|Home | About | Journals | Submit | Contact Us | Français|
Senile plaques consisting of β-amyloid (Aβ) and neurofibrillary tangles composed of hyperphosphorylated tau are major pathological hallmarks of Alzheimer’s disease (AD). Elucidation of factors that modulate Aβ generation and tau hyperphosphorylation is crucial for AD intervention. Here we identify a novel mouse gene Fg01 that originated through retroposition of ribosomal protein S23. We demonstrate that FG01 protein reduces the levels of Aβ and tau phosphorylation by interacting with adenylate cyclases to activate cAMP/PKA and thus inhibit GSK-3 activity. The function of Fg01 is demonstrated in cells of various species including human, and in transgenic mice overexpressing FG01. Furthermore, the AD-like pathologies of triple transgenic AD mice were improved and levels of synaptic maker proteins increased after crossing them with Fg01 transgenic mice. Our studies reveal a new target/pathway for regulating AD pathologies and uncover a novel retrogene and its role in regulating protein kinase pathways.
Alzheimer’s disease (AD) is featured by extracellular neuritic plaques, intracellular neurofibrillary tangles (NFTs), synaptic dysfunctions and neural degeneration in vulnerable brain regions (Tanzi and Bertram, 2005). Neuritic plaques are composed of aggregates of heterogeneous β-amyloid (Aβ) peptides, which are derived from β-amyloid precursor protein (APP) through sequential cleavages by β-secretase (BACE1) and the γ-secretase complex (consisting of at least four components: presenilin, nicastrin, APH-1 and PEN-2) (Cole and Vassar, 2007; De Strooper, 2003; Zhang and Xu, 2007). Multiple lines of evidence suggest that overproduction/aggregation of Aβ in the brain is a causative factor for AD pathogenesis (Hardy and Selkoe, 2002). NFTs are composed of hyperphosphorylated microtubule associated protein tau (Buee et al., 2000; Lee et al., 2001). Numerous studies have shown that pathogenic APP metabolism/Aβ generation and tau phosphorylation are highly regulated via various signal transduction pathways, e.g., protein kinases and phosphatases (Buxbaum et al., 1994; Fang et al., 2000; Xu et al., 1996) and steroid and peptide hormones (Gasparini et al., 2001; Xu et al., 1998). Among these regulatory pathways, glycogen synthase kinase-3 (GSK-3, α and β isoforms), a serine/threonine kinase essential for a variety of cellular functions including cell adhesion, cell-division, transcription (Frame and Cohen, 2001), has been demonstrated in regulating both Aβ generation and tau phosphorylation (Flaherty et al., 2000; Phiel et al., 2003). This unique feature renders manipulation of GSK-3 activity an attractive therapeutic approach for AD (Frame and Cohen, 2001; Martinez et al., 2002; Medina and Castro, 2008). Hence identification of new genes involved in these processes will be instrumental in developing novel AD therapeutics.
The creation of genetic novelty by the formation of new genes has an important role in evolution. New genes can originate through different mechanisms that include exon shuffling, gene duplication, gene fusion/fission, mobile element integration, lateral gene transfer, and retroposition (Long et al., 2003). Retroposition is a process in which a parental mRNA is reverse-transcribed and inserted into the organism’s genome, creating duplicate genes in new genomic positions (Hollis et al., 1982; Karin and Richards, 1982; Ueda et al., 1982). Although these intronless retroposed gene copies commonly lack the regulatory elements of parental genes and thus rountinely have been classified as processed pseudogenes (Jeffs and Ashburner, 1991; Mighell et al., 2000; Zhang et al., 2004), occasionally, these retroposed gene copies can recruit regulatory elements as well as protein-encoding sequences at or near the retroposition site and become expressed and functional (Babushok et al., 2007; Kaessmann et al., 2009; Long et al., 2003; Vinckenbosch et al., 2006). Nevertheless, studies to elucidate the functions of these newly originated genes, especially the functions related to diseases, are limited (Kaessmann et al., 2009; Vinckenbosch et al., 2006).
Random Homozygous Gene Perturbation (RHGP, previously called Random Homozygous Knockout, RHKO) is a genome-wide genetic approach that identifies genes based on their biological functions (Li and Cohen, 1996; Liu et al., 2000a; Liu et al., 1999; Liu et al., 2000b). The design of RHGP enables the inactivation of both alleles of randomly addressed chromosomal genes within populations of mammalian cells using gene search vector cassettes that contain a regulated antisense promoter. This strategy has been used successfully to identify genes whose functional homozygous inactivation leads to reversible tumorigenesis (Li and Cohen, 1996; Liu et al., 2000a; Liu et al., 1999; Liu et al., 2000b), or altered sensitivity to chemotherapeutic agents (Lih et al., 2006).
