Evidence of Functional Modulation of the MEKK/JNK/cJun Signaling Cascade by the Low Density Lipoprotein Receptor-related Protein (LRP)*

, Lipoprotein receptors, such as LRP, have been shown to assemble multiprotein complexes containing intracellular signaling molecules; however, in vivo , their signaling function is poorly understood. Using a novel LRP receptor fusion construct, a type I transmembrane protein chimera, termed sIgG-LRP (bearing the intracellular COOH-terminal tail of human LRP as recombinant fusion to a transmembrane region plus the extracellular IgG-F c domain), we here investigated LRP signal trans- duction specificity in intact cells. First and similar to activated (cid:1) 2-macroglobulin as agonist of endogenous LRP, expression of sIgG-LRP demonstrated significant apoptosis protection. Second and similar to (cid:1) 2-macro-globulin-induced endogenous LRP, sIgG-LRP is sufficient to negatively modulate mitogen-induced Elk-1 and cJun (but not NF- (cid:2) B) transcriptional activity. Third, expression of sIgG-LRP also impaired cJun transactivation mediated by constitutive active mutants of Rac-1 and MEKK-1. Fourth and unexpectedly, sIgG-LRP expression was found to be associated with

Low density lipoprotein receptor-related protein (LRP) 1 is one member of the LDL receptor family that also includes the LDL receptor, the very low density lipoprotein receptor, megalin, LRP5, LRP6, and apoER2 receptor (see Ref. 1 for review). These multifunctional lipoprotein receptors are established cargo transporters, but expression level at the cell surface and/or (ant)agonistic binding of diverse biological ligands are now thought to potentially evoke signaling pathways involved in cell fate determination (see Ref. 2

for review).
LRP is expressed abundantly in neurons and microglia of the central nervous system (3,4). Disruption of the LRP gene in mice blocks development of LRP Ϫ/Ϫ embryos around the implantation (5). However, the complex phenotype of the few malformed LRP-deficient embryos that survive until E10 (6), similar to the very low density lipoprotein ApoER2 receptor double knockout phenotype (7), postulated some LRP receptor signaling function(s). Consistently, LRP and several of its ligands, including ␣2-macroglobulin, tissue plasminogen activator (tPA), apoE-containing lipoproteins, and the amyloid precursor protein (APP) (8,9), have been implicated in various cellular functions including the neuropathogenesis of Alzheimer's disease (see Ref. 10 for review).
Based on yeast two-hybrid and co-immunoprecipitation analysis, lipoprotein receptors assemble intracellular multiprotein complexes containing the adapter and scaffold proteins Dab-1 (7), FE65 (11) and Shc (12,13), the non receptor tyrosine kinases Src and Fyn (14), and the JNK-interacting proteins (JIP-1 & 2) (15,16), which act as molecular scaffolds for the JNK signaling pathway (see Ref. 17 for review). Quite similarly, such intracellular signaling protein complexes have been found for APP (7, 11, 18 -20) and presumably for its closely related homologues amyloid precursor-like proteins 1 and 2. Nevertheless, the cellular signaling function of these receptor families are not well understood, and no systematic investigation on downstream effector functions has yet been performed.
Studies presented here are focused on the elucidation of intracellular key signaling pathways that occur downstream of LRP (as one representative member of this family) and in direct comparison downstream of APP activation. Given the complex situation of the LRP ectodomain binding protein network, we here employed the established chimeric receptor approach (21,22), fusing a transmembrane region and the extracellular sIgG F c -domain to the intracellular tails of APP and LRP, respectively. Our results indicate that in the transfected Jurkat TAg cell line ectopic expression of the cytoplasmic domain of LRP (but not or at least much less of APP), as a specific constituent downstream or parallel of MEKK-1, exerts modifying signaling functions and that active JNK is sequestered by plasma membrane resident LRP multiprotein complexes, preventing JNK from translocation into and transactivation of Elk-1 and cJun within the nucleus. Thus, this study formally demonstrates, for the first time, that lipoprotein receptors (via the established interaction with JIP) (15,16) modulate mitogenic signaling function(s), e.g. scaffold JNK in intact cells, providing a novel insight into the role of JNK sequestration and subsequent inhibition of nuclear JNK target substrates, cJun-and Elk-1 transactivation by LRP.
sIgG-Receptor Cross-linking-For reporter gene analysis, anti-F c antibody-mediated cross-linking was done for 16 h employing Dynabeads-Protein A (Dynal, Hamburg Germany) coated with rabbit anti-human IgG. For apoptosis detection soluble goat anti-human IgG was used for cross-linking.
