The Nogo receptor 2 is a novel substrate of Fbs1.

Members of the Nogo66 receptor family (NgR) are closely associated with nerve growth inhibition and plasticity in the CNS. All three members, NgR1, NgR2 and NgR3, are GPI anchored and highly glycosylated proteins. The binding and signaling properties of NgR1 are well described, but largely unknown for NgR2. At present the only known ligands are myelin associated glycoprotein (MAG) and amyloid beta precursor protein (APP). Despite the requirement of co-receptors for signaling no other binding partner has been uncovered. To learn more about the interactome of NgR2 we performed pull down experiments and were able to identify F-box protein that recognizes sugar chain 1 (Fbs1) as binding partner. We confirmed this finding with co-immunoprecipitations and in vitro binding assays and showed that the binding is mediated by the substrate recognition domain of Fbs1. As a substrate recognition protein of the SCF complex, Fbs1 binding leads to polyubiquitination and finally degradation of its substrates. This is the first time a member of the Nogo receptor family has been connected with an intracellular degradation pathway, which has not only implications for its production, but also for amyloid deposition in Alzheimer's disease.

Introduction

As a result of a various number of inhibitory molecules and their cognate receptors the regeneration capacity of the CNS is very limited [1]. One of the first described nerve growth inhibitors was Nogo-A [2-4], which signals through binding to Nogo receptor 1 (NgR1) and its complex members Lingo and p75NTR/Troy [5-9]. The very same receptor complex is also used by myelin associated glycoprotein (MAG) and oligodendrocyte myelin glycoprotein (OMgp) [6,10]. This signaling results in the activation of RhoA followed by growth cone collapse [11,12]. Subsequently, other CNS expressed members of the Nogo receptor family were identified and denoted NgR2 and NgR3 [13-15]. Despite the high degree of homology neither NgR2 nor NgR3 bind Nogo-A. The only common ligands are MAG for NgR1 and NgR2 [14,16,17] and amyloid beta precursor protein (APP) for all three family members [18]. In fact, unlike NgR1, for NgR2 and NgR3 no co-receptor has been identified yet.

All members of the Nogo receptor family share a very similar domain structure. They consist of a leucine rich repeat domain and a C-terminal stalk region, which shows the highest degree of variation. Furthermore, all three proteins are GPI anchored and glycosylated proteins [16]. The binding of MAG by NgR2 was shown to be sialic acid dependent [14].

In our study we describe the binding of F-box protein that recognizes sugar chain 1 (Fbs1), also called Fbx2, Fbxo2 and OCP1, to NgR2. It is expressed mainly in the brain [19] and in the organ of Corti [20]. Fbs1 is a substrate recognizing member of the SCF complex, which is a multi-protein E3 ubiquitin ligase containing Skp1, Cullin1 and a variable F-box protein. It recognizes high-mannose type polysaccharides and either functions as a chaperon [21] or leads to polyubiquitination and finally to degradation [22].

Furthermore, it was shown that Fbs1 also interacts with another protein known for glycoprotein homeostasis, CHIP, a co-chaperon with ubiquitin ligase properties. This interaction increased the turnover rate of Fbs1 bound glycoproteins [23]. In addition overexpression of Fbs1 in Alzheimer disease model mice reduced BACE1 levels and led to reduced synaptic deficits [24]. This is particular interesting because the NgRs are also known to bind APP [18,25,26] and NgR2 knock out reduced plaque formation [18]. Here we provide the first evidence that NgR2 is a substrate of Fbs1, which finally could also regulate its level.

Materials and methods

Plasmid construction

Mouse NgR2 mRNA sequence (91-1167) was subcloned in the plasmid pIgκ-V5-NgR [27] using the BamHI – PmeI sites and the primers, which encode a single strep-tag, fw 5′ CTG TCT AGA GGA CCT ATG CTC TGC ACC TGC TAC TCC TCC 3′ and re 5′ CAG GTT TAA ACT CAT TTT TCG AAC TGC GGG TGG CTC CAC AGC GCT CCC GGG CAC GCT TGG AAA TCG GAG TCG CTC 3′. C-terminal a synthesized double strep-tag was inserted using the AfeI site and the oligos 5′ TAG CGG CGG CGG ATG GAG CCA CCC GCA GTT CGA AAA AGG 3′ and 5′ CCT TTT TCG AAC TGC GGG TGG CTC CAT CCG CCG CCG CTA 3′. Sequencing revealed a double insertion of the oligo resulting in a triple strep-tag, which showed higher affinity compared to single tag.

