HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 14, 12053-12060, April 5, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/14/12053    most recent M110872200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Amaar, Y. G. Articles by Mohan, S. Search for Related Content PubMed PubMed Citation Articles by Amaar, Y. G. Articles by Mohan, S. Social Bookmarking                      What's this?

Insulin-like Growth Factor-binding Protein 5 (IGFBP-5) Interacts with a Four and a Half LIM Protein 2 (FHL2)^*

Yousef G. Amaar^§, Garrett R. Thompson^¶, Thomas A. Linkhart^¶, Shin-Tai Chen^¶, David J. Baylink^§, and Subburaman Mohan^§¶**

From the  Musculoskeletal Disease Center, Jerry L. Pettis Veterans Affairs Medical Center, Loma Linda, California 92357 and ^¶ Department of Biochemistry,  Department of Pediatrics, ^§ Department of Medicine, and ^** Department of Physiology, Loma Linda University, Loma Linda, California 92350

Received for publication, November 13, 2001, and in revised form, January 22, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Recent studies using insulin-like growth factor I (IGF-I) knockout mice demonstrate that IGF-binding protein (IGFBP)-5, an important bone formation regulator, itself is a growth factor with cellular effects not dependent on IGFs. Because IGFBP-5 contains a nuclear localization sequence that mediates transport of IGFBP-5 into the nucleus, we propose that IGFBP-5 interacts with nuclear proteins to affect transcription of genes involved in bone formation. We therefore undertook studies to identify proteins that bind to IGFBP-5 using IGFBP-5 as bait in a yeast two-hybrid screen of a U2 human osteosarcoma cDNA library. Five related clones that interacted strongly with the bait corresponded to the FHL2 gene, which contains four and a half LIM domains. Co-immunoprecipitation studies with lysates from U2 cells overexpressing FHL2 and IGFBP-5 confirmed that interaction between IGFBP-5 and FHL2 occurs in whole cells. In vitro interaction studies revealed that purified FHL2 interacted with IGFBP-5 but not with IGFBP-3, -4, or -6. Northern blot analysis showed that FHL2 was strongly expressed in human osteoblasts. Nuclear localization of both FHL2 and IGFBP-5 was evident from Western immunoblot analysis and immunofluorescence. The role of FHL2 as an intracellular mediator of the effects of IGFBP-5 and other osteoregulatory agents in osteoblasts will need to be verified in future studies.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

IGFs¹ are growth-promoting peptides that play an important role in growth, remodeling, and repair of skeletal tissues as well as in modulating development, growth, and cell survival in many other tissues (1-4). IGFs are the most abundant growth factors stored in bone, and they are the most abundant growth factors produced by osteoblast cells in culture (5). IGF-I serum levels correlate with serum levels of bone formation markers (osteocalcin and alkaline phosphatase) as bone formation increases during puberty and after growth hormone therapy or decreases with growth hormone deficiency, menopause, and aging (1, 6-8). Bone formation is severely compromised in mice lacking a functional IGF-I gene (9).

IGFs stimulate bone formation by regulating proliferation, differentiation, and apoptosis of osteoblasts (1, 10). Actions of IGFs on osteoblasts depend not only on the amounts of IGFs but also on the other components of the IGF system including type-I and -II IGF receptors, IGF binding proteins (IGFBPs), and IGFBP proteases (10, 11). IGFBPs either stimulate (e.g. IGFBP-3 and -5) or inhibit (e.g. IGFBP-4 and -6) IGF effects on target cells, and hence, they are an important regulatory part of the IGF system in bone (1, 10, 12, 13).

Of the various IGF system components, IGFBP-5 has several features that suggest it is a key component of the IGF system. IGFBP-5 is the most abundant IGFBP stored in bone, because it is also the only IGFBP that binds avidly to hydroxyapatite (10). Decreased skeletal content of IGFBP-5 has been shown to correlate with decreased skeletal content of IGF-I that may contribute to the impairment in coupling of bone formation to resorption (14). Among the IGFBPs known, IGFBP-5 has been shown to stimulate both osteoblast cell proliferation and activity in vitro (15-17).

Recent findings demonstrate that IGFBP-5 itself is a growth factor with cellular effects that are not dependent on IGFs (18). In this regard, IGFBP-5 treatment increased bone formation parameters in vitro and in vivo in osteoblasts derived from IGF-I knockout mice. IGFBP-5 binds to a putative receptor on the osteoblast cell surface, which may induce downstream signaling pathways (10, 19, 20). IGFBP-5 also contains a nuclear localization sequence that mediates transport of IGFBP-5 to the cell nucleus (21, 22), where it may affect gene transcription. Based on these exciting findings, we have proposed the concept that IGFBP-5 interacts with transcription factors to stimulate transcription of genes that lead to increased osteoblast proliferation.

The idea that IGFBPs may affect cells by IGF-independent mechanisms is not restricted to IGFBP-5. For instance, IGFBP-3 has been shown to mediate its effects on a variety of cell types in part via an IGF-independent mechanism (23-26). Several IGFBP-3-interacting proteins have been discovered using the yeast two-hybrid assay (27, 28). For example, it has been shown that IGFBP-3 interacts with retinoid X receptor-, and this interaction results in the modulation of the transcriptional activity of retinoid X receptor-, which is essential for mediating IGFBP-3 effects on apoptosis (28). IGFBP-3-induced apoptosis was abolished in retinoid X receptor- knockout cells, and IGFBP-3 and retinoid X receptor ligands enhanced apoptosis in prostate cancer cells.