Here using the RHGP approach, we identified a novel gene Fg01 that originated through retroposition of the mouse ribosomal protein S23 (Rps23) mRNA. The Fg01 gene is reversely transcribed relative to its parental gene, expressing a structurally unrelated, yet functional protein FG01. More importantly, we demonstrated both in vitro and in vivo that overexpression of the FG01 protein decreases the levels of Aβ and tau phosphorylation and increases synaptic marker proteins in AD transgenic mice, by inhibiting GSK-3 activity via the adenylate cyclase/protein kinase A (PKA) pathway.
It has been shown that reduction of Aβ levels is accompanied by cell surface accumulation of APP βCTF (the product of β-cleavage and immediate substrate for γ-cleavage), which is readily detectable in cells deficient in PS1 (Chen et al., 2000); and these cells can be identified using an antibody specifically recognizing the N-terminus of APP βCTF (FCA18) (Ancolio et al., 1999). Based on this observation, we adapted RHGP as a high throughput screen to search for genes that regulate Aβ generation.
We modified the original vector pLLGSV (Li and Cohen, 1996) to increase efficiency of retroviral integration and gene recovery. The new RHGP gene search vector contains modified LTRs and utilizes the Cre-LoxP mediated recombination to minimize promoter interference in provirus and to facilitate genomic DNA cloning (Figure 1A). This vector was transfected into Phoenix-Ampho cells for viral packaging. Harvested retrovirus was used to infect mouse neuroblastoma N2a cells stably expressing the human APP Swedish mutation (N2aSwe). After random insertion, the provirus (Figure 1B) expressed Cre recombinase for recognition and recombination of the two LoxP sites located in the 5’LTR and 3’LTR, respectively, generating the final integrated provirus (Figure 1C). A tetracycline regulated promoter (TRE-CMV) promoter in the final integrated provirus drives the expression of Pac for puromycin selection and initiates transcription into flanking chromosomal gene that can either overexpress, when TRE-CMV is in the same orientation, or suppress (by expressing antisense transcripts) when TRE-CMV is in the opposite orientation relative to the flanking gene. Moreover, transcription of the tetracycline-regulated (tet-off) transactivator was reversed in the presence of tetracycline (or doxycycline).
Therefore, we first acquired N2aSwe cells with stable integration of the RHGP search vector by puromycin selection. These cells were live-immunostained with fluorescently labeled FCA18 antibody followed by multiple rounds of FACS sorting to enrich for cells showing surface APP βCTF accumulation. Less than 0.01% of cells showing βCTF accumulation after first round of FACS sorting were enriched up to 75% following another two rounds of FACS sorting (Figure 2A). These cells were then sorted with additional FACS in the presence of doxycycline for a reversion of cell surface APP βCTF to background level to eliminate potential false positive. The final sorted cells were cloned individually, propagated and assayed for both accumulated cell surface APP βCTF and reduced Aβ generation. This screening strategy is shown in Figure 1D.
One clone, FG01, was isolated for the high level of cell surface APP βCTF and the significant reduction of Aβ secretion (Figures 2B and 2C). Both phenotypes were reversed by doxycycline treatment, validating that the effects were indeed a result of RHGP rather than cell-cell/clonal variation in APP/βCTF expression or any random mutagenesis (Figure 2C). Subcloning and sequence analyses revealed that the RHGP vector was inserted into chromosome 8 at a site ~1.2Kb upstream of the C330021F23Rik gene (GenBank ID: 546049), which has no known function. We herein designated this gene as Fg01. The upstream location and the same orientation of the inserted RHGP vector strongly suggested that Fg01 was likely overexpressed in this cell clone. This notion was supported by real-time reverse transcription-PCR (RT-PCR) using RNAs from parental N2aSwe and the FG01 RHGP cell clone, which showed that Fg01 was overexpressed in the FG01 RHGP cell clone and its overexpression was reversed by doxycycline treatment (Figure 2D).