Cells and Transfections-PC12 cells were cultivated in RPMI 1640 medium complemented with 2 mM glutamine, 100 units/ml penicillin, 5% fetal calf serum, 0.1 mg/ml streptomycin, and 10% heat-inactivated horse serum. Cells were fed every second day and split once a week (split ratio 1:10). Culture plates had to be coated with collagen S (120 g/ml in distilled water). 24 h after medium exchange cells were harvested and washed in prewarmed PBS (10 min, 200 g, room temperature). 2 ϫ 10 6 cells (resuspended in RPMI medium without supplements) were mixed with up to 45 g of plasmid DNA (total volume 400 l) and transferred into cuvettes and incubated at room temperature for 15 min. The electroporation was performed in an Electro Cell Manipulator 600 using predetermined optimal conditions (250V, 48⍀, 1100 microfarads) (28). After transfection cells were incubated at room temperature for an additional 15 min before cultured in RPMI medium with supplements at 37°C.
CD4 ϩ Jurkat TAg T cells (a kind gift from Dr. G. Crabtree, Stanford University, Stanford, CA) that stably express the SV40-derived large T antigen were cultivated in RPMI 1640 medium complemented with 2 mM glutamine, 100 units/ml penicillin, 0.1 mg/ml streptomycin, and 10% heat-inactivated fetal calf serum. 24 h after medium change Jurkat TAg cells were harvested by centrifugation, washed three times in medium without supplements, resuspended at a concentration of 2 ϫ 10 7 cells/400 l, and incubated on ice for 15 min. Transient transfection of cells was performed by electroporation in a BTX T820 ElectroSquare-Porator (ITC, Biotech, Heidelberg, Germany) apparatus using predetermined optimal conditions (volume, 400 l of cell suspension; mode, 99 s; voltage, 450 V; number of pulses, 5). After electroporation cells were incubated for 15 min at room temperature before cultured in RPMI medium with supplements at 37°C. Promoter Reporter Gene Analysis-Reporter gene expression was measured in co-transfection assays using 20 -30 g of the sIgG expression constructs and 10 g of NF-B-Luc or 16 g of pFRLuc promoter firefly luciferase reporter (RLU1) plus 4 g of pFA-Jun or pFA Elk-1, respectively, as described in Ref. 29. After 30 h cells were stimulated with 50 ng/ml PMA and 1 g/ml ionomycin for up to 16 h or left unstimulated (Me 2 SO buffer control), as indicated. To allow a comparison between different experiments, which vary with regard to transfection efficiencies, the different cell populations to be tested were co-transfected with a housekeeping expression plasmid (0.3 g of the Renilla luciferase reporter vector pTK-Renilla-Luc, RLU2, Promega, Madison, WI) driven by the Herpes simplex virus thymidine kinase promoter, and results of enhancer/promoter activation (i.e. firefly luciferase reporter protein expression) were normalized to the corresponding Renilla luciferase reporter. Luciferase activities were determined using the Dual-Luciferase Reporter Assay System (Promega) and a 1450 Micro-Beta WALLAC Jet Liquid Scintillation and Luminescence Counter. PC12 cells were stimulated in triplicates with 5 g/ml NGF-␤ for 16 h for reporter gene analysis. 50 g/ml activated or native ␣2macroglobulin were added. The results shown are obtained with different preparations of expression plasmids and represent the mean Ϯ S.E. from representative experiments done in duplicates.
FACS Analysis-sIgG surface expression of transfected cells was detected with goat anti-human IgG (Dianova, Hamburg, Germany) and, as secondary reagent, donkey anti-goat IgG-fluorescein isothiocyanate (Jackson ImmunoResearch), each employed for 30 min at 4°C and analyzed by FACS scan.
Western Blot Analysis-SDS-PAGE was performed under reducing conditions using 4 -12% Bis/Tris-buffered gels (Novex, San Diego, CA) and MOPS-running buffer for 90 min at 150 V. Proteins were transferred onto a polyvinylidene difluoride membrane (Millipore, Bedford, MA) by semi-dry blotting (90 mA/112 cm 2 , 90 min, room temperature). Blots were blocked in Tris-buffered saline containing 0.05% Tween 20 and 5% non-fat dry milk for 1 h at room temperature, incubated with primary antibodies over night at 4°C and peroxidase-conjugated secondary antibodies (Pierce, 1:5.000) for 1 h at room temperature. For antigen detection enhanced chemiluminescence was used (Super Signal, Pierce).
