Molecular Vision 2000; 6:222-231 <http://www.molvis.org/molvis/v6/a30/> Received 22 August 2000 | Accepted 31 October 2000 | Published 6 November 2000

Download Reprint

Subcellular localization of the expressed 18 kDa FGF-2 isoform in corneal endothelial cells

Jongkyu Choi,¹ Minhee K. Ko,¹ EunDuck P. Kay^1,2
 
 

¹Doheny Eye Institute and ²Department of Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles, CA

Correspondence to: EunDuck P. Kay, D.D.S., Ph.D., Doheny Eye Institute, 1450 San Pablo Street, DVRC 203, Los Angeles, CA, 90033; Phone: (323) 442-6625; FAX: (323) 442-6688; email: ekay@hsc.usc.edu

Abstract

Purpose: To determine the subcellular localization of 18 kDa FGF-2 after synthesis and before secretion into the extracellular matrix.

Methods: Corneal endothelial cells (CEC) were transfected with an expression vector coding for green fluorescent protein (GFP) and 18 kDa FGF-2. Expression of the fusion protein was determined by immunoblot analysis and the subcellular localization of the fusion protein was examined by immunocytochemical analysis.

Results: When the expression of the fusion protein was determined by immunoblot analysis, the expressed fusion protein had a molecular weight of 45 kDa, resulting from the 27 kDa GFP and 18 kDa FGF-2. Following a 90 min exposure of cells to the vector, the expressed 18 kDa FGF-2 was completely translocated to the nucleus within a 24 h incubation. When cells were further incubated for another 24 h, one-half of the fusion protein was retro-transported from the nucleus to the cytoplasm, largely to the membrane and focal adhesion site, while the other half remained in the nucleus. During a 72 h incubation, the fusion protein was completely translocated to the cytoplasm, where it was diffusely distributed and its staining potential was greatly lost. Transfected cells showed both a slight increase in cell proliferation and a down-regulation in the expression of the high affinity receptors of FGF.

Conclusions: These results indicate that the 18 kDa FGF-2 is directly translocated from its synthetic site to the nucleus. The nuclear 18 kDa FGF-2 is then retro-transported to membrane/focal adhesion sites, after which the molecule may be secreted. When 18 kDa FGF-2 remains in the nucleus, there is a slight stimulatory activity of cell proliferation and a down-regulation of its receptor. These data suggest an intracellular action of 18 kDa FGF-2 through mechanisms independent of the receptor-mediated signaling pathways.

Introduction

Fibroblast growth factor-2 (FGF-2) is a ubiquitous multifunctional growth factor that is present in Descemet's membrane [1,2], as well as in other tissues and cells [3,4]. FGF-2 induces angiogenesis, mesoderm induction and tumorigenesis in vivo, and stimulates proliferation, morphogenic induction, migration, and wound healing in vitro [2-11]. However, the mere presence of FGF-2 in Descemet's membrane is not sufficient to induce cell proliferation of corneal endothelium in vivo; the capacity for regeneration in human corneal endothelial cells (CEC) is severely limited [12]. Unlike this physiological situation, our study shows that FGF-2 becomes a potent modulator for endothelial mesenchymal transformation in CEC under the conditions that may accompany atypical wound repair [2,8,11]. As a consequence of the transformation, CEC convert to fibroblast-like cells that resume cell proliferation and produce fibrillar collagens, leading to the formation of a fibrillar extracellular matrix (ECM). One clinical example of this process is the development of a retrocorneal fibrous membrane in Descemet's membrane [13-15], the presence of which blocks vision, thereby causing blindness.

Five isoforms of FGF-2, with molecular weights of 18, 22, 22.5, 24 and 34 kDa, have been identified, all derived from a single messenger RNA [16-19]. The 18 kDa FGF-2 is primarily a cytosolic protein without a signal sequence. This molecule is exported out of the cells by an unknown mechanism and stored as an ECM isoform, whereas the high molecular weight (HMW) isoforms are predominantly located in the nucleus [1,2,20-22]. It has been shown that NH₂-terminal sequences are required for nuclei localization [18,19,23,24]. The different subcellular localization of HMW isoforms of FGF-2 and 18 kDa FGF-2 supports the concept of potentially differential roles for these FGF-2 isoforms [18,19,25]. However, there is evidence that exogenous 18 kDa FGF-2 can also be recovered from the nucleus of vascular endothelial cells [26], suggesting that at least two nuclear localization sequences (NLS) may exist in FGF-2 molecules, one within 18 kDa FGF-2 and a second within the amino-terminal extension present in the HMW isoforms. Chondrocytes from the fetal growth plate synthesize 18 kDa and 23 kDa FGF-2 isoforms, both of which are capable of nuclear translocation [27]. CEC accumulate the 18 kDa form in the nucleus during the early stage of cell growth [28]. Thus, the assumption that different subcellular localization of FGF-2 isoforms influences the mechanisms of action and cellular activities warrants reinvestigation. In the present study, we have taken advantage of fusion with green fluorescent protein (GFP), which has been proven to be a useful tag for monitoring the subcellular distribution and trafficking of various proteins in living cells [29,30]. We investigated whether 18 kDa FGF-2 is translocated to the nucleus immediately after synthesis or is first secreted into ECM before internalization and subsequent translocation to the nucleus. We addressed this question by transfecting CEC with cDNAs harboring GFP and the full length of 18 kDa FGF-2 and by examining the subcellular localization of the expressed fusion protein after a brief period of transfection.