The Fg01 gene is predicted to encode a 141 amino acid-long hypothetical protein that we designate FG01. Interestingly, analyses of multiple genome databases (GenBank, UCSC Genome Browser and Ensemble Genome Browser) with the FG01 protein sequence identified no FG01 homologs in other species including humans and rats. Further analysis with the Fg01 gene sequence showed that the predominant protein-encoding region of the Fg01 gene was highly homologous to the reverse and complementary sequence of the mouse ribosomal protein S23 (Rps23) mRNA (Figure 3A). The similarity between Fg01 and mouse Rps23 was even higher than those between mouse Rps23 and rat or human Rps23 (Figure 3B), suggesting that Fg01 originated from mouse Rps23 after the divergence of mice and rats. New genes can originate through different mechanisms (Long et al., 2003). However, the presence of the mouse Rps23 untranslated regions (UTRs) and the absence of the mouse Rps23 introns in the homologous regions between Fg01 and mouse Rps23 clearly suggest that Fg01 originated through retroposition of the mouse Rps23 mRNA, which recruited regulatory units and additional protein-encoding sequence near the retroposition site. But transcription of Fg01 is reversed compared to Rps23. To search for human homologs of Fg01, we scanned the human genome with the human Rps23 cDNA sequence and identified several Rps23 retroposition sites (Figure S1). However, computational gene prediction of these sites revealed no functional Fg01-like genes. We also carried out RT-PCR with primers binding regions right next to these human Rps23 retroposition sites and failed to obtain positive amplification (data not shown).
Bioinformatics analysis using the FG01 amino acid sequence predicted a helical transmembrane domain near the C-terminus but no obvious signal peptide sequence. We constructed a vector expressing recombinant FG01 with a Myc tag at the N-terminus and a His6 tag at the C-terminus (Myc-FG01-His6, Figure 4A). Both Myc and His6 antibodies recognized a product of approximately 17 kDa in transfected Myc-FG01-His6 cells, consistent with the predicted molecular weight, indicating that there is no cleavable signal peptide sequence within FG01 (Figure 4B). Furthermore, after transfection of Myc-FG01-His6 vector into N2a cells, fractionation of cell lysates into cytosolic and membrane components indicated that the majority of FG01 protein was located in membrane fractions (Figure 4C). Biotinylation assays also revealed that FG01 was delivered to the cell surface (Figure 4D). To determine FG01 topology, we transfected N2a cells with the Myc-FG01-His6 vector and immunostained either live cells or cells after permeabilization, using antibodies against Myc or His6. Our results show that although both antibodies were immunoreactive in permeabilized cells, only the Myc antibody positively stains the membranes of live cells, whereas the His6 antibody does not, suggesting that the FG01 N-terminus is extracellular (Figure 4E). Hence these results suggest that FG01 is a type Ib transmembrane protein that has a normal type I transmembrane protein orientation but no signal peptides. Immunoprecipitation combined with live-immunostaining also confirmed the type Ib transmembrane topology of FG01 (Figure S2). We derived an antibody against the N-terminus of FG01 (Figure S3) and using this antibody for immunoprecipitation followed by Western blot analysis, we confirmed expression of FG01 in both cortex and hippocampus (Figure 4F). In situ hybridization analysis of Fg01 expression in mouse brain revealed that Fg01 is indeed expressed primarily in hippocampus, dentate gyrus, and cortex (Figure 4G).
The FG01 RHGP cell clone exhibits increased APP βCTF accumulation and reduced Aβ levels (Figure 2C), suggesting that FG01 regulates APP processing. To corroborate this, we overexpressed FG01 in N2aSwe cells. The results showed that, although the levels of total APP were not affected, the levels of extracellular and intracellular Aβ were significantly reduced by FG01 overexpression, and the levels of βCTF and sAPPα were significantly increased (Figure 5A). When human HeLa cells stably expressing the human APP Swedish mutation (HeLaSwe) were transfected with mouse FG01, we also observed reduced Aβ levels and increased accumulation of βCTF and sAPPα (Figure 5A). ELISA analysis confirmed that both Aβ40 and Aβ42 levels were significantly reduced following FG01 expression in HeLaSwe cells (Figure 5B). These data demonstrated that mouse FG01 can function not only in mouse cells, but also in human cells.
To determine whether FG01 reduces Aβ levels by modulating β-secretase activities, we examined β-secretase (BACE1) activity and the protein level of BACE1 in vitro in FG01-overexpressing cells and found them both unchanged (Figures S4A and S4B). APP βCTF accumulation can also be attributed to a decrease in γ-secretase-mediated cleavage. However, cleavage of Notch, another important γ-secretase substrate (Kopan and Goate, 2000), was not altered by FG01 overexpression (Figure S4C). Protein levels of nicastrin, an important component of the γ-secretase complex, and ADAM10 and TACE, two putative α-secretases (Zhang and Xu, 2007), were also not affected by FG01 overexpression (Figure S4B).