The flag tagged NgR2 construct was cloned into pAPtag5 vector using the restriction enzymes HindIII and XbaI (Fermentas) and the Primers fw 5′ CGA AGC TTA CGA TTA CAA GGA TGA CGA CGA TAA GTC CGT GAC CCC CAG CTG TCC 3′ and re 5′ TAT TCT AGA TCA GAG GTG ATG GAG CGC CAG 3′. The forward primer encodes the flag tag. The derived construct uses the signal peptide from the pAPtag5 followed by the flag tag and the NgR2 coding sequence starting with amino acid 30. The C-terminus including the GPI anchor sequence is not changed.

Fbs1 full length was amplified from mouse brain cDNA using the primers fw 5′ ATT GCT AGC GCG ATG GAT GGA GAT GGT G 3′and re 5′ ATT CTC GAG GGG TTC CAC CCA CAC GCT AC 3 and cloned into pAPtag5 vector (Genehunter) using NheI and XhoI sites. The construct thus has a C-terminal myc/his tag.

Plasmids containing the wild type sugar binding domain of Fbs1 and mutant F177A were kindly provided by Yoshida (Tokyo Metropolitan Institute of Medical Science) [28].

All constructs were sequenced (LGC) to verify the successful cloning. Schemes of all used constructs can be seen in Fig. 1(A).

Fig. 1

(A) Wild type NgR2 consists of a signal peptide (SP), a leucine-rich repeat domain (LRRs) flanked by a N-terminal flank (NF) and a C-terminal flank (CF) and a Stalk region. The cloned NgR2-strep constructs uses the signal peptide and V5 tag of the vector backbone, avoids the GPI anchor sequence and has a C-terminal triple-strep-tag (STrEP). The NgR2-Flag has a N-terminal flag tag (F) and the wild type C-terminus. Fbs1 consists of a PEST, F-Box and the substrate recognition sugar binding domain (SBD). We added a C-terminal myc tag for the Fbs1-myc construct. The constructs SBD and F177A have a N-terminal HIS tag (H). (B) Pull down experiment was separated on gradient SDS–PAGE (5%–15%) and silverstained. Recombinant bait protein NgR2-strep is marked with black arrow and band identified to be Fbs1 marked with arrow head. (C) Analysis of the single band from B resulted in the identification of the peptides marked with underlined bold letters. The sequence coverage is 36%.

Pull down assay

CHO K1 cells adapted to grow in suspension were transiently transfected with NgR2 strep using Amaxa™ Nucleofector™ (Lonza). 4 days after transfection conditioned medium was harvested and dialysed extensively against PBS. It was then applied to a mini column containing 150 μl strep-Tactin®Sepharose (IBA, 2-1201-010). After extensive washing 10 mg total protein of whole mouse brain lysate was applied. Adult C57Bl6 mice were decapitated and whole brain was lysed in buffer A (50 mM Hepes pH 7.6, 150 mM NaCl, 10% glycerol, 1% triton x-100, 1 mM PMSF, 5 μg/ml leupeptin and 10 μg/ml aprotinin) and cleared by centrifugation. Protein concentration was measured using Protein assay (Bio-Rad, 500-0006). Column was washed three times with buffer A and eluted using elution buffer (IBA, 2-1019-025). Sample was either separated by SDS–PAGE followed by MALDI-TOF analysis of single silverstained bands, run on SDS–PAGE followed by coomassie staining and analysis by LC/MS/MS, or TCA precipitated and analyzed by LC/MS/MS. Silverstaining was performed as described [29]. Experiments were controlled by use of empty beads for pull downs. In addition also the purified NgR2 was analyzed, to ensure only binding partners from brain lysate are considered.

