The intracellular domain of the amyloid precursor protein (AICD) is generated following cleavage of the precursor by the γ-secretase complex and is involved in membrane to nucleus signaling, for which the binding of AICD to the adapter protein FE65 is essential. Here we show that FE65 knockdown causes a downregulation of the protein Bloom syndrome protein (BLM) and the minichromosome maintenance (MCM) protein family and that elevated nuclear levels of FE65 result in stabilization of the BLM protein in nuclear mobile spheres. These spheres are able to grow and fuse, and potentially correspond to the nuclear domain 10. BLM plays a role in DNA replication and repair mechanisms and FE65 was also shown to play a role in DNA damage response in the cell. A set of proliferation assays in our work revealed that FE65 knockdown in HEK293T cells reduced cell replication. On the basis of these results, we hypothesize that nuclear FE65 levels (nuclear FE65/BLM containing spheres) may regulate cell cycle re-entry in neurons as a result of increased interaction of FE65 with BLM and/or an increase in MCM protein levels. Thus, FE65 interactions with BLM and MCM proteins may contribute to the neuronal cell cycle re-entry observed in brains affected by Alzheimer’s disease.
Almost 15 years ago, the interaction of the amyloid precursor protein (APP) and the FE65 adapter protein family was described (Fiore et al., 1995; Guénette et al., 1996). It was suggested that FE65 might have an important impact on APP related mechanisms and in the pathogenesis of Alzheimer’s disease (AD). FE65 consists of three functional domains, two phosphotyrosine binding domains (PTB) and a WW domain (Ermekova et al., 1997), pointing to a highly interesting molecule potentially involved in various intracellular pathways. The YENPTY sequence in APP constitutes the interaction motif, which is responsible for binding to the PTB2 domain of FE65 (Bressler et al., 1996; Fiore et al., 1995; McLoughlin and Miller, 1996). In a similar fashion, the two other members of the FE65 family, FE65L1 and FE65L2, bind to APP (Duilio et al., 1998; Guénette et al., 1996). A role for FE65 in AD may result from the effect of FE65 proteins on APP processing, with FE65 proteins contributing to Aβ generation in neurons and in the brain (Sabo et al., 1999; Suh et al., 2011; Wang et al., 2004). Alternatively, the observations that phosphorylation of the APP intracellular domain (AICD) at T668 is necessary for its interaction with FE65 in the nucleus (Chang et al., 2006), that phosphorylation of T668 leads to FE65 nuclear translocation (Nakaya et al., 2008), and that pT668 APP is elevated in AD brain (Shin et al., 2007) suggest a role for nuclear FE65 and AICD in AD. Furthermore, FE65 knockout mice are impaired for the hidden platform in the Morris water maze test and showed defective early-phase LTP (Wang et al., 2009). Although the underlying molecular mechanisms are unknown, these data suggest a possible role for FE65 proteins in the learning and memory deficits found in AD. In support of a role for the FE65/APP interactions in AD, there is evidence for a protective effect of an FE65 polymorphism in the AD population over 75 years (Hu et al., 1998). This bi-allelic polymorphism in intron 13 generates an altered FE65 protein lacking part of the APP binding site and binds APP less efficiently than FE65 (Guénette et al., 2000; Hu et al., 2002; Lambert et al., 2000). At present whether FE65 proteins contribute to AD is unclear, but a better understanding of the role of nuclear FE65 is necessary to determine whether it plays a role in AD pathophysiology.
FE65 forms a multimeric complex with AICD and the histone acetyltransferase TIP60, which activates transcription of a reporter gene construct raising the possibility that this complex regulates transcription (Cao and Südhof, 2001). Furthermore, FE65 and FE65L1 stimulate APP intracellular domain (AICD) generation (Chang et al., 2003; Wiley et al., 2007; Xie et al., 2007). However, the identification of several putative target genes have been controversially discussed (Baek et al., 2002; Bimonte et al., 2004; Müller et al., 2007; Telese et al., 2005; von Rotz et al., 2004). Elucidation of the role of FE65 in a gene expression complex is further complicated by the observation that FE65 itself triggers robust gene expression in a reporter assay (Yang et al., 2006). Microscopy studies revealed that the above described complex consisting of AICD, FE65 and TIP60, is visible as speckles in the nucleus (Muresan and Muresan, 2004; von Rotz et al., 2004). Further evidence for FE65 function in the nucleus was obtained by X-ray treatment of mouse embryonic fibroblasts isolated from FE65 knockout mice, which revealed significantly higher levels of the histone gamma-H2AX than in controls, a typical cellular response to DNA damage (Minopoli et al., 2007). Notably, increased nuclear DNA oxidation is known to occur in AD caused by elevated reactive oxygen species (Santos et al., 2012). Although our present knowledge is limited, the presented facts point to a nuclear function for this adapter protein, as well as to a putative role in neurodegeneration.