Methods

Cell cultures

Isolation and establishment of rabbit CEC were performed as previously described [2]. Briefly, the Descemet's membrane-corneal endothelium complex was treated with 0.2% collagenase and 0.05% hyaluronidase (Worthington Biochemical, Lakewood, NJ) for 60 min at 37 °C. Cultured cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Gibco-BRL, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS; Irvine Scientific, Santa Ana, CA) and 50 mg/ml of gentamicin (DMEM-10) in a 5% CO₂ incubator. This method has been shown to promote cell proliferation during the early phase of culture and to maintain the culture as a contact-inhibited monolayer when the cells reach confluence. First passage CEC were used for all experiments. For subculture, confluent cultures were treated with 0.2% trypsin and 5 mM EDTA for 3 to 5 min.

Reverse Transcriptase-Coupled Polymerase Chain Reaction (RT-PCR) and Cloning of 18 kDa FGF-2 cDNA

RNA was isolated from pituitary gland using a QIAGEN RNeasy Mini Kit (Qiagen Inc., Valencia, CA). Two mg of total RNA was reverse-transcribed with Moloney murine leukemia virus reverse transcriptase (Ambion, Austin, TX) using oligo-dT as the primer. PCR was performed using a RoboCycler (Stratagene, La Jolla, CA) with a Perkin-Elmer/Cetus Gene Amp kit, 1.5 mM MgCl₂, 1 unit of Taqpolymerase, and 5 ml of RT reaction products. The sequences of the gene-specific sense (33-nt) primer and antisense (29-nt) primer of 18 kDa FGF-2 were taken from the referenced studies by Winkles et al. [31] and Kurokawa et al. [32] to cover the full length of 18 kDa FGF-2 (471 bp). PCR was performed as follows: Denaturation for 1 min at 95 °C, annealing for 1 min at 65 °C, elongation for 2 min at 72 °C, for a total of 35 cycles. Electrophoresis and direct sequencing using a Cyclist Exo pfu sequencing kit (Stratagene) identified the amplified DNA fragments. The 471 bp PCR product, which has the potential to generate the full length of 18 kDa FGF-2 from the methionine (AUG) site, was ligated to PCR2.1-TOPO vector (Invitrogen, Carlsbad, CA). The 471 bp ApaI-KpnI fragment from the identified recombinant plasmid was cloned into pcDNA3.1/Myc-His(-) B (Invitrogen), and the resultant clone was designated p18kcDNA. The clone was sequenced by the dideoxy chain-termination method, using 2 units of T7 sequence (Version 2.0) and a-[³⁵S]dATP (Amersham, Piscataway, NJ). Analysis of the sequences was performed using the software program DNASIS (Version 2.5) and the blast network of NCBI for homology search. In parallel, the DNA fragment with the reversed sequence of 18 kDa FGF-2 was cloned and designated p18kcDNA/R. When the expression of the fusion protein was analyzed by immunofluorescent staining using anti-Myc antibody, the antibody demonstrated a nonspecific staining pattern, which would hamper the interpretation of the subsequent experiments. Therefore, the 471 bp ApaI-KpnI fragment was cloned into the corresponding sites of the pEGFP-C3 (Clontech Laboratories, Inc., Palo Alto, CA) to obtain pEGFP-18-kDa FGF-2. In parallel, a fragment isolated from p18kcDNA/R was cloned into the multicloning site of pEGFP-C3 to obtain pEGFP-18-kDa FGF-2/R. Plasmid pcDNA3.1/Myc-His(-)/lacZ (Invitrogen), which encodes the b-galactosidase enzyme regulated by the same cytomegalovirus promoter as that of pEGFP-C3, was used to assess transfection efficiency.