Although the possibility that FG01 may modulate substrate specificity or accessibility to γ-secretase can not be excluded, it is equally possible that FG01 regulates other proteins/pathways functioning in APP processing and Aβ generation. One of those could be GSK-3, which has been shown to affect Aβ generation (without affecting γ-secretase-mediated Notch cleavage) and tau phosphorylation (Phiel et al., 2003; Takashima et al., 1995). Thus, we studied the activities of GSK-3α/β in the presence or absence of FG01 overexpression by examining levels of phospho-GSK-3α (Ser 21) and phospho-GSK-3β (Ser 9) (which represent inactivated forms of GSK-3) and by in vitro kinase assays. In cells overexpressing FG01 we observed that GSK-3α/β was more highly phosphorylated at these sites in mouse N2aSwe (Figure 5A), rat PC12 (data not shown), and human HeLaSwe (Figure 5A) and HEK293 (data not shown) cells, indicating that FG01 can decrease GSK-3 kinase activity in cells of various types and species including human. In vitro kinase assays also demonstrated an approximately 50% decrease in GSK-3α and a 40% decrease in GSK-3β activities in FG01-overexpressing cells (Figure 5C). Remarkably, in the presence of lithium, a general inhibitor of both GSK-3α and GSK-3β, FG01 overexpression could not further reduce GSK-3α/β activity or Aβ production (Figure 5D). These data suggest that FG01 reduces Aβ levels by downregulating GSK-3 activity.
GSK-3 is a major kinase that phosphorylates tau in AD (Flaherty et al., 2000). Hence, we asked whether FG01 affects tau phosphorylation by inhibiting GSK-3. We cotransfected N2aSwe cells with the human tau splice variant T40 and with FG01 (or control vectors), and examined tau phosphorylation at the threonine 205 site (pT205) and the PHF-1 tau epitope sites (serine 396 and serine 404), which are GSK-3 phosphorylation targets and the major paired helical filament (PHF) sites found in NFTs. Overexpression of FG01 significantly decreased tau phosphorylation at these GSK-3 target sites and increased unphosphorylated tau levels without affecting total tau levels (Figure 5E). These results suggest a role for FG01 in reducing tau phosphorylation, in addition to its effect on Aβ levels. We also analyzed protein levels and activity of CDK5, another kinase mediating tau phosphorylation in AD (Flaherty et al., 2000), following FG01 overexpression and observed little change, suggesting that CDK5 is not involved in FG01-regulated tau phosphorylation (Figure S5).
Inhibition of GSK-3 activity via phosphorylation of serine 21 in GSK-3α and serine 9 in GSK-3β can be mediated by protein kinase A (PKA) (Fang et al., 2000), so we studied whether FG01 regulates PKA activity. In vitro kinase assays revealed that FG01-transfected cells had significantly more PKA activity than control cells (Figure 6A), consistent with the observation that FG01 overexpression increased phosphorylation of CREB, a PKA substrate (Figure 6B). In addition, FG01 failed to inhibit GSK-3 activity and Aβ generation when PKA activity was suppressed by the specific inhibitor H89 (Figure 6B). Furthermore, downregulation of endogenous FG01 expression in N2aSwe cells by RNA interference (Figures 6C and 6D) dramatically reduced CREB phosphorylation, and increased GSK-3 activity and Aβ generation (Figure 6D). These data indicate that FG01’s effects on GSK-3 activity and Aβ levels require PKA activation. Moreover, increased sAPPα secretion upon FG01 overexpression (Figure 5A) is also likely due to PKA activation, because PKA can stimulate budding of APP-containing vesicles from the Trans-Golgi Network (TGN) to cell surface, the major site for APP cleavage by α-secretase, therefore facilitating sAPPα generation (Xu et al., 1996).
Since cAMP binds to and activates PKA (Taylor et al., 2008), we investigated whether FG01 overexpression had any effect on cAMP levels. We found that FG01 overexpression significantly increased cAMP levels in both mouse N2a (38%) and rat PC12 (75%) cells (Figure 6E). We next examined potential interaction between FG01 and adenylate cyclases, enzymes responsible for cAMP synthesis (Kamenetsky et al., 2006). Co-immunoprecipitation studies showed that FG01 interacts with both overexpressed (data not shown) and endogenous (Figure 6F) adenylate cyclases in N2a cells overexpressing FG01, suggesting a possible modulation of enzymatic activity for cAMP production.
