We used co-immunoprecipitation experiments to test whether FE65 interacted with VLDLR. COS7 cells were transfected with VLDLR and empty vector, VLDLR and FE65, or FE65 and empty vector. Full-length VLDLR co-precipitated with FE65 and was not detectable in the absence of FE65 (Figure 1A). Western blot analysis of COS7 cell extracts confirmed that levels of VLDLR and FE65 were consistent across transfections (Figure 1A). We also performed the reverse experiment, which showed that full-length FE65 co-precipitated with VLDLR and was not detectable in the absence of VLDLR (Figure 1B).

To test whether VLDLR CTF interacted with FE65, we transfected COS7 cells with full length VLDLR and empty vector, full length VLDLR and FE65, or VLDLR CTF and FE65 (Figure 1C, D), and performed co-immunoprecipitations. We found that FE65 co-precipitated with both full length VLDLR and VLDLR CTF (Figure 1E). Consistent with these findings, the reverse experiment resulted in co-precipitation of full length VLDLR and VLDLR CTF with FE65 in COS7 cells (Figure 1F).

We then examined whether there was a physical association between FE65 and VLDLR in vivo. To test this, we performed co-immunoprecipitations from whole brain lysates, using anti-5F3 to recognize VLDLR or a nonspecific IgG as a negative control. Immunoprecipitation of VLDLR resulted in the co-precipitation of FE65 (Figure 1G). In the reverse experiment, we performed co-immunoprecipitation from whole brain lysates using anti-FE65 and then probed with anti-5F3. We found that FE65 co-immunoprecipitated with both the mature and immature forms of VLDLR in brain lysates (Figure 1H). Overall, these results suggest that VLDLR interacts with FE65 both in vitro and in vivo. To further examine whether VLDLR interacts with FE65, we incubated wild-type brain lysates with purified immobilized GST or GST-VLDLR CTF protein and probed for FE65 (Figure 1 I-K). We found that VLDLR CTF interacted with FE65 in vivo. No signal was detected in lanes of brain lysates incubated with GST alone.

To test whether endogenous FE65 co-localizes with VLDLR during early neuronal development, primary hippocampal neurons (DIV 3) were fixed and immunostained with anti-5F3 and anti-FE65 antibodies. VLDLR and FE65 immunoreactivities were strong in the cell body and punctuate throughout neuronal processes (Figure 2A). The immunostainings overlapped suggesting that VLDLR co-localized with FE65 within the cell bodies and partially co-localized in neuronal processes (Figure 2A).

To test whether FE65 and VLDLR can still co-localize during the peak of synaptogenesis, primary hippocampal neurons (DIV 14) were fixed and immunostained with anti-5F3 and anti-FE65 antibodies. Interestingly, FE65 expression was up-regulated on DIV 14 compared to DIV3, consistent with previous findings 1. Moreover, VLDLR and FE65 immunoreactivity was strong in the cell body and punctuate throughout neuronal processes with partial co-localizations, consistent with what we observed on DIV 3 (Figure 2B).

To determine which domain of FE65 interacts with VLDLR, COS7 cells were co-transfected with full length VLDLR and FE65 deletion constructs containing a c-terminal myc tag (Figure 3A). Each FE65 construct resulted in protein expression at the anticipated sizes, as determined by western blot (Figure 3B, upper panel). VLDLR was expressed to similar levels in all transfected cells (Figure 3B, lower panel). Immunoprecipitation with an anti-HA antibody (for VLDLR) and probing with an anti-myc antibody (for FE65) resulted in VLDLR immunoprecipitation with all three FE65 constructs containing the PTB1 domain, but not the FE65 containing only the PTB2 domain construct (Figure 3C, upper panel). Interestingly, VLDLR interacted strongly with the FE65 construct lacking the WW domain (construct #1) compared to full length FE65 (construct #4) and the FE65 construct containing only the WW and PTB1 domains (construct #2) (Figure 3C). However, the FE65 WW domain alone does not co-precipitate with VLDLR (data not shown). Since it has been shown that the WW and PTB domains of FE65 can interact with each other 3, the FE65 WW domain may induce conformational changes in full length FE65 which reduce the exposure of the FE65 PTB1 domain for interaction with VLDLR.