Transfection

Cells were transfected using SuperFect transfection Reagent (Qiagen Inc). The plasmid and the transfection reagent were mixed in DMEM containing no serum or antibiotics and incubated for 10 min at room temperature to allow the DNA-SuperFect complex formation. The complexes were added drop-wise onto the cells, and transfection of the cells was then carried out for 90 min, after which the cells were maintained in fresh growth medium (DMEM-10) for a designated time. To determine the optimal conditions of transfection, cells were plated at a density of 5 x 10⁴ cells per well in a 24 well plate and incubated for approximately 24 h prior to transfection. The plasmid pcDNA3.1/Myc-His(-)/lacZ was used in concentrations ranging from 0.1 to 2.0 mg; the ratios of DNA to transfection reagent were 1:2, 1:5, 1:10, and 1:15. The transfected cells were harvested 24 h after transfection, and b-galactosidase activity was measured in the cell lysates using a High Sensitivity b-galactosidase Assay Kit (Stratagene), according to the procedures recommended by the manufacturer. Briefly, chlorophenol red-b-D-galactopyranoside substrate (130 ml) was added to the 96 well microtiter dish with 20 ml of cell lysates. The plate was incubated for 60 min at 37 °C until the sample turned dark red. The reaction was terminated by adding 80 ml of 500 mM Na₂CO₃ and the optical density was measured at 570 nm, using the 96 well plate reader. The highest b-galactosidase activity was achieved with a mixture of 1.0 mg plasmid and 2 ml of the SuperFect transfection reagent incubated for 90 min. Therefore, this condition was used throughout the study. To measure the transfection efficiency of pEGFP-18-kDa FGF-2, cells were transfected and maintained in growth medium for 24 h. Cells were then trypsinized; total cell numbers and the GFP-expressing cells were respectively counted using the hemocytometer. The transfection efficiency was represented as a percentage of GFP-expressing cells relative to the total cell numbers.

Protein Preparation and Protein Determination

Cells transfected with pEGFP-18-kDa FGF-2 or control vectors were washed three times with phosphate-buffered saline (PBS) and scraped. Cells were then lysed with lysis buffer (500 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% NP-40, 50 mM EDTA, 0.5 mM gelatin, 1 mM paramethylsulfonyl fluoride, 1 mM N-ethylmaleimide, 1 mg/ml leupeptin, and 1 mg/ml aprotinin) on ice for 30 min. Concentration of the resultant lysates was assessed with a Bio-Rad DC protein assay system (Bio-Rad Laboratories, Hercules, CA).

SDS-Polyacrylamide Gel Electrophoresis (SDS-PAGE)

The conditions of electrophoresis were as described by Laemmli, using the discontinuous Tris-Glycine buffer systems [33]. Thirty mg of protein was loaded on a 12.5% SDS-polyacrylamide gel and separated on the gel under reducing conditions.

Cell Proliferation Assay

CEC were plated in 96 well tissue culture plates. When the cells reached about 60% confluency, they were transfected with pEGFP-18-kDa FGF-2, pEGFP-18-kDa FGF-2/R, and pEGFP-C3, respectively. Cells were then maintained further for 24 to 72 h before assay. At the end of each incubation period, 20 ml of CellTiter 96^RAQ_ueousOne Solution Reagent (Promega, Madison,WI) was added to the well. The plates were incubated for 3 h at 37 °C in a humidified 5% CO₂ atmosphere, after which the absorbancy was measured at 490 nm, using the 96 well plate reader.

Immunoblot Analysis

The proteins separated by SDS-PAGE were transferred to a polyvinylidene difluoride (PVDF) at 0.22 ampere for 10 h in a semidry transfer system (transfer buffer; 25 mM Tris-HCl, pH 8.3, 190 mM glycine, 20% MeOH). Immunoblot analysis was performed using a commercial ABC Vectastain kit (Vector Laboratories, Inc., Burlingame, CA). All washes and incubations were carried out at room temperature in TTBS (0.9% NaCl, 100 mM Tris-HCl, pH 7.5, 0.1% Tween 20). Briefly, PVDF membrane was immediately placed in the blocking buffer (5% nonfat milk in TTBS) and kept overnight at 4 °C. The incubations were carried out with monoclonal mouse anti-FGF-2 antibodies (1:2500 dilution; Upstate Biotechnology, Lake Placid, NY) for 1 h, with biotinylated secondary antibody (1:5000 dilution; Vector Laboratories, Inc.) for 1 h, and with ABC reagent for 30 min. The enhanced chemiluminiscence (ECL) reagent (Amersham Life Science, Buckinghamshire, England) was used and the ECL-treated membrane was exposed to ECL film.

Immunofluorescent Staining

Cells (5 x 10⁴) were plated in 4 chamber slides and allowed to achieve about 60% confluency. Cells were then transfected with pEGFP-18-kDa FGF-2 or pEGFP-C3 for 90 min and maintained in the growth medium for 24, 48, or 72 h. At the end of each incubation, cells were fixed with 4% paraformaldehyde in PBS for 15 min and simultaneously permeabilized and blocked with buffer A (0.1% Triton X-100, 1% bovine serum albumin [BSA] in PBS) for 15 min at room temperature. The subsequent incubation was carried out with buffer A and all washes were carried out in PBS at room temperature. Cells were incubated with the primary antibody (1:25 dilution for FGF-2 type II antibody; 1:50 dilution for FGF-R1 antibody; Upstate Biotechnology Inc., Lake Placid, NY) for 1 h at 37 °C and then incubated with Texas red-conjugated horse anti-mouse immunoglobulin (1:100 dilution, Vector Laboratories Inc.) for 1 h at 37 °C in the dark. After extensive washing, the slides were mounted in a drop of Vectashield mounting medium (Vector Laboratories Inc.) to reduce photobleaching. The control experiments were performed in parallel with the omission of the secondary antibodies and these controls did not show the activity.