In order to gain further mechanistic insights into FE65 function, we established a FE65 knockdown cell culture model, which was screened for effects by state-of-the-art proteomics and functional studies. We were able to show that FE65 plays an important role in the localization of a certain set of proteins, including the Bloom syndrome protein (BLM), which is involved in DNA replication and repair. In FE65 knockdown cells, BLM and other proteins are massively aggregated in the ER. Functional studies demonstrate that FE65 knockdown cells are characterized by a significantly lower level of proliferation and DNA replication. In contrast, FE65 overexpression results in a re-localization of BLM to nuclear mobile spheres. Size, growth and fusion of these spheres may be associated with the cell cycle and DNA replication, for which FE65 is a key regulator. We hypothesize that FE65 plays a crucial role in neuronal cell cycle re-entry, which is reported to occur in AD.
Knockdown of FE65 significantly disturbs the cellular proteome
FE65 is an APP binding protein, whose physiological role as well as pathophysiological role in AD is poorly understood. In order to gain further functional insights, we generated five stable FE65 HEK293T knockdown clones using selectable shRNAmir constructs. As a control, we established five cell lines transfected with random shRNAmir constructs. A significant knockdown was achieved as demonstrated by immunoblotting of cell lysate samples (Fig. 1A). Densitometric analysis of five stable knockdown versus five control cell lines revealed an effective downregulation of the FE65 protein with high statistical significance (Fig. 1B, downregulation to 39.6%, P<10−2). Moreover, significant downregulation of the FE65 transcript (using GAPDH as reference protein) was also proven by quantitative real-time PCR (qPCR) analysis (Fig. 1C, P<10−4, higher dCt values correspond to lower RNA levels in comparison to the control) with a calculated remaining gene expression of 40.3% (+3.9%, −3.6%). As a redundancy of the FE65 protein family is discussed in the literature (Guénette et al., 2006) and in order to preclude silencing of the FE65-like proteins, we also determined the expression of FE65L1 (Fig. 1D) and FE65L2 (Fig. 1E) by qPCR. However no significant differences of the like protein expression in FE65 knockdown cells compared to controls were evident. Subsequently, total protein extracted from the five established stable knockdown clones and the five stable control clones were used for a differential proteome approach (Fig. 2A), which we successfully applied before (Müller et al., 2011b; Müller et al., 2011a; Spitzer et al., 2010). Using the LTQ Orbitrap Velos mass spectrometry instrument, we identified 10,819 proteins in total. Quantification of these proteins was performed using the label-free spectral count based approach (Müller et al., 2011b; Müller et al., 2011a; Spitzer et al., 2010). Among the total identified proteins, 136 revealed significant abundance changes (supplementary material Table S3, P<0.05), whereas 30 proteins correspond to the so called black-and-white list, which are proteins identified only in one group with at least three measured values (supplementary material Table S3, last section). Black-and-white proteins might be highly interesting candidate proteins but might also correspond to false-positive results. Thus, we focused on the proteins with significant abundance changes for further experiments. The choice of candidates for subsequent validation experiments was derived from a ternary strategy: initially, we evaluated the significance of the results from the label-free study (statistical significance, ratio of abundance changes as given in supplementary material Table S3). Next, protein candidate selection was performed with the help of the pathway analysis software IPA (Ingenuity Pathway Analysis) and STRING (Search Tool for the Retrieval of Interacting Genes/Proteins). Detailed data analysis strategy is described in Materials and Methods. Finally, recent literature findings with respect to FE65 and AD were manually assessed and aligned with the pathway results. Data upload in the mentioned software tools resulted in the pathway cartoons that are given in Fig. 2B (IPA) and Fig. 2C (STRING, excerpt, please find the full pathway cartoon in supplementary material Fig. S6), respectively. Data analysis in IPA pointed to an impairment of a network involved in DNA replication, recombination and repair (called Top Network in IPA) including the MCM (minichromosome maintenance) protein family (MCM2–MCM7), which are all quantified as significantly less abundant proteins in FE65 knockdown cells. Another top network identified by IPA (not shown) was similarly entitled with DNA replication, recombination and repair (cell death, post-translational modifications) and pointed to the proteins BLM (Bloom syndrome), TERF2 (telomeric repeat-binding factor 2) as potentially interesting candidates. The interplay of MCM proteins was also evident using STRING for data interpretation (Fig. 2C). BLM and TERF2 were illustrated as associated proteins. The candidate KAT7 (TIP60_2) was another interesting candidate as it belongs to the family of histone acetyltransferases of which KAT5 (TIP60) is known to interact with FE65 (Cao and Südhof, 2001).
Abundance of proteins involved in DNA replication and repair are significantly reduced upon FE65 knockdown
For subsequent experiments we selected the MCM3 protein as representative for the MCM family as this protein was identified with high abundance (spectral index >45 on average). Moreover, MCM3 was found 1.5-fold less abundant in the knockdown cells with high significance (P<0.025). As a second candidate and representative of proteins involved in DNA repair, we selected the BLM protein, which was 7.6-fold less abundant in the FE65 knockdown cells (P<0.01). Finally, KAT7 was selected, as this protein might be able to bind FE65 similar to its family member protein KAT5.