Apoptosis Detection-Cerebellar neurons were cultivated 4 days in full BME medium (on poly-L-lysine-coated glass cover slips) and then preincubated with 50 g/ml activated or inactivated ␣2-macroglobulin for 16 h. Cells were washed once with prewarmed PBS and incubated with supplemented medium containing 8 M Hoechst 33342 for 15 min at 37°C in the dark. 5 g/ml propidium iodide were added, and cells incubated for an additional 5 min. Cells were washed twice in PBS and counted in an axioplan fluorescence microscope (PI: excitation: 570 nm, emission: 590 nm/Hoechst: excitation: 400 nm, emission: 420 nm) and pictures were taken with an Olympus U-CMAD-2 camera.
Jurkat TAg cells were co-transfected with 20 g of the indicated pEF-neomycin empty vector control, bcl-2, or sIgG expression constructs and the surface marker expression plasmid encoding the murine major histocompatibility complex class I H2-K k , respectively, and incubated for 48 h. Activation of mitogen-activated protein kinase ERK and JNK-2 ϫ 10 7 Jurkat TAg cells per assay point were transiently co-transfected with 10 g of epitope-tagged reporter kinase constructs plus 30 g of the various sIgG cDNA expression plasmids, encoding sIgG-receptor chimera or receptor control as indicated (sIgG-receptor expression plasmid was used in 3-fold molar excess to statistically ensure cotransfection in each cell, which harbors the reporter kinase construct). 48 h posttransfection (with or without stimulation with 50 ng/ml PMA and 1 g/ml ionomycin for 15 min), cells were harvested, solubilized in lysis buffer (5 mM NaP 2 P, 5 mM NaF, 5 mM EDTA, 50 mM NaCl, 50 mM Tris, pH 7.3, 2% Nonidet P-40, and 50 g/ml each aprotinin and leupeptin) for 30 min on ice and centrifuged at 16,000 ϫ g for 15 min at 4°C. After preclearing, the epitope-tagged protein kinases were immunoprecipitated by incubation overnight at 4°C with 2 g of the respective anti-tag antibody and protein-G-Sepharose (Pharmacia Biotech Inc.) for 1 h at 4°C. The immunoprecipitates were washed six times with lysis buffer. The reporter protein kinase was then resolved by immunoblotting with anti-active and pan mitogen-activated protein kinase (ERK and JNK) antibodies (New England Biolabs, Inc.). A representative result of at least five experiments is shown.
Fluorescence Microscopy-2 ϫ 10 7 Jurkat TAg cells per assay point were co-transfected with 1 g HA-tagged JNK constructs plus 20 g of the sIgG-receptor chimera or receptor control as indicated. 40 h posttranfection cells were stimulated with 50 ng/ml PMA and 1 g/ml ionomycin for 60 min and seeded on poly-L-lysine-coated glass cover slips and incubated at 37°C and for 10 min. Cover slips were washed, and the cells fixed and permeabilized with the Fix and Perm Kit (An der Grub Bioresearch), according to the manufacturer's instructions. Rhodamine-conjugated rabbit anti-human IgG-F c and fluorescein conjugated rat anti-HA antibodies were added to the permeabilization solution, and cells incubated for 20 min at 25°C. The cover slips were washed four times with PBS and mounted on slides with mowiol mounting solution (Calbiochem). Microscopy was performed with the Olympus BX50 Fluorescence Microscope, and pictures were taken with the digital camera Microview TE/CCD1317-K/1 (Princeton Scientific Instruments) using the metamorph imaging software (Universal Imaging Corporation). Finally the pictures were processed with the AutoDeblur 7.5 software (Autoquant Imaging, Inc.).
Cell Fractionation-2 ϫ 10 7 Jurkat TAg cells per assay point were transfected with 30 g of cDNA expression plasmids encoding chimeric sIgG-receptors, sIgG or sIgG-LRP respectively, plus 1 g of JNK-2-HA for subcellular fractionation of JNK. After incubation for 44 -46 h cells were stimulated with 50 ng/ml PMA and 1 g/ml ionomycin for up to 60 min at 37°C or left unstimulated (Me 2 SO buffer control). Cell fractionation was performed by subsequent lysis and centrifugation in equivalent amounts of different buffers: "n, nuclear fraction" (pelleted by 10 min and 1000 ϫ g centrifugation) and "s, cytoplasmic (post-1000 ϫ g soluble) fraction" were resolved by lysis buffer without detergent (5 mM Na 3 VO 4 , 5 mM NaF, 2 mM EDTA, 5 mM EGTA, 1 mM dithiothreitol, 20 mM Tris pH 7.3, and 50 g/ml each aprotinin and leupeptin), "pt, particulate/membrane fraction" was resolved in lysis buffer containing 1% Nonidet P-40, as described (29). Data are expressed as the means Ϯ S.E. (n ϭ 6).