Confocal analysis

Antibody labeling was examined using a Zeiss LSM-20 laser scanning confocal microscope equipped with a barrier filter for fluorescein (DTAF filter and Argon 488 nm as light source) and Cy3 epi-fluorescence (Helium Neon 543 nm as light source). A plan-neofluar x63 oil immersion objective (N.A. 1.3) was used for imaging of fluorescently labeled samples. Separate optical images of FITC or Texas-red were captured from the same optical section. The captured images were then pseudocolored for colocalization. Regions of codistribution appear in yellow, reflecting the additive effect of superimposing green and red pixels. Image analysis was performed using the standard operating system software provided with the Zeiss LSM microscope (Version 2.08).

Results

Establishment of GFP/18-kDa FGF-2 fusion protein

We first cloned an expression vector for 18 kDa FGF-2 using pcDNA3.1/Myc-His(-) vector. In parallel, a control vector with the reversed sequence of 18 kDa FGF-2 was cloned in the same vector. To our disappointment, anti-Myc antibody demonstrated a nonspecific staining pattern, which would hamper the interpretation of the expressed fusion proteins; therefore, a 471 bp Apa1-Kpn1 fragment, which retained the potential to generate the 18 kDa FGF-2 using the methionine (ATG) initiation site, was prepared from the p18kcDNA clone. This fragment was cloned into the multicloning site of pEGFP-C3 that has the potential to generate GFP (Figure 1). In parallel, a negative control vector, pEGFP-18-kDa FGF-2/R, was cloned into pEGFP-C3 using the same protocols. Prior to performing the transfection experiments, we established optimal conditions for transfection with pEGFP-18-kDa FGF-2. The highest transfection efficiency was achieved by the mixture of 1.0 mg of plasmid and 2.0 ml of transfection reagent for a 90 min incubation. Under these conditions, GFP-expressing cells were detected in 20% of cells transfected with pEGFP-18-kDa FGF-2 (data not shown). To examine whether GFP/18-kDa FGF-2 fusion protein was expressed in the transfected cells, cellular proteins were prepared from the cells transfected with pEGFP-18-kDa FGF-2 or with control vectors and analyzed by immunoblotting using FGF-2 antibody. The fusion protein was detected as a single band with a molecular weight of 45 kDa, resulting from the 27 kDa GFP and 18 kDa FGF-2 in the cells transfected with pEGFP-18-kDa FGF-2 (Figure 2, lane 4). Western blot analysis further showed that cells transfected with pEGFP-18-kDa FGF-2 had a high expression level of the fusion protein. Neither untransfected cells nor CEC transfected with either pEGFP-C3 or pEGFP-18-kDa FGF-2/R showed the presence of the 45 kDa fusion protein. The identity of the higher molecular weight peptides shown in the lanes 3 and 4 is unknown.

Subcellular localization of FGF-2 in the transfected corneal endothelial cells

Subcellular localization of the GFP/18-kDa FGF-2 fusion protein was examined in CEC transfected with pEGFP-18-kDa FGF-2 in comparison with the cells transfected with pEGFP-C3 control vector. After cells were transfected for 90 min, the medium was changed with DMEM-10 and cells were maintained in culture for 24, 48, or 72 h prior to immunofluorescence analysis. CEC transfected with the control plasmid and maintained for 24 h showed a nuclear and a diffuse but faint cytoplasmic GFP signal (Figure 3A). The staining profile is similar to the previous report in which GFP alone was distributed throughout the cells [29,30]. On the other hand, cells transfected with the pEGFP-18-kDa FGF-2 and maintained for 24 h showed a nuclear localization of the fusion protein (Figure 3D). When these transfected cells were stained with anti-FGF-2 antibody, CEC transfected with the control vector barely stained with the FGF-2 antibody (Figure 3B), whereas CEC transfected with pEGFP-18-kDa FGF-2 showed strong nuclear staining (Figure 3E). When the two signals were superimposed, it was clear that the expressed fusion protein was composed of GFP and 18 kDa FGF-2 and that both are present in the nuclei (Figure 3F). The untransfected cells were not stained with anti-FGF-2 antibody, suggesting that there is no endogenous production of the growth factor (Figure 3E).

After 48 h in culture following transfection, CEC transfected with pEGFP-18-kDa FGF-2 demonstrated a dual distribution of the fusion protein in the nuclei and cytoplasm (Figure 4D). Of interest, there was strong staining at the membrane and focal adhesion sites. When cells were stained with FGF-2 antibody, the staining pattern was identical to the fluorescent signal. The expressed 18 kDa FGF-2 was distributed in nuclei, cytoplasm and membrane/focal adhesion sites (Figure 4E,F). Within the 48 h incubation period following transfection, a nuclear to cytoplasm export of the fusion protein began to take place. Unlike these positive staining profiles, cells transfected with the control vector showed a very faint staining for GFP in the nuclei and cytoplasm, and a negative staining with FGF-2 antibody was observed in the cells (Figure 4A-C). The untransfected cells showed no staining with anti-FGF-2 antibody, suggestive of the absence of production of endogenous FGF-2 (Figure 4E).