To understand the molecular mechanism of how IGFBP-5 stimulates bone formation by an IGF-independent pathway, it is essential to identify cellular proteins that interact with IGFBP-5. These proteins could be IGFBP-5 receptors as well as nuclear proteins that regulate transcription. We therefore utilized a yeast two-hybrid assay (29) screen to identify proteins that bind to IGFBP-5 using human IGFBP-5 as bait for screening a human osteosarcoma U2 cDNA library. We have identified clones encoding the full or partial coding sequence of the LIM-only protein FHL2 (30, 31) and have shown that FHL2 binds IGFBP-5 but not IGFBP-3, IGFBP-4, or IGFBP-6.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- FHL2 monoclonal antibody was a kind gift from Dr. Muller, University of Freiberg, Germany (32). Recombinant human IGFBP-4 and -5 were expressed in Escherichia coli and purified as previously described (17, 18). IGFBP-6 was purchased from Upstate Biotechnology, Lake Placid, NY. Recombinant human IGFBP-3 was a gift from Dr. A. Sommer (Celtrix Corp., Palo Alto, CA). The yeast Matchmaker two-hybrid system 3 was purchased from CLONTECH, Palo Alto, CA.

Construction of the GAL4 BD:Bait Gene Fusion-- BamHI and SalI double-digestion excised the cDNA fragment encoding the entire IGFBP-5 from the plasmid pGEM-T. Then the cDNA fragment was gel-purified and cloned into the BamHI/SalI-digested pGBKT7 vector, and DNA sequencing confirmed that the cDNA encoding the IGFBP-5 (bait) was in-frame with GAL4 BD (amino acids 1-147).

The bait vector was transformed into the yeast reporter strain AH109, and protein samples were prepared from selected yeast cell lysates using standard protocols or as recommended by the supplier of the Matchmaker two-hybrid system 3 kit (CLONTECH). The expression level of the bait IGFBP-5/GAL4 BD fusion protein was determined by Western blot analysis using a GAL4 BD monoclonal antibody (CLONTECH).

Testing the BD:IGFBP-5 Plasmid for Auto-Activation-- It is crucial to test each BD:bait plasmid construct for auto-activation of reporter gene promoters before performing the two-hybrid screen. To check for auto-activation, the yeast reporter strain AH109 was transformed with the BD:IGFBP-5 plasmid using the Lithium acetate method as described in the Matchmaker two-hybrid system 3 user manual. The AH109 reporter contains three reporter genes, ADE2, HIS3, and MEL1, with GAL4-responsive upstream-activating sequences (AUS) and TATA boxes. ADE2, HIS3, and MEL1 are expressed from GAL1 and GAL2 promoters, which strongly respond to GAL4.

Transformed cells were plated on synthetic complete medium (CM) minus adenine and histidine as well as on CM minus tryptophan (bait plasmid contains a tryptophan as a selection marker for growth control). -X-Gal was added to the medium at 20 mg/ml to screen for auto-activation of MEL1.

Activation Domain (AD):cDNA Library Amplification-- The GAL4 AD:cDNA library constructed in the yeast plasmid pACT2 from U-2Os (U2) human osteosarcoma cell (ATCC HTB-96) mRNA was purchased from CLONTECH. Because a library screen requires up to 500 µg of AD:cDNA plasmid DNA, depending on the yeast strain and the BD:bait plasmid, the AD:cDNA plasmid library in pACT2 was amplified according to the supplier's instructions. We prepared 6 mg of plasmid DNA from amplified E. coli cells using the Nucleobond Mega plasmid prep kit (CLONTECH). The plasmid contains the ampicillin gene for selection in E. coli and a leucine marker for selection in the yeast host.

Library Screen-- The reporter strain AH109 containing the BD:IGFBP-5 plasmid was transformed with the 0.5 mg of AD:cDNA plasmid library using the lithium acetate method as outlined in the THS3 manual. The transformed cells were plated onto low stringency minimal selection medium (lacking histidine, leucine, and tryptophan) and incubated for 4-21 days at 30 °C. The plates were checked for colonies after 4 days of incubation at 30 °C, and positive colonies were picked over a 3-week time period. Positive colonies were transferred to high stringency minimal selection medium plates (lacking Ade, His, Leu, and Trp and containing -X-gal) plates and incubated at 30 °C until sufficient growth was achieved and colonies turned blue. Positive colonies were maintained on high stringency selection media that select for reporter gene activation.

Isolation of AD:cDNA Plasmid-- Colonies that activated all of the reporter genes in the AH109 stain were further analyzed. The AD:cDNA plasmid encoding the interacting protein was isolated from yeast cells using a modified Qiagen (Valencia, CA) plasmid prep method obtained from Qiagen technical support. The plasmid DNA prepared was then used to transform E. coli (HB101), and colonies containing AD:cDNA plasmid were selected with ampicillin (50 µg/ml). The Amp^R colonies were picked and inoculated into 5 ml of LB-Amp medium and grown overnight at 37 °C with shaking. Plasmid DNA was isolated using a Qiagen kit.

Reconfirmation of Positive Clones-- AD:cDNA plasmids isolated from the primary screen were used to transform the AH109 strain containing the BD:IGFBP-5 plasmid to test for activation of reporter genes. Transformed cells were plated on high stringency selection medium containing -X-gal. AD:cDNA clones that confirmed positive were further characterized by DNA sequencing with an automated Applied Biosystems 373A genetic analysis system. Clones that failed to grow in the reconfirmation screen were not pursued for any further analysis.