Characterization of sIgG-APP and sIgG-LRP Receptor Fusion
Mutants-Two distinct sIgG-receptor chimeric expression constructs, termed sIgG-LRP and sIgG-APP, were constructed by fusing transmembrane region and extracellular IgG-F c domain to the cytoplasmic COOH termini of the human LRP and APP cDNA, respectively (see Fig. 1A for illustration). Ectopic Jurkat TAg cell transfection with these chimeric sIgG constructs (and the truncated sIgG-receptor control protein, expressing no cytoplasmic fusion tail) followed by FACS employing the anti-IgG F c antibody demonstrated cell surface expression of sIgG chimera in transfected cells (Fig. 1B). Interestingly (and most likely due to altered kinetics of processing, surface expression and/or endocytotic rates exerted by the internalization signals of the LRP and APP COOH-tail domains), reduced levels of surface expression of these chimeric sIgGreceptors versus sIgG-receptor control could be reproducibly detected (48 h posttransfection, 20% versus 50% sIgGreceptor-expressing cells, respectively). Immunoblot analysis of plasma membrane fractions employing F c -specific antisera detected protein smears of 54 -65 kDa proteins in cells transfected with the corresponding sIgG-expression plasmids but not in cells transfected with empty vector control (Fig. 1C). The distinct molecular appearance of the different sIgGreceptor proteins indicates posttranslational glycoyslation of the F c domain.
Activated ␣2-Macroglobulin Exerts Strong Apoptosis Protection during ex Vivo Culture of Primary Neuronal Cells-First, and for comparison purposes with our chimeric sIgG-LRP receptor, we have examined the ability of endogenous LRP signaling pathways to modulate cellular survival. Activated ␣2macroglobulin, an established ligand of LRP (1) reduces (by 40%) apoptosis of the ex vivo cultured rat cerebellar granule cells ( Fig. 2A). Native ␣2-macroglobulin, which is not a ligand for LRP, served as control. Similarly, decreased susceptibility to DNA damage, were reproducibly observed by sIgG-LRP expression (29% increase in survival, relative to sIgG control; Fig.  2B). As positive control, transient bcl-2 overexpression also inhibited DNA damage-induced apoptosis in parallel experiments (Fig. 2C). This provides evidence that ␣2-macroglobulin-exerted survival effects (through the endogenous LRP-mediated pathway) can be bypassed by ectopic sIgG-LRP expression.

Activated ␣2-Macroglobulin Modulates Signaling Function on Mitogen-induced Elk-1 Enhancer/Promoter Reporter Gene
Transactivation in PC12 Cells-We then extended our analysis by testing the signaling role of endogenous LRP on mitogenic signaling function employing transient transfection with promoter reporter constructs. Activated ␣2-macroglobulin by itself did not affect signals to Elk-1. Interestingly, however, ␣2-macroglobulin treatment altered the cellular response to NGF, e.g. significantly abrogated mitogen-induced Elk-1 and cJun transcriptional activity (Fig. 3 and not shown), indicating that the ERK and/or JNK pathways operate downstream of LRP in the regulation of NGF signaling. This very reproducible result prompted us to examine mitogenic signaling modulation employing our sIgG-LRP chimeric receptor approach.
sIgG-Receptor Chimera by Itself Are Not Sufficient to Induce Downstream Elk-1, cJun, or NF-B Enhancer/Promoter Reporter Gene Transactivation-To investigate the potential role of sIgG-LRP and sIgG-APP (in comparison to sIgG-control) in induction of the Elk-1, cJun, and NF-B-promoter reporters, cells were transiently co-transfected with sIgG-chimeric cDNA constructs and the promoter reporters and subsequently stimulated for up to 16 h with PMA/ionomycin or left unstimulated, lysed, and assayed for reporter gene expression. Consistent to the endogenous LRP function (Fig. 3), sIgG-LRP, and sIgG-APP by itself (with or without bead-bound anti-F c antibody-mediated clustering) are not sufficient to induce signaling function on these pathways under investigation (Fig. 4A). Additionally, no change in basal phosphotyrosine staining of total cellular proteins could be detected upon expression of sIgG-APP and sIgG-LRP (with or without anti-F c cross-linking, not shown). However, expression of the established receptor sIgG-Syk, a protein tyrosine kinase isoenzyme of the ZAP-70 family, induced strong changes of total cellular phosphotyrosine protein staining (not shown), strictly dependent on anti-F c cross-linking (26).