MCM3 downregulation was also evident by immunoblotting samples from the five knockdown clones versus five controls (Fig. 3A). Using β-actin as reference, densitometric analysis revealed significance of our findings for MCM3 (Fig. 3B, P<0.05). Gene expression changes caused by FE65 and its interacting proteins have been discussed in several publications (for review see Müller et al., 2008). In order to identify whether MCM3 is regulated at the mRNA level, we performed quantitative real-time PCR with GAPDH as expression control. Indeed, qPCR revealed a significant reduction in MCM3 gene expression to 73.6% (+2.2%, −2.2%; Fig. 3C, P<10−3). In order to study the subcellular localization of the MCM3 protein, we next stained fixed FE65 knockdown and control cells using immunofluorescence (Fig. 3D). MCM3 was predominantly localized to the nucleus, and a differing localization in knockdown and control cells was not evident. In a similar fashion, we were able to validate our findings of different BLM protein abundance using immunoblotting (Fig. 3E). Densitometry revealed significant lower BLM abundance in the knockdown cells (Fig. 3F, P<0.05). BLM regulation was also evident at the mRNA level supporting the significant downregulation of BLM in KD cells to 71.5% (+4.3%, −4%; Fig. 3G, P<10−3; higher dCt values indicate downregulation). However in contrast to the MCM3 protein, a significantly different subcellular localization of the BLM protein was observed in the FE65 knockdown cells by immunofluorescence staining (Fig. 3H, lower row, right). Stable control cells (upper row) revealed a nuclear background staining of BLM with distinct nuclear spots (nuclear foci) in all cells (Hoechst stain was used to visualize the nucleus of cells) similar to the phenotype of AICD-, FE65-, TIP60-transfected cells. In contrast, nuclear spots were completely absent in the FE65 knockdown cells (lower row). Instead, a single distinct perinuclear signal was evident potentially localized to the ER.
In order to test whether the ER conformation might be affected by FE65 in general, we used an ER live cell tracker in control and knockdown cells (supplementary material Fig. S1A). The knockdown cells revealed a prominent focused perinuclear ER staining in contrast to the control cells, in which the ER-tracker showed a more uniform pattern around the nucleus.
To further investigate the hypothesis of ER protein localization changes caused by the FE65 knockdown, we investigated the protein PRDX4 (peroxiredoxin 4), a family member of PRDX proteins known to be located in the ER (Tavender et al., 2008). PRDX4 signal was preferentially detected in the ER in control cells (supplementary material Fig. S1B, upper row). Similar to the BLM protein, a prominent change in PRDX4 localization was evident in the fixed knockdown cells with structures compatible to the ER live cell tracker stain. To confirm that these changes in protein distribution were dependent on FE65, we overexpressed a FE65-dsRED construct in the knockdown cells in order to evaluate PRDX4 localization following FE65 reconstitution (supplementary material Fig. S2). As expected, PRDX4 localization in rescued KD cells was similar to control cells (green arrow).
In contrast to MCM3 and BLM, validation experiments for the candidate KAT7 failed (detailed results are given in supplementary material Fig. S3). Taken together, our mass spectrometry based approach is not errorless but was able to identify FE65 knockdown dependently regulated proteins.
FE65 knockdown resulted in a significant decrease in cell proliferation
The redistribution of BLM, which is involved in DNA replication, from the nucleus to the ER may have functional consequences for cell proliferation. In addition, a significantly reduced abundance of MCM proteins in FE65 knockdown cells might impact DNA replication and repair. In order to examine cell proliferation in FE65 knockdown cells, we plated equal numbers of cells in culture dishes and visualized the same field over 96 hours of growth (Fig. 4A). Knockdown cells revealed less proliferation within this timeframe, whereas control cells were confluent after 96 hours. As control and knockdown cells were established with a similar vector background containing a turbo GFP cassette, we next evaluated and quantified cell proliferation using the GFP signal. As outlined in Fig. 4B the GFP fluorescence ratio (control/knockdown) significantly increased over time (96 hours) providing evidence for elevated proliferation of the control cells compared to knockdown cells. The difference in GFP signal between control and FE65 KD after 96 hours of growth was highly significant (Fig. 4B, P<10−8). Finally, we used the xCELLigence System (Roche, Switzerland) to quantify the cell doubling time in control versus knockdown cells (Fig. 4C). Calculated for a timeframe from 12 to 96 hours, the cell index doubling time was significantly higher in FE65 KD than control cells (P<10−4). Since proliferation is accompanied by synthesis of new DNA, we next studied DNA replication using the EdU (BrdU analogue) cell assay. Indeed, knockdown cells revealed less EdU incorporation than control cells (Fig. 4D, right) and Hoechst stain was unaffected in both conditions (Fig. 4D, middle). Collectively, these data indicate that DNA replication/proliferation is significantly reduced in FE65 knockdown cells.
BLM and MCM3 might play a crucial role in FE65-dependent DNA replication/repair mechanisms. Interestingly, the nuclear localization of BLM resembles the nuclear spot-like structures known to originate in AICD-, FE65-, TIP60-transfected cells (von Rotz et al., 2004).