To confirm that it is indeed FG01 protein rather than Fg01 mRNA that mediates these effects, we constructed an Fg01 cDNA mutant vector with a stop codon at the beginning of the protein-coding region. Cells transfected with this mutant vector showed mRNA expression (detected by RT-PCR) but no protein expression of the mutant Fg01 (Figure S6). In addition, overexpression of this mutant Fg01 did not affect the activity of PKA or GSK-3, or Aβ levels (Figure S6), excluding any potential RNA interfering effects arising from anti-sense interaction with Rps23 RNA.
To validate FG01 function in vivo, we generated an Fg01 transgenic mouse model specifically overexpressing Myc-tagged FG01 in the brain. A transgenic expression cassette driven by the human Thy-1 promoter (Figure S7A) was microinjected into C57Bl6 mice and we used primers specifically amplifying exogenous Fg01 to genotype transgenic mice (Figure S7B). Reverse transcription-PCR revealed that mRNA of the exogenous gene was indeed expressed in transgenic mouse brains (Figure S7C). Protein expression of exogenous Myc-tagged FG01 was also confirmed in brain tissues of transgenic mice (Figure S7D). In addition, immunoprecipitation/Western blot showed that total (including exogenous and endogenous) protein levels of FG01 in the transgenic mice were about two-fold higher than endogenous FG01 levels in control mice (Figure S7E). We generated two mouse lines with similar FG01 expression levels and results obtained from the two lines (including their crossing with 3XTg mice as described below) were similar. Herein we only presented results from line 2. Our results showed that levels of phosphorylated and therefore inactive GSK-3α/β were increased (Figure S7D), accompanied by reduced GSK-3β activity in Fg01 transgenic mouse brain (Figure S7F). Increased CREB phosphorylation indicative of upregulated PKA activity, as well as decreased phosphorylation of endogenous mouse brain tau, was also seen in Fg01 transgenic mouse brain (Figure S7D). In addition, preliminary observation detected no obvious aberrant behavioral phenotypes in Fg01 transgenic mice (data not shown).
We next crossed Fg01 transgenic mice with triple transgenic (3XTg) AD mice harboring mutations in human App (APP), Mapt (tau) and Psen1 (presenilin 1) genes (Oddo et al., 2003). As expected, FG01 overexpression in 3XTg mice dramatically increased the levels of cAMP and PKA activity (Figure S8A), resulting in elevated CREB activity, and reduced GSK-3 activity, tau phosphorylation, and Aβ levels in mouse brains (Figures 7A and S8B). Consistently, the numbers of both Aβ-immunostaining-positive (by Aβ40-specific antibody and 6E10 antibody) and phosphorylated tau-immunostaining-positive (by PHF-1 and pT205 antibodies) neurons in the 3XTg mouse brain (both hippocampus and cortex) were significantly decreased following FG01 overexpression (Figures 7B, 7C, 7E, S8C, S8D, and S8F). Interestingly, protein levels of the synaptic marker PSD-95 were markedly increased following FG01 overexpression in 3XTg mouse brains (Figure 7A). Immunostaining of PSD-95 (Figures 7D and 7E) and synapsin (Figures S8E and S8F), another synaptic marker, also revealed significantly higher immunoreactivity in the hippocampus of 3XTg mouse brains with FG01 overexpression than that seen in 3XTg mice without FG01. These results imply that FG01 may rescue synapse impairment seen in 3XTg mice (Oddo et al., 2003), in addition to, or as a consequence of, its effects on reducing Aβ generation and tau phosphorylation. Consistent with the results found in cell cultures (Figure S4B), protein levels of ADAM10 and TACE were not affected by FG01 overexpression in the brain of 3XTg mice (Figure 7A).
Caloric restriction and environmental enrichment have been shown to reduce AD-like pathologies and behavior deficits in animal models (Halagappa et al., 2007; Lazarov et al., 2005). Since elevated CREB activity in the brain may affect animal behaviors such as food intake and daily activity, we compared body weight of FG01-overexpressing mice to that of control mice and found no difference at 3 months of age. At 7 and 11 months of age, we noticed that FG01-overexpressing mice are slightly (but not significantly) lighter than controls (data not shown), even though visual inspection of daily activity and food intake between these mice showed no obvious differences. Therefore, there is a possibility that FG01 exerts its effect on alleviating AD-like pathologies by altering mouse behaviors and this possibility deserves further investigation.
Using the RHGP assay to screen for genes involved in regulating Aβ generation, we identified the functional retroposed Fg01 gene on mouse chromosome 8. The FG01 protein is a type Ib transmembrane protein and is expressed in the brain. In the present report, we provide compelling evidence to show that FG01 overexpression can reduce both Aβ levels and tau phosphorylation, two major pathological hallmarks of AD. We also reveal the underlying mechanism, i.e. FG01 interacts with adenylate cyclases to upregulate cAMP levels, which activates PKA activity, hence inhibiting GSK-3 activity, tau phosphorylation and Aβ generation. These results elucidate an important link between adenylate cyclases and AD, which had not been illustrated previously.