sIgG-LRP Expression Mediates Negative Signaling Function on Mitogen-induced Elk-1 and cJun, but Not NF-B, Enhancer/ Promoter Reporter Gene Transactivation-Therefore we decided to further investigate the cellular signaling cascade(s) involving sIgG-APP and sIgG-LRP vis-à -vis of mitogenic activation leading to Elk-1, c-Jun, and NF-B-transactivation employing pleiotropic phorbol ester and ionomycin stimulation. As shown in Fig. 4B, transient overexpression of sIgG-LRP (relative to the sIgG-receptor control) caused a significant decrease in the transcriptional activation of the Elk-1 (Fig. 4, B and C) and cJun (Fig. 4D) promoter reporter gene, (with or without bead-bound anti-F c antibody-mediated clustering) indicating interference function(s) of LRP on the given signaling pathways. This phenomenon was also observed with endogenous LRP signaling (see Fig. 3).
Interestingly, and in contrast to similar binding results of JIP to both, LRP and APP (1), sIgG-APP mutant was much less effective (cJun, Fig. 4D) and even failed (Elk-1, Fig. 4, B and C) to significantly abrogate mitogen-induced promoter reporter activity even though both transgenes have been properly expressed (not shown and Fig. 1, B and C). Again, inhibition of cJun-and Elk-1-regulated reporter gene expression was not further enhanced by anti-F c antibody-mediated cross-linking (data not shown), suggesting that expression levels of LRP and not LRP-clustering is the predominant factor of LRP signaling modifier function(s). Importantly, sIgG-LRP-mediated inhibition of the transcription factor Elk-1 and cJun-activity appeared to be selective since sIgG-LRP failed to significantly abrogate the NF-B reporter (Fig. 4E). Additionally, the transfection control Herpes simplex virus-thymidine kinase-promoter reporter was not significantly modulated by sIgG-LRP and sIgG-APP expression compared with vector controls (not shown), excluding sIgG-LRP expression as a general inhibitor of transcription. These data indicate that sIgG-LRP, but not or at least much less sIgG-APP, suppresses signaling by the JNK/ cJun and the JNK/Elk-1 and/or ERK/Elk-1 pathways. (17). To define a potential role of LRP and/or APP for the activation of these two distinct MAPK signaling pathways, cells have been transiently co-transfected with the different sIgG-receptor constructs plus epitope-tagged reporter MAPK constructs (e.g. Myc-tagged ERK2 or HAtagged JNK2). The active phosphorylation status of these reporter kinases was then determined by immunoprecipitation via their epitope tags and subsequent immunoblotting with the anti-active JNK and ERK antibodies (recognizing only their dual pTyr and pThr phosphorylation status, Fig. 5, A and B). As a result, analysis of sIgG-LRP-expressing cells reproducibly demonstrated significant enhanced mitogen-mediated activation of JNK (estimated 5-fold increase relative to the sIgGreceptor control). Consistent with results shown in Fig. 4, the activity of JNK was only significantly enhanced in cells stimulated with PMA plus ionomycin, and no sIgG-LRP-mediated effect has been observed in non-stimulated cells (Fig. 5A). Expression of sIgG-APP did not induce any enhancement on mitogen-activated protein JNK-2 (Fig. 5A) and both sIgG-LRP and sIgG-APP did not enhance the activation of extracellular signal-regulated kinase 2 (ERK-2), even though PMA/ionomycin treatment increased the activity of this protein kinase (Fig.  5B). Similar expression of the kinase reporter proteins has been confirmed (see Fig. 5, C and D), excluding artifacts by different reporter kinase expression levels. Therefore, signaling function of sIgG-LRP (but not sIgG-APP) may be mediated, in part, by the JNK signaling pathway.

sIgG-LRP Chimera-transfected Cells Demonstrate a Significant Enhancement of Mitogen-induced JNK but Not ERK Activation-Mitogenic stimulation with pleiotropic phorbol ester and ionomycin induces the activation of the MAPK family members, ERK and JNK
sIgG-LRP and JNK Are Found to Significantly Co-localize in Co-transfected Cells-Although JNK is known to be located in both the cytoplasm and the nucleus of quiescent cells, activation of JNK is associated with enhanced accumulation of JNK in the nucleus (see for review Ref. 17). To investigate how plasma membrane resident sIgG-LRP may affect the nuclear transactivation function of JNK, we examined the subcellular distribution of JNK in transfected cells. First, expression of sIgG-LRP (but not sIgG-receptor control) caused significant co-localization of JNK with the sIgG-LRP receptor protein at FIG. 4

. Selective inhibition of GAL4-cJun and GAL4-Elk-1-mediated reporter gene expression in cells expressing sIgG-LRP.
