Jun activation domain–binding protein 1 binds Smad5 and inhibits bone morphogenetic protein signaling

Authors


Abstract

Objective

Bone morphogenetic proteins (BMPs) play an important role in the development and the homeostasis and pathology of cartilage tissue, particularly in the differentiation and anabolic activity of chondrocytes. The present study was undertaken to identify binding partners of the Smad proteins, the intracellular mediators of BMP activity, which might actively modify BMP signaling in chondrocytes.

Methods

Yeast 2-hybrid technology was used to screen a complementary DNA library, constructed from human adult articular cartilage, for molecular binding partners of Smad5, a major intracellular mediator of BMP signaling. Primary interaction partners were verified by coimmunoprecipitation, and the relevance of the interactions to BMP signaling was evaluated by transcriptional reporter assay. Additionally, messenger RNA expression analysis (conventional and quantitative polymerase chain reaction) and immunostaining were performed in adult normal and osteoarthritic articular cartilage.

Results

We identified a novel Smad5 interactor, Jun activation domain–binding protein 1 (Jab1), expressed in adult cartilage. The interaction was confirmed in coimmunoprecipitation experiments. Overexpression of Jab1 resulted in an attenuation of BMP-dependent transcriptional responses, suggesting that Jab1 acts as an inhibitor of BMP signaling.

Conclusion

Jab1 is a newly identified intracellular (negative) modulator of BMP signaling in chondrocytes and other cells. Jab1 represents an interesting molecule for understanding anabolic signaling in chondrocytes, as well as a potential therapeutic target for anabolic activation. Most interestingly, Jab1 appears to crosslink the BMP and interleukin-1 pathways.

Balance of matrix turnover is crucial for maintaining the integrity and functionality of normal articular cartilage. In osteoarthritis an imbalance of anabolic and catabolic events in favor of matrix catabolism ultimately leads to the destruction of articular cartilage. The anabolic activity of chondrocytes is thought to be maintained by factors such as insulin-like growth factor 1, transforming growth factor β (TGFβ), and bone morphogenetic proteins (BMPs). In fact, BMP/TGFβ signaling is important for the maintenance of articular cartilage homeostasis ( 1, 2), and the results of recent antisense inhibition experiments suggest that shutting down intrinsic BMP activity leads to disintegration of articular cartilage (1).

The action of BMP is mediated intracellularly by the receptor-associated Smad proteins (R-Smads) 1, 5, and 8, whereas Smads 6 and 7 act as inhibitors (I-Smads) (for review, see ref. 3). In previous studies, we have shown that R- and I-Smads are expressed in human adult articular chondrocytes, but no regulation was observed in normal or osteoarthritic samples (4, 5). In this study, we wished to identify Smad-binding molecules with the potential to modulate intracellular signaling of the BMPs. We used the yeast 2-hybrid (Y2H) method to screen a chondrocyte library for interaction partners with the major R-Smad Smad5, which might potentially modulate chondrocyte behavior and BMP signaling.

MATERIALS AND METHODS

For this type of study no formal approval is required at our institution. There is a formal allowance for performing molecular studies of the kind presented herein, as well as approval for cell culture and histologic studies of human material.

Plasmid construction.

Plasmids pGBKT7-Smad5, pcDNA3.1(-)-FLAG-Smad4, pcDNA3.1(-)-FLAG-Smad5, and pCMV-Myc-Jab1 were constructed as described in the supplementary text, available online at http://www.mrw.interscience.wiley.com/suppmat/0004-3591/suppmat/.

Cell culture and transfection.

Cell lines T/C-28a2, HCS-2/8, and 293 EBNA were cultured as detailed in the supplementary text, available online at http://www.mrw.interscience.wiley.com/suppmat/0004-3591/suppmat/. For transient DNA transfection studies, FuGENE 6 transfection reagent (Roche Diagnostics, Mannheim, Germany) was used at a reagent:DNA ratio of 3:1 (293 EBNA cells) or 4:1 (T/C-28a2 cells).

Cartilage samples for RNA isolation.