Osteoblast Cell Culture-- Normal human osteoblasts were isolated as described (33) from calvaria and rib bone specimens obtained from the Cooperative Human Tissue Network, which is supported by the National Cancer Institute, National Institutes of Health. For the present study, cells isolated from calvaria and rib were grown from frozen stocks made at the second passage and were used at passage 3-4. These cells maintain an osteoblastic phenotype for more than six passages (34). U-2 Os (HTB-96) and MG63 (CRL-1427) human osteogenic sarcoma cells were from the American Type Culture Collection (Manassas, VA). SaOs human osteosarcoma cells are a low alkaline phosphatase subline developed by Farley et al. (35). Cells were grown at 37 °C in humidified incubator with 5% CO₂. Growth medium consisted of Dulbecco's modified Eagle's medium (Invitrogen), 10% iron-supplemented newborn calf serum (Hyclone, Logan, UT), and 1% antibiotics (Cellgro).

RNA Isolation and Northern Analysis-- Total RNA from untransformed normal human osteoblasts derived from calvaria and rib and human osteosarcoma cell lines (SaOs-2 and U2) was isolated using the Trizol reagent (Invitrogen). 20 µg of total RNA was loaded on a 1.2% agarose gel, and the gel was blotted using standard techniques (36) after electrophoresis. FHL2 full-length cDNA was randomly ³²P-labeled using a commercial kit (New England BioLabs) for Northern hybridization. A glyceraldehyde-3-phosphate dehydrogenase cDNA was used as a control probe. Probes were hybridized at 42 °C in the presence of 50% formamide using standard protocols.

Preparations of Nuclear and Cytoplasmic Extracts-- Nuclear and cytoplasmic extracts from normal human calvaria and rib osteoblasts and from U2 cells were prepared as described (37). The nuclear extracts were then used for Western blot analysis using the FHL2 monoclonal antibody kindly provided by Dr. Muller, University of Freiburg, Freiburg, Germany.

Immunofluorescent Microscopy-- MG63 cells were plated at 2000 cells/well in 96-well plates in Dulbecco's modified Eagle's medium supplemented with 10% calf serum. The next day the medium was replaced with serum-free medium, and cells were incubated for an additional 24 h before 10 nM human recombinant IGFBP-5 was added. After 48 h of incubation in the presence or absence of recombinant IGFBP-5, cells were fixed, rendered permeable with ethanol, and rinsed with PBS. IGFBP-5 protein localization was detected using IGFBP-5 guinea pig polyclonal antibody and fluorescent anti guinea pig secondary antibody. After staining, cells were rinsed three times with PBS, stored in 50% glycerol, and visualized using an Olympus IX70 epifluorescence microscope.

Bacterial Expression of cDNA Clones of Interest-- A full-length cDNA clone corresponding to the FHL2 was cloned into the expression vector pQE32 (Qiagen) in-frame with the 6-His tag. The FHL2 coding sequence was amplified by PCR from the selected AD:cDNA plasmid template. We used a primer pair to create BamHI and SalI restriction enzyme sites at 5' and 3' ends, respectively. The forward primer was 5'-CGCGGATCCTGACTGAGCGCTT-3' (the bold sequence corresponds to BamHI site, and the underlined sequence correspond to the N terminus of FHL2), and the reverse primer was 5'-ACGCGTCGACAAGTGAACTTGCGGGGTTTTCAGTATCTACG-3' (the bold sequence corresponds to the SalI restriction site, and the underlined sequence corresponds to the pACT2 vector 3' sequence). The PCR product was subsequently purified using QIAXII (Qiagen), digested with BamHI and SalI, and ligated to the vector pQE32. A selected clone was confirmed by DNA sequencing to be in-frame with the 6-His tag. The pQE32 expression vector was transformed into E. coli host M15, and expression was induced with isopropyl--D-thiogalactopyranoside. The rFHL2 was purified according to instructions in the Qiagen Expressionist handbook.

Construction of FHL2 and IGFBP-5 Murine Leukemia Retroviral Vectors-- To overexpress the FHL2 and IGFBP-5 cDNAs in bone cells, we constructed a retroviral expression vector. We cloned the FHL2 and IGFBP-5 cDNA coding sequences of 850 and 750 bp, respectively, in place of the -galactosidase gene in a murine leukemia virus-based retroviral vector plasmid, pCLSA-gal (38). The FHL2 850-bp cDNA fragment was generated by PCR using a pair of oligonucleotides; the forward primer was 5'-ACGCGTCGACATGACTGAGCGCTTT-3', and the reverse primer was 5'-GCGCGGATCCAATTCAGATGTCTTTCCCAC-3'; the IGFBP-5 cDNA fragment was generated by PCR using the forward primer 5'-ACGCGTCGACATGGGCTCCTTCGTGCAC-3' and the reverse primer 5'-CGCGGATCCATCACTCAACGTTGCTGCTG-3' (restriction sites noted in bold). The PCR product was digested SalI/BamHI, purified, and ligated to SalI/BamHI-digested retroviral vector plasmid, pCLSA-gal. The transcription of the FHL2 and IGFBP-5 cDNAs in pCLSA-FHL2 and pCLSA-IGFBP-5 was under the control of the murine leukemia virus long terminal repeat promoter. To produce retroviral expression vector, 293T cells were co-transfected with the pCLSA-FHL2 or pCLSA-IGFBP-5 plasmid and a viral envelope expression plasmid pCMV-G using the calcium phosphate method (39). The viral vectors were harvested 36 h post-transfection, and the titer of the viral vector was determined by end point dilution and by transducing HT1080 cells as described (38).