Sequence analyses demonstrated that Fg01 originated through retroposition of mouse Rps23, which recruited regulatory units and additional protein encoding fragments at the retroposition site and became functional (Figure 3A). The reversal in transcriptional direction of Fg01 relative to the parental Rps23 gene explains why there is no protein sequence similarity between FG01 and RPS23. Rps23 belongs to the ribosomal protein family and is highly conserved among species (Hori et al., 1993). Since human ribosomal protein genes have been found to generate a large number of processed pseudogenes through retroposition (Zhang et al., 2002), there is a possibility that human Rps23 may have also retroposed in humans and generated new functional genes with orientations and functions similar to that of Fg01. However, although we indeed identified several human Rps23 retroposition sites in the human genome (Figure S1), neither computational gene prediction nor RT-PCR with primers binding adjacent regions of these human Rps23 retroposition sites revealed any Fg01-like genes (data not shown), suggesting that the possibility that humans possess functional Fg01 homologs is low.
During aging, humans are susceptible to AD pathogenesis, typically characterized by Aβ overproduction/aggregation and tau hyperphosphorylation. In contrast, wild type mice rarely develop AD pathologies (De Strooper et al., 1995; Jankowsky et al., 2007; Johnstone et al., 1991). The differences in AD susceptibility between humans and mice have been attributed to the sequence disparity between human and mouse Aβ (and possibly tau) that underlie different aggregation properties (De Strooper et al., 1995; Jankowsky et al., 2007; Johnstone et al., 1991), to the short lifespan of mice relative to humans (Jankowsky et al., 2004; Jankowsky et al., 2007), and to the differences in processing of human and mouse APP by BACE1 (Cai et al., 2001). Should humans lack Fg01 homologs, our results would provide an alternative explanation, i.e. some genetic factors in mice, such as Fg01, protect them against an AD-like disease by preventing Aβ over-production and tau hyperphosphorylation.
On the other hand, mouse FG01 also exerts its functions in human cells, suggesting that FG01-mediated signaling pathways are active in humans. Further scrutiny of these pathways, especially upstream events involving FG01’s effects on adenylate cyclases, is critical and underway. While it is not yet known whether there are functional analogs of FG01 in humans, further elucidation of FG01 functions and mechanism of action may prove to be important for developing new strategies for combating AD and other diseases including cancer and diabetes, in which the PKA and GSK-3 signaling pathways are centrally involved (Martinez et al., 2002; Naviglio et al., 2009).
Maintenance of mouse neuroblastoma N2a cells, N2a cells stably expressing human APP Swedish mutation (N2aSwe), human HeLa cells stably expressing human APP Swedish mutation (HeLaSwe), and rat PC12 cells has been described (Lin et al., 2007; Wang et al., 2006; York et al., 2000). Phoenix-Ampho helper cells were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum. Antibodies used were: anti-Myc (9E10), anti-adenylate cyclases, anti-His, anti-ADAM10 and anti-TACE from Santa Cruz Biotechnology; anti-GSK-3α, anti-GSK-3β, anti-phospho-GSK-3α/β (Ser21/9), anti-CREB, anti-phospho-CREB (Ser133), anti-PSD-95, and anti-synapsin from Cell Signaling Technology; anti-Aβ40, anti-pT205 tau and anti-total tau from Abcam; anti-Aβ (6E10) from Covance; anti-tau-1 from Chemicon; anti-α-tubulin from Sigma; anti-PHF-1 tau from P. Davies at Albert Einstein School of Medicine; and the FCA18 antibody specifically recognizing the N-terminus of APP βCTF from F. Checler at Institut de Pharmacologie Molecularie et Cellulaire du CNRS (Ancolio et al., 1999). The rabbit polyclonal antibody 369 against the APP C-terminus (Xu et al., 1997) and the anti-FG01 antibody were developed in our laboratory. PKA inhibitor H89 and GSK-3 inhibitor lithium chloride were from Sigma.