For the study of messenger RNA (mRNA) expression levels, cartilage from human femoral condyles was processed as described previously ( 6). Macroscopically normal human articular cartilage was obtained from 8 subjects (mean age 64.5 years [range 48–83]) at autopsy, within 48 hours of death. Osteoarthritic cartilage was obtained from 9 patients undergoing total knee replacement (mean age 70.0 years [range 60–83]). Cartilage was considered to be normal if it showed no substantial softening or surface fibrillation. Cartilage from patients with rheumatoid arthritis was excluded from the study. Only primary degenerated cartilage, and not regenerative cartilage (osteophytic tissue), was used for RNA isolation.

Isolation and stimulation of primary articular chondrocytes.

Normal human knee articular cartilage was obtained from 3 subjects (ages 61–72 years) at autopsy, within 48 hours of death, and osteoarthritic knee articular cartilage was obtained from 3 patients (ages 57–78 years) undergoing total knee replacement. Cells were enzymatically isolated and cultured for 2 days in high-density monolayer cultures (2 × 106 cells/well) in 6-well tissue culture plates with or without serum, as described previously ( 7). Subsequently, chondrocytes were treated for a further 2 days with 200 ng/ml BMP-7 (Stryker Biotech, Hopkinton, MA) or 1 ng/ml interleukin-1β (IL-1β; Biomol, Hamburg, Germany) and cells were harvested for RNA isolation.

Isolation of total RNA from cartilage tissue and isolated chondrocytes.

Total RNA from cartilage tissue and cultured chondrocytes was isolated and complementary DNA (cDNA) synthesized as described previously ( 8).

Y2H screening.

For Y2H screening, the Matchmaker GAL4 Two-Hybrid System 3 (Clontech, Heidelberg, Germany) was used. A human chondrocyte Matchmaker cDNA library, prepared from pooled unstimulated, IL-1β–stimulated, and TGFβ1-stimulated human chondrocytes, was purchased from Clontech. Library screening for Smad5 interactors and analysis of clones were carried out according to the instructions of the manufacturer (http://www.clontech.com/clontech/techinfo/manuals/PDF/PT3247-1.pdf).

Coimmunoprecipitation and immunoblotting.

T/C-28a2 and 293 EBNA cells were transfected with Myc-tagged Jun activation domain–binding protein 1 (Jab1) expression constructs (pCMV-Myc-Jab1) and FLAG-tagged Smad constructs (pcDNA3.1[-]-Smad4/5). Twelve hours after transfection the cells were serum starved (1% fetal calf serum [FCS]) for 24 hours, stimulated with 200 ng/ml BMP-7 for 6 hours, and lysed in Triton-containing lysis buffer. FLAG-tagged Smad proteins were immunoprecipitated overnight with anti-FLAG M2 affinity gel (Sigma, Taufkirchen, Germany). Bound proteins were analyzed by Western blot using anti-Myc or anti-FLAG antibodies.

Transcriptional reporter assays.

Transcriptional reporter assays were performed using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI). The 293 EBNA cells were transfected with the BMP-responsive luciferase reporter plasmid p(SBE)4 and increasing amounts of pCMV-Myc-Jab1. Twelve hours after transfection the medium was changed to fresh serum-deprived medium (Dulbecco's modified Eagle's medium/1% FCS). After 16 hours the cells were stimulated for 24 hours with 200 ng/ml BMP-7. The activation of p(SBE)4 was measured according to the instructions of the manufacturer (Promega). The transcriptional activation of the p(SBE)4 reporter was normalized to the activity of the control plasmid pRL-TK. All assays were performed in triplicate, and the results are presented as the mean (±SEM) from 3 independent transfections.

Immunolocalization studies.

Conventional immunohistochemical experiments were performed on paraformaldehyde-fixed and paraffin-embedded specimens of normal (n = 5) and osteoarthritic (n = 10) articular cartilage, using a streptavidin–biotin complex technique, as described in detail in the supplementary text (http://www.mrw.interscience.wiley.com/suppmat/0004-3591/suppmat/).

For immunofluorescence staining of native Jab1 in cultured cells and of Myc-tagged Jab1 in transfected T/C-28a2 cells, primary antibodies against Myc (diluted 1:200; Clontech) and human Jab1 (sc-23719, diluted 1:100; Santa Cruz Biotechnology, Heidelberg, Germany) were applied, and detection performed using immunofluorescence technology (see supplementary text). For double-labeling of Myc-tagged Jab1 and native Smad5, cells were processed using both antibodies in parallel, and detection was again performed using 2 different fluorescence labeling systems (see supplementary text).