U2 cells were seeded at 1 × 10⁵ cells/well in 6-well plates. After 24 h of incubation, the cells were transduced with 100 µl each of the viral stocks of pCLSA-FHL2 plus pCLSA-gal, pCLSA-IGFBP-5 plus pCLSA-gal, or pCLSA-FHL2 plus pCLSA-IGFBP-5 vectors as described (38). Multiplicity of infection was 10 based on viral vector titer of ~1 × 10⁷ transforming units/ml. Twenty-four hours after transduction, the cells were rinsed and passaged twice for expansion. The transduction efficiency was determined by staining the cells for -galactosidase expression was estimated to be ~90% after two passages. U2 cells transduced with pCLSA-FHL2 were used to prepare cytoplasmic and nuclear extracts for Western blot analysis using the FHL2 monoclonal antibody. U2 cells transduced with pCLSA-FHL2 and pCLSA-IGFBP-5 were used to prepare whole cell lysates.

Co-immunuoprecipitation of FHL2/IGFBP-5 in Whole Cell Lysates-- U2 cells transduced with pCLSA-FHL2 and pCLSA-IGFBP-5 were used to prepare whole cell lysates using a procedure recommended by Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). 250 µl of cell lysates were incubated in presence of 25 µl of protein A-Sepharose conjugated to FHL2 monoclonal antibody or to only 25 µl of protein A-Sepharose at 4 °C for 14 h on a rotary shaker. After incubation, samples were centrifuged for 1 min at 14,000 rpm. The protein A-Sepharose pellets were washed four times with the lysis buffer (150 mM NaCl, 50 mM Tris, pH 8.0, 1.0% Triton X-100) and resuspended in 50 µl of SDS sample buffer. The immune complex was dissociated from protein A by boiling, and 20 µl was removed and analyzed by SDS-PAGE in 12% acrylamide gel. The IGFBP-5 in the immune precipitate was detected by IGFBP-5 antibody in Western blot analysis according to standard protocols (42).

Co-immunoprecipitation of FHL2/IGFBP-5 in Vitro-- IGFBP-5, IGFBP-4, and IGFBP-6 were ¹²⁵I-labeled as described previously (40). The ability of FHL2 to bind to IGFBP-5 was analyzed by immunoprecipitation. The FHL2 protein (100 ng/ml) was incubated with FHL2 monoclonal antibody (1:500 dilution) and ¹²⁵I-IGFBP-5 (100,000 cpm/ml) for 14 h at 4 °C in 250 µl of incubation buffer (30 mM sodium phosphate, pH 7.4, 10 mM EDTA, 0.1% bovine serum albumin, and 0.5% Tween 80) (41). For competitive binding experiments, 1 µg/ml unlabeled IGFBP-5 was included in addition to ¹²⁵I-IGFBP-5. After incubation, 10 µl of protein A-Sepharose (Upstate Biotechnology) was added, and the samples were incubated at room temperature for 1 h with mixing every 10 min, then were centrifuged for 10 min at 12,000 × g. The pellets were washed three times with incubation buffer and re-suspended in 50 µl of SDS sample buffer. The immune complex was dissociated from protein A by boiling, and 20 µl were removed and analyzed by SDS-PAGE in 12% acrylamide gel. The ¹²⁵I-IGFBP-5 precipitated by FHL2 anti-body was detected by autoradiography. The same procedure was used to test binding of FHL2 to ¹²⁵I-IGFBP-4 and -6.

IGFBP-5/FHL2 interaction using unlabeled proteins was also performed with IGFBP-5 polyclonal antibodies to immune precipitate and the 6-His-tag monoclonal antibody to detect FHL2 using the same procedure as mentioned above. FHL2 in the immune precipitate was detected by Western blot according to standard protocols (42).

Surface-enhanced Laser Desorption Ionization (SELDI) ProteinChip Analysis-- PS1 ProteinChip arrays were used for the analysis of FHL2 interactions with IGFBPs. The spots in PS1 ProteinChip were pre-activated with iminodiacetate chemistry that covalently binds to the free primary amine groups (Ciphergen Biosystems, Inc., Fremont, CA). Briefly, 200 ng of FHL2 (50 µl) was added to each spot, incubated for 2 h at room temperature on an orbital shaker, and blocked with 0.1 M Tris, pH 8.0, for 30 min to remove nonspecific binding. After rinsing the spots with PBS, 50 µl of PBS or PBS containing 200 ng of IGFBP was added to each spot and incubated for 1 h at room temperature on an orbital shaker. The spots were are then rinsed with 50 mM Tris, pH 8.0, containing 150 mM NaCl and 1% Triton X-100 before the addition of energy absorbing molecule (-cyano-4-hydroxy cinnamic acid) according to the manufacturer's instructions. The spots were air-dried and analyzed using the SELDI ProteinChip system (PBS-1, Ciphergen) as described previously (43). Data were collected using laser intensity of 80% and mass calibrated with protein standards (43).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Verification of the Expression of GAL4 DNA Binding Domain-IGFBP-5 Fusion Protein in the Yeast Strain AH109-- We used the THS3 to identify candidate proteins that interact with our protein of interest, IGFBP-5. The IGFBP-5 cDNA sequence was fused to the DNA binding domain sequence of GAL4 in the expression vector pGBDT7. The yeast reporter strain AH109 was transformed with this plasmid, and cells were plated on the appropriate selection medium. Expression of the GAL4-IGFBP-5 fusion protein in AH109 cells was subsequently confirmed by Western blot analysis using an antibody (CLONTECH) to the GAL4 DNA binding domain of the fusion protein (Fig. 1). AH109 cells containing the bait but not control AH109 cells expressed BD-IGFBP-5 fusion protein of the anticipated molecular mass.

View larger version (46K): [in this window] [in a new window]   Fig. 1.   Western immunoblot analysis of the GAL4 BD-IGFBP-5 bait using GAL4 BD antibody. Yeast strain AH109 protein extracts were prepared as outlined in the Yeast Protocol Hand Book (CLONTECH) and were separated using SDS-PAGE. The antibody detected a single band in extracts of AH109 transformed with BD-IGFBP5 (baits 1 and 3) construct, but no band was detected in untransformed AH109 cells.