We constructed a new RHGP gene search vector from the original pLLGSV vector (Li and Cohen, 1996). In both the 5’LTR and the 3’LTR regions of the new RHGP gene search vector, there is a sequence containing a puromycin N-acetyl-tranferase gene (pac), a TRE (tetracycline-regulated element, tet-off) regulated CMV promoter driving the pac gene, a plasmid replication origin and a chloramphenicol resistance marker (Ori-CAT), and a LoxP site. In addition, there is a Cre recombinase gene (Cre) between the 5’LTR and the 3’LTR (Figure 1A). This new RHGP gene search vector was transfected into Phoenix-Ampho help cells. Generated infectious retrovirus in the cell culture supernatant was harvested and used to infect N2aSwe cells.
The infected N2aSwe cells with RHGP vector integration were selected with puromycin, live-stained with fluorescence-labeled APP βCTF antibody FCA18 (Ancolio et al., 1999), and subjected to multiple rounds of FACS sorting for cells with accumulated cell surface APP βCTF. The positive sorted cells were then treated with doxycycline (a derivative of tetracycline) and sorted for cells whose surface APP βCTF level was reversed back to background level in the presence of doxycycline. Resultant cells were cloned individually and assayed by ELISA and Western blotting to confirm surface accumulation of APP βCTF and reduction of Aβ generation. Positive candidate cell clones were further characterized and used for gene isolation.
Genomic DNA was extracted from FG01 cells, digested with restriction enzyme BamHI or HindIII, and self-ligated overnight with T4 ligase. The ligated DNA was precipitated, dissolved in TE buffer and electroporated into DH10B ElectroMax competent cells. The plasmid DNA from individual colonies was prepared for DNA sequencing. The target gene was identified by using UCSC Genome Browser Program.
We blasted GenBank database with Fg01 cDNA sequence to explore its origin. Homologous sequences between Fg01 and mouse Rps23 were aligned manually. Sequence similarity between Fg01 and Rps23 cDNA sequences of humans, mice and rats were compared using their homologous regions. Potential transmembrane region in the FG01 protein was predicted using PredictProtein (Rost et al., 2004).
N2a cells were transfected with FG01, APP or SMAD3 expression vectors (all Myc-tagged). After 48 hrs, cells were washed with ice-cold phosphate-buffered saline, collected with homogenization buffer (10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 200 mM sucrose, 1 mM phenylmethylsulfonyl fluoride) and homogenized with a ball bearing cell cracker. Samples were centrifuged at 900 × g for 10 min to remove cell debris and nuclei. Supernatants were centrifuged at 100,000 × g for 60 min at 4°C. After transferring the supernatant (cytosol) to a new tube, the pellet was washed and re-suspended with an equal volume (to that of cytosol) of homogenization buffer.
FG01 transfected N2a cells were washed with ice-cold phosphate-buffered saline containing 1 mM each of CaCl2 and MgCl2 and incubated at 4 °C with 0.5 mg/ml Sulfo-NHS-LC-biotin (Pierce) for 20 min and the process repeated once. Cell lysates were prepared in Nonidet P-40 lysis buffer. After affinity precipitation with streptavidin beads (Pierce), biotinylated proteins were eluted with SDS-PAGE sample buffer (Invitrogen) and loaded directly on SDS-PAGE gels for electrophoresis followed by Western blot analysis with the Myc antibody.
For cell surface immunostaining of FG01, N2a cells were first transfected with the Myc-FG01-His6 plasmid. Cells were then directly incubated with Myc or His6 antibody at 4°C for 2 hrs, followed by washing, fixation, and permeabilization. In some experiments, cells were permeabilized before incubating with antibodies. Treated cells were incubated with Alexa Fluor 488-conjugated secondary antibody and DAPI. Specimens were examined and fluorescence images collected using a Zeiss fluorescence microscope with AxioVision software.
HeLaSwe cells were transfected with FG01 or controls. Conditioned media and lysates from these cells were collected. The levels of Aβ40 and Aβ42 were quantified using ELISA kits (Invitrogen), following the manufacturer’s protocols.
N2aSwe cells were transfected with control vector or FG01 and then equally split. Four hours before collection, cells were treated with the GSK-3 inhibitor lithium chloride (5 mM) or sodium chloride (5 mM, as control). Alternatively, cells were treated with a PKA inhibitor H89 (10 µM) or with DMSO for control.
The mouse FG01 siRNA used was: 5’-UACUGUUUGUCAUGCCACUUCUGAU-3’. The control siRNA was from Invitrogen. siRNA was transfected into N2a cells using Lipofectamine RNAiMAX reagent (Invitrogen), following the manufacturer’s protocol. After FG01 RNA interference, total RNA was extracted from N2a cells by Trizol reagent (Invitrogen). After reverse transcription into first strand cDNA using standard conditions, samples were analyzed independently by real-time PCR using an iCycler iQ with SYBR green supermix (Bio-Rad). The FG01 primer pair used for real time PCR was: FG01–5’ (5’-TGTTGCATACACATACATGC-3’) and FG01–3’ (5’-TCATTAAGAACGGGAAGAAG-3’). A pair of β-actin primers served as controls (Zhang et al., 2007).