Qualitative and quantitative real-time polymerase chain reaction (PCR).

Primers specific for Jab1 were selected using Primer Express software (Applied Biosystems, Darmstadt, Germany), and PCR was performed as described in the supplementary text. All experiments were performed in triplicate using titrated standard curves. Data obtained were normalized to GAPDH.

Statistical analysis.

For the in vivo investigations, the significance of differences in expression levels was evaluated with the nonparametric Wilcoxon-Mann-Whitney test. For the in vitro probes, the t-test for pairwise comparison was used, due to the limited number of comparisons.

RESULTS

Findings of the Y2H assay.

Screening of the human chondrocyte cDNA library with Smad5 as bait resulted in the identification of many clones belonging to a group comprising partial cDNA inserts of types I and III collagen. These clones were regarded as false-positive and were not further analyzed. One of the interacting clones was found to contain a partial cDNA (amino acids 1–227) coding for Jab1. The occurrence of this interaction in mammalian cells was confirmed by coimmunoprecipitation of overexpressed Jab1 and FLAG-Smad5 in 293 EBNA and T/C-28a2 cells (Figure 1A). The Jab1–Smad5 interaction could be detected in FLAG precipitates from unstimulated and BMP-7–stimulated cells and thus was shown to be independent of BMP-7 activation. While the amount of coprecipitated Jab1 in 293 EBNA cells was reduced to a certain extent after BMP-7 stimulation, no change was detectable in T/C-28a2 cells. A similar interaction was observed with Smad4 ( 9), which was used as a positive control (results not shown).

Figure 1.

A, Interaction of Jun activation domain–binding protein 1 (Jab1) with Smad5. Myc-Jab1 and a FLAG-Smad5 expression construct were coexpressed in 293 EBNA and T/C-28a2 cells. Bone morphogenetic protein 7 (BMP-7) (200 ng/ml) was added to the cells in some experiments. Cell lysates were immunoprecipitated (IP) with an anti-FLAG antibody, and the interaction between Jab1 and Smad5 was analyzed by Western blot (WB) using an anti-Myc antibody. Arrowhead indicates bands corresponding to coprecipitating proteins. Expression of transfected DNA was monitored by direct immunoblotting using anti-Myc antibody. The nonspecific bands in the anti-FLAG blots correspond to the Ig heavy chain (Ig). B, Jab1-induced inhibition of BMP-dependent transcription. Cells of the 293 EBNA line were transfected with the BMP/transforming growth factor β–dependent reporter plasmid p(SBE)4 together with increasing amounts of Jab1 expression construct. In some experiments, cells were treated with BMP-7 for 24 hours prior to cell lysis. BMP-7–dependent transcriptional responses were determined based on expression of the luciferase reporter gene. As a control protein, pCMV-Myc-lamin, which does not interact with Smad proteins, was used. Values are the mean and SEM fold induction of luciferase expression in cells after BMP-7 stimulation compared with untreated cells. C, Jab1 mRNA expression in chondrocytes and chondrocytic cell lines. I, Expression in normal and osteoarthritic knee cartilage. Total RNA was directly isolated from cartilage. Polymerase chain reaction (PCR) results in 3 normal and 3 osteoarthritic joints are shown. Molecular weight markers (M) are shown at the left. C = negative control. II, Quantitation, by real-time PCR, of Jab1 mRNA expression in human knee cartilage. Total RNA was directly isolated from normal (N) cartilage (n = 8) and osteoarthritic (OA) cartilage (n = 9). III, Quantitation, by real-time PCR, of Jab1 mRNA expression in the chondrocytic cell lines HCS-2/8 and T/C-28a2. Cells were left unstimulated (control) or were stimulated for 48 hours with 200 ng/ml BMP-7. IV, Quantitation, by real-time PCR, of Jab1 mRNA expression in cultured primary chondrocytes. Chondrocytes were isolated from normal (n = 3) and osteoarthritic (n = 3) human knee cartilage, cultured for 48 hours in high-density monolayer cultures, and left unstimulated (control) or stimulated for 48 hours with 200 ng/ml BMP-7. Values are the mean and SEM ratio of Jab1 transcripts to GAPDH transcripts.

Dose-dependent inhibition of BMP signaling by Jab1.