After the 72 h incubation following transfection, the fusion proteins were completely translocated into the cytoplasm, where the GFP signal was diffusely observed (Figure 5D). When FGF-2 antibody was used to stain the expressed 18 kDa FGF-2, the staining potential was almost negligible (Figure 5E,F). Loss of the staining potential in the transfected cells suggests that most of the fusion protein is either secreted or is degraded intracellularly. On the other hand, untransfected cells showed a faint positive staining for FGF-2 in the nuclei and cytosol, although it is not known what isoforms are produced by the untransfected cells (Figure 5B,E,F). Cells transfected with the control vector showed a faint positive staining of GFP in the nuclei and cytoplasm (Figure 5A-C).

Mitogenic effect of the expressed 18 kDa FGF-2 in CEC

In contrast to the conventional idea that HMW isoforms of FGF-2 are preferentially localized in the nucleus and 18 kDa FGF-2 is found predominantly in the cytoplasm and ECM, the expressed 18 kDa FGF-2 is localized in the nucleus immediately after transfection in this study. During the later stage of transfection, a nuclear to cytoplasm export of the fusion protein was observed. We, therefore, investigated whether the expressed 18 kDa FGF-2 exerted mitogenic activity on CEC and whether cellular location of the growth factor influenced its intracellular activity. Cells were transfected with pEGFP-18-kDa FGF-2, pEGFP-18-kDa FGF-2/R, or pEGFP-C3 for 90 min and further incubated in culture for 24, 48, or 72 h. At the end of each incubation period, cell proliferation assay was performed. The proliferation activity of the transfected cells with either pEGFP-18-kDa FGF-2 or pEGFP-18-kDa FGF-2/R was compared to that of the cells transfected with pEGFP-C3 vector. The 18 kDa FGF-2 expressed during the 24 h or 48 h incubation showed a stimulatory activity on cell proliferation of CEC, albeit at a low level, whereas the transfected cells that were incubated for 72 h did not show any stimulatory activity on cell proliferation (Figure 6). Since the expressed growth factor was primarily located in the nucleus during the early phase of the transfection and there is no endogenous FGF-2, it is likely that the nuclear 18 kDa FGF-2 may be responsible for the stimulatory activity shown in 24 h and 48 h incubation periods, although this stimulatory activity is not strong. As the expressed growth factor was retro-transported to the cytoplasm, the stimulatory activity began to fade. The cells transfected with pEGFP-18-kDa FGF-2/R showed no stimulatory activity, regardless of the incubation time following transfection.

Expression of FGF receptors in the transfected cells

It has been reported that FGF receptors were down regulated in NIH 3T3 cells transfected with cDNA coding for 18 kDa FGF-2 [25]. To investigate whether such a phenomenon was observed in CEC transfected with pEGFP-18-kDa FGF-2, the transfected cells were stained with antibody to the high affinity receptor of FGF (anti-FGF R-1 antibody) for 24 or 48 h following transfection. The transfected cells demonstrated the characteristic subcellular localization of the expressed protein as a function of the incubation time. Cells incubated for 24 h showed the nuclear location (Figure 7A), whereas those incubated for 48 h showed a dual distribution in the nucleus and cytoplasm (Figure 7D). On the other hand, the expression of the FGF receptor appeared to be down-regulated, regardless of the incubation time. During the 24 h incubation, the transfected cells showed a faint membrane staining for the receptor (Figure 7B), while the expression of the receptors was markedly down-regulated in the transfected cells that were maintained for 48 h in culture (Figure 7E). On the other hand, the untransfected cells expressed a much higher level of FGF receptors, which was diffusely located in the cytoplasm during the 48 h incubation (Figure 7E).

Discussion

Recent discoveries of FGF-2 isoforms and their unique subcellular location have suggested that there may be differential biologic activities among these isoforms [18,19,25]. The conventional knowledge of the subcellular localization of these isoforms is that the HMW isoforms initiated using the CUG codons are predominantly localized in the nucleus, whereas the AUG-initiated 18 kDa form is localized primarily in the cytoplasm. Our study indicates that only a minor proportion of FGF-2 exists as the ECM form [28]. Recently, others reported that exogenous 18 kDa FGF-2 is internalized via high- and low-affinity FGF receptors [34-36], and that it is translocated to the nucleus [26]; fetal chondrocytes synthesize 18 kDa and 23 kDa FGF-2, both of which are capable of nuclear translocation [27]; CEC accumulate the 18 kDa form in the nucleus during the early stage of cell growth [28]. When this information is taken together, the unique subcellular localization of 18 kDa FGF-2 in the cytoplasm and ECM may not be accepted as a general rule. Since the mechanism of release for the 18 kDa FGF-2 that lacks signal peptides remains to be elucidated, it is equally unknown whether the 18 kDa FGF-2 is translocated after synthesis directly from cytoplasm to the nucleus or whether it is secreted first, internalized later and subsequently translocated to the nuclei. The presence of the 18 kDa FGF-2 in the nucleus further suggests that more than one NLS exists in FGF-2; one within the 18 kDa FGF-2 and a second within the amino-terminal extension present in the HMW isoforms. This may explain why mutagenesis of individual putative nuclear localization domains in FGF-1 and FGF-2 has not always prevented nuclear localization [37,38]. Or, this means that the intracellular 18 kDa FGF-2 can be targeted to the nucleus via NLS-independent pathways.