We conducted an additional experiment to ensure that the lack of co-immunoprecipitation between VLDLR and the FE65 containing only the PTB2 domain was not due to the decreased expression level of the FE65 PTB2 domain (construct #3) in cell lysates. To test this, we used a different set of FE65 deletion constructs, which have a GFP c-terminal tag. COS7 cells were co-transfected with full length VLDLR-myc and GFP, VLDLR-myc and FE65 PTB2-GFP (construct #1), or VLDLR-myc and full length FE65-GFP (construct #2) (Additional file 1: Figure S1A). VLDLR and each FE65 construct resulted in similar protein expression in all transfected cells (Additional file 1: Figure S1B). Immunoprecipitation with an 5F3 antibody (for VLDLR) and probing with an anti-GFP antibody (for FE65) resulted in full length FE65 immunoprecipitation with the VLDLR but the FE65 construct containing only the PTB2 domain did not. Consistent with these findings, the reverse experiment resulted in co-precipitation of VLDLR with the full but not with the truncated PTB2 construct (Additional file 1: Figure S1B).

Our previous studies have shown that VLDLR undergoes α- and γ- secretase cleavage similar to APP and ApoER2 22. Because VLDLR CTFs were undetectable with overexpression of full length VLDLR, we hypothesized that VLDLR CTF may undergo proteasome degradation. To test this possibility, COS7 cells were transfected with full length VLDLR and treated with the proteasomal inhibitor, MG132 (10 uM) or vehicle (10% DMSO) for 24 hours. We found that VLDLR CTFs were detectable when full length VLDLR-transfected cells were treated with MG132 (Figure 4A). Interestingly, there was also a large increase in full length VLDLR suggesting that both VLDLR CTFs and full length VLDLR undergo proteasomal degredation.

To test whether FE65 could modulate VLDLR processing in vitro, COS7 cells were transfected with VLDLR-HA and empty vector or VLDLR-HA and FE65, and the levels of sVLDLR, total VLDLR, and VLDLR CTF were measured. Co-transfection of FE65 increased sVLDLR (Figure 4B, upper panel) and had no effect on total VLDLR levels in COS7 cells (Figure 4B, lower panel). VLDLR CTFs were still undetectable in the presence of FE65. These data suggest that FE65 may regulate VLDLR processing.

To test whether FE65 could affect VLDLR trafficking, we transfected COS7 cells with full length VLDLR and empty vector or full length VLDLR and FE65 for 24 hours. Cell surface proteins were biotinylated, isolated with avidin beads, and immunoblotted for VLDLR. We found that FE65 significantly increased cell surface levels of VLDLR by 118% (Figure 4C, D, n = 4/condition). To verify our findings, we conducted live cell surface staining by overexpressing GFP, VLDLR and empty vector or GFP, VLDLR and FE65 in primary hippocampal neurons (DIV 14-16). FE65 increased cell surface levels of VLDLR by 120%, a 1.2 fold increase, in primary hippocampal neurons (Figure 4E, F, p < 0.05; n = 10). However, total VLDLR protein level was unchanged in the presence of FE65, consistent with our previous in vitro and in vivo data (Figure 4G, H). Thus, two independent assays suggest that FE65 can modulate cell surface expression of VLDLR.

Several studies have shown that FE65 and the cytoplasmic domain of APP form a complex and translocate into the nucleus in COS7 and H4 cells 232324. Consistent with previous findings, we observed that APP CTF was present in nuclear fractions when co-expressed with FE65 compared to controls (Figure 5A). We then examined whether FE65 could also translocate VLDLR CTFs to the nucleus. To test this, we transfected COS7 cells with full length VLDLR or VLDLR CTF with either FE65 or empty vector. We found that full length VLDLR and FE65 were present in the cytosol/membrane fractionation, but were not present in the nucleus (Figure 5B). Similar to APP CTF and FE65 complex, VLDLR CTF and FE65 were expressed in both the cytosolic/membrane and in the nucleus (Figure 5C).

ApoE Receptors, including LRP1 and ApoER2, have been shown to interact with APP 1719202125, and thus we wanted to investigate whether VLDLR can interact with APP. For this experiment, we performed co-immunoprecipitations from whole brain lysates using anti-VLDLR antibody or an anti-IgG antibody and probed for APP. We observed that APP co-precipitated with VLDLR in vivo (Figure 6A, upper panel). We also conducted the reverse experiment and found that VLDLR co-precipitated with APP (Figure 6B). APP and VLDLR were expressed to similar levels in all conditions (Figure 6A, B lower panels).

To examine the effect of APP on VLDLR processing, we transfected COS7 cells with full length VLDLR and empty vector or full length VLDLR and APP, and then the levels of sVLDLR, total VLDLR, VLDLR CTF, and total APP were measured. Co-transfection with APP resulted in increased sVLDLR and total VLDLR compared to empty vector (Figure 6C). However, VLDLR CTF levels remained undetectable.