BLM colocalizes with FE65 in nuclear structures and binds to FE65 in co-immunoprecipitation assays
We hypothesized that the similar nuclear staining pattern of BLM and FE65 may be due to a physical interaction between these two proteins. To test this hypothesis, we overexpressed different combinations of AICD, FE65 and TIP60 in HEK293 cells until we obtained the reported nuclear spot phenotype (Müller et al., 2013). As transfection of FE65-EGFP/TIP60 resulted in the largest percentage of cells with nuclear spots, we selected these cells to determine whether FE65 and BLM colocalized. As demonstrated in Fig. 5A, FE65–EGFP/TIP60-positive cells (GFP, green) reveal a clear colocalization with BLM (red), which is also evident in the overlay as well as in the zoom image in Fig. 5C. A similar phenotype is also observed by single overexpression of a BLM-EGFP construct (Fig. 5D) and the colocalization of BLM and TIP60 is also evident in FE65-EGFP/TIP60-transfected cells (Fig. 5B). We next sought evidence for a BLM–FE65 interaction using co-immunoprecipitation (Fig. 5E). Using combined lysates from FE65 and BLM–EGFP expressing cells, we precipitated BLM with anti-BLM or anti-GFP antibody. In both samples FE65 was co-immunoprecipitated. As the control revealed some unspecific BLM interaction to the beads we performed another co-immunoprecipitation for FE65. BLM could be identified in the FE65 IP as well supporting the interaction of both proteins.
BLM and FE65 localize to dynamic nuclear spheres
In order to better understand the nature of the nuclear foci in which FE65 localizes, we re-examined FE65-GFP/TIP60-transfected cells (Fig. 6A). A higher magnification of FE65-positive nuclear foci revealed a ring-like structure, which was also evident in the brightfield microscopy image. The shape of the FE65-positive nuclear structure was further examined with serial digital confocal microscopy sections. The sections from top to bottom (Fig. 6BI) and a profile view image (Fig. 6BII) of the EGFP structure show a hollow sphere. The diameter of the largest spheres were measured and the average size is of 2.5 µm. Microscopy observations of these structures in living cells showed that these sphere-like structures are highly mobile. In order to evaluate the dynamics of these circular objects, we used live cell fluorescence imaging and tracked the GFP signal in FE65/TIP60-transfected cells (Fig. 6C and supplementary material Movie 1). Video microscopy demonstrated that smaller dots are able to grow and fuse to larger compartments appearing as spheres. In addition to the mobility of the nuclear spheres, the different size of these objects in cells was striking. We have classified the FE65–EGFP spheres into two different types, either small (supplementary material Fig. S4, red arrow) or large, with the later showing a clear hollow sphere structure (supplementary material Fig. S4, yellow arrow). Each cell shows similar types of FE65–EGFP-positive spheres, but the average sphere size varies from cell to cell (supplementary material Fig. S4). Finally, we monitored the growth and fusion of different spheres as shown in frames 1 to 4 in Fig. 6C (see also supplementary material Movie 1). The fusion of three different spheres (indicated by the yellow arrow) occurs within 6 minutes (movie was recorded for 3 hours in total). Following fusion, the size of these objects in frame 4 is clearly enlarged in comparison to frame 1. The lower images correspond to upper images processed by the EMBOSS filter in order to further improve image quality by relief transformation. The size of the hollow is also enlarged after sphere fusion. Using electron microscopy we subsequently monitored the nuclear spheres in more detail (Fig. 6D). The first image in the upper row demonstrates large circular structures with a reduced electron density in the center of some structures. Other cells bear a greater number of smaller spheres (middle), whereas control-vector-transfected cells (right) did not show any nuclear sphere-like structures. Our brightfield microscopy data are in agreement with the EM data (lower row). Both types of spheres (small, large with inner hole) are evident in FE65-EGFP-transfected cells (Fig. 6E, left). However, the extent of sphere-positive cells is remarkably increased by co-transfection of FE65 with TIP60 (Fig. 6E, right). The nuclear sphere-like structures demonstrate highly dynamic changes within seconds, growth and fusion within a few minutes as well as different sizes in different nuclei. The identification of BLM as an FE65-binding protein whose nuclear localization and dynamics within the nucleus is FE65 dependent, provides interesting insights into the possible consequences of nuclear FE65.
In the present study, the proteomic analysis of stable FE65 knockdown cells revealed the downregulation of MCM3 and BLM proteins. Furthermore, FE65 and BLM are able to interact in the co-immunoprecipitation assay. In FE65 knockdown cells, BLM localization was shifted from the nucleus presumably to the ER and overexpression of FE65 restored BLM localization to nuclear spheres. These nuclear spheres are able to grow and fuse and are highly dynamic. FE65 knockdown also produced a reduction in cellular proliferation due to an increase in the doubling time possibly due to slower DNA replication resulting from BLM mislocalization.
BLM is a DNA helicase present in nuclear foci that are sites of DNA synthesis during the late S and G2/M phases of the cell cycle (Bhattacharyya et al., 2009). The MCM (minichromosome maintenance) protein family plays an important role in S-phase genome stability in the context of DNA replication, damage and repair (Bailis and Forsburg, 2004; Mincheva et al., 1994). MCM is a hexamer of six related proteins (MCM2–7) forming a nuclear circular structure (Cortez et al., 2004) and loss of MCM function causes DNA damage and genome instability (Bailis and Forsburg, 2004). Both BLM and members of the MCM family have been identified with significantly lower abundance in FE65 knockdown cells suggesting a functional relationship between these proteins and FE65.
The present knowledge for FE65 function includes a role for FE65 in DNA repair, especially following DNA double strand breaks (Minopoli et al., 2007; Stante et al., 2009). However, the molecular mechanism is unknown. Our work suggests that the contribution of FE65 to DNA repair may be mediated by its effects on BLM and MCM proteins, which is further strengthened by the finding that the localization of BLM is changed from a nuclear spot-like phenotype to localization presumably in the ER in the FE65 knockdown cells. In cells lacking FE65, BLM mislocalization likely prevents its participation in important nuclear DNA replication or repair processes (Fig. 7). Our work also demonstrates that BLM colocalizes with FE65 in nuclear sphere-like structures and that the two proteins interact. As BLM contains a PPKP amino acid motif near its C-terminal end, we hypothesize that the interaction occurs at the WW domain in FE65, which is already known to interact with the PPLP motif in Mena (Ermekova et al., 1997). Thus, the reduced BLM protein levels in nuclear spheres appears to result from the downregulation of its interacting protein FE65 and may also be due to reduced BLM expression. FE65 also seems to influence the expression of MCM3 at the mRNA level by an as yet unknown mechanism.
Different microscopic techniques revealed the existence of FE65/TIP60 and BLM in highly mobile nuclear spheres that are able to grow and fuse with each other. Nuclear AICD, FE65, TIP60-dependent spot-like structures have already been described (Konietzko et al., 2010). Here, we show colocalization of the BLM protein with FE65/TIP60 in nuclear foci suggesting that these foci correspond to the nuclear domain 10 (ND10), which is a structure of 0.2 to 1 µm size composed of proteins such as PML (promyelocytic leukemia protein) and BLM (Yankiwski et al., 2001). However, the largest spheres we observed correspond to a diameter of 2.5 µm on average. The discrepancy in the size of these structures might point to an unknown nuclear structure, or a consequence of overexpression of FE65/Tip60 on ND10 foci. The possibility of these structures being different from ND10 is further supported by the finding that BLM is not part of ND10 during DNA synthesis when it colocalizes instead with the Werner syndrome gene product (Yankiwski et al., 2000). Thus, the size as well as the fusion of FE65-positive nuclear spheres may depend on the cell cycle with tiny spheres being part of ND10 and large spheres being separate from ND10. Collectively, our data suggest that nuclear FE65 plays a pivotal role in cellular DNA replication/repair mechanisms mediated by BLM and MCM proteins.
DNA replication and repair appears to be of central relevance in brain development since BLM expression is closely related to neuronal development (Hachiya et al., 2001; Morimoto et al., 2002). FE65 proteins also play an important role during brain development, in cortical plate formation and axonal path finding (Guénette et al., 2006). DNA replication and repair may also be of central relevance to AD pathogenesis. Given that higher nuclear levels of FE65 protein result from T668 phosphorylation of APP and that pT668 APP is increased in AD, higher nuclear FE65 levels may occur in AD and may impair BLM and the MCM protein family function in DNA replication and repair.
Functionality of DNA repair is highly important in AD, where DNA damage by oxidation possibly involving Aβ or reactive oxygen species, belongs to the earliest detectable events during the progression from healthy aging to dementia (Coppedè and Migliore, 2010; Kruman et al., 2002). Notably, the number of BLM foci increases in response to ionizing radiation (Wu et al., 2001), which is able to induce reactive oxygen species (Yamamori et al., 2012), and BLM overexpression exacerbates the sensitivity to DNA damaging agents (Mirzaei et al., 2011). DNA damage is normally repaired by mechanisms involving BLM and the MCM family, responsible for genome replication but also repair. Presently, the relationship between oxidative damage in AD and nuclear FE65 protein levels in neurons is unknown.
Slower proliferation and a reduction in DNA replication were observed in the FE65 knockdown cells. The molecular mechanism may involve lower BLM and MCM protein levels in the nucleus. Putatively elevated nuclear FE65 levels in AD might promote proliferation (and thereby cell cycle re-entry), for which the interaction of BLM with FE65 might be essential. Cell cycle re-entry in neurons occurs in AD (Herrup, 2012; Lopes et al., 2009b) and results in neuronal cell death (Folch et al., 2011). Cell cycle re-entry in AD is characterized by the association of MCM2 with neurofibrillary tangles (Bonda et al., 2009), kinase upregulation, cytoskeletal alterations (Lopes et al., 2009b) and re-expression of cell cycle related proteins like CDK11 (Bajić et al., 2011). The re-entry was also reported not to be induced by Aβ or Tau pathology (Lopes et al., 2009a). Finally tetraploidy, a putative initial result of the re-entry is also a known AD feature (Frade and López-Sánchez, 2010). Notably, BLM colocalizes with TERF2 (telomeric repeat-binding factor 2) in foci actively synthesizing DNA during late S and G2/M phases of the cell cycle, which has been described for immortalized cells using the alternative lengthening of telomere pathway (Bhattacharyya et al., 2009; Yankiwski et al., 2000). TERF2 was also identified as a differentially abundant protein in FE65 knockdown cells supporting a role for FE65 in cell cycle regulation. We hypothesize that elevated or diminished nuclear FE65 levels may contribute to neuronal cell cycle re-entry in AD causing neurodegeneration as neurons are hardly able to proliferate. Thus, the intervention of FE65 nuclear translocation may be a promising therapeutic approach for AD. Targeting FE65 localization or BLM–FE65 interaction might also correspond to a strategy for Bloom syndrome treatment – a disease that is characterized by a high risk of cancer in affected individuals caused by mutations in the BLM gene (Amor-Gueret, 2006).
Materials and Methods
The FE65 knockdown was established by transfection with a GIPZ lentiviral shRNAmir vector (Open Biosystems, USA) in HEK293T cells. For the control state, a non-silencing GIPZ lentiviral shRNAmir vector (Open Biosystems, USA) was used. First 180×104 cells were seeded on 10 cm cell culture dishes (TPP, Switzerland) using five dishes for control and five dishes for the knockdown condition (n = 5). Cells were cultivated at 37°C and 5% CO2 in 10 ml minimum essential medium (MEM; Sigma-Aldrich, USA) supplemented with 1% penicillin/streptomycin and 10% fetal calf serum (FCS; Invitrogen, USA). 48 hours after seeding, HEK293T cells were transfected using Lipofectamine reagent METAFECTENE™ (Biontex, Germany). For each cell culture dish, 36 µl METAFECTENE™ and 36 µg total DNA was used whereby 5 ml MEM was removed before transfection. The transfection procedure was performed according to the manufacturer’s instructions. The hairpin sequences of the applied GIPZ lentiviral shRNAmir vectors (Open Biosystems, USA) are shown in supplementary material Table S1. Puromycin was used to improve transfection efficiency and to get rid of untransfected cells. 24 hours after transfection, the final MEM dish volume was adjusted to 7 ml total MEM and 12 µg/ml puromycin was added in each cell culture dish. Cells were selected for 96 hours followed by MEM filtration through a Filtropur S 0.2 sterile filter (Sarstedt, Germany). The filtered MEM was added again to the selected cells. 12 hours after filtration procedure, cells were washed with 7 ml cold (4°C) GIBCO™ phosphate-buffered saline +/+ (PBS+/+; Invitrogen, USA) to initiate the generation of stable cell lines. Single colonies were picked and transferred to larger cell culture dishes (TPP, Switzerland). Cells were cultivated at 37°C and 5% CO2 in 10 ml MEM supplemented with 1% penicillin/streptomycin and 10% FCS for 1–1.5 months. After this incubation period, cells were washed with 7 ml cold (4°C) PBS+/+ and harvested with 8 ml cold PBS+/+ . 2 ml were used for RNA and 6 ml for protein isolation. Subsequently, PBS+/+ was removed by centrifugation at 200×g at 4°C. For protein isolation, cell pellets were sonicated in DIGE buffer [7 M urea, 2 M thiourea, 2% CHAPS, 130 mM dithiothreitol (DTT), 30 mM Tris-HCl, pH 8.5] and the cellular extracts were centrifuged at 15.500 g for 15 minutes at 4°C. The supernatants were used for subsequent experiments and the protein concentration was determined by amino acid analysis (ASA) on an HPLC Alliance 2695 instrument (Waters, USA).
Quantitative PCR (qPCR)
To quantify mRNA levels of FE65, FE65L1 and FE65L2, SYBR Green real-time PCR assays were performed on a RotorGene RG-3000 device (Corbett Life Science, Australia). RNA lysates from cell culture were gained using the NucleoSpin RNA II kit (Machery Nagel, Germany) according to the manufacturer’s protocol. Template cDNA was synthesized from 2 µg total RNA using the RevertAid™ First Strand cDNA Synthesis Kit (Thermo Scientific, USA) and random hexamer primers following manufacturer’s instructions. Cycling conditions were 95°C for 15 minutes, followed by 45 cycles of 95°C for 15 seconds, 56°C for 30 seconds, and 72°C for 30 seconds. The dCt values were calculated using GAPDH as control. Experiments were performed in triplicates for each clone analyzed. Melting curve analysis confirmed that only one product was amplified. For statistical analysis of all quantitative PCR experiments, normal distribution of data was assured and remaining gene expression was calculated by the method of Livak and Schmittgen (Livak and Schmittgen, 2001). For quantification of mRNA levels of MCM3, BLM and KAT7, the same qPCR procedure was used. All sense and antisense primer sequences used in qPCR experiments are shown in supplementary material Table S2.
Total protein was separated by SDS-PAGE using 4–12% NuPAGE™ Bis-Tris gels (Invitrogen, USA) and proteins were transferred to nitrocellulose. The Odyssey Infrared Imaging System (LI-COR Biosciences, USA) was used for protein detection. For the FE65 blot, 25 µg total protein were used for each lane, the blot was probed with a FE65 antibody (cat. no: sc-33155; Santa Cruz Biotechnology, USA) and an antibody dilution of 1∶200 in StartingBlock™ TBS blocking buffer (Thermo Scientific, USA) was applied. The MCM3 blot was probed with a MCM3 antibody (cat. no: ab4460; Abcam, UK) using a dilution of 1∶3000 in StartingBlock™ TBS blocking buffer and 20 µg total protein. The BLM blot was incubated with a BLM antibody (dilution factor 1∶1000, 20 µg total protein; cat. no: ab2179; Abcam, UK). The KAT7 blot was probed with a KAT7 antibody (dilution factor 1∶200 in StartingBlock™ TBS blocking buffer, 20 µg total protein; cat no: ab37289; Abcam, UK). Additionally, all immunoblots were incubated with a β-actin antibody (dilution factor 1∶10000; cat. no: A1978; Sigma Aldrich, Germany). The β-actin signal intensity was used for normalization. Secondary antibodies were used as follows: IRDye™ 800CW antibody (dilution factor 1∶15000; cat. no: 926-32211; LI-COR Biosciences, USA). IRDye™ 680CW antibody (dilution factor 1∶15000; cat. no: 926-32220; LI-COR Biosciences, USA). Quantification was carried out by densitometry with Odyssey Application Software version 3.0.21 (LI-COR Biosciences, USA).