Jurkat cells were electroporated with distinct promoter/luciferase reporter plus the indicated sIgG-LRP, sIgG-APP, or sIgG-control expression constructs. A, at 48 h posttransfection, cells were treated with or without anti-F c cross-linking antibody, bound to beads (for an additional 16 h, as indicated by ϩ), and luciferase reporter gene expression was determined. B-E, the specific transreporting system was utilized to investigate Elk-1 (B and C), cJun (D), and NF-B (E) transactivation. At 48 h posttransfection, cells were split and incubated in the absence (Me 2 SO buffer control) or presence of PMA (50 ng/ml) and ionomycin (1 g/ml) and reporter activity was determined. Normalized luciferase activity was quantified. Statistical analysis of two (C-E) and four (B) independent experiments is shown. the plasma membrane as indicated by the yellow staining in the overlay (Fig. 6A) presumably via its postulated interaction with endogenous JIP (15), an established cytoplasmic anchor for JNK. This already indicates to us that the LRP/JIP-1 complex may inhibit the biological effects of the JNK signaling pathway; e.g. interaction of JNK with the LRP/JIP complex may account for the retention of JNK in the plasma membrane regions of the cell. Second and consistently, expression of sIgG-LRP significantly reduced the amount of nuclear JNK in stimulated cells: JNK was mostly detected in the cytoplasm of untreated and mitogen-treated cells transfected with sIgG-LRP (Fig. 6A). In contrast, JNK was detected in both the cytoplasmic and the nuclear compartments of mitogen-treated cells, transfected with sIgG-control (Fig. 6A) as well as sIgG-APP expressing cells (data not shown). This result was independently confirmed by subcellular fractionation assays, demonstrating a 52% reduction of nuclear JNK-HA reporter kinase in stimulated sIgG-LRP versus sIgG control-transfected cells (Fig. 6B). In contrast, no significant changes of cytoplasmic JNK protein levels of resting or mitogen-treated cells, transfected with sIgG-control or sIgG-LRP-expressing cells have been observed (data not shown). Nuclear targets of the JNK signal transduction pathway include e.g. the transcription factors cJun and Elk-1 (17,30). Thus, as a working hypothesis, overexpression of sIgG-LRP may cause cytoplasmic retention of JNK, thereby potentially inhibiting Elk-1 and cJun phosphorylation and, subsequently, gene regulation by JNK.

sIgG-LRP Abrogates Gain-of-Function Mutants Rac-1-and MEKK-1-induced cJun-Enhancer/Promoter Reporter Gene
Transactivation-This observation prompted us to further examine the modulatory effects mediated by expression of sIgG-LRP and in direct comparison of JIP-1 (as well as the combination of sIgG-LRP plus JIP-1 expression) on oncogenic small GTPase Rac-1V12 and MEKK-1 (17) catalytic fragmentinduced cJun transactivation. As a result, expression of sIgG-LRP (relative to the sIgG-receptor control) was capable of significantly blocking CA-Rac-1-and CA-MEKK-1-induced cJun promoter reporter activity (Fig. 7), suggesting that LRP function is parallel or downstream with MEKK-1 for cJun transactivation. JIP-1 overexpression also inhibited CA-Rac-1 and CA-MEKK-1 mutant-induced cJun transactivation. Nevertheless, no additive inhibition of the Rac1/MEKK1/JNK/cJun signaling cascade could be observed by co-transfection of both sIgG-LRP and JIP-1, consistent with a role of sIgG-LRP and JIP-1 in the same pathway. DISCUSSION Mouse genetic (5,31) as well as in vitro biochemical evidence (7,15,16) over the last few years strongly argued that lipoprotein receptors might couple receptor ligand binding to intracellular signaling events (see for review Ref. 32). Evidence in support of these findings are reports that native LDL, as in-FIG. 6. sIgG-LRP and JNK colocalize in co-transfected cells, and sIgG-LRP abrogates nuclear location of JNK in mitogenstimulated cells. A, indirect immunofluorescence analysis of the subcellular distribution of transfected JNK2-HA and sIgG and sIgG-LRP, respectively. The cells were exposed by PMA/ionomycin for 60 min before fixation. sIgG-receptors were detected with the anti-F c rhodamine-conjugated antibody (red). JNK2-HA was detected with the fluorescein-isothiocyanate-labeled anti-HA-tag antibody (green). Images were collected by digital imaging microscopy. For confirmation of the nuclei, DAPI staining was used in parallel experiments (data not shown). B, Nuclear distribution of JNK2-HA was detected by biochemical fractionation assays and anti-HA-tag immunoblotting as described in M&M section. Data are expressed as the means Ϯ S.E. (n ϭ 6). teractor of the closely related LDL receptor, also modulated cJun/AP-1 (but not NF-B) activation in endothelial cells (33,34). More relevant, LRP (next to ApoER2 and megalin) is a critical receptor for apoE, whose allelic status is a major risk factor for late-onset Alzheimer's disease (AD) (35) and other neurodegenerative disorders. Consistently, apoE-mediated functional modulation (potentially via agonistic binding of lipoprotein receptors) of cellular responses to mitogens as well as apoE-mediated inhibition of cell proliferation has been described (36). Recent work in mice also established LRP6 as critical co-receptor for WNT signal transduction (37,38,39) known to mediate cell differentiation, cell polarity, and cell adhesion. However, the intracellular mechanisms and downstream effector function(s) of this lipoprotein receptor family in vivo are still poorly understood.