Next, COS7 cells were transfected with APP and empty vector or APP and VLDLR in order to examine the effect of VLDLR on APP processing. VLDLR increased the levels of total APP, sAPPα and APP CTF (Figure 6D). These data suggest that the interaction between APP and VLDLR affects the metabolism of both proteins.

We next examined whether APP alters cell surface expression of VLDLR. COS7 cells were transfected with VLDLR and empty vector or VLDLR and APP, and cell surface biotinylation was performed. We found that APP increased cell surface levels of VLDLR (Figure 7A). We also examined whether VLDLR can regulate cell surface expression of APP. COS7 cells were transfected with APP and empty vector or APP and VLDLR. We found that VLDLR increased cell surface levels of APP (Figure 7B). To further examine the effects of VLDLR on APP trafficking, primary hippocampal neurons were transfected with GFP, APP, and empty vector or GFP, APP, and VLDLR and live cell surface staining was conducted. Consistent with our findings, VLDLR significantly increased cell surface levels of APP by 24% (p < 0.05, n = 10) (Figure 7C, D).

We and others have shown that FE65 forms tripartite complexes with APP and LRP1 or ApoER2, modulating the interaction of these proteins 1720. We investigated whether FE65 can affect the interaction between VLDLR and APP in vitro. COS7 cells were transfected with VLDLR, APP, and empty vector or VLDLR, APP, and FE65. Immunoprecipitation with an anti-VLDLR antibody and probing for APP revealed that FE65 increased the interaction between VLDLR and APP in COS7 cells (Figure 8A). In the reverse experiment, co-transfection with FE65 increased the association between APP and VLDLR (Figure 8B).

To verify whether FE65 can modulate the interaction between APP and VLDLR, we transfected COS7 cells with APP, VLDLR and either full length FE65 or FE65 PTB2 domain, which interacts with APP but not VLDLR. Cell lysates were immunoprecipitated with an anti-5F3 antibody (for VLDLR) and probed with an anti-22C11 antibody (for APP) (Figure 8C). We found that FE65 PTB2 domain construct significantly decreased the association between APP and VLDLR compared to full-length FE65 (Figure 8C). To examine whether FE65 can alter the association between APP and VLDLR in vivo, we immunoprecipitated VLDLR from brain lysates and found that an APP immunoreactive band was decreased in FE65 knockout brain lysates compared to wild-type littermates (Figure 8D). These data further demonstrate that FE65 is a linker between APP and VLDLR.Total levels of VLDLR were unchanged in FE65 knockout mice compared to wildtype littermates. (Figure 8D, lower panel). Interestingly, FE65 knockout mice had significantly increased total APP and APP CTFs compared to wild-type littermates (Figure 8D, Additional file 2: Figure S2). These data indicate that FE65 may also differentially regulate the processing of APP and VLDLR.

We generated C-terminal tagged myc and C-terminal tagged HA for full length VLDLR and C-terminal of VLDLR. Recombinant DNA sequences were confirmed by sequencing, and expression of correctly sized proteins was confirmed by Western blot.

COS7 was maintained in Opti-MEM (Invitrogen) with 10% fetal bovine serum (FBS, Life Technologies, Inc.) in a 5% CO₂ incubator. COS7 cells were transiently transfected with 0.5 -1 ug of plasmid in FuGENE6 (Roche) according to the manufacturer's protocol and cultured 24 h in DMEM containing 10% FBS. For co-transfections, cells were similarly transfected with 0.5-1 ug of each plasmid in Fugene 6 (Roche) and cultured 24 hr in DMEM with 10% FBS. After 24 hr the cells were transferred to Opti-MEM serum free media (Invitrogen) and treated with indicated compounds.

For isolation of nuclear fraction, cells were harvested and 200 μl of ice-cold CER1 was added to the cell pellet, vortexed vigorously to fully resuspend the cell pellet. The tube was incubated on ice for 10 min, 11 μl of ice-cold CER II was added, vortexed for 5 sec, and centrifuged for 5 min (16, 000 × g, 5 min). Immediately after the supernatant (cytoplasmic/membrane extract) fraction was transferred, and the insoluble (pellet) fraction was resuspended in 100 ul of ice-cold NER. This was then vortexed for 15 sec, and returned to ice for continued vortexing for 15 sec every 10 min, for a total of 40 min. The sample was then centrifuged for 10 min (16, 000 × g, 10 min) and the supernatant (nuclear extract) fraction was immediately transferred.