Expression of increasing amounts of Jab1 (pCMV-Myc-Jab1) in 293 EBNA cells resulted in a dose-dependent reduction of Smad-dependent transcription of the TGFβ/BMP-responsive luciferase reporter p(SBE)4, while the control protein lamin, which does not interact with Smad proteins, had only a minor effect (Figure 1B). A comparable attenuation of BMP-7–dependent transcription by Jab1 was seen in the chondrocytic cell line T/C-28a2 (data not shown).

Immunolocalization of Jab1 in normal and osteoarthritic cartilage in vivo.

Conventional immunohistochemistry studies showed that Jab1 staining was largely confined to the cytoplasm in most normal and osteoarthritic chondrocytes throughout all cartilage zones, with no significant difference detectable between normal and osteoarthritis samples (Figures 2H and I). These results were confirmed by confocal scanning microscopy (Figures 2K and L). In fetal cartilage (n = 6) moderately strong expression was found in all cells except very late hypertrophic cells (Figures 2G and J).

Figure 2.

Immunolocalization of Jun activation domain–binding protein 1 (Jab1) and transfected Myc-Jab1 in chondrocytes in vitro and in vivo. A–F, Jab1 immunolocalization in cultured primary articular chondrocytes (A and D) and T/C-28a2 cells (B and E) with (D and E) and without (A and B) bone morphogenetic protein 7 (BMP-7) stimulation, and Myc-Jab1 immunolocalization in transfected T/C-28a2 cells with (F) and without (C) BMP-7 stimulation. G–L, Native Jab1 immunolocalization in fetal articular cartilage (G and J [J is a higher-magnification view of the lower portion of G]), normal adult articular cartilage (H and K [confocal imaging]), and osteoarthritic adult articular cartilage (I and L [confocal imaging]). Bars = 100 μm in A, B, D, E, and G–J, and 10 μm in C, F, K, and L.

Localization of endogenous Jab1 in cultured primary chondrocytes and the chondrocytic cell line T/C-28a2.

Both primary chondrocytes and T/C-28a2 cells exhibited Jab1 expression within the cytoplasm. No difference in signal strength or localization was observed after BMP-7 stimulation (Figures 2A, B, D, and E).

Subcellular localization of Myc-tagged transfected Jab1 in T/C28a2 cells with and without BMP-7 stimulation.

Using anti-Myc monoclonal antibodies, a mostly cytoplasmic distribution of Jab1 was detected, with no difference in localization or quantity between BMP-stimulated and non–BMP-stimulated cells (Figures 2C and F). Double labeling for Myc–Jab1 and endogenous Smad5 protein revealed partial colocalization of the 2 proteins in the cytoplasm (Figure 3). The degree of colocalization was unaffected by BMP-7 stimulation of the cells (results not shown).

Figure 3.

Colocalization of Jun activation domain–binding protein 1 (Jab1) and Smad5 in T/C-28a2 cells. T/C-28a2 cells transfected with Myc-Jab1 were analyzed for colocalization of Jab1 and Smad5 by indirect immunofluorescence, using anti-Myc antibodies (green) and anti-Smad5 antibodies (red) (nuclei are shown in blue). Yellow areas represent colocalization of Jab1 and Smad5. A, Jab1 staining. B, Smad5 staining. C–F, Overlays of Jab1 and Smad5 signals. Bars = 20 μm.

Expression of Jab1 in normal articular chondrocytes and chondrocytic cell lines.

Conventional PCR confirmed the expression of Jab1 in T/C28a2 and 2 other chondrocytic cell lines (HCS-2/8 and SW1353) (data not shown) as well as in normal adult articular chondrocytes (3 separate specimens) (Figure 1CI). Quantitative PCR showed low, but substantial, Jab1 mRNA expression in all normal articular cartilage samples (Figure 1CII) and chondrocytic cell lines (Figure 1CIII).

Expression of Jab1 in normal articular chondrocytes in vitro and regulation by anabolic and catabolic agents.

Quantitative PCR revealed a similar level of Jab1 mRNA expression in normal articular chondrocytes in vitro and in vivo (Figures 1CII and 1CIV). No regulation of Jab1 mRNA levels was found after stimulation of the cells with anabolic (BMP-7) or catabolic (IL-1β) stimuli, either in the primary chondrocytes or in the chondrocytic cell lines (Figures 1CIII and 1CIV and data not shown).

