/alpha, eta and gamma) were identified in total cell extracts. 14-3-3 colocalized with ANKRD55 in IF microscopy, and ANKRD55 was detected in reverse IP of 14-3-3ε, the major ANKRD55 partner ( Table 1). Being pS/pT-binding proteins, their presence led us to identify and confirm five phosphorylation sites in ANKRD55 with S436 as most likely 14-3-3 binding phosphosite. 14-3-3 proteins modulate the action of proteins that are involved in cell cycle and transcriptional control, signal transduction, intracellular trafficking and regulation of ion channels (62) through a variety of mechanisms: (i) direct conformational change of the target protein; (ii) physical occlusion of sequencespecific or structural features; and (iii) scaffolding that anchors proteins within close proximity of one another (63).
Eleven ribosomal proteins (RPs) were identified in ANKRD55 interactomes. Of these, RPS3 was identified in both nuclear and total protein extract interactomes as one of the top 10 hits, and was confirmed by western blot of IPs and by colocalization with ANKRD55 by IF microscopy. ANKRD55 was also detected in reverse IP of RPS3. Although RPs are known for playing an essential role in ribosome assembly and protein translation, their ribosome-independent functions are also increasingly being appreciated (64). RPS3, specifically, has been shown to induce apoptosis by collaborating with E2F1 (65). Phospho-RPS3 translocates into the nucleus and upregulates prosurvival gene expression via association with NF-κB in non-small cell lung cancer cells (66,67). RPS3 is also involved in immune signaling by selectively modulating NF-κB target gene expression. Given the fact that RPS3 also activates the p53 tumor suppressive pathway (68), this RP is regarded as one of the most fascinating RP with pivotal multifunctions (69).
ANKRD55 colocalized with TUBB in IF microscopy, and their interaction was confirmed by western blot; ANKRD55 was also identified in reverse IPs of TUBB. Inspection by IF microscopy revealed that ANKRD55 is detectable in the nucleus mainly during mitosis. Throughout the different phases of mitosis, βtubulin presented a specific marking in the mitotic spindle that colocalized with ANKRD55. Other proteins were reproducibly identified in the nuclear ANKRD55 interactome that have been reported to fulfill roles associated with mitotic spindle dynamics or structure. These include SMC3 and SMC1A that are members of the cohesin complex that is crucial for chromatid cohesion. By holding the two sister chromatids together from their formation during replication until their separation in anaphase, cohesin creates a counterforce to the pulling of the mitotic spindle, which allows correct chromosome alignment and segregation (70). The association between the ANKRD55-interacting proteins and the mitotic spindle is also supported by the presence of RPS3 protein which was reported to be involved in spindle dynamics (71). Similarly, Ran GTPase binding, enriched in the GO analysis, has been related to the spindle formation (72). MCM3, 6 and 7, DNA helicases involved in the initiation of DNA replication, were also found in ANRD55 interactomes. Of these, MCM7 is known to co-localize with tubulin in mitotic cells with MCM7 depletion resulting in aberrant mitosis. This is in line with observations that MCM7 exerts certain functions on spindle formation to prevent cytokinesis during early mitosis by regulating CDK1 activity (73).
GO categories from total protein extracts were enriched in nuclear transport terms providing a possible explanation for the nucleocytoplasmic transport of ANKRD55, which seems to be mediated by the importins/exportins transport system enriched in the interactome (CSE1L, XPO1, IPO4, XPOT, IPO5, NUP160, NUP205, KPNA2, COPG1, CDK5, YWHAH, CLTC). However, a number of AR-containing proteins, such as IκBα, ASPP2, and GABPβ, have been reported to enter the nucleus via an unknown mechanism independent of a canonical nuclear localization signal (74). As demonstrated by Lu et al. (75), this mechanism is characterized by the presence of components of the RanGDP/AR pathway which represents a general importin-independent nuclear import pathway frequently used by AR-containing proteins (75). Ran GTPase binding, which obtained high significance in the GO, is known to regulate protein trafficking through the nuclear envelope and is an active component of both nuclear transport systems.
In conclusion, the large number of protein interactors identified in this study suggests that ANKRD55 may act as FIGURE 10 | Colocalization of nuclear recombinant ANKRD55 iosoform 001 with endogenous β-tubulin in HEK293 cells following antigen retrieval treatment. Representative microphotographs of immunostaining for ANKRD55 (FLAG Ab; red), tubulin beta chain (TUBB Ab; green) and nuclei (DAPI; white) in non-transfected (CTRL) and transfected (24 h or 48 h) HEK293 cells with ANKRD55 isoform 001 subjected to antigen retrieval treatment. Original image included zoom-in region (yellow), the signal intensity in zoom-in images was increased. Colocalization of ANKRD55 with TUBB in the mitotic spindle is indicated with colored triangles. a scaffold, probably exerting function(s) in the formation or architecture of multiple protein complexes, that may be regulated via (de)phosphorylation reactions occurring at up to 5 sites in the protein.

DATA AVAILABILITY
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD013332.