Once the expression of our bait in yeast AH109 cells was confirmed, we tested for auto-activation by plating the reporter strain AH109 containing the IGFBP-5 construct on selection medium lacking the amino acids adenine, histidine, and tryptophan. In addition, the selection medium contained -X-gal as a color indicator for the expression of the MEL1 reporter gene that encodes -galactosidase. The results indicated that the IGFBP-5 construct did not auto-activate the reporter genes since no colonies appeared on the test plates.

Identification of Proteins That Potentially Interact with IGFBP-5-- We transformed the AH109-IGFBP-5 with an amplified U2 human osteosarcoma cDNA library fused to the GAL AD in the expression vector pACT2. We picked several clones from our initial low stringency screen of the human osteosarcoma cDNA library, which were subsequently transferred to high stringency medium for a second screen. Positive clones were then picked for AD:cDNA plasmid DNA isolation as described under "Experimental Procedures."

Sequence Analyses of Positive AD:cDNA Clones-- The positive AD:cDNA plasmids that activated all of the three reporter genes in the AH109 strain were isolated and were used to transform the E. coli strain HB101. We then partially sequenced the positive clones and assessed the cDNA sequences using the BLAST program. Five clones matched a known gene sequence in the GenBank^TM (Table I); of these we chose a 1.4-kb EcoRI/XhoI full-length (clone 41) and a 0.8-kb cDNA (clone 2) for further characterizations. Upon obtaining the entire sequence of clone 41, we found it encodes a protein product of 279 amino acids (Fig. 2). The BLAST search identified our sequence as a LIM-only protein that contains four and a half LIM domains (FHL2) (31, 32). The 0.8-kb cDNA clone encoded amino acids 158-279 of the FHL2 protein and strongly interacted with IGFBP-5 as demonstrated by the two-hybrid assay.

View this table: [in this window] [in a new window]   Table I Five positive clones obtained using the yeast two-hybrid system 3 to screen the human osteosarcoma cDNA library The five AD-cDNA clones listed contained different lengths of cDNA sequence identical to human FHL2 cDNA. Co-transfection of the reporter yeast strain with each isolated clone together with the BD:bait plasmid activated all three reporter genes. pGBKT7-BP-5, -Lam, and -53 encode fusions of Gal4 BD and human JGFBP-5, human lamin C, and murine p53, respectively. pGADT7-T is a positive control that encodes a fusion of Gal4 AD and SV40 large T-antigen that is known to bind p53. + indicates that the transformed AH109 strain grew on high stringency selection medium.  indicates that the transformed AH109 strain did not grow on high stringency medium.

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Sequence alignments of the five clones recovered from the yeast two-hybrid screen and full-length FHL2. The four and a half LIM domains of FHL2 are shown in the top line. The full-length clone 41 encoding 279 amino acids of the four and a half LIM domain protein and the four partial clones are aligned below. A half LIM domain sequence contains CX2CX17CX2C, and a complete LIM domain sequence contains CX2CX17-21CX2CX17-21CX2(H/D/C) (X represents variable amino acids). UTR, untranslated region.

Reconfirmation of IGFBP-5/FHL2 Interaction in Yeast-- After identifying the FHL2 AD:cDNA clones by DNA sequencing, we transformed AH109 yeast cells with the BD-IGFBP-5 and AD:cDNA plasmids to reconfirm the IGFBP-5/FHL2 interaction. Table I summarizes the reconfirmation screen and essential control plasmid transformations. The results indicate that both the full-length and the truncated cDNA clones of FHL2 strongly interact with IGFBP-5 as judged by growth of the AH109 reporter strain on high stringency medium. No growth was observed when either the BD-IGFBP-5 or the AD:FHL2 plasmids were tested with control plasmids pGADT7-T and pGBKT7-53 to transform the reporter strain, respectively. Hence, the IGFBP-5/FHL2 interaction was specific.

The FHL2 Gene Is Strongly Expressed in Bone Cells-- To determine whether FHL2 is expressed in other bone cell types besides U2 human osteosarcoma cells, total RNA (10 µg/lane) from HBC, HBR, U2, and SaOs-2 human osteosarcoma cells was probed with the EcoRI-XhoI FHL2 cDNA fragment. Northern analysis shows that the FHL2 is expressed in all of the human osteoblast cell types tested. The FHL2 cDNA probe detected a 1.4-1.5-kb band, the size of which is similar to that reported in other cell types (Fig. 3).

View larger version (53K): [in this window] [in a new window]   Fig. 3.   Northern blot analysis of total RNA extracted from various human osteoblast cell preparations using the FHL2 cDNA as a probe. RNA was extracted from 70-80% confluent cultures of normal human osteoblasts derived from calvaria (HBC) and rib (HBR) and from SaOs-2 and U2 osteosarcoma cells. 20 µg of total RNA was loaded per lane, and the blot was probed with 32P-labeled FHL2 cDNA. The probe detected a 1.4-1.5-kb band in all cell lines tested. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression detection was used as an internal control to monitor loading and transfer.

FHL2 Protein Localizes in the Nucleus and Cytoplasm of Bone Cells-- Western blot analysis of nuclear and cytoplasmic extracts from U2, HBC, and HBR cells were performed using the FHL2 monoclonal antibody. The FHL2 protein was detected in the nuclear and cytoplasmic extracts of both normal human osteoblasts and U2 osteosarcoma cells (Fig. 4, A and B). In addition to the 32-kDa band that corresponds to FHL2, high molecular weight bands were seen in the nuclear extracts. These additional bands may represent higher molecular weight forms of FHL2 (dimer or complex of FHL2 with another protein) or other proteins that cross-react with the FHL2 monoclonal antibody used in this study. No band was detected when normal mouse IgG was used for Western blotting (data not shown).