MALDI TOF

Relevant protein band was excised from gel, destained, and in-gel digested with modified porcine trypsin, sequence grade (Promega) as described in [30]. Micro-ZipTipC18 (Millipore) were utilized for digest desalting, and peptides were eluted with the acetonitrile solution containing the CHCA matrix (Fluka) directly onto the target. Mass spectra were acquired using a MALDI-TOF/TOF Ultraflex instrument (Bruker Daltonics). Peptide mass fingerprintings (PMF) were interpreted with the MASCOT by scanning the SwissProt database. A peptide mass tolerance was set to 0.1 Da and one missed cleavage was allowed; a complete carbamidomethylation of cysteines and partial oxidation of methionines were considered as modifications.

LC-MS-MS

Protein digests were analyzed using nanoLC-ESI-MS (LTQ Orbitrap XL; ThermoScientific) equipped with a nanospray interface and coupled to an UltiMate 3000 system (Dionex). The nano-HPLC separation was done as described by Faserl et al. [31]. Protein identification was performed via Sequest, Proteome Discoverer (Version 1.1, ThermoScientific) and a mouse database. The criteria for positive identification of peptides were Xcorr > 2.3 for doubly charge ions, and Xcorr > 2.8 for triply charge ions and Xcorr > 3.3 for fourfold and higher charged ions.

TCA precipitated and acetone washed protein pellets were reduced with TCEP, alkylated with jodoacetamide and digested with trypsin. Peptides were analyzed by nanoLC-MSMS with a 1200 HPLC (Agilent) connected to a LTQ Orbitrap Velos (Thermo Scientific). The proteins were identified with Mascot searching Swiss-Prot 2010_09 [32].

Co-Immunoprecipitations

NgR2-flag and Fbs1-myc/his were co-transfected in CHO K1 cells using Lipofectamine 2000 (Invitrogen, 11668019). Two days after transfection cells were lysed in buffer A and used for immunoprecipitation. 1 mg of lysate (5 μg/μl total protein), 2 μg of flag antibody (Sigma–Aldrich, F1804), 2 μg of myc (Sigma–Aldrich, M4439) antibody and as control 2 μg of anti strep antibody (IBA, 2-1507-001) was used per reaction. Samples were incubated 1 h with antibodies followed by 1 h with protein A/G agarose (Pierce, 20421). After washing with buffer A samples were boiled in 1× laemmli buffer for 10 min and analyzed by Western blotting using 3% skim milk in TBST for blocking, anti-flag antibody 1:1000 (Sigma–Aldrich, F1804) and polyclonal anti-myc antibody 1:1000 (Santa Cruz, SC-789) as primary antibodies. As secondary antibodies HRP labeled goat anti mouse IgG 1:20000 (Thermo Scientific-Pierce, 31430) and goat anti rabbit IgG 1:20000 (Thermo Scientific-Pierce, 31460) were used.

In vitro binding assays

NgR2 strep was purified as described in previous paragraph. The sugar binding domain and the F177A mutant, both His tagged, were produced in E.coli BL21 and purified using TALON® Metal Affinity Resin (Clontech, 635501) according to manufacturer. 1 μg of purified sugar binding domains were incubated with 500 ng NgR2 bound to 15 μl bed volume of strep-tactin Sepharose. After washing with strep-tag washing buffer (IBA, 2-1003-100) the sample was eluted using biotin containing elution buffer (IBA, 2-1019-025) and analysed by SDS–PAGE and silverstaining [29] and by western blotting using HRP labeled anti-HIS antibody (Novagen, 71840-3) and anti V5 antibody (Invitorgen, R0960-25).

Results

NgR2 is known to be a receptor for MAG, but also influences the plaque formation in Alzheimer’s disease by interacting with APP. To increase our knowledge about this receptor we performed pull down assays. We used suspension CHO K1 culture to produce truncated NgR2 protein containing a C-terminal triple-strep-tag (Fig. 1(A)). The mammalian system was chosen to ensure proper glycosylation. The protein was secreted into the protein free medium (CD CHO, Invitrogen) and subsequently purified using the strep-tactin system. The combination of secretion in protein free media and the use of triple-strep-tag allowed a one step purification. The recombinant bait protein was bound by strep-tactin Sepharose and used directly for pull down assays using 10 mg total protein of adult mouse brain lysate. Two separate negative controls were included to reveal for non specific protein binding. Firstly, the same amount of empty beads were used for pull down to ensure only specific binding partners were analyzed. In addition NgR2 beads without brain lysate was used to ensure only binding partners coming from the brain lysate were considered.