We used antibodies anti-HA (Abcam), anti-c-myc (Abcam), anti-22C11 (Chemicon), anti-V5 (Chemicon), and anti-FE65 (From Dr. Andre Goffinet). The anti-5F3 antibody was a kind gift of Dr. Dudley Strickland, the C1/6.1 antibody was a kind gift from Dr. Paul Matthew, and the VLDLR IIII antibody was a kind gift of Dr. Guojun Bu. For analysis of secreted APP, we used 6E10 (identifying sAPPα) (Signet).

Secreted fragments were identified by western blot analysis of the media (sVLDLR, 5F3 antibody; sAPP, 6E10 antibody). CTF were measured by western blots of cell lysates (VLDLR CTF, myc antibody; APP CTF, C1/6.1 antibody).

Primary hippocampal neurons from embryonic day 18-19 Sprague-Dawley rats were cultured at 150 cells/mm² as described 38. Neurons were transfected at 14 days in vitro (DIV) with GFP, APP-HA, VLDLR-Myc or empty vector by lipofectamine 2000 (Invitrogen) (2 μg DNA per well) according to manufacturers instructions. Transcription of each insert was driven by the CMV promoter.

COS7 cells were transiently transfected with VLDLR and vector or VLDLR and FE65 in Fugene 6 (Roche) and cultured 24 hrs in DMEM containing 10% FBS. After 24 hr, cells were washed twice with PBS, and surface proteins were labeled with Sulfo-NHS-SS-Biotin 500 ul at 500 ug/ml PBS (Pierce) under gentle shaking at 4°C for 30 min. 50 ul of quenching solution was added to cells at 4°C, which were then washed twice with TBS. Cells were lysed in 500 ul lysis buffer, collected with a cell scraper, disrupted by sonication on ice, incubated for 30 min on ice, and clarified by centrifugation (10, 000 × g, 2 min). To isolate biotin-labeled proteins, lysate was added to immobilized NeutrAvidin TM Gel (50 ul) and incubated 1 hr at room temperature. Gels were washed five times with wash buffer and incubated 1 hr with SDS-PAGE sample buffer including 50 mM DTT. Elutions were analyzed by immunoblotting.

Hippocampal cultured neurons were fixed in methanol at -20°C for 10 min (for immunostaining of endogenous synaptic markers). Antibodies for immunostaining were incubated in GDB buffer (0.1% gelatin, 0.3% Triton X-100, 16 mM sodium phosphate pH 7.4, 450 mM NaCl).

Cell surface expression levels of VLDLR were performed as described 17. Live neuronal cultures were briefly incubated (10 min) with the 5F3 antibody directed against extracellular N-termini of VLDLR (10 μg/mL in conditioned medium) to specifically label surface receptors, then lightly fixed for 5 min in 4% paraformaldehyde (non-permeabilizing conditions). After fixation, the surface-remaining antibody-labeled protein was measured with Alexa Fluor 555-conjugated anti-mouse secondary antibodies for 2 hr. Immunostaining was quantified using Metamorph analysis of immunostaining intensity or punctate number from Z-stacked images obtained with a Zeiss LSM510 confocal microscope. Surface localization of staining was also confirmed visually from these images.

Brain Lysates from 13 month old FE65 knockout mice and wild-type littermate were homogenized in buffer containing 50 mm Tris-HCl, pH 8.0, 0.15 m NaCl, 1% Nonidet P-40, and phosphatase and protease inhibitors (IP Buffer). For immunoprecipitations, lysates were incubated overnight at 4°C with APP or VLDLR antibody and protein G-Sepharose beads (Amersham Biosciences). The precipitates were washed five times with lysis buffer and resuspended in SDS sample buffer.

The recombinant GST or GST-VLDLR CTF protein was expressed in Escherichia coli BL21 strain, using the pGEX-4B system as previously described (Kim, et al 2011). The GST or GST-VLDLR CTF fusion protein was then purified using glutathione-agarose beads (Sigma), in accordance with the manufacturer's instructions. An equal amount of GST or GST-VLDLR CTF fusion protein was incubated overnight with brain lysates of wild-type mice (n = 3 per each condition). After incubation, protein-A agarose was added, and the samples were incubated for 3 hours at 4°C on a rotator. Following incubation, the beads were washed three times in ice-cold PBS and boiled with Laemmli sample buffer.

Experiments were repeated a minimum of four times unless otherwise noted. Data were analyzed using t-tests with significance determined as p < 0.05. Descriptive statistics were calculated with StatView 4.1 and displayed as an expressed mean ± S.E.M.