For immunofluorescence microscopy, cells were washed with PBS+/+, fixed with 4% paraformaldehyde (PFA), permeabilized with standard 0.5% Triton X-100 and incubated with the corresponding antibody as follows: MCM3 antibody (dilution factor 1∶50; cat. no: ab4460; Abcam, UK); BLM antibody (dilution factor 1∶50; cat. no: ab2179; Abcam, UK); KAT7 antibody (dilution factor 1∶50; cat. no: ab37289; Abcam, UK), KAT5/TIP60 antibody (1∶50; cat. no: sc-5727, Santa Cruz, USA), secondary TRITC antibody (dilution factor 1∶200; cat. no: T5268 or T6028; Sigma-Aldrich, USA), secondary DyLight405 (1∶200; cat. no: 3063-1; Epitomics, USA); nuclear visualization was carried out with Hoechst Staining. Images were obtained using the fluorescence microscope IX51 (Olympus Microscopy, UK). To confirm presumed endoplasmic reticulum (ER) signals, a commercially available ER-Tracker™ Blue-White DPX (cat. no: E-12353; Invitrogen, USA) was used. Cell preparation and ER live-cell staining was performed according to the manufacturer’s instructions. For confocal laser scanning microscopy, HEK293T cells, 48 hours after seeding, were transfected with FE65-GFP/TIP60 (equal-molar mixture of pFE65-EGFP and p2N3T-TIP60 vectors), pBLM-EGFP vector, or pFE65 (without EGFP fusion) vector using Lipofectamine reagent METAFECTENE™ (Biontex, Germany) according to the manufacturer’s instructions. FE65-GFP/TIP60-transfected cells were analyzed by confocal laser scanning microscopy (LSM 510, Zeiss) in combination with Zeiss 63× (Plan-Neofluar, NA 1.4) oil immersion lenses. Time-lapse imaging was done for 3 hours, frames were taken every 2 minutes. Progress of moving nuclear spheres was monitored by phase-contrast optics using a phase-contrast microscope (Zeiss) equipped with a CCD camera (Princeton Instruments, USA) and Metamorph image software (Vistron, Germany). For transmission electron microscopy, specimen were fixed in 2.5% GA in PB, postfixed in OsO4, dehydrated, embedded in Epon and examined (Philips EM 410).
Mass spectrometry, spectral count based label-free proteomics
Total cell lysates (20 µg each lane) were separated by SDS-PAGE with 4–12% NuPAGE™ Bis-Tris Gel (Invitrogen, USA) and stained with Imperial™ Protein Stain (Thermo Scientific, USA). Following destaining, the complete gel was reduced using DTT and alkylated with iodacetamide (IAA). The gel was cut into three horizontal slices and afterwards every lane was excised (compare Fig. 2A). In-gel digestion was performed overnight at 37°C with trypsin (Promega, USA) in 10 mM HCl and 50 mM ammonium hydrogen carbonate (NH4HCO3) at pH 7.8. Resulting peptides were extracted once with 100 µl of 1% FA, and twice with 100 µl of 5% FA, 50% ACN. Extracts were combined and ACN was removed in vacuo. For LC-MS analysis, a final volume of 40 µl was prepared by addition of 0.1% TFA.
Nano-HPLC-MS/MS was performed on an UltiMate 3000 RSLCnano LC system (Dionex, Idstein, Germany). Samples were loaded on a trap column (Dionex, 75 µm×2 cm, particle size 3 µm, pore size 100 Å) with 0.1% TFA (flow rate 10 µl/minute). After washing, the trap column was connected with an analytical C18 column (Dionex, 75 µm×25 cm, particle size 2 µm, pore size 100 Å). Peptides were separated with a flow rate of 400 nl/minute using the following solvent system: (A) 0.1% FA; (B) 84% ACN, 0.1% FA. In a first step, a gradient from 5% B to 40% B (95 minutes) was used, followed by a second gradient from 40% B to 95% B within 5 minutes and finally a gradient from 95% to 5% B within 25 minutes. ESI-MS/MS was performed on a LTQ Orbitrap Velos (Thermo Fisher Scientific), which was directly coupled to the HPLC system. MS spectra were scanned between 300 and 2000 m/z with a resolution of 30,000 and a maximal acquisition time of 500 ms. The m/z values initiating MS/MS were set on a dynamic exclusion list for 35 seconds. Lock mass polydimethylcyclosiloxane (m/z 445.120) was used for internal recalibration. The 20 most intensive ions (charge >1) were selected for MS/MS-fragmentation in the ion trap. Fragments were generated by low-energy collision-induced dissociation (CID) on isolated ions with collision energy of 35% and maximal acquisition time of 50 ms.