Expression of Jab1 in osteoarthritic chondrocytes in vivo.

Expression of Jab1 mRNA was demonstrated in all samples of arthritic chondrocytes investigated by PCR (n = 3) (Figure 1CI). Quantitative PCR showed that levels of expression of Jab1 mRNA in osteoarthritic chondrocytes were similar to those in normal cells (Figure 1CII).

DISCUSSION

This study demonstrates for the first time that Jab1 physically interacts with Smad5 and is able to inhibit BMP signaling. Previously, Jab1 was reported to physically interact with Smads 4 and 7 ( 9, 10) but not with Smads 1, 2, 3, and 6; Smad5 was not tested in those studies. The fact that Jab1 interacts with Smad5 but not Smad1 is of interest because it supports the notion that, despite all of the similarities between members of the Smad subfamilies (e.g., Smads 1 and 5 are both BMP receptor–associated Smads), there are functional differences between them in terms of signal processing.

Jab1 represents subunit 5 of the COP9 signalosome (CSN) (for review, see ref. 11). Although the exact function of CSN is still unclear, the data are consistent with the notion that it has a substantial role as an interface between signal transduction and ubiquitin-mediated proteolysis. Thus, findings suggest that CSN controls signaling pathways in 2 ways: by interacting with certain receptors and their coactivators and by controlling the molecular stability of various transcription factors.

Several interaction partners of Jab1 have been previously described ( 12–14). Of note, some of these interactions, e.g., with choriogonadotropin receptor, p27, and p53, lead directly to ubiquitin-mediated degradation of bound molecules by the 26S proteasome machinery. In contrast, other interactions, e.g., with activator protein 1 (AP-1), lead to cooperative enhancement of signaling activity.

The functional relevance of Jab1 and/or the COP9 complex to the skeleton, and specifically to cartilage, is unclear at present. Jab1-knockout mice die soon after implantation, most likely due to impaired general proliferative activity and increased apoptosis of cells ( 15). In accordance with this, heterozygous animals show reduced skeletal growth. Our results suggest that Jab1 might have a role during skeletal development, at least in part by negatively modulating BMP signaling, which is important for skeletal growth.

In articular cartilage, the anabolic/catabolic balance is crucial for maintenance of the structural integrity of the tissue and its functional properties. In this scenario the underlying balance of anabolic factors such as the BMPs/TGFs and major catabolic agents such as the IL-1/TNF-system ( 16) is thought to be crucial. This is clearly the case at the effector level, since the 2 types of growth factors/cytokines exert largely opposite effects regarding both anabolic and catabolic genes. Also, Fukui and colleagues were able to show that IL-1β stimulates BMP transcription (17), indicating a compensatory loop. Results of this study provide evidence that there is also an intimate linkage of the 2 pathways at the intracellular level; Jab1, which interacts with Smads 4, 5, and 7 and thereby modulates BMP signaling, was originally identified as coactivator of c-Jun/AP-1 activity (14), one major intracellular mediator of IL-1 activity. It will be very important, but also very difficult, to elucidate the intracellular links of these major anabolic and catabolic pathways within the chondrocytes and to determine their importance with regard to the homeostasis of cartilage tissue.

In this study we have identified a new molecule that potentially has an important role in cartilage and chondrocyte biology. It is initially difficult to evaluate the relevance of the newly identified factors using approaches such as Y2H. However, these factors are novel and innovative players, and in cancer research the same type of approaches have ultimately opened up completely new perspectives regarding pathogenesis and treatment. Although in the present study we were not able to clearly demonstrate differences in terms of cartilage degeneration, our findings provide interesting information on the biologic background of chondrocyte behavior. Even if Jab1 is not actively involved in enhanced blocking of BMP signaling in cartilage degeneration, its constant presence and BMP blocking properties, together with its IL-1–modulatory ability, make this molecule an interesting target for therapeutic intervention, i.e., targeting the anabolic/catabolic (im)balance at the intracellular mediator level.

Acknowledgements

We thank Drs. M. Goldring and M. Takigawa for providing the cell lines, Dr. C.-H. Heldin for providing the plasmid p(SBE)4, and Stryker Biotech for providing recombinant human BMP-7.

Ancillary