To examine by molecular cell biology the discussed role of one representative of these lipoprotein receptors in mitogenic-mediated signal transduction, sIgG-receptor chimera encoding the cytoplasmic domains of LRP (as representative family member) as well as APP were constructed and biochemically characterized (Fig. 1). Ectopic expression of sIgG-LRP and signal transduction studies revealed that sIgG-LRP is capable of transducing signals, e.g. it attenuated Elk-1 reporter gene activation induced by mitogen (Fig. 4). This is consistent to endogenous LRP signaling function, triggered by activated ␣2-macroglobulin (Fig. 3). Alternatively, sIgG-LRP attenuated cJun reporter gene activation induced by constitutively active mutants of the small GTPase Rac-1 and MEKK1, an upstream kinase that preferentially activates JNK (Fig. 7). Overexpression of JIP-1, a cytosolic scaffold protein that binds JNK (40), also suppressed transcriptional activity of cJun (Fig. 7) in parallel experiments, consistent with its established inhibition of JNKregulated gene expression (41).
The ability of LRP to modulate cJun/Elk-1 and therefore AP-1 transcriptional activity suggests that it may also be sensitive to negative signals that are involved in the induction of apoptosis. Decreased susceptibility to DNA damage, known to depend on p53 and JNK action (42)(43)(44)(45), were reproducibly observed by sIgG-LRP expression (29% increase in survival, relative to sIgG control; Fig. 2B). In contrast, sIgG-LRP expression did not alter promotion of apoptosis by staurosporine (data not shown). Therefore, expression of sIgG-LRP selectively reduced DNA damage-induced apoptosis. Even so, the in vivo relevance of this finding remains speculative; these data indicate that LRP may modulate the biological actions of the JNK signal transduction pathway, known to be involved in apoptosis, neuronal damage, and stress responses. Consistent with this hypothesis, activated ␣2-macroglobulin triggered an antiapoptogenic program involving LRP (46) (Fig. 2A).
It appears to be of physiological importance that although both lipoprotein receptors and APP are reported to bind and co-localize with JIP (15,19,20,47), only overexpression of sIgG-LRP (but not, or at least much less, sIgG-APP) delivers a signal contributing to Elk-1 and cJun enhancer modulation (Fig. 4, B-D) or apoptosis protection (Fig. 2), indicating a functional specialization rather than a simple functional redundancy between sIgG-LRP and sIgG-APP. The fact that transient transfection of sIgG-LRP and sIgG-APP induces similar surface expression (Fig. 1B) excludes the possibility that the level of sIgG-APP expression or its protein stability in intact cells is significantly lower than that of sIgG-LRP. Along this line, sIgG-APP (but not sIgG-LRP) chimera-transfected cells demonstrated a significant enhancement (150% induction relative to sIgG-control, set as 100%) of mitogen-induced NF-B transactivation (Fig. 4E). Therefore, the functions of these two receptors appear not to be simply overlapping. The data sug-gest that the distinction between the two receptor chimera lies in intrinsic differences in the cytoplasmic tails. Such functional divergence of distinct receptors is biologically expected and provides a rationale to further explain the presence of multiple APP and lipoprotein receptor members in a given cell.
Because JIP intracellularly interacts with lipoprotein receptors such as LRP, our observed subcellular retention of activated JNK by sIgG-LRP chimera (Fig. 6) support a protein sequestration model (as illustrated in Fig. 8), i.e. LRP/JIPmediated plasma membrane targeting of JNK as a mechanism that regulates its nuclear translocation. Importantly, however, while sIgG-LRP (via JNK sequestration) suppressed PMA/ ionomycin-induced transcriptional activation of Elk-1 and cJun-promoter reporter gene constructs, it also strongly enhanced PMA/ionomycin-mediated JNK induction (Fig. 5A). This reproducibly observed enhancement of JNK activation might very well be due to co-recruitment of additional signaling molecules (involved in JNK stimulation) to the plasma membrane-resident sIgG-LRP/JIP/JNK signaling complex. Consistent with this theory, tyrosine phosphorylation of the intracellular domain of LRP was recently found to involve first a functional upstream role of phosphatidylinositol 3-kinase pathway (14) and second a downstream physical Shc association (13). Shc and phosphatidylinositol 3-kinase are both established upstream signaling molecules in the activation of JNK, e.g. by growth factors (48 -50).
The JNK signal transduction pathway is known to be activated in response to environmental stress and by the engagement of several classes of cell surface receptors, including cytokine receptors, serpentine receptors, and receptor tyrosine kinases (51,52). The results of this study now confirm and extend previous biochemical work and add LRP (as representative of the lipoprotein receptors) to the list of JNK recruiting The role of the cytoplasmic domain of LRP is to couple to intracellular signal transduction mechanisms, e.g. exerting physical and functional cross-talk with the JNK pathway. Mitogen-activated JNK may be sequestered into the plasma membrane resident LRP/JIP multiprotein complexes, preventing activated JNK from translocation into the nucleus and, subsequently, nuclear transactivation of the JNK-dependent transcription factors, cJun and/or Elk-1. receptors. Consistently, genetic studies of Drosophila have demonstrated that JNK is required, similarly to lipoprotein receptors in mice (5,31,38), for early embryonic development (53). Nevertheless, and despite the observed role of sIgG-LRP/ JIP-1 to suppress signal transduction by the JNK pathway, LRP/JIP-1 may have a more direct role in targeting JNK to specific regions of the cell (i.e. microdomains at the plasma membrane), i.e. thereby targeting activated JNK to specific substrates. Indeed, disruption of the JIP-1 gene in mice by homologous recombination abrogated stress-induced JNK activation (52), identifying JIP-1 as critical positive component of a mitogen-activated kinase signal transduction pathway. Thus the LRP/JIP/JNK complex may serve to actively regulate e.g. cytoskeletal rearrangements (via p150-Spir, an established downstream target of JNK and direct link between JNK and actin organization (54,55) but also other yet undefined substrates/effectors of JNK), potentially affecting the intracellular vesicle transport, cell migration, axon guidance, and synaptic plasticity.
In human brain, neurons and microglia are known to express high levels of LRP (3,4). ApoE isoforms differentially modulate neurite outgrowth (apoE4 inhibits and apoE3 promotes neurite outgrowth) (56), suggesting a potential mechanism whereby apoE4 may cause regenerative failure of neurons and therefore accelerate their degeneration. Thereby, LRP has been directly implicated to neurite outgrowth (56 -60) and synaptic plasticity (34). Similarly, a pathogenic role of microglia cells has long been suggested since microglia is activated at sites of neuronal degeneration (61,62). One hypothetical scenario is that the reported role of apoE isoforms in nervous system injury and AD disease progression may be mediated via apoE:LRP ectodomain interaction, potentially altering LRP signaling function. Because in this putative scenario, apoE only modulates LRP function, i.e. in regenerative synaptic plasticity this would be consistent with the nature of apoE as susceptibility factor, as observed in late-onset AD. Interestingly, the silent C776T polymorphism in the LRP gene (a discussed allelic risk factor for late-onset AD) significantly lowers levels of LRP in the brain (63). Decreased LRP expression then may affect critical signaling thresholds of LRP receptor (e.g. in response to agonistic ligand binding) and impair effective protection from neuronal degeneration. Additionally, intramembranous proteolytic processing of LRP (similar to APP) (64, 65) from the plasma membrane may involve concerted subcellular translocation of preformed signaling complexes (6); however, the biochemical pathways thereby transduced remain unclear. Therefore, it remains to be shown whether the signaling competence of the cytoplasmic domain of LRP within the chimera sIgG-LRP is indistinguishable from that generated by the intact LRP receptor.
Taken together, we conclude from this study, that the intracellular tail of LRP by itself, i.e. expressed in the context of the sIgG-LRP chimera, is competent to modulate transcriptional activation of Elk-1 and cJun and differs from sIgG-APP (encoding the COOH-tail of APP) in its competence to participate in this signaling cascade. Due to the chimeric receptor approach, any role of lipid transport properties as well as co-receptors, e.g. interacting via the large LRP ectodomain, can be ruled out, thereby directly establishing the intracellular tail of human LRP, by itself as a necessary and sufficient component affecting mitogenic effector pathway(s) that occur downstream of LRP. These mechanistic data strongly suggest that the role of the cytoplasmic domain is to exert functional cross-talk with the JNK pathway.