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Western immunoblot analysis of FHL2 protein in normal human osteoblasts using FHL2 monoclonal antibody in various human osteoblast cell types. Nuclear and cytoplasmic extracts from U2 (A and B), calvaria (HBC; B), and rib (HBR; B) cells were separated by SDS-PAGE and transferred to a nylon membrane, and the blot was developed with FHL2 monoclonal antibody. Note that U2 nuclear and cytoplasmic extracts in B were prepared from cells overexpressing FHL2 from the retroviral vector pCLSA-FHL2, and hence, a much stronger signal was detected compared with that of calvaria and rib. CE, cytoplasmic extract; NE, nuclear extract.

Evidence of IGFBP-5 Localization in the Nucleus-- Localization of exogenous IGFBP-5 in human osteoblast was determined in MG63 osteosarcoma cells that have low endogenous levels. IGFBP-5 protein (10 nM) added to the cells found to localize to the nucleus of MG63 cells as determined by IGFBP-5 antibody and fluorescent-conjugated secondary antibody staining (Fig. 5). There was no evidence of nuclear localization in the absence of exogenous recombinant IGFBP-5 in MG63 cells, which do not produce detectable levels of IGFBP-5 (10). The addition of IGFBP-5 resulted in accumulation of IGFBP-5 in nuclei.

View larger version (94K): [in this window] [in a new window]   Fig. 5.   Nuclear localization of IGFBP-5 in MG63 cells. IGFBP-5 protein (10 nM) added to the cells was found to localize to the nucleus of MG63 cells as determined by IGFBP-5 antibody and fluorescent-conjugated secondary antibody staining (b). No IGFBP-5 was detected in cells without the addition of IGFBP-5 protein (d). Cells were stained with propidium iodide to visualize nuclei (a and c).

Untransformed normal human osteoblasts derived from calvaria, which express IGFBP-5 (43), contained IGFBP-5 in the nucleus in the absence of exogenous IGFBP-5. The addition of exogenous IGFBP-5 further increased accumulation of IGFBP-5 in the nucleus (data are not shown).

Recombinant Expression of FHL2-- An FHL2 full-length coding sequence PCR product was fused in-frame with the 6-His-tag sequences in the expression vector pQE32. The fusion protein was expressed in the E. coli M15 and was purified using Nickel beads. Expression of FHL2 was confirmed by Western blot analysis using a His-tag antibody (Fig. 6A) and the specific FHL2 monoclonal antibody (Fig. 6B).

View larger version (54K): [in this window] [in a new window]   Fig. 6.   Characterization of recombinant FHL2 protein expressed in E. coli using the pQE32 expression vector. Purified FHL2 protein was separated on SDS-PAGE, transferred to a nylon membrane, and detected as a 33-kDa band by both FHL2 (A) and His tag (B) monoclonal antibodies.

FHL2-IGFBP-5 Interaction Determined by Co-immunoprecipitation-- We performed co-immunoprecipitation studies by incubating the FHL2 protein with ¹²⁵I-labeled IGFBP-5 in the presence of FHL2 monoclonal antibody in test tubes. Fig. 7 shows that FHL2 interacts specifically with IGFBP-5 in an in vitro reaction. Binding of ¹²⁵I-IGFBP-5 to FHL2 was competed with unlabeled IGFBP-5. FHL2 specificity for binding IGFBP-5 was tested using ¹²⁵I-IGFBP-4 and -6 proteins in the co-immunoprecipitation assay (Fig. 7). FHL2 did not interact with IGFBP-4 or -6, suggesting that FHL2 binding to IGFBP-5 is specific.

View larger version (65K): [in this window] [in a new window]   Fig. 7.   In vitro binding of FHL2 with 125I-IGFBP5. FHL2/125I-IGFBP5 binding was tested by co-immunoprecipitation with the FHL2 monoclonal antibody. 250 ng of 125I-IGFBP-5 and FHL2 protein were incubated at 4 °C overnight in the presence of FHL2 monoclonal antibody (ab). Controls lacked 125I-IGFBP5 or FHL2 antibody. The protein complexes were precipitated using protein A-agarose beads. Precipitates were analyzed by direct exposure of x-ray film. Lane 1 of each gel contained an aliquot of the 125I-labeled IGFBP used in the immune precipitation. Lanes 2-4 contained the precipitated proteins. The co-immunoprecipitation results indicate that FHL2 interacts specifically with IGFBP-5, and the interaction was competed for with unlabeled IGFBP-5. FHL2 was incubated with IGFBP-4 and -6, but no binding was detected.

We further examined in vitro the IGFBP-5/FHL2 interaction by incubating IGFBP-5/rFHL2 in the presence of protein A-bound IGFBP-5 polyclonal antibodies. The immunoprecipitated complex was analyzed by Western blots using IGFBP-5 antibodies and the 6-His-tag monoclonal antibody (Fig. 8, A and B). The results indicate that FHL2 was precipitated by IGFBP-5 antibody but not by normal guinea pig IgG.