We attempted to determine whether the 18 kDa FGF-2 is secreted prior to re-entry to the cells or the molecule is first translocated to the nucleus, followed by retro-transport to the cytoplasm and subsequent secretion to ECM. To address this question, we cloned an expression vector harboring GFP and the full length of 18 kDa FGF-2. Following transfection of CEC with this vector, the expressed fusion protein was identified by immunoblot analysis. The fusion protein has a molecular weight of 45 kDa, resulting from the 27 kDa GFP and 18 kDa FGF-2. The expressed 18 kDa FGF-2 was identified with coincidental localization of FGF-2 with GFP. We employed a protocol in which the exposure of cells to the plasmid was very brief (90 min). Our data demonstrate that the expressed 18 kDa FGF-2 is translocated immediately after synthesis. During the 90 min transfection followed by the subsequent 24 h incubation, the fusion protein is completely translocated to the nucleus; no staining of the fusion protein was observed in the cytoplasm. Our unpublished data further indicate that the translocation of the fusion protein was observed from the cells incubated for 15 h following transfection. Since GFP does not contain any known NLS and GFP alone demonstrates a faint nuclear and cytoplasmic staining, the nuclear localization of the fusion protein represents a property of FGF-2 rather than a property of GFP. During the next 24 h incubation period, one half of the fusion protein is retro-transported from the nucleus to the cytoplasm, while the other half remains in the nucleus. The fusion protein that is exported from the nuclei is, interestingly, targeted to membrane and focal adhesion sites, the punctated staining of which is most likely vinculin, as shown in our earlier report [39]. The specific targeting of the fusion protein to the membrane and focal adhesion site may rule out the possibility that the cytoplasmic location results from continuous expression of the fusion protein between 24 and 48 h. Cells incubated for 72 h demonstrated that the fusion protein was completely translocated in the cytoplasm and that the signals for the fusion protein became very weak. The nucleus was completely devoid of the fusion protein; the unique features of the 18 kDa FGF-2 at the focal adhesion sites shown in the cells transfected for 48 h are completely lost. It is not known whether the fusion protein is released from the cells or is intracellularly degraded. In this culture, the untransfected cells show endogenous production of FGF-2, albeit at a low level (isoforms unknown).

Our data further show that transfected cells demonstrate a marked down-regulation of the expression of the high affinity receptors of FGF when compared to the untransfected CEC. This observation is in agreement with the previous report [25], in which high affinity receptors of FGF-2 were down regulated in cells expressing 18 kDa FGF-2 but not in cells expressing HMW forms. The transfected cells also demonstrate a slight increase in cell proliferation activity, during which the expressed 18 kDa FGF-2 remains in the nucleus. Considering the transfection efficiency, the stimulatory activity observed in the transfected cells may represent an apparent increase in cell proliferation mediated by the nuclear 18 kDa FGF-2. Taken together, the translocation of the 18 kDa FGF-2 to the nucleus prior to release from the cells suggests that there is, in part, an intracellular action of 18 kDa FGF-2 through mechanisms independent of the receptor-mediated signaling pathways. An alternative explanation is that the expression of the 18 kDa FGF-2 in the transfected cells may cause transformation of CEC, which subsequently alter the distribution of the newly synthesized fusion protein. However, our previous study [28] demonstrates the nuclear localization of the 18 kDa isoform during early growth stage of CEC in the absence of transformation; thus confirming the present result that translocation of the 18 kDa isoform is the event taken place during early stage of growth in CEC.

For a great deal of time, it was thought that the sole way in which the physiologic regulation of cellular activities such as proliferation took place was through second messenger systems and kinase cascades stimulated by the binding of a growth factor to a cell surface receptor. However, it is known that there are certain hormones and growth factors that can be found in the nuclei of their target cells [40,41]. Furthermore, the translocated peptide hormones and growth factors exert a biochemical effect as a result of this translocation. Bouche et al. [26] showed that exogenously added FGF-2 can indeed translocate to and accumulate in the nucleolus of adult bovine aortic endothelial cells during G0/G1 transition. Apart from the internalization process of the 18 kDa FGF-2, we demonstrate a direct translocation of 18 kDa FGF-2 to the nucleus immediately after its synthesis in the present study. As a consequence of the translocation, there is a stimulatory activity of cell proliferation and a down-regulation of the high affinity receptors of FGF-2. The mechanisms by which the 18 kDa FGF-2 is translocated from the cytoplasm to the nucleus and by which the nuclear 18 kDa FGF-2 exerts its intracellular mode of action are yet to be elucidated. Furthermore, the mechanism by which the nuclear 18 kDa FGF-2 is retro-translocated to the cytoplasm is also to be determined. This information, altogether, suggests that the biological activity of FGF-2 is highly modulated, resulting from regulatory events occurring at each step of its synthesis, translocation, secretion, receptor binding, and internalization.