Histological sections from 2-month-old C57Bl6 mice were used for in situ hybridization reactions. Digoxygenin-labeled sense and antisense probes were generated for FG01 (corresponding to nucleotides 1–641 of NM_001024728), and the hybridization signal was detected using an alkaline-phosphatase-conjugated anti-digoxygenin antibody and BCIP/NTB (Roche).
We generated brain-specific Fg01 transgenic mice (Figure S7). Hemizygous Fg01 transgenic mice were crossed with homozygous triple transgenic (3XTg) AD mice harboring mutations in human App and Mapt (tau) genes on a presenilin 1 (PS1) mutant background (Oddo et al., 2003). Procedures involving animals and their care conformed to institutional guidelines (Animal Resources Department at Burnham Institute for Medical Research).
Fg01/3XTg mice and littermate controls on a 3XTg background were sacrificed at 11 months of age. Half of the brain was used for immunoblot analysis and the other half was paraffin-embedded for immunohistochemistry. Coronal brain sections (4 µm) were deparaffinized, hydrated, and then immunostained with anti-Aβ antibodies (an anti-Aβ40 specific antibody and 6E10), anti-phosphorylated tau antibodies (PHF-1 and pT205), or antibodies against PSD-95 and synapsin. After additional incubation with biotinylated secondary antibody, samples were incubated in ABC Elite (HRP) reagent (Vector Laboratories). Reactions were visualized by developing in DAB substrates (Vector Laboratories). All samples were visualized under a light microscope.
For immunohistochemistry comparison of Aβ and tau, immunostained neurons (>400) in were counted from five randomly selected cortical regions. Ratios of Aβ-positive and phosphorylated tau-positive neurons to total neurons were determined and normalized to those of controls. For immunohistochemistry comparison of synapse markers, five hippocampal regions were randomly selected and the images captured. After converting the images to grayscale, the optical density (darkness) of molecular layer staining was measured as an average of the gray value between white (0) and black (255) as described (Mathern et al., 1997) for comparison, by a computer-based image analysis using the Photoshop software.
Commercial kits were used to assay in vitro activities of GSK-3β (Sigma) and PKA (Upstate). For GSK-3α activity, a commercial GSK-3β activity assay kit was used but the procedure to immunoprecititate GSK-3β was replaced with immunoprecipitation of GSK-3α using an anti-GSK-3α antibody (Cell Signaling). cAMP levels were assayed using a commercial kit (Biovision).
Cells transfected with FG01 were lysed in either CHAPSO buffer (1% CHAPSO, 25 mM HEPES, pH7.4, 150 mM NaCl, and 2mM EDTA supplemented with protease inhibitors) or in NP40 buffer (1% NP40 in phosphate buffered saline, supplemented with protease inhibitors). Lysates were immunoprecipitated using mouse IgG, rabbit IgG, and antibodies against Myc or adenylate cyclases and Trueblot™ IP beads (eBioscience), followed by Western blot with antibodies against Myc or adenylate cyclases.
We thank H. Zheng for providing the pHZ04 construct and helpful discussion, Y. M. Li for technical help, F. LaFerla for providing the 3XTg AD mice, F. Checler for providing the FCA18 antibody, and P. Davies for providing the PHF-1 tau antibody. This work was supported in part by National Institutes of Health grants (R01 AG021173 to H.X. and L.L.; R01 NS046673 and R01 AG030197 to H.X.; R01 NS054880 to F-F. L.; P01 AG009464 to P.G.), and grants from the Alzheimer’s Association (to H.X. and F-F.L.), the American Health Assistance Foundation (to H.X.), US Department of Health and Human Services (AOA 90AZ2791 to P.G.), Fisher Center for Alzheimer’s Research Foundation (to P.G.), Cure Alzheimer’s Fund (to P.G.), National Natural Science Foundation of China (30672198 and 30840036 to Y.-w.Z.), National S&T Major Project (2009ZX09103-731 to Y.-w.Z.), and Natural Science Funds for Distinguished Young Scholar of Fujian Province (2009J06022 to Y.-w.Z.). Y.-w.Z. is supported by the Program for New Century Excellent Talents in Universities (NCET) and the Program for the New Century Excellent Talents in Fujian Province Universities (NCETFJ).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.