View larger version (34K): [in this window] [in a new window]   Fig. 8.   Co-immunoprecipitation of FHL2/IGFBP-5 by IGFBP-5 antibody in vitro. Lane 1, 0.6 µg of IGFBP-5 protein was incubated with 0.2 µg of His-tagged rFHL2 protein overnight. Then 100 µl of IGFBP-5 polyclonal antibodies conjugated to protein A was used to pull down the IGFBP-5/FHL2 protein complex. After SDS-PAGE and immunoblotting, the IGFBP-5 protein in lanes 1 and 4 was detected using affinity-purified IGFBP-5 polyclonal antibody (A). Lane 2 is the same as lane 1, except that normal guinea pig serum conjugated to protein A was used instead of IGFBP-5 antibodies. This lane showed nonspecific immunoreactive bands but no IGFBP-5 band. Lane 3 is the rFHL2 standard, and lane 4 is recombinant IGFBP-5 standard. After the first immunodetection, the blot was stripped and redeveloped with the His-tag monoclonal antibody. The HIS-tag monoclonal antibody detected rFHL2 protein in lanes 1 and 3 (B). In addition to the band that corresponds to intact FHL2, additional bands were detected by the His-tag monoclonal antibody in lane 3. Because FHL2 preparation used in this experiment contained both intact and fragment forms of FHL2, the lower molecular weight bands probably represent the proteolytic fragments of FHL2.

To evaluate if the interaction between FHL2 and IGFBP-5 occurs in whole cells, cell lysate from U2 cells overexpressing FHL2 and IGFBP-5 was immunoprecipitated with FHL2 monoclonal antibody and then probed with IGFBP-5-specific antibody by western immunoblot analysis. Fig. 9 shows that IGFBP-5/FHL2 was co-immunoprecipitated by FHL2 antibody in whole U2 cell lysate. These data together with the yeast two-hybrid data provide evidence that the interaction between FHL2 and IGFBP-5 occurs in whole cells after transfection and overexpression of both FHL2 and IGFBP-5.

View larger version (14K): [in this window] [in a new window]   Fig. 9.   Co-immunoprecipitation of FHL2/IGFBP-5 in whole U2 cells over-expressing FHL2 and IGFBP-5. Lane 1, 250 µl of cell lysate incubated with 25 µl protein A-Sepharose conjugated to FHL2 monoclonal antibody (mab) were incubated for 14 h at 4 °C on a rotary shaker. Then the protein complex was washed and subjected to SDS-PAGE and immunoblotting using IGFBP-5 polyclonal antibody. Lane 2 is the same as lane 1 except that protein A-Sepharose used was not conjugated to FHL2 monoclonal antibody, and hence, no IGFBP-5 was co-immunoprecipitated.

SELDI ProteinChip Analysis-- The specificity of FHL2-IGFBP-5 interaction was verified using an alternate technology, namely SELDI ProteinChip technology (43). The ProteinChip system detects proteins captured on ProteinChip arrays and provides accurate molecular weight determinations with deviations less than 0.2%. FHL2 was immobilized covalently on the surface of the PS-1 ProteinChip array and used to determine if the 29-kDa IGFBP-5 could be captured on the array. The bound IGFBP released by laser was detected, and the mass was analyzed by a time of flight mass spectrophotometer. Fig. 10 shows that 29-kDa IGFBP-5 was captured by FHL2 covalently linked to the PS1 surface. None of the other IGFBPs (IGFBP-3, -4, or -6) was captured by FHL2 covalently linked to PS1 surface under identical conditions. These data together with the co-immunoprecipitation data provide evidence that the binding of FHL2 to IGFBP-5 is specific.

View larger version (26K): [in this window] [in a new window]   Fig. 10.   IGFBP-5 was captured by FHL2 covalently bound to pre-activated ProteinChip array. FHL2 was immobilized covalently on the surface of the PS1 ProteinChip array and incubated with 50 µl of PBS containing IGFBP-3 (A), IGFBP-4 (B), IGFBP-5 (C), IGFBP-6 (D), or buffer control (E). The bound protein was released by laser and detected using ProteinChip Reader (SELDI ProteinChip System, Ciphergen). IGFBP-5 but not other IGFBPs (-3, -4, or -6) was retained by the immobilized FHL2.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Using the yeast two-hybrid system and IGFBP-5 as bait, we isolated several FHL2 cDNAs from a human osteosarcoma cDNA library. This evidence for IGFBP-5-FHL2 interaction was further supported by in vitro protein-protein data and by localization of both FHL2 and IGFBP-5 in the nucleus. Although the functional significance of the association between IGFBP-5 and FHL2 has not been established, the findings that IGFBP-5 itself is a growth factor that translocates to nucleus (21, 22) and that FHL2 acts as a coactivator of the androgen receptor (32) supports the conclusion that FHL2-IGFBP-5 interaction may play a significant role in the modulation of osteoblast cell proliferation and/or differentiation. To our knowledge, this report is the first one to describe identification of an IGFBP-5 binding partner in any cell type using the yeast two-hybrid system.

We chose to screen for IGFBP-5-interacting proteins in osteoblasts based on the established importance of IGFBP-5 in bone (15-18) and on the discoveries that IGFBP-5 has IGF-independent mitogenic effects on cells and translocates to the nucleus (18, 20-22). We hypothesized that IGFBP-5 interacts with other cellular proteins, perhaps nuclear proteins. Consistent with this hypothesis, five independent clones selected under high stringency conditions corresponded to a four and a half LIM domain gene 2, FHL2, which is also known as SLIM-3. (30, 31). One clone encoded the entire open reading frame of 297 amino acids of FHL2, whereas the other four were partial clones. The finding that the smallest clone identified in the yeast-two hybrid screen encoded only LIM domains 3 and 4 suggests that the C-terminal region of FHL2 containing the last two LIM domains is sufficient for FHL2 interaction with IGFBP-5.

We have confirmed the interaction between FHL2 and IGFBP-5 observed in yeast with in vitro co-immunoprecipitation studies using purified recombinant human FHL2, purified recombinant human IGFBP-5, and antibodies directed against FHL2 or IGFBP-5. FHL2/IGFBP-5 was also co-immunoprecipitated in whole cell lysate from U2 cells overexpressing FHL2 and IGFBP-5. Furthermore, IGFBP-5 binding to FHL2 was verified using a ProteinChip array in which IGFBP-5 was captured using FHL2 covalently linked to preactivated PS1 surface. We believe that the interaction between IGFBP-5 and FHL2 is specific based on the following findings. 1) FHL2 interacted with IGFBP-5 but not with either IGFBP-3, -4, or -6, which are known to be produced by human osteoblasts (44). The lack of FHL2 binding to IGFBP-3, which is highly homologous to IGFBP-5 in the N-terminal and C-terminal domains suggests that mid-region of IGFBP-5 may be critical for FHL2 binding. Alternatively, IGFBP-3 may bind to FHL2 but with a reduced affinity compared with IGFBP-5 and that this binding could not be detected by the experimental technique used in this study. 2) Although rat osteoblasts in culture have been shown to express another LIM domain-containing protein, namely LMP-1 (45), we did not identify any homologs of LMP-1 in our yeast two-hybrid screen.

In this study, co-immunoprecipitation experiments were performed in lysates of U2 cells overexpressing both FHL2 and IGFBP-5. Although these data provide evidence that the interaction between FHL2 and IGFBP-5 could occur in whole cells that are induced to overexpress both of these proteins, they do not prove that these two proteins bind under normal physiological conditions. Further interaction studies using normal human osteoblasts without forced overexpression of either FHL2 or IGFBP-5 are needed to establish that the interaction between FHL2 and IGFBP-5 occurs under physiological conditions.

To determine that the FHL2 interaction with IGFBP-5 could occur in other osteoblast cell types besides U2 human osteosarcoma cells, we evaluated FHL2 expression in untransformed normal human osteoblasts as well as in other human osteosarcoma cell types. Northern analysis showed that the FHL2 is strongly expressed in human osteoblasts derived from calvaria and rib and in U2 cells. Of the various human osteoblast cell types tested, FHL2 expression was low in SaOs-2 human osteosarcoma cells compared with other human osteoblast cell types. Based on the findings that SaOs-2 cells lack functional p53 and that FHL2 expression is up-regulated in cell lines expressing functional p53 and down-regulated in cell lines expressing p53 mutants (46), it is possible that p53 is an important regulator of FHL2 expression. In addition, our data that FHL2 is strongly expressed in normal human osteoblasts argue against the previous conclusion that FHL2 expression is restricted to the cardiovascular system (45-48).

Besides IGFBP-5, FHL2 has been shown to interact with other proteins. In this regard, it is known that FHL2, a member of the LIM domain-only proteins, can participate in protein-protein interactions by forming homodimers (LIM-LIM) (49) or heterodimers (LIM-non-LIM) (50). By using the yeast two-hybrid assay, FHL2 has also been shown to interact with androgen receptor, hCDC7 (51), - and -integrin subunits (52), the polypyrimidine tract-binding protein-associated splicing factors (53), and Alzheimer's disease-associated presenilin 2 (54). Because FHL2 contains four and a half LIM domains, each of which contains zing finger motifs, it is possible that different LIM domains may be involved in FHL2 interaction with the various binding partners (52-53). In this regard, our data on the identification of a partial FHL2 cDNA clone, which encodes for amino acids 158-279 in the yeast two-hybrid screen, suggests that the last two LIM domains may be sufficient for interaction with IGFBP-5.

The significance of a potential interaction between FHL2 and IGFBP-5 in osteoblasts can only be speculated at this time. In this regard, it has been shown by Muller et al. (32) that FHL2 binds and selectively activates the transcriptional activity of androgen receptors in an agonist- and AF-2-dependent manner. Furthermore, Boden et al. (45) show that LIM domain-containing protein (LMP-1) is an essential intracellular positive regulator of rat osteoblast differentiation, which acts to mediate BMP-6 effects on bone formation in osteoblasts, thus raising the possibility that FHL2 could function as an intracellular mediator of IGFBP-5 actions. Future demonstration that the interaction between IGFBP-5 and FHL2 occurs under physiological conditions and that these two proteins are co-localized in the nucleus of normal human osteoblasts could provide the basis for the hypothesis that IGFBP-5 may bind to FHL2, a transcription modulator, to stimulate transcription of putative IGFBP-5 target genes that may be involved in regulation of osteoblast cell proliferation and differentiation.

	ACKNOWLEDGEMENTS

We acknowledge the technical assistance provided by Joe Rung-Aroon, Rongqing Guo, and Melanie Hamilton-Ulland. We thank Dr. John Farley for providing SaOs-2 cells.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grants AR31062 and AR07543 and grants from Veterans Affairs and Loma Linda University.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Musculoskeletal Disease Center (151), Jerry L. Pettis Veterans Affairs Medical Center, 11201 Benton St., Loma Linda, CA 92357. Tel.: 909-825-7084 (ext. 2932); Fax: 909-796-1680; E-mail: mohans@lom.med.va.gov.

Published, JBC Papers in Press, January 30, 2002, DOI 10.1074/jbc.M110872200

	ABBREVIATIONS

The abbreviations used are: IGF, insulin-like growth factor; IGFBP-5, IGF-binding protein 5; -X-gal, 5-bromo-4-chloro-3-indolyl -D-galactopyranoside; AD, activation domain; BD, binding domain; PBS, phosphate-buffered saline; rFHL2, recombinant FHL2; bp, base pair(s); SELDI, surface-enhanced laser desorption ionization; kb, kilobase.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike