Cellular Adhesion Mediated by Factor J, a Complement Inhibitor

Factor J (FJ) is a complement inhibitor that acts on the classical and the alternative pathways. We demonstrated FJ-cell interactions in fluid phase by flow cytometry experiments using the cell lines Jurkat, K562, JY, and peripheral blood lymphocytes. FJ bound to plastic plates was able to induce in vitro adhesion of these cells with potency equivalent to fibronectin. As evidence for the specificity of this reaction, the adhesion was blocked by MAJ2, an anti-FJ monoclonal antibody, and by soluble FJ. Attachment of the cells required active metabolism and cytoskeletal integrity. The glycosaminoglycans heparin, heparan sulfate, or chondroitin sulfates A, B, and C inhibited to varying degrees the binding of FJ to cells, as did treatment with chondroitinase ABC. In the search for a putative receptor, a protein of 110 kDa was isolated by affinity chromatography, and microsequence analysis identified this protein as nucleolin. Confocal microscopy evidenced the presence of nucleolin in cell membrane by immunofluorescence with monoclonal (D3) and polyclonal anti-nucleolin antibodies in Jurkat cells. The interaction FJ-nucleolin was evidenced by Western blot and enzyme-linked immunosorbent assay. Furthermore, purified nucleolin and D3 inhibited adhesion of Jurkat cells to immobilized FJ, suggesting that the interaction was specific and that nucleolin mediated the binding.

Factor J (FJ) 1 is a complement inhibitor that is able to regulate both the classical and the alternative pathways of complement (1,2). FJ was initially found as a soluble molecule in urine and serum (1,3). FJ is a peculiar protein having a high sugar content and a pI Ն 9.6 (4). In addition, the existence of FJ-related molecules has been described on the surface of human circulating cells and lymphoid cell lines (5). Flow cytometry analysis with an anti-FJ monoclonal antibody (mAb) has revealed staining in a small but consistent population of lymphocytes (mean: 11%) but not in monocytes, granulocytes, erythrocytes, or platelets. Moreover, we have also found that anti-FJ stained several cell lines (Jurkat, U937, K562, and Ramos) at varying percentages.
Relationships between complement and adhesion processes have been described previously; moreover, complement regulators such as vitronectin (VN) (6), clusterin (7), and more recently, factor H-like protein 1 (8) and factor H (9) have been implicated in adhesion processes. Affinity purification identified Mac1 (CD11b/CD18) as a factor H binding receptor (9). Moreover, CD11b/CD18 (Mac1) binds several soluble ligands including the complement fragment iC3b (10). Some characteristics of FJ resemble VN. Among them are the ability to regulate complement, the tendency to form aggregates, and the potential to bind heparin (11). VN is a ligand for ␤ 1 and ␤ 3 integrins. Clusterin, a different complement regulator, is capable of promoting the aggregation and adhesion of renal epithelial cells.
FJ also strongly binds heparin (12). Heparin-binding proteins constitute a diverse group that includes extracellular matrix molecules, growth factors, degradative and lipolytic enzymes, protease inhibitors, nuclear proteins, and lipoproteins. The molecular characteristics of FJ and its presence on cell membranes could make this protein a candidate for cell-cell or cell-matrix interactions.
The aim of this work was to investigate whether FJ could interact with different cell lines and peripheral blood lymphocytes (PBL). We analyzed FJ-cell interactions by flow cytometry and in adhesion assays using FJ immobilized on plastic plates. We observed that soluble FJ binds to several cell lines and PBL. Furthermore, immobilized FJ was able to induce cell adhesion in vitro. We isolated an FJ receptor protein by affinity chromatography and identified it by microsequencing analysis as nucleolin. In addition, competition experiments with nucleolin, its fragments, and with an anti-nucleolin mAb were performed. Metabolic requirements for FJ-cell interactions were also investigated.

MATERIALS AND METHODS
Cell Lines and Blood Cells-Cell lines: Jurkat (human T cell line), K562 (human chronic myelogenous leukemia), U937 (human histiocytic lymphoma), JY (human B cell line), and Vero (monkey kidney cell line) (American Type Culture Collection, Rockville, MD) were cultured in *This work was supported by Grants FIS 94/0119 and FIS 96/2072 from the Fondo de Investigaciones Sanitarias and Grant AE 00256/95 from the Comunidad de Madrid. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. FJ was purified from human urine following our original scheme with minor modifications (3). Briefly, human urine (1l) was collected in the presence of protease inhibitors and was sequentially chromatographed on ion exchange and heparin affinity columns. The presence of FJ was followed up throughout the procedure by means of an ELISA using MAJ2.
FACS Analysis; Binding of FJ to Cell Lines-Several cell lines and PBL were tested for FJ-cell binding in fluid phase, and the reaction was analyzed by fluorescence flow cytometry on a FACSsort Cytometer using Cell Quest software (Becton Dickinson). First, Jurkat cells (2 ϫ 10 5 ) were treated with different concentrations of FJ (0.25, 0.50, 1, 2.5, 5, and 10 g/ml) for 30 min at 4°C in a volume of 100 l. Then they were incubated with 2.5 l of ascitic fluid containing IgM (8 mg/ml) anti-FJ mAb MAJ2 (5) for an additional 30 min at 4°C, immunostained with 1:500 phycoerythrin-conjugated goat anti-mouse IgM F(abЈ) 2 (Immunotech), and analyzed by flow cytometry. Three thousand cells per sample were counted. The same experiment was carried out with K562, JY, and U937 cell lines and PBL except that 0 and 5 g/ml FJ were used. The specificity of this interaction was demonstrated by competition experiments performed by incubation of the cells in a final volume of 100 l with a constant amount of fluorescein isothiocyanate-labeled FJ (FITC-FJ) (5 g/ml) and in the presence of several concentrations of unlabeled FJ (5-100 g/ml).
Cell Attachment Assays-The adhesion assays were performed as described previously (13). Briefly, 96-well flat-bottomed plates (Costar, Cambridge, MA) were coated overnight at 4°C with 1 g of FJ diluted in 0.1 M carbonate-bicarbonate pH 9.6 (100 l) per well. In most of the experiments, fibronectin (FN) was used as a positive control; plates were coated using the same buffer. After that, plates were washed and saturated with 1% BSA in PBS (200 l) for 1 h at 37°C. The plates were washed twice with PBS, and cells (2 ϫ 10 5 /well) from PBL or different cell lines (Jurkat, K562, JY, U937, and Vero) were added in RPMI and incubated for 30 min at 37°C. For Vero cell line (fibroblast-like cell line), cells were removed from tissue culture plates with trypsin-EDTA, washed, and resuspended in RPMI. To quantify cell attachment, plates were washed twice with RPMI, and cells were fixed with a mixture of acetone:methanol (1:1) and stained with crystal violet 0.5%. Absorbance at 550 nm was measured in an ELISA reader (Anthos Labtec Instruments). Optical density was found to be a nearly linear function of the number of cells as assessed by a calibration curve (optical density versus number of cells) made for each cell type used in our assays. There was a linear relation in a range from 6 ϫ 10 3 -2 ϫ 10 5 cells (0.06 -1.4 OD units) for PBL, 8 ϫ 10 3 -2.5 ϫ 10 5 cells (0.09 -2.09 OD units) for Jurkat, 1.5 ϫ 10 4 -2.5 ϫ 10 5 cells (0.035-1.13 OD) for K562, 4 ϫ 10 3 -2.5 ϫ 10 5 cells (0.07-1.9 OD units) for JY, 3 ϫ 10 3 -2 ϫ 10 5 cells (0.12-0.9 OD units) for U937, and 12.5 ϫ 10 3 -4 ϫ 10 5 cells (0.10 -2.41 OD units) for Vero. For Jurkat, JY, and Vero cells the precipitate was dissolved in 95% methanol before reading OD. To calculate the percentage of attachment, basal adherence to BSA (cell binding to BSA-coated wells was always Ͻ1% for each cell type) was subtracted from attachment values, and the final results were expressed as a percent of the total number of cells initially added.
Enzymatic Treatment of Cells-Jurkat cells (5 ϫ 10 4 ) in RPMI were treated with heparinase I (20 units/ml) or chondroitinase ABC (0.5 unit/ml) for 30 min at room temperature. After that, the cells were washed to remove the enzymes and tested in cell adhesion experiments.
Affinity Chromatography-FJ (500 g) was covalently coupled to Affi-Prep (Bio-Rad) according to the manufacturer's instructions. A blank column was prepared in the same conditions using the matrix alone, without FJ. Jurkat cells (2 ϫ 10 7 ) were labeled with sulfo-NHSbiotin at 1.34 mg/ml in PBS, pH 8 for 30 min at 4°C and washed extensively. Then cells were lysed with 0.5% Nonidet P-40, 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM CaCl 2 , 1 mM MgCl 2 , and 1 mM phenylmethylsulfonyl fluoride and pipetted for 2 min at room temperature and 8 min at 0°C. The lysates were centrifuged at 22,000 ϫ g for 30 min at 4°C, and the supernatant was collected and applied to the FJ-Affi-Prep matrix or to the blank column, previously equilibrated with column buffer (0.125% Nonidet P-40, 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 1 mM MgCl 2; conductivity: 15 mS at 20°C). The lysates were incubated for 30 min at room temperature with both FJ and blank matrices. After that, the matrices were washed with 6 volumes of column buffer, and bound proteins were eluted with the same buffer, adjusting the conductivity to 30 mS at 20°C with NaCl. The column was washed with 6 M urea and re-equilibrated with column buffer. Retained proteins were dialyzed against water, concentrated with a Speed Vac concentrator (Savant, Farmingdale, NY), and loaded onto 8% SDS-PAGE followed by Western blotting. Transferred proteins were revealed using avidin-biotinylated horseradish peroxidase complex (Vectastain Elite, Vector Laboratories, Inc., Burlingame, CA) according to the manufacturer's instructions.
Microsequencing and Ligand Blot of the Purified Receptor-Various amounts of purified receptor were electroblotted onto a polyvinylidene difluoride (PVDF) membrane (Immobilon P, Millipore, Bedford, MA). The protein band detected on the primary blotting sheet by Ponceau S staining was excised and subjected to automated Edman degradation on an Applied Biosystems 473A pulsed-liquid phase protein sequencer by Dr. A. Marina (Servicio de Química de Proteínas, Centro de Biología Molecular "Severo Ochoa," Madrid).
Purified receptor was assayed for FJ binding by a ligand blot assay. For that, unlabeled proteins from the Affi-Prep column were run in two different lanes on 8% SDS-PAGE, transferred to a nitrocellulose membrane, and probed with purified FJ at 300 ng/ml or without FJ. After washing the membranes, both lanes were incubated with MAJ2, followed by affinity isolated goat F(abЈ) 2 anti-(mouse IgM Mu chain) peroxidase conjugate. The reaction was developed with 4-chloronaphthol.
Effect of Nucleolin or an Anti-nucleolin mAb on Cellular Adhesion Mediated by FJ-To prove the involvement of nucleolin in FJ binding, competition experiments were performed in the presence of purified nucleolin (18) or its GAR, p40, and p50 fragments (19). For that, either nucleolin at concentrations ranging from 0.1 to 10 g/ml or fragments (10 g/ml of each) were added to the FJ-coated plates before adding the cells at a volume of 100 l. After washing the plates, cells were added and incubated for 30 min at 37°C. Attachment values were expressed as a percent of control. Furthermore, Jurkat cells were preincubated with different concentrations (10 -80 g/ml) of anti-nucleolin mAb (D3) for 30 min at room temperature. Then cells were assayed for attachment on plates coated with 1 g of FJ by well for 30 min at 37°C and tested for cellular adhesion as described above.
Nucleolin Interaction with FJ-The availability of FJ and purified nucleolin allowed us to design experiments to characterize FJ-nucleolin interaction. For that, 200 ng of purified nucleolin (18) was subjected to 8% SDS-PAGE transferred to PVDF and probed for FJ as described above for purified FJ receptor. Moreover ELISA experiments were performed. For that, 96-well flat-bottomed plates were coated with 100 l of FJ (0.5 g/ml) in 0.1 M carbonate-bicarbonate. pH 9.6. After blocking with BSA (1%) in PBS, increasing concentrations of nucleolin were added and incubated for 45 min at 37°C. The reaction was developed with anti-nucleolin polyclonal Ab (20), followed by the addition of affinity isolated goat anti-rabbit Ig peroxidase conjugate (Biosource International, Camarillo, CA). ABTS was used as substrate for developing the enzymatic reaction. A at 405 nm was measured and plotted against nucleolin concentration. This system was used for competition experiments with heparin and other GAGs. For that, wells were coated with FJ (100 ng) and blocked with BSA as described above. Subsequently, several GAGs were added and incubated for 45 min, followed by the addition of nucleolin (50 ng/well). The reaction was developed with anti-nucleolin polyclonal Ab (20). The binding percentage was compared with the binding in the absence of competitor.
Immunofluorescence Microscopy-We have examined the presence of nucleolin on the membrane of the Jurkat cell line with two Abs, a mAb (D3) or a rabbit polyclonal Ab (20). For that, cells were grown on poly-L-lysine-coated coverslips. After this step, cells were washed with PBS and fixed with 2% paraformaldehyde or with methanol at Ϫ20°C. After washing the slides with PBS, D3 and rhodamine-labeled rabbit Igs anti-CD71 generated at the laboratory (Centro de Biología Molecular "Severo Ochoa") were added and incubated for 60 min at room temperature. The coverslips were extensively washed with PBS and incubated on a drop of the fluorescein-labeled goat antibodies specific for mouse Igs (Pierce) for 60 min. Furthermore, we analyzed the presence of nucleolin on the membrane with a rabbit polyclonal anti-nucleolin Ab (20) followed by the addition of a FITC-labeled second Ab (Pierce). Finally, the coverslips were extensively washed with PBS and mounted on slides with mowiol solution mixed with ProLong antifade kit (Molecular Probes, Leiden, The Netherlands). The preparations were examined in a Zeiss Axioskop microscope and photographed on Kodak TMAX 400 or Kodak Ektachrome P800/1600 film.

RESULTS
FJ Binds Specifically to Cells-The presence of a putative FJ receptor was analyzed by flow cytometry. A dose-dependent binding of FJ to cells was observed when the Jurkat cell line was incubated with different amounts of FJ (Fig. 1A). When several cell lines were compared, we observed that approximately 90% of the Jurkat, K562, JY, and PBL cells were able to join FJ. In contrast, only 23% of the U937 cells bound FJ (Fig.  1B). The competence between FJ and FITC-FJ is shown in Fig.  1C. Fig. 2A shows the attachment of different cell lines and PBL to plates coated with 1 g of FJ. Maximal attachment was observed with Vero, JY, and Jurkat cell lines, and the lowest value corresponded to U937 (3%). These results correlated with those observed in flow cytometry analysis for hematopoietic cells, with the exception of PBL. Moreover, FJ induced a dosedependent adhesion to the cells (Fig. 2B). Attachment was comparable with that induced by FN, and it did not seem to be due to the positive charge of FJ since a cationic control protein (cytochrome c) did not induce cell binding. The kinetics of cell adhesion to FJ are shown in Fig. 2C; saturation (maximum) binding is reached in 20 -30 min.
The specificity of this adhesion was tested by the incubation of the FJ-coated plates, with an excess of purified MAJ2 (5 g/well) or with an irrelevant mAb (MsIgM, used as control at the same concentration), prior to the addition of the cells (Jurkat). MAJ2 inhibited cell adhesion to FJ (Fig. 3A). In contrast, the control mAb had no effect on FJ-induced cell adhesion. Furthermore, when several amounts of FJ were incubated with Jurkat cells in fluid phase before cells were added to the plates, the attachment was reduced in a dose-response manner (Fig.  3B). In parallel experiments, adhesion of the Jurkat cells to FN remained unaffected.
FJ-mediated Adhesion Is Dependent on Cytoskeletal Integrity and Is Inhibited by Heparin-The requirements of FJ-mediated cell adhesion were investigated in Jurkat cells (Table I). Adhesion required active metabolism since it did not take place at 4°C, and it was inhibited by 2-deoxyglucose in the presence of 0.1% sodium azide. To determine whether FJ-mediated adhesion depended on an intact cytoskeleton, binding assays were performed after treatment with, and in the presence of, cytochalasin B or nocodazole, drugs that disrupt actin filaments and microtubules, respectively. The cytochalasin B concentrations were chosen to be comparable with those shown to be effective in the disruption of FN-␤ 1 integrin interaction (13). For nocodazole treatment, the concentrations were chosen to be comparable with those used in other cell adhesion experiments as for CD4/MHC Class II protein interaction (21). Both treatments inhibited FJ-mediated cell adhesion in a dose-response manner. In addition to cytoskeletal integrity, a functional Na ϩ /H ϩ antiporter was also required since an amiloride derivative (EPA) blocked adhesion (see Table I). Simultaneously, FN/cell interaction was analyzed as control, the results of these experiments being similar to those previously described (13).
Next we examined the effect of certain glycosaminoglycans (GAGs), including heparin, heparan, or chondroitin sulfate (CS) A, B, or C on cell attachment to plates coated with FJ. For this experiment, FJ-coated plates were incubated with several concentrations of GAGs in a final volume of 100 l. After that, the plates were washed, and cells were added. The strongest inhibition was observed with heparin. However, a partial dosedependent inhibition was also observed with heparan sulfate and with chondroitin sulfate A, B, or C (Fig. 4).
When Jurkat cells were treated with chondroitinase ABC prior to adding them to FJ-coated plates, 30% adhesion inhibition was observed. In contrast, inhibition was not observed after heparinase treatment. Moreover, when a mixture of heparinase and chondroitinase was used, we did not find a further decrease in cell adhesion over that observed with chondroitinase alone (data not shown).
Nucleolin Constitutes the Cell Surface Receptor for FJ-The possibility that FJ could bind cells through an integrin or selectin molecule was considered. Cells were incubated with different anti-␤ 1 , -␤ 2 , and -␤ 3 integrins or anti-CD62L selectin mAbs or with anti-CD71 and anti-HLA-A and -B mAbs as controls and analyzed for binding to plates coated with FJ. None of these mAbs inhibited cell binding to FJ. Similar negative results were obtained using cells incubated with anti-CD43 or anti-CD44 (data not shown).
The above results suggested the presence of a different molecule as FJ receptor. To identify the receptor, intact cells were labeled with biotin and cell lysates passed through a FJ affinity column. Fig. 5 shows a Western blot of the biotinylated material retained and later eluted from a FJ-Affi-Prep column or from a control column. A protein band of approximately 110 kDa was specifically eluted from the FJ column. Ligand blot assay verified the specificity of this interaction because proteins eluted from the Affi-Prep column and electroblotted into a nitrocellulose membrane were probed with purified FJ. MAJ2 bound to the membrane only when FJ was previously added (Fig. 6A, lane 2) but not when FJ was absent (Fig. 6A, lane 1). Furthermore, FJ binds selectively to this 110-kDa species above all other cellular proteins as shown in Fig. 6B.
In addition, the purified 110-kDa protein was transferred to PVDF membrane, excised, and sequenced. The microsequence analysis showed that 10 N-terminal amino acid residues had a 100% homology with human nucleolin. Moreover, D3 antinucleolin mAb bound to the 110-kDa band in Western blot experiments (Fig. 7). A rabbit anti-nucleolin polyclonal Ab (20) also bound to this band (data not shown).
The involvement of nucleolin in FJ binding was also tested by competition cell adhesion experiments in the presence of purified nucleolin (18), or its GAR, p40, and p50 fragments (19). There was a dose-dependent inhibition when nucleolin, from 0.1 to 10 g/ml, was added to the plates before the cells (Fig.  8A). In contrast, GAR, p40, and p50 fragments did not inhibit cell binding (Fig. 8B). The possibility that FJ interacted with nucleolin because of its negative charge (22) was ruled out because BSA, with a pI slightly lower than nucleolin, was used to saturate the plastic plates at 10 mg/ml. As shown in Fig. 8A, adhesion mediated by FJ was completely inhibited by nucleolin at doses 1000-fold lower than those of BSA. D3, an anti-nucleolin mAb incubated with cells, also inhibited cellular adhesion in a dose-dependent manner (Fig. 8C).
The availability of FJ and purified nucleolin allowed characterization of the FJ-nucleolin interaction by Western blot and ELISA. After Western blot, purified nucleolin was probed for FJ binding and a 110-kDa band was revealed with MAJ2 (Fig.  9A). In ELISA experiments, we observed that FJ and nucleolin interacted in a dose-dependent way (Fig. 9B) and also that heparin and other GAGs could modify FJ-nucleolin interaction (Fig. 9C).
We examined the presence of nucleolin on the membrane of the Jurkat cell line with two Abs, D3 mAb, or a rabbit polyclonal Ab (20). A positive staining was observed on the plasma   6. Ligand-blot for FJ receptor. A, FJ binds the receptor isolated by affinity chromatography. Jurkat cells (2 ϫ 10 7 ) were lysed (see text) and passed through an FJ-Affi-Prep column. Two lanes containing retained proteins from the FJ-Affi-Prep column were loaded on 8% SDS-PAGE and transferred to nitrocellulose membrane. After that, the nitrocellulose membrane containing transferred receptor was incubated with PBS-Tween (lane 1) or with purified FJ (300 ng/ml) in PBS-Tween (lane 2). After the membranes were washed, both lanes were incubated with MAJ2 (mAb anti-FJ), followed by affinity-isolated goat F(abЈ) 2 anti-(mouse IgM mu chain) peroxidase conjugate. The reaction was developed with 4-chloronaphthol. B, FJ binding to the complete cell lysate. Jurkat lysates (10 g) were subjected to 8% SDS-PAGE and transferred to PVDF membrane. After that, the membrane was either incubated with PBS-Tween (lane 1) or with purified FJ in PBS-Tween (300 ng/ml, lane 2). Finally, the binding of FJ was detected using MAJ2, followed by the addition of affinity-isolated goat antimouse IgM alkaline phosphate conjugate. The reaction was developed with nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate as precipitating substrate. membrane of the cells with D3 and with rhodamine-labeled anti-CD71 (Fig. 10, panels A and B). Anti-CD71 Ab recognized the transferrin receptor, a cellular surface marker. Staining with rabbit polyclonal anti-nucleolin (20) was also observed on the membrane (Fig. 10, panel C) and on the nucleolus of the cells permeabilized with methanol (Fig. 10, panel D). DISCUSSION For the first time, our results described the ability of FJ, a complement inhibitor, to induce cell adhesion. The specificity of the interaction was demonstrated by competition with anti-FJ mAbs or with soluble FJ. This adhesion was not mediated by many of the presently known adhesion molecules, including ␤ 1 , ␤ 2 , or ␤ 3 integrins, and CD62L, CD43, or CD44 molecules. Temperature dependence and the requirement for cytoskeletal integrity make FJ adhesion similar to that meditated by FN-␤ 1 integrin interaction (13). Furthermore, the percentages of attachment of several cell lines were similar for both molecules (Fig. 2B). The interaction was inhibited to varying degrees by several glycosaminoglycans including heparin, heparan sulfate, and chondroitin sulfate A, B, and C, as well as by the treatment of the cells with chondroitinase ABC. The interaction of FJ and cells was also evident in fluid phase as assessed by FJ binding to cells in flow cytometry experiments. Moreover, in these experiments, a dose-dependent effect was observed.
To characterize the molecule(s) involved in FJ-cell interaction we detected, by elution from immobilized FJ, a major protein band of approximately 110 kDa. This protein was able to bind FJ when immobilized on nitrocellulose membrane. Microsequence analysis identified it as nucleolin. Nucleolin, also known as C23, is a major nucleolar protein in eukaryotic cells (23,24) and one of the proteins associated with the nucleolar organizer region (22,25). Nucleolin contains RNA recognition motifs and has been demonstrated to bind to RNA (24,26) and DNA (27). It is phosphorylated by casein kinase II during interphase (28) and by Cdc2 kinase during mitosis (29). The nucleolin sequence data suggest a modular organization in functional domains. The first N-terminal domain contains repeated motifs (TPXKK) surrounded by acidic stretches and is the site of multiple phosphorylation (24,30). This domain also interacts with chromatin (31) and is likely to be implicated in protein-protein interactions. The FJ-nucleolin interaction also seems to be mediated by this domain since p40, p50, or GAR domains do not block the adhesion.
Despite being a nuclear protein, nucleolin may be located at other unexpected sites, including the cell surface (16,17,(32)(33)(34)(35)(36)(37)(38). Cell surface nucleolin has been reported to bind lipoproteins, laminin, and growth factors (34,35,37). Kibbey et al. (35) suggest that nucleolin binding to the IKVAV site of laminin may play a role in neurite differentiation, while Gillery et al. (39) suggest that extracellular matrix interactions with nucleolin may directly affect protein synthesis. Thus, experimental studies indicate a functional role for nucleolin at the cell surface. Deng et al. (36) demonstrate cell surface nucleolin on Hep-2 and HepG2 cell lines and show that antibody to surface nucleolin is rapidly internalized; therefore, a pathway exists for cellular uptake of molecules bound to nucleolin.
We showed that nucleolin was biotinylated in intact cells using a method reported to give selective biotinylation of the cell surface proteins of leukocytes (40). Furthermore, we evidenced nucleolin on the cell membrane by confocal microscopy so our data agree with previous reports of cell surface nucleolin. Many putative cell surface receptors are identical to abundant intracellular multifunctional proteins (32); while some of the proposed receptor activities may be artifactual, many such proteins appear to be truly expressed at the cell surface and to exhibit developmentally regulated patterns of cell surface expression.
The specificity of FJ-nucleolin interaction was evidenced by several experiments such as ligand blot and ELISA. We also showed that some GAGs can modify this interaction. The association of nucleolin to casein kinase 2 was also inhibited by heparin, an inhibitor of the enzyme (41). We showed that FJ-nucleolin interaction was also modified for several GAGs. The FJ-nucleolin interaction may also be involved in cell control mechanisms. There are many similarities between basic fibroblast growth factor (bFGF) and FJ. bFGF is a cationic protein involved in the regulation of many cellular processes. FGFs are a family of heparin-binding polypeptides that play a role in a wide array of biological processes, including cell growth, differentiation, angiogenesis, tissue repair, and transformation (42,43).
Recently, Maher (44) has found nuclear translocation of FGF receptors in response to bFGF. Different sequences containing basic amino acids have been implicated in protein transport to the nucleus (45). FJ is a cationic protein; attempts to sequence it have been unsuccessful, probably because this protein is resistant to proteolytic enzymes. Other strategies, including different purification schemes or its digestion in biological fluids, are in progress. If the analogy to bFGF holds, FJ may also be internalized after membrane interactions. This hypothesis is currently being tested.
The cellular distribution of receptors for FJ in cell lines supports the analogy of FJ receptors with FGF receptors. FGF receptor expression in human leukemia cell lines is heteroge- FIG. 9. Nucleolin interaction with FJ. A, FJ binds purified nucleolin. Two hundred ng of nucleolin was run on an 8% SDS-PAGE and transferred to Immobilon-P. After that, the membrane was blocked with 3% non-fat milk and probed with 300 ng of purified FJ (lane 1) or without FJ (lane 2), followed by the addition of rabbit polyclonal anti-FJ and an alkaline phosphatase-conjugated secondary Ab. B, the interaction FJ-nucleolin analyzed by ELISA. Flat-bottomed plates were coated with 100 l of purified FJ (0.5 g/ml) in 0.1 M carbonate-bicarbonate buffer, pH 9.6. After blocking with BSA (1%) in PBS, increasing concentrations of nucleolin were added, and the reaction was developed with a polyclonal anti-nucleolin Ab. The reaction was followed by the addition of peroxidase-labeled secondary Ab and developed with ABTS as substrate. C, modification of FJ-nucleolin interaction by GAGs. Plates were coated with purified FJ followed by the addition of BSA for blocking nonreactive sites. After that, different GAGs were added at several concentrations and incubated for 45 min at 37°C. The plates were washed, and purified nucleolin was added to the plates. The interaction FJ-nucleolin was measured by the reactivity with an antinucleolin polyclonal Ab. neous as shown by mRNA analysis (46). Monocytic cell lines U937 and RC-2A do not express mRNA for FGF receptors (FGFR-1, FGFR-2, FGFR-3, FGFR-4). U937 also binds FJ to a lower extent than other cell lines (Fig. 1B). FGFs interact with two classes of FGF receptors, a low and a high affinity (47,48). bFGF interacts with cells through heparan sulfate, and the interaction is mediated by an increase in tyrosine phosphorylation. The interaction of FJ with cells also induces an increase in tyrosine-phosphorylated proteins (data not shown). However, we have ruled out the identity of these two molecules, FJ and bFGF, by ELISA using anti-FJ and anti-bFGF mAbs (data not shown). A regulatory relationship between bFGF expression and nucleolin phosphorylation is demonstrated by Bonnet et al. (49). The reduction on cell adhesion induced on cells after chondroitinase treatment suggests that other molecules, probably involving more than one receptor, are implicated in the cell adhesion.
FJ might act as an extracellular matrix protein similar to FN, since FJ also strongly binds both heparin (12) and heparan sulfate and is highly glycosylated. Although initially described as a complement inhibitor, FJ may mediate cell-cell or cellmatrix interactions. Therefore, adhesion might act in the regulation of complement inhibition, sequestering FJ in a functionally inert form until it is required by the body. Data from our laboratory indicate the presence of high levels of FJ in synovial fluid from patients having different inflammatory arthropathies (50). FJ could modify cell traffic to synovial tissue as other adhesion molecules do (51,52).
Clusterin, Factor H, and Factor H-like protein 1 can simultaneously regulate complement and are involved in the adhesion mechanism (7-9). Cellular adhesion induced by complement regulators could contribute to focal tissue repair at sites of complement activation. Interactions of cells with components of the extracellular matrix, such as a FN, laminin, or tenascin, play an important role in morphogenesis, tissue repair, regeneration, or metastasis; by analogy FJ may also play a role in these processes.