Acknowledgements

Supported by grants EY06431 and EY03040 from the National Institutes of Health, Bethesda, Maryland, and by an unrestricted grant from Research to Prevent Blindness, New York, NY.

References

1. Vlodavsky I, Folkman J, Sullivan R, Fridman R, Ishai-Michaeli R, Sasse J, Klagsbrun M. Endothelial cell-derived basic fibroblast growth factor: synthesis and deposition into subendothelial extracellular matrix. Proc Natl Acad Sci U S A 1987; 84:2292-6.

2. Kay EP, Gu X, Ninomiya Y, Smith RE. Corneal endothelial modulation: a factor released by leukocytes induces basic fibroblast growth factor that modulates cell shape and collagen. Invest Ophthalmol Vis Sci 1993; 34:663-72.

3. Basilico C, Moscatelli D. The FGF family of growth factors and oncogenes. Adv Cancer Res 1992; 59:115-65.

4. Bikfalvi A, Klein S, Pintucci G, Rifkin DB. Biological roles of fibroblast growth factor-2. Endocr Rev 1997; 18:26-45.

5. Slack JM, Darlington BG, Heath JK, Godsave SF. Mesoderm induction in early Xenopus embryos by heparin-binding growth factors. Nature 1987; 326:197-200.

6. Kimelman D, Abraham JA, Haaparanta T, Palisi TM, Kirschner MW. The presence of fibroblast growth factor in the frog egg: its role as a natural mesoderm inducer. Science 1988; 242:1053-6.

7.Sherman L, Stocker KM, Morrison R, Ciment G. Basic fibroblast growth factor (bFGF) acts intracellularly to cause the transdifferentiation of avian neural crest-derived Schwann cell precursors into melanocytes. Development 1993; 118:1313-26.

8. Kay EP, Gu X, Smith RE. Corneal endothelial modulation: bFGF as direct mediator and corneal endothelium modulation factor as inducer. Invest Ophthalmol Vis Sci 1994; 35:2427-35.

9. Kandel J, Bossy-Wetzel E, Radvanyi F, Klagsbrun M, Folkman J, Hanahan D. Neovascularization is associated with a switch to the export of bFGF in the multistep development of fibrosarcoma. Cell 1991; 66:1095-104.

10. Couderc B, Prats H, Bayard F, Amalric F. Potential oncogenic effects of basic fibroblast growth factor requires cooperation between CUG and AUG-initiated forms. Cell Regul 1991; 2:709-18.

11. Gu X, Seong GJ, Lee YG, Kay EP. Fibroblast growth factor 2 uses distinct signaling pathways for cell proliferation and cell shape changes in corneal endothelial cells. Invest Ophthalmol Vis Sci 1996; 37:2326-34.

12. Van Horn DL, Sendele DD, Seideman S, Buco PJ. Regenerative capacity of the corneal endothelium in rabbit and cat. Invest Ophthalmol Vis Sci 1977; 16:597-613.

13. Michels RG, Kenyon KR, Maumence AE. Retrocorneal fibrous membrane. Invest Ophthalmol 1972; 11:822-31.

14. Brown SI, Kitano S. Pathogenesis of the retrocorneal membrane. Arch Ophthalmol 1966; 75:518-25.

15. Waring GO, Laibson PR, Rodrigues M. Clinical and pathologic alterations of Descemet's membrane: with emphasis on endothelial metaplasia. Surv Ophthalmol 1974; 18:325-68.

16. Florkiewicz RZ, Sommer A. Human basic fibroblast growth factor gene encodes four polypeptides: three initiate translation from non-AUG codons. Proc Natl Acad Sci U S A 1989; 86:3978-81.

17. Prats H, Kaghad M, Prats AC, Klagsbrun M, Lelias JM, Liauzun P, Chalon P, Tauber JP, Amalric F, Smith JA, et al. High molecular mass forms of basic fibroblast growth factor are initiated by alternative CUG codons. Proc Natl Acad Sci U S A 1989; 86:1836-40.

18. Arnaud E, Touriol C, Boutonnet C, Gensac MC, Vagner S, Prats H, Prats AC. A new 34-kilodalton isoform of human fibroblast growth factor 2 is cap dependently synthesized by using a non-AUG start codon and behaves as a survival factor. Mol Cell Biol 1999; 19:505-14.

