Heparin’s Roles in Stabilizing, Potentiating, and Transporting LEDGF into the Nucleus

Abstract

purpose. Lens epithelium-derived growth factor (LEDGF) is a 60-kDa protein that dramatically enhances cellular survival, growth, adhesiveness, and resistance to heat and oxidative stress. Full-size recombinant LEDGF is degraded during prokaryotic preparation. Heparin’s capacity to stabilize recombinant LEDGF in the face of various stresses (heat, pH, proteolysis), to potentiate its growth-enhancing properties, and to enable transport of LEDGF into the nucleus of mouse lens epithelial cells has been characterized.

methods. LEDGF-cDNA was cloned in a pGEX-2T expression vector to produce a fusion protein, GST-LEDGF. Porcine heparin was used to stabilize GST-LEDGF. Heparin-Sepharose was used to characterize heparin-GST-LEDGF binding, and GST-LEDGF or heparin-GST-LEDGF was used to quantitate heparin’s stabilization of LEDGF in the face of heat, pH, and proteolytic stresses. Fluorescein isothiocyanate–labeled GST-LEDGF and heparin-GST-LEDGF were incubated with cultured mouse lens epithelial cells (LECs). Fluorescence microscopy and immunostaining techniques were used to monitor heparin’s potentiation of LEDGF’s growth stimulation and heparin’s role in the translocation of GST-LEDGF from the extracellular space into the cytoplasm and nucleus.

results. Heparin, at concentrations as low as 7.1 mg/ml, protected GST-LEDGF from degradation and increased the yield of the full-size fusion protein in a prokaryotic system. It also protected GST-LEDGF from heat, acid-base deactivation, and proteolytic degradation with trypsin and chymotrypsin and greatly potentiated LEDGF’s enhancement of mouse LEC growth in culture. It also increased nuclear uptake of exogenous GST-LEDGF and endogenous LEDGF.

conclusions. Heparin protected GST-LEDGF from degradation under various stress conditions and facilitated transport of GST-LEDGF into the nucleus.

In 1998 we reported ¹ a novel growth, survival, and adhesive factor cloned from a human lens epithelium cDNA library. We named it lens epithelium-derived growth factor (LEDGF). LEDGF is found in many other tissues ^{1 2} and in the nuclei of early stem cells (Wendy A. Bickmore, personal communication, March 1999). We defined the DNA sequence and the intron–exon organization of the LEDGF gene. ³ The primary amino acid sequence of a splicing cofactor protein p75 ⁵ was identical with that of LEDGF, and the 325 N-terminal amino acids of a transcriptional coactivator p52 were found to be identical with the N-terminal region of p75 ⁴ and LEDGF. ^{1 3} We showed also that LEDGF and p52 are derived from a single gene by alternative splicing. ³ The protein p52, but not p75, interacts with transcriptional coactivators, general transcription factors, and splicing factors to modulate pre-mRNA splicing of class II genes. ⁵ LEDGF also markedly enhances the resistance of lens epithelial cells (LECs) to heat and oxidative stress, ⁶ and it upregulates the expression of Hsp-27 andα B-crystallin. ⁶  

To expand our studies of the extracellular–intracellular trafficking of LEDGF and LEDGF’s effects on cell growth, survival, resistance to stress, transcription, and other cellular processes, we needed milligram amounts of full-size protein. Unfortunately, LEDGF, like other highly charged proteins, was degraded in a prokaryotic expression system. To protect against this degradation we chose heparin, a sulfated polysaccharide known to stabilize many other growth factors. ^{7 8} This choice offered us potential additional benefits. Heparin may not only enhance yields of recombinant full-size protein, but it also may play a role in the trafficking and function of LEDGF. Published investigations with eukaryotic cells indicate that heparin potentiates growth factors ^{9 10} and complexes with growth factors to facilitate their transport into the nucleus. ^{11 12} In view of evidence that p75 (and consequently LEDGF) is a transcriptional coactivator, ^{4 5} that LEDGF upregulates the expression of Hsp27 andα B-crystallin, ⁶ that LEDGF is translocated from the cytoplasm into the nucleus, and that glycosaminoglycans (GAGs) may regulate nuclear gene expression, ¹³ the study of heparin’s growth-enhancing properties, and its role in intracellular trafficking of LEDGF seemed intriguing. 