We separated the samples using SDS–PAGE and analyzed single bands migrating at 42 kDa using a MALDI-TOF mass spectrometry, which led to the identification of Fbs1 as a unique band in the pull down lane (Fig. 1(B)). In total, all peptides that were identified correspond to 36% of Fbs1 (Fig. 1(C)).

To analyze the pull down experiments in more detail we used additional approaches. The samples were either run on SDS–PAGE and extracted from gel, or TCA precipitated followed by LC/MS/MS using either a LTQ Orbitrap XL or a LTQ Orbitrap Velos. With the gain in sensitivity we found in addition to Fbs1 its binding partner Skp1 in the pull down samples. However other members of the SCF complex were not identified, which is in accordance with the finding that Fbs1 forms a complex mainly with Skp1 and not necessarily the full SCF complex including Cul1 [33]. The sequence coverage was 75% for Fbs1 and 76% for Skp1. The unique peptides identified for both proteins can be seen in Table 1. In all experiments not a single spectra of both Fbs1 and Skp1 was found in the controls. In addition to this novel interaction we were also able to pull down MAG, reported previously to interact with NgR2 [14], supporting the specificity of the pull-down assay. The LC/MS/MS data for MAG are shown in the Supplementary data (Table S1). Interestingly no other sugar binding F-box protein was identified. In particular, Fbs2 (Fbxo6), which is highly expressed in brain and is known to bind partly the same substrates [34], was not seen in any pull down samples analyzed.

Table 1

Peptides of Fbs1 (A) and Skp1 (B) identified by LC/MS/MS. The sequence coverage is 75% for Fbs1 and 76% for Skp1 as indicated by underlined and bold amino acids.

Peptide mass	Sequence
2771.33	EAEEEEEAEAVEYLAELPEPLLLR
1668.89	VLAELPATELVQAcR
1666.95	WKELVDGAPLWLLK
1353.77	ELVDGAPLWLLK
3282.48	cQQEGLVPEGSADEERDHWQQFYFLSK
2500.99	NPcGEEDLEGWSDVEHGGDGWR
1498.72	DHWQQFYFLSK
2206.03	VEELPGDNGVEFTQDDSVKK
1352.58	YFASSFEWcR
3056.61	KAQVIDLQAEGYWEELLDTTQPAIVVK
1323.67	TDAGSLYELTVR
4394.07	LLSENEDVLAEFATGQVAVPEDGSWMEISHTFIDYGPGVR
1452.65	FEHGGQDSVYWK
1	MDGDGDPESV SHPEEASPEE QPEEAGAEAS AEEEQLREAE EEEEAEAVEY
51	LAELPEPLLL RVLAELPATE LVQACRLVCL RWKELVDGAP LWLLKCQQEG
101	LVPEGSADEE RDHWQQFYFL SKRRRNLLRN PCGEEDLEGW SDVEHGGDGW
151	RVEELPGDNG VEFTQDDSVK KYFASSFEWC RKAQVIDLQA EGYWEELLDT
201	TQPAIVVKDW YSGRTDAGSL YELTVRLLSE NEDVLAEFAT GQVAVPEDGS
251	WMEISHTFID YGPGVRFVRF EHGGQDSVYW KGWFGARVTN SSVWVEP
B
Peptide mass	Sequence
2317.21	PTIKLqSSDGEIFEVDVEIAK
1877.93	LQSSDGEIFEVDVEIAK
2535.31	LQSSDGEIFEVDVEIAKQSVTIK
3124.5	TMLEDLGMDDEGDDDPVPLPNVNAAILKK
1760.87	RTDDIPVWDQEFLK
2135.15	VDQGTLFELILAANYLDIK
904.47	GLLDVTcK
2068.97	TFNIKNDFTEEEEAQVR
1465.64	NDFTEEEEAQVR
1250.54	KENQWcEEK
1	MPTIKLQSSD GEIFEVDVEI AKQSVTIKTM LEDLGMDDEG DDDPVPLPNV
51	NAAILKKVIQ WCTHHKDDPP PPEDDENKEK RTDDIPVWDQ EFLKVDQGTL
101	FELILAANYL DIKGLLDVTC KTVANMIKGK TPEEIRKTFN IKNDFTEEEE
151	AQVRKENQWC EEK

To further validate the authenticity of the proteins identified by mass spectrometry, we performed co-immunoprecipitation experiments. Myc tagged Fbs1 was co-expressed with NgR2-flag in CHO cells. When precipitating Fbs1 using an anti-myc antibody, we observed a strong signal for flag tagged NgR2. Similarly for precipitation of NgR2 using an anti-flag antibody a clear signal for myc tagged Fbs1 was seen. An unrelated anti-strep antibody failed to precipitate both proteins (Fig. 2(A)). This confirmed our initial finding of the pull down experiments. Due to the lack of a specific NgR2 antibody, we were not able to also perform co-immunoprecipitation experiments in endogenous material.

Fig. 2

(A) Fbs1-myc co-precipitated NgR2-flag and vice versa. An unrelated antibody against the strep-tag did not precipitate Fbs1 nor NgR2. (B) In vitro binding assay was analyzed by silver staining. While the wild type sugar binding domain (SBD) (arrow head) bound the bait protein (arrow), the F177A mutant lost this binding completely. (C) Analyzing the in vitro binding assay with western blotting, using an anti V5 antibody to detect the NgR2-strep (arrow) and a HRP labeled anti HIS antibody to detect the sugar binding domains (arrow head), showed binding of the wild type construct and only weak background signal for F177A mutant.

Fbs1 is known to bind mannose rich polysaccharides without any specificity concerning amino acid sequence. NgR2 is strongly glycosylated [14,35] and therefore it is likely that the Fbs1 binding depends on its sugar binding domain. To address this hypothesis we performed in vitro binding assays using recombinant NgR2 produced in CHO cells and the sugar binding domain (SBD) of Fbs1 and a mutant (F177A) expressed in E.coli. The mutation F177A is sufficient to completely abolish the binding of Fbs1 to substrates [28]. NgR2 was bound to strep-tactin Sepharose and incubated either with purified recombinant SBD or F177A. Subsequently samples were washed and separated by SDS–PAGE, followed either by silver staining (Fig. 2(B)) or Western blotting (Fig. 2(C)). As expected, only the wt SBD was able to bind, demonstrating that NgR2 binding is mediated via the substrate binding domain of Fbs1 and not via its PEST and F-box domains. Therefore NgR2 is considered to be a novel substrate of Fbs1.

Discussion

Although NgR2 is known to be a receptor of MAG its physiological role is still unclear. Recently the ability of NgR2 to alter the plaque formation in Alzheimer’s disease model mice [18] renewed the interest in NgR2.

Our results show binding of NgR2 via the sugar binding domain of Fbs1. Interestingly, only Fbs1 and not Fbs2 was identified in our pull down assay. Both proteins are known to bind partly the same substrates and are both expressed in the brain [19,22,34].

Fbs1 is a cytoplasmic protein and can bind its substrates only in this compartment. The substrate needs to be retro-translocated before being recognized. This transport is known to take place from the ER and endosomes to the cytoplasm [36,37]. After binding the substrates are polyubiquitinated and degraded via the proteasome. Interestingly several substrates of Fbs1 show decreased levels when Fbs1 is overexpressed [23,24,37,38]. One explanation would be that the degradation process is limited by this F-box protein.

Our finding has relevance in Alzheimer’s disease. Recently a study showed that overexpression of Fbs1 in Alzheimer’s disease models resulted in reduced BACE1 and most importantly in a rescue of synaptic deficits [24]. Here we show that NgR2 is also bound by the substrate recognition site of Fbs1. As there is no amino acid specificity of this special F-box protein, we propose that NgR2 levels might also be reduced by Fbs1 overexpression. Reduction of NgR2 was shown to reduce amyloid deposition and therefore is considered to be beneficial [18]. The reduction of BACE1 by Fbs1 overexpression might well be accompanied by reduced NgR2 levels, which both leads to reduced plaque formation and rescue of synaptic deficits.

To consider Fbs1 as therapeutic target is however risky and needs further investigation. In our opinion the substrate recognition is not ideal, because there is no amino acid sequence specificity. This means that there are probably hundreds of N-glycosylated proteins, that might be degraded via Fbs1 binding. This makes a reasonable calculation of side effects challenging. That is why it is important to identify as many substrates of Fbs1 as possible.