Data processing and database searches
Raw files were transformed to *.mgf-files (ProteomDiscoverer 1.3, Thermo Scientific Fisher), imported in ProteinScape™ (version 2.1, Bruker Daltonics, Bremen, Germany), and analyzed using Mascot (Matrixscience, London, UK) with a peptide mass tolerance of 10 ppm and a fragment mass tolerance of 0.5 Da. Searches were performed allowing one missed cleavage site after tryptic digestion. Carbamidomethylation (C), oxidation (M), and phosphorylation (S,T,Y) were considered as variable modifications. All data were searched against a database created by DecoyDatabaseBuilder (Reidegeld et al., 2008) containing the whole human ipi (release 2011/06, v3.84 human, 90166 entries) with one additional shuffled decoy for each protein.
After peptide identification, an algorithm using a given minimal peptide score (minPepScore) and a minimal number of peptides per protein (minNrPeps) was applied. The algorithm performs the following steps: score calculation for all proteins by adding up the Mascot ion scores of the protein’s peptides, which have a score of at least minPepScore. Here, a peptide is defined by an amino acid sequence and its modifications. If two peptides are equal except for the score, only the higher score is taken. Reporting the highest scoring protein group (a group consists of all proteins in the database containing the same set of identified peptides) which has at least minNrPeps peptides not yet flagged as used up and flag all the peptides of the reported proteins as used up. Repetition of step 2 was done until no more protein groups get reported. A local false discovery rate (FDR) was calculated for each protein group, regarding a group as decoy, if it consists of decoy proteins only. With this strategy the minPepScore was calculated, which yielded the list with the most target (opposed to decoy) groups beneath an FDR-threshold of 5%. For the given data a minNrPeps of 2 was used to exclude ‘one hit wonders’, which yielded a minPepScore of 22 for the longest list. Among the proteins in this list, every peptide spectrum match (PSM) was extracted.
These PSMs were further processed using the Pivot table function of Microsoft Excel resulting in a table representing spectral counts for every peptide belonging to a certain protein. Processed spectral counts (PSC) based on spectral and peptide counts were calculated as described previously (22,23) and subsequently used as basis for label-free quantification. In brief, PSC calculation was performed by summing up all spectral counts belonging to the respective protein. To identify differentially expressed proteins, the ratio between the averaged spectral indices of the knockdown samples and controls was calculated and Student’s t-test was conducted for each protein. In order to control the FDR, the resulting p-values were adjusted for multiple testing according to Benjamini and Hochberg (Benjamni and Hochberg, 1995). However, adjusted P-value calculation was inappropriate for our approach as the number of identified proteins (which is a determining factor for the correction calculation) is extremely large in our study caused by the use of a sensitive MS instrument. Thus, all proteins reported to be relevant in this work have been validated by independent methods. Initially, a significant t-test (<0.05) and a spectral index ratio >1.8 was used to assign a protein as potential candidate for subsequent experiments.
The pathway analysis was done with two independent softwares, IPA 9.0 (Ingenuity Pathway Analysis, Ingenuity Systems, USA, http://www.ingenuity.com) and STRING 9.0 (Search Tool for the Retrieval of Interacting Genes/Proteins, EMBL Institute, Europe, http://string-db.org) according to our experience of pathway tools in proteomics (Müller et al., 2011c).
Cell proliferation and DNA replication assay
At first, cell proliferation was examined visually over 96 hours using a 96-well microplate format. Equal well position was ensured by centering the black shading visible in the mid of wells in brightfield microscopy (Fig. 4A). Simultaneously GFP fluorescence of control and knockdown cells (both stable clones contain a turbo GFP cassette) was monitored from 12 to 96 hours with the Infinite™ 200 PRO device (excitation at 395 nm, emission at 509 nm, Tecan Group, Switzerland) and plotted with a Boltzmann sigmoidal fit. To determine cell index doubling times, we used the xCELLigence System (Roche, Switzerland). 5000 cells were seeded on an xCELLigence System microplate and cell index doubling times was monitored from 12 to 96 hours. DNA replication assay was done by EdU labeling (BrdU analogue) with the Click-iT EdU Alexa Fluor 594 Imaging Kit (Invitrogen, USA) according to manufacturer’s protocol with Hoechst counterstaining.
We thank Prof. T. Russo (Dipartimento di Biochimica e Biotecnologie Mediche, Università di Napoli Federico II, Italy) for FE65 expression vectors and Prof. J. Groden (Department of Molecular Virology, Immunology and Medical Genetics, Ohio State University, USA) for providing the BLM expression vector. We thank Prof. Suzanne Guénette from MGH, Harvard for her critical revision of the manuscript.
A.S. performed the experiments assisted by T.M., F.M.N., M.N., C.L., K.P., and F. El M.; T.M. designed the experiments, analyzed the data and wrote the manuscript. All authors discussed the data.
This work was funded by Forschungsförderung Ruhr-Universität Bochum Medizinische Fakultät [grant numbers AZ-F616-2008 and AD F680-2009 to T. Müller]; and the federal state of North Rhine-Westphalia within the Protein Research Unit Ruhr within Europe project.