Materials and Methods

Reagents and Media

Porcine heparin sodium (140 U/mg, molecular weight [MW] unspecified), Dulbecco’s modified Eagle’s medium (DMEM), isopropyl-d-thiogalactopyranoside (IPTG), and fetal calf serum were obtained from Gibco (Grand Island, NY). Two low molecular weight porcine heparins (MWs: 3000 and 6000 Da), porcine fluorescein isothiocyanate (FITC)–labeled heparin, trypsin, chymotrypsin, heparan sulfate, dextran sulfate, phenylmethylsulfonyl fluoride (PMSF), lysozyme, and Triton X-100 were obtained from Sigma (St. Louis, MO). Heparin-Sepharose CL-6B, and the GST purification modules (including glutathione-Sepharose 4B) were purchased from Pharmacia Biotech (Piscataway, NJ). Trifluoroacetic acid (TFA) was obtained from Applied Biosystems (Foster City, CA); EcoRI, XhoI, and BamHI restriction enzymes from New England Biolabs (Beverly, MA); Escherichia coli from Stratagene (La Jolla, CA); The polyclonal anti-GST antibody from Pharmacia Biotech (Piscataway, NJ); and the ABC (avidin-biotin conjugate) kit for immunostaining of cells from Santa Cruz Biotechnology (Santa Cruz, CA). 

Construction and Purification of Recombinant LEDGF

LEDGF cDNA was inserted into a pGEX-2T vector to produce a glutathione-S-transferase (GST)-LEDGF fusion protein. First, a 564-bp PCR fragment was generated that covered the initiation codon ATG and extended to an internal EcoRI site of LEDGF. Two primers (the 5′-primer 5′-ccccggatcccatgactcgcgatttcaaacct-3′, and the 3′-primer 5′-tcttgaattctgtagctgcaggtcgtcctct-3′ were used with LEDGF cDNA to generate the fragment. After the PCR product was cleaved with restriction enzymes BamHI and EcoRI, the PCR fragment was ligated between the BamHI and EcoRI sites of pGEX-2T. Next, a XhoI and an EcoRI fragment (2153 bp) was generated from LEDGF cDNA and ligated between the XhoI and EcoRI sites of the previous construct. E. coli (BL21) was transformed with the construct and incubated in 500 ml of Luria broth (LB) medium (tryptone 10 g, yeast extract 5 g, and NaCl 5 g/l containing 100 μg ampicillin per milliliter) at 37°C with shaking until the optical density of the culture reached 0.6 (OD_600nm). IPTG was then added at a final concentration of 100 μM and the incubation continued for 5 to 6 more hours. To make the lysate, the pellet was suspended in 25 ml lysis buffer (final concentrations: 50 mM Tris-HCl [pH 8.0] 200 mM NaCl, 1.5 mM EDTA) and 1 mM PMSF. Lysozyme was added (final concentration, 1 mg/ml) and the mixture kept on ice for 15 minutes The lysate was sonicated with short bursts, Triton X-100 was added to a final concentration of 1%, and the lysate was mixed gently for 30 minutes to solubilize the fusion protein. The lysate was centrifuged at 12,000 rpm for 10 minutes at 4°C. To purify the fusion protein, the supernatant was incubated overnight with 200 μl of a 50% slurry of glutathione-Sepharose 4B at 4°C. The suspension was then centrifuged at 500g for 5 minutes. The pellet was washed four times in lysis buffer and the fusion protein eluted with glutathione elution buffer. The protein was dialyzed against 2000 volumes of phosphate-buffered saline (PBS) at 4°C, the protein concentration was determined by the Bradford method, ¹⁴ and identity of the eluted protein was confirmed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), immunoblot analysis using a polyclonal anti-GST, and an anti-(C-terminal) LEDGF antibody. 

Generation of Anti-(C-Terminal) LEDGF Antibodies

A peptide (LYNKFKNMFLVGEGDSVIT; 419–437) from the C terminus of LEDGF was commercially synthesized. Four milligrams of peptide in 500μ l PBS was emulsified with 500 μl complete Freund’s adjuvant (CFA) and injected subcutaneously into a rabbit. Two milligrams of this peptide was emulsified in incomplete Freund’s adjuvant (IFA) and injected subcutaneously at 15 and 30 days after the first injection. Blood (for serum) was taken 7 days after the last injection. The antibody titer was measured through an enzyme-linked immunosorbent assay (ELISA), and this immune serum was used during the study. 

SDS-PAGE and Western Blot Analysis

SDS-PAGE (10%) was used to separate and visualize the protein. ¹⁵ The gels were stained with Coomassie brilliant blue. Western blot analysis (immunoblot analysis) was performed to confirm the identity of the protein. ¹⁶  

Assay of Growth Stimulation by GST-LEDGF

Mouse LECs from confluent cultures were subcultured in tissue culture flasks (growth area 75 cm²; Falcon; Becton Dickinson, Bedford, MA) at 37°C in DMEM with 10% FCS in a 5% CO₂ environment. For the biologic assays of the growth-enhancing potency of GST-LEDGF, cells were trypsinized (0.25% trypsin and 1 mM EDTA-Na in PBS) for 5 to 10 minutes at room temperature, separated from the bottom of the flask, washed with DMEM plus 10% FCS, and then washed again with DMEM without FCS. Five thousand cells/well in 96-well culture plates were used. 

MTS Assay

This colorimetric assay of cellular proliferation uses 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2 to 4-sulfophenyl)-2H-tetrazolium salt (MTS; Promega, Madison, MI). When added to medium containing viable cells, MTS is reduced to a water-soluble formazan salt. ¹⁷ Twenty microliters MTS was added to each well, and OD_490nm was measured in the well after 4 hours with an ELISA reader. 

Localization of FITC-Heparin and FITC-Heparin-GST-LEDGF Complex

FITC-labeled heparin and GST-LEDGF were mixed together in 1:4 ratio in binding buffer (10 mM NaH₂PO₄, pH 7.3) containing 150 mM NaCl and incubated with occasional shaking at room temperature for 1 hour. The unbound FITC-labeled heparin was removed on a Centricon YM-30 (Amicon, Beverly, MA) by washing three times with binding buffer. The washed heparin-GST-LEDGF complex and the FITC-heparin were added to incubated cells grown on glass coverslips. Periodically, the coverslips were mounted in PBS, observed, and photographed with a fluorescence microscope. 

Monitoring Extracellular–Intracellular Trafficking of LEDGF

Five thousand mouse LECs from confluent cultures were trypsinized, washed with DMEM containing 10% FCS, and transferred to and grown on glass coverslips overnight at 37°C in 6% CO₂ in an incubator. On the following day, adherent cells were washed gently three times with serum-free DMEM, fasted for 5 hours, placed in fresh serum-free DMEM containing heparin-GST-LEDGF or GST-LEDGF (100 ng/ml), and incubated for 2 more days. Cells were then washed three times with DMEM, treated with 1:1 ratio of methanol-acetone at −20°C for 5 minutes, fixed in 1% formalin in PBS for 10 minutes, and immunostained using the ABC kit. Briefly, endogenous peroxidase was blocked with 0.5% hydrogen peroxide. For the nonspecific antibody-binding control, specimens were washed twice in PBS and incubated for 1 hour at room temperature in 2% normal blocking serum. After the blocking serum was removed, the specimens were incubated overnight at 4°C with primary polyclonal anti-GST (diluted 1:500) or anti-(C-terminal) LEDGF antibodies (diluted 1:200). After three washes, the biotin-conjugated secondary antibody (approximately 1 mg/ml) in 1.5% goat serum was applied for 30 minutes. Specimens were washed three times again and incubated with avidin-biotin reagents. Antibody binding was visualized with diaminobenzidine. Specimens were washed, mounted, and microphotographed. 

All animal research was conducted according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, revised August 1998. 

Statistical Methods

The unpaired Student’s t-test was used to assess the statistical significance of differences between samples with and without heparin. 

Results

LEDGF-Heparin Binding

We investigated whether GST-LEDGF bound to heparin and whether heparin protected GST-LEDGF against degradation. GST-LEDGF was isolated and purified with glutathione-Sepharose beads. The purified GST-LEDGF eluted from glutathione-Sepharose beads was added to heparin-Sepharose beads in binding buffer (10 mM NaH₂PO₄ and 150 mM NaCl[ pH 7.3]) and incubated for 0.25, 0.50, 1, 2, and 4 hours at room temperature and at 4°C overnight. The beads were then separated by centrifugation (6000 rpm). The washed beads were boiled in sample buffer to release bound GST-LEDGF, and samples of supernatant were applied to an SDS gel (Fig. 1) . The supernatant from the mixture of purified GST-LEDGF and heparin-Sepharose beads that had been incubated for 0.25 hours (Fig. 1 , lane 2), showed no band. This indicated that the heparin-GST-LEDGF binding had occurred rapidly (within 0.25 hours). Similar results were observed after 30 minutes, 1 hour, 2 hours, 4 hours, and overnight (data not shown). The GST-LEDGF forms released from heparin-Sepharose are shown in lane 3. The GST-LEDGF forms present in the mixture before adding to heparin-Sepharose are shown in lane 4. Multiple forms (full size and degraded) of GST-LEDGF are present. Full-size GST-LEDGF (MW, 89 kDa) is the first band in lanes 3 and 4. The other bands are GST-containing LEDGF degradation products. To show that these other bands originated from GST-LEDGF degradation, a separate experiment (data not shown) was performed in which the SDS gel slice containing the full-size GST-LEDGF was cut out, the protein therein was eluted electrophoretically ¹⁸ and the eluted protein kept at room temperature for 48 hours and then run again on SDS-PAGE. Six to seven bands corresponding to GST-LEDGF degradation products were obtained. In lane 5 the proteins were immunostained with an anti-GST antibody (diluted 1:1000). It is clear that multiple bands with molecular sizes ranging from 89 to 25 kDa were present. In lane 6 the proteins were immunostained with the anti-(C-terminal) LEDGF antibody (diluted 1:500). Full-size GST-LEDGF, two large (MWs, approximately 60 kDa) degradation products, and a smaller protein (presumably a small peptide fragment containing the C-terminal epitope) were immunostained. 

To measure the affinity of binding, 1 mg purified GST-LEDGF was applied to a 1.0-ml heparin-Sepharose CL-6B column. The eluate was reapplied three to four times to maximize binding. The column was then washed with binding buffer to remove the unbound LEDGF, and bound GST-LEDGF was eluted with a NaCl gradient. The samples were dialyzed overnight against PBS and concentrated on a Centricon 30. The protein concentration was estimated, and aliquots were run on SDS-PAGE. There was little or no elution of GST-LEDGF with 0.1 to 0.4 M NaCl (data not shown). Twenty percent to 25% of the bound GST-LEDGF eluted with 0.6 M NaCl and 40% to 50% with 1.0 M NaCl. Only 75% of GST-LEDGF applied to the column was recovered. Assuming that the 25% nonrecovered LEDGF remained tightly bound to heparin-Sepharose and could have been released with higher salt concentrations, the binding ratio of heparin-to-GST-LEDGF was approximately 1:4. This is not a molar ratio, because Sepharose CL-6B beads contain heparins with a wide range of molecular weights. 

Heparin Prevents GST-LEDGF Degradation

We did not know the time during prokaryotic production of LEDGF when the addition of heparin would be most beneficial in preventing its degradation. When we added heparin only during the bacterial culture, the proteolytic degradation of GST-LEDGF was reduced in a direct concentration-dependent manner (Fig. 2A ). The higher the concentration of heparin, the less degradation of full-size GST-LEDGF there was. As a percentage of total protein in each lane determined by scanning densitometry, the amount of full-size GST-LEDGF increased from 32% (without heparin) to 43% (with heparin at 7.1 mg/ml) to 56% (with heparin at 71 mg/ml; lanes 2,3 and 4, respectively). We also ran GST on this gel, and it migrated to a position corresponding to a 29-kDa protein (data not shown). 

To confirm the protective effect of the lowest final concentration of heparin (7.1 mg/ml) and to extend this protective effect throughout the prokaryotic process, we added heparin at a final concentration of 7.1 mg/ml at each of the three stages of recombinant production (at the beginning, at extraction [sonication], and before elution). In Figure 2B , note the protection of GST-LEDGF against degradation and the enhanced accumulation of full-size GST-LEDGF when heparin (7.1 mg/ml) was present. Heparins of low molecular weight (i.e., 3000 or 6000 Da) and mixed molecular weight also protected against degradation (data not shown). 

Two other highly sulfated polysaccharides, heparan sulfate (up to 0.2μ g/ml) and dextran sulfate (up to 5 mg/ml), however, did not protect GST-LEDGF from degradation. The gel patterns from samples containing heparan sulfate or dextran sulfate did not differ from samples containing no additive (data not shown). 

IPTG stimulates the expression of a transfected gene by blocking the lacI repressor of E. coli. To ascertain whether heparin enhances the accumulation of GST-LEDGF without IPTG, IPTG was omitted from the E. coli culture. We wanted also to know whether the heparin enhancement was dose-dependent. Bacterial cultures without IPTG were grown until the OD_600nm of the culture was 0.6; then heparin was added at final concentrations of 0.71, 7.1, and 71 mg/ml. In a control culture IPTG (1 μg/ml) without heparin was used as the positive control. Heparin without IPTG enhanced accumulation of GST-LEDGF even at final concentrations as low as 0.71 mg/ml, and the enhancement with heparin did not appear to be dose-dependent in these dose ranges (Fig. 3) . These results strongly suggest that heparin, in a manner as yet undefined, significantly increased the yield of full-size GST-LEDGF in the prokaryotic expression system used. 

Effect of Heparin on the Ability of LEDGF to Enhance Cellular Growth

To assess heparin’s effect on GST-LEDGF-enhanced growth with LECs in culture, two experiments were performed. In the first, cell growth as a function of increasing amounts (1.0, 10, and 100 ng) of GST-LEDGF and the heparin-GST-LEDGF complex were compared. In the second, cell growth with varying concentrations of exogenous heparin alone or exogenous heparin with a constant amount (100 ng) of GST-LEDGF were compared. 

The results of the first experiment are illustrated in Figure 4 . At each of the three concentrations tested, heparin increased significantly (P < 0.001) the number of viable cells at the end of the 96-hour incubation. The effects were similar at all three concentrations of LEDGF. These results indicate that heparin potentiated the growth-enhancing effect of LEDGF. 

The second experiment was of interest because of the possibility that a beneficial effect of heparin could be independent of LEDGF or could result from its effect on endogenous LEDGF. One member of the pair had exogenous heparin alone; the other had a constant concentration of exogenous GST-LEDGF (100 ng/ml) and increasing concentrations of heparin (3.125, 6.25, 12.5, 25, and 50 μg/ml). The results are presented in Figure 5 . After a 96-hour culture, heparin alone approximately doubled the number of viable cells, and the increase was not dose-dependent. GST-LEDGF (without heparin) increased the number of viable cells approximately threefold in comparison with control. Heparin-GST-LEDGF increased the number of viable cells approximately sixfold (P < 0.001) with the potentiation plateauing at a concentration of 12.5 μg heparin/ml. 

Heparan sulfate and dextran sulfate in experiments similar to those described both failed to enhance cell growth. The rates of growth of cells exposed to GST-LEDGF with and without heparan sulfate or dextran sulfate were identical (data not shown). 

Effect of Heparin on the Stability of GST-LEDGF under Stressed Conditions

That heparin protected GST-LEDGF from proteolytic degradation suggested to us that heparin might also be able to protect GST-LEDGF against other stresses (heat, pH changes, and proteolysis by trypsin and chymotrypsin). For these stress experiments, 100 μg of GST-LEDGF were incubated with increasing concentrations of heparin (12.5, 25, and 50 μg/ml) for 1 hour at room temperature to form heparin-GST-LEDGF complexes. A sample without heparin served as control. The heparin-GST-LEDGF complex and the unbound heparin were separated on a Centricon 30 and washed twice with PBS buffer. The complexes were then exposed to a variety of stresses. After each stress incubation, samples of GST-LEDGF or heparin-GST-LEDGF were dialyzed against PBS to remove degraded products, acid or base; then 0.05% BSA was added to each sample to stabilize the LEDGF as described elsewhere. ⁷ Then each sample was assayed for growth-enhancing potency (with mouse LECs) and subjected to SDS-PAGE. Appropriate controls showed that BSA did not affect the growth of LECs. 

Accordingly, we first tested the effect of heparin on heat-inactivation of GST-LEDGF. Aliquots with or without heparin (12.5–50 μg/ml) were exposed to three temperatures: 37°C for 1 hour and 24 hours, 41°C for 24 hours, and 65°C for 5 minutes. Heparin, even at its lowest concentration (12.5 μg/ml), preserved the growth-enhancing potency of GST-LEDGF at 37°C, 41°C, and 65°C (Fig. 6) . Results at higher concentrations of heparin (up to 50 μg/ml) were similar. Without heparin, the potency of GST-LEDGF decreased at all three temperatures. 

Next we assessed the ability of heparin to protect against high or low pH. To lower the pH in an aliquot of the stock solution of GST-LEDGF, trifluoroacetic acid, diluted 1:10, was added until the final pH was 3.0. To increase the pH of the solution, 100 mM sodium bicarbonate was added until the pH was 9.0 (final concentration of sodium bicarbonate was 5.0 mM). All samples were incubated for 60 minutes. The control sample was without either preparation of GST-LEDGF. GST-LEDGF without heparin lost nearly 40% of its growth-enhancing potency at the high and low pHs. Again, even at the lowest concentration (12.5 μg/ml) heparin stabilized GST-LEDGF at all pHs tested (Fig. 7) . 

Next, we tested the ability of heparin to stabilize GST-LEDGF against proteolytic digestion by trypsin and chymotrypsin. Heparin at concentrations of 12.5, 25, and 50 μg/ml failed to protect GST-LEDGF from proteolytic degradation by trypsin or chymotrypsin (10 U/ml; data not presented), and higher concentrations of heparin were therefore tried (100, 200, 400, 800 μg/ml). 100 μg GST-LEDGF was used and the wt/wt ratios of GST-LEDGF/heparin were: 1:1, 1:2, 1:4, and 1:8 according to the method of Sommer and Rifkin. ¹⁹ The mixtures were incubated at room temperature for 1 hour, and the heparin-GST-LEDGF complexes were then purified on a Centricon 30, washed twice with PBS buffer, and incubated for 60 minutes with trypsin or chymotrypsin (final concentrations of 1 or 10 U/mg). Also SDS gel electrophoresis was performed. At a 1:4 wt/wt ratio of GST-LEDGF/heparin, GST-LEDGF was protected from trypsin or chymotrypsin proteolysis (Fig. 8) . 

Cellular Uptake and Intracellular Localization of Heparin and Heparin-GST-LEDGF

Preliminary studies indicated that LEDGF accumulated within the nucleus; this accumulation was markedly affected by temperature. ^{1 2} To study whether the transport of LEDGF into the nucleus was modified by heparin, we used LECs, immunostaining techniques, FITC-labeled heparin, and FITC-labeled heparin-LEDGF. By linking FITC to heparin and to GST-LEDGF we could monitor separately, and thereby distinguish, heparin-dependent from GST-LEDGF-dependent transport. In Figures 9 FITC-labeled heparin was easily visualized first in the cytoplasm and later around the nucleus (Fig. 9A) . FITC-heparin-GST-LEDGF also accumulated in the cytoplasm, but it was found later inside the nucleus (Fig. 9B) . These results indicate that heparin plays an important role in the nuclear uptake of exogenous GST-LEDGF. 

Further Studies on Intracellular Trafficking of GST-LEDGF and Heparin-GST-LEDGF

To confirm with another marker the nuclear localization of exogenous GST-LEDGF and the influence of heparin on this uptake, LECs were incubated without any form of LEDGF (Figs. 10A 10D ), with GST-LEDGF (Figs. 10B 10E) , or with heparin-GST-LEDGF (Figs. 10C 10F) . The LECs were then immunostained with the ABC kit with a polyclonal anti-GST antibody (Figs. 10A 10B 10C) or immune serum containing a polyclonal anti-(C-terminal) LEDGF antibody (Figs. 10D 10E 10F) . The anti-(C-terminal) LEDGF pAb could recognize exogenous GST-LEDGF, endogenous LEDGF, and two of their larger and one of their smaller degradation products, whereas the anti-GST pAb could visualize only exogenous GST-LEDGF and its degradation products. In Figure 10A , the absence of nuclear or cytoplasmic staining suggests that there was little or no GST in the nucleus or cytoplasm. In Figure 10B the faint staining of nuclei and cytoplasm indicates slight cytoplasmic and nuclear uptake of exogenous GST-LEDGF. In Figure 10C strong staining of the cytoplasm and nuclei of all cells is shown, indicating that the cell had taken up exogenous heparin-GST-LEDGF into the cytoplasm and transported it into the nucleus. In Figure 10D the light staining of cytoplasm and nuclei suggests that nuclear and cytoplasmic levels of endogenous LEDGF were low. In Figure 10E variably positive staining of nuclei and cytoplasm indicate increased levels of LEDGF (compared to that in 10B) confirming low-level cytoplasmic and nuclear uptake of exogenous GST-LEDGF. In Figure 10F , however, we see the effect of heparin; the cytoplasm and nuclei of all cells manifest strong immunostaining for GST-LEDGF. The increased staining with both antibodies suggests that the sources for this increased cytoplasmic and nuclear LEDGF are endogenous LEDGF and exogenous GST-LEDGF. These data strongly suggest that LEDGF is present in varying low levels in the cytoplasm and nuclei of mouse LECs. In the presence of heparin, both cytoplasmic and nuclear full-size LEDGF levels are increased. 

Discussion

We have shown that GST-LEDGF had a high affinity for heparin and that heparin protected LEDGF against proteolytic cleavage and heat and acid inactivation. Heparin protected LEDGF in both prokaryotic and eukaryotic systems. When added to a prokaryotic system, the yield of full-size LEDGF was increased. In eukaryotic systems, heparin increased the functional activity of LEDGF. We do not yet know the mechanism of this functional effect, but our results suggest that possible mechanisms include stabilization of LEDGF and/or modification of its trafficking within the cell. Although we are certain that the heparin-GST-LEDGF complex is transported into the nucleus, we do not yet know whether or how this transport is linked to the change in cellular functions. 

What started as a search for a practical means to increase the yield of a new growth factor, LEDGF, evolved into a study of a physiologically significant interaction between LEDGF and heparin. Heparin is a polymer made up of repeating units of sulfated and nonsulfated d-glucosamine, l-iduronic acid, and d-glucuronic acid. Heparin’s molecular weights range between 5,000 and 30,000 Da. The many O- and N-sulfate linkages and carboxyl groups make it the strongest organic acid in the body. In heparan sulfate the sulfated regions are sparsely clustered along the molecule. ²⁰ In heparin, however, more than 80% of the molecule is sulfated, and the charge density is much higher. 

Our ability to purify GST-LEDGF with heparin-Sepharose affinity chromatography illustrated GST-LEDGF’s high affinity for heparin. LEDGF contains many potential heparin-binding sites. Arginine, for example, has the highest affinity for heparin; it is 2.5 times greater than that of lysine. ⁸ There are 27 arginine residues and 84 lysine residues among the 530 amino acids of LEDGF, and undoubtedly some of these bind strongly to heparin (e.g., RRGRKRK; 146–152). In contrast, histidine was found to be unimportant in GAG’s binding of protein. ²¹ That a 0.4 to 0.6 M NaCl gradient was able to disrupt the heparin-GST-LEDGF attraction suggests that the bond was electrostatic, a suggestion consistent with arginine and lysine being key binding sites. However, the nature of this interaction is complex. Coulombic forces between basic amino acids and anionic groups on the polysaccharides are of major importance. Indeed, coulombic forces appear to dominate the interaction of sulfated polysaccharides with proteins. 

GAG–protein interaction regulates hemostasis, cell adhesion, lipid metabolism, and growth factor signal transduction. ¹³ Although defining the mechanism of the heparin-GST-LEDGF effect on LECs was not our objective, our findings are consistent with published reports of heparin’s impact on other tissues. Heparin modulates the function of other growth factors, ²² increases the affinity and interaction of growth factors to their receptors, ^{23 24} confers preferential binding to extracellular matrix proteins, ²⁵ and may be taken up directly by cells and carried to perinuclear ²⁴ or intranuclear locations. ²⁵ Heparin’s potentiation of LEDGF’s effect as a growth factor may be due in part to the accelerated transport of heparin-GST-LEDGF into the nucleus. Fibroblast growth factor (FGF) also binds heparin, forming a complex that is transported into the cytoplasm and nucleus, where heparin in the complex is digested by nuclear heparinase. ^{26 27} Heparin and heparan sulfate proteoglycans (HSPGs) also serve as low-affinity binding sites for growth factors, ²⁸ and newly synthesized growth factors may be released from cells bound to soluble GAGs. ^{13 25}  

Other glycosaminoglycans have functions similar to those we have found for heparin with GST-LEDGF. Heparan sulfate protects bFGF from proteolytic degradation. ²⁹ HSPGs are involved in the internalization and degradation of lipoprotein lipase in endothelial cells and in avian adipocytes. ³⁰ They may also bind to the cell surface and regulate the interaction of activated growth factor receptors with their intracellular mitogenic signaling pathways. ³¹ One of the most intriguing concepts expressed by Jackson et al. ¹³ in 1991 was that GAGs may regulate gene expression in the nucleus by binding to transcription factors and modifying specific gene promoter regions. Our findings and those of Ge et al. ^{4 5} support the concept that p75 and LEDGF are identical proteins, and along with a third highly homologous protein p52, are transcriptional coactivators and also bind to pre-mRNA splicing factors. These findings suggest that LEDGF and heparin may play important roles in the growth, differentiation, and perhaps the death of LECs. Further experiments in these areas are under way. 

Our study of GST-LEDGF’s interaction with sulfated polysaccharides was limited to heparin, heparan, and dextran sulfate. The latter two did not protect LEDGF from proteolytic degradation and did not have any effect on the growth-enhancing effect of LEDGF. Although the heparin effect is intriguing, heparin may not be the GAG that interacts with LEDGF in the ocular lens. There are no published measurements of heparin in the aqueous humor, and the lens, being avascular, would not have the heparin found in blood vessels elsewhere in the body. ^{28 29 32 33 34} There is, however, a rich literature on glycosaminoglycans (GAGs) and the lens, and one of these heparin-like GAGs may bind with LEDGF endogenously. GAGs and glycopeptides are synthesized by lens epithelium. ³⁵ Chondroitin-4- and 6-sulfates, dermatan sulfate, and heparan sulfate are induced in lens epithelium by retinal growth factor, and these GAGs are present in the extra-, peri-, and intracellular compartments of the lens. ^{36 37 38 39 40} A wide variety of GAGs are secreted by cultured rabbit lens epithelium. ^{41 42} It remains to be determined which of these endogenous GAGs in the lens may bind, transport, and potentiate LEDGF in the lens. 

In our study heparin, but not heparan sulfate, potentiated GST-LEDGF growth stimulation. There is precedent for variability in cellular responses to individual GAGs. Heparin and