19. Delrieu I. The high molecular weight isoforms of basic fibroblast growth factor (FGF-2): an insight into an intracrine mechanism. FEBS Lett 2000; 468:6-10.

20. Renko M, Quarto N, Morimoto T, Rifkin DB. Nuclear and cytoplasmic localization of different basic fibroblast growth factor species. J Cell Physiol 1990; 144:108-14.

21. Florkiewicz RZ, Baird A, Gonzalez AM. Multiple forms of bFGF: differential nuclear and cell surface localization. Growth Factors 1991; 4:265-75.

22. Mason IJ. The ins and outs of fibroblast growth factors. Cell 1994; 78:547-52.

23. Quarto N, Finger FP, Rifkin DB. The NH2-terminal extension of high molecular weight bFGF is a nuclear targeting signal. J Cell Physiol 1991; 147:311-8.

24. Amalric F, Bouche G, Bonnet H, Brethenou P, Roman AM, Truchet I, Quarto N. Fibroblast growth factor-2 (FGF-2) in the nucleus: translocation process and targets. Biochem Pharmacol 1994; 47:111-5.

25. Bikfalvi A, Klein S, Pintucci G, Quarto N, Mignatti P, Rifkin DB. Differential modulation of cell phenotype by different molecular weight forms of basic fibroblast growth factor: possible intracellular signaling by the high molecular weight forms. J Cell Biol 1995; 129:233-43.

26. Bouche G, Gas N, Prats H, Baldin V, Tauber JP, Teissie J, Amalric F. Basic fibroblast growth factor enters the nucleolus and stimulates the transcription of ribosomal genes in ABAE cells undergoing G0----G1 transition. Proc Natl Acad Sci U S A 1987; 84:6770-4.

27. Hill DJ, Logan A. Cell cycle-dependent localization of immunoreactive basic fibroblast growth factor to cytoplasm and nucleus of isolated ovine fetal growth plate chondrocytes. Growth Factors 1992; 7:215-31.

28. Gu X, Kay EP. Distribution and putative roles of fibroblast growth factor-2 isoforms in corneal endothelial modulation. Invest Ophthalmol Vis Sci 1998; 39:2252-8.

29. Michigami T, Suga A, Yamazaki M, Shimizu C, Cai G, Okada S, Ozono K. Identification of amino acid sequence in the hinge region of human vitamin D receptor that transfers a cytosolic protein to the nucleus. J Biol Chem 1999; 274:33531-8.

30. Moede T, Leibiger B, Pour HG, Berggren PO, Leibiger IB. Identification of a nuclear localization signal, RRMKWKK, in the homeodomain transcription factor PDX-1. FEBS Lett 1999; 461:229-34.

31. Winkles JA, Friesel R, Alberts GF, Janat MF, Liau G. Elevated expression of basic fibroblast growth factor in an immortalized rabbit smooth muscle cell line. Am J Pathol 1993; 143:518-27.

32. Kurokawa T, Sasada R, Iwane M, Igarashi K. Cloning and expression of cDNA encoding human basic fibroblast growth factor. FEBS Lett 1987; 213:189-94.

33. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 227:680-5.

34. Roghani M, Moscatelli D. Basic fibroblast growth factor is internalized through both receptor-mediated and heparan sulfate-mediated mechanisms. J Biol Chem 1992; 267:22156-62.

35. Rusnati M, Urbinati C, Presta M. Internalization of basic fibroblast growth factor (bFGF) in cultured endothelial cells: role of the low affinity heparin-like bFGF receptors. J Cell Physiol 1993; 154:152-61.

36. Murono EP, Washburn AL, Goforth DP, Wu N. Evidence that both receptor- and heparan sulfate proteoglycan-bound basic fibroblast growth factor are internalized by cultured immature Leydig cells. Mol Cell Endocrinol 1993; 98:81-90.

37. Friedman S, Zhan X, Maciag T. Mutagenesis of the nuclear localization sequence in FGF-1 alters protein stability but not mitogenic activity. Biochem Biophys Res Commun 1994; 198:1203-8.

38. Gualandris A, Coltrini D, Bergonzoni L, Isacchi A, Tenca S, Ginelli B, Presta M. The NH2-terminal extension of high molecular weight forms of basic fibroblast growth factor (bFGF) is not essential for the binding of bFGF to nuclear chromatin in transfected NIH 3T3 cells. Growth Factors 1993; 8:49-60.

39. Park SY, Barron E, Suh PG, Ryu SH, Kay EP. FGF-2 facilitates binding of SH3 domain of PLC-gamma 1 to vinculin and SH2 domains to FGF receptor in corneal endothelial cells. Mol Vis 1999; 5:18 <http://www.molvis.org/molvis/v5/a18/>.

40. Levine JE, Prystowsky MB. Polypeptide growth factors in the nucleus: a review of function and translocation. Neuroimmunomodulation 1995; 2:290-8.

41. Jans DA, Hassan G. Nuclear targeting by growth factors, cytokines, and their receptors: a role in signaling? Bioessays 1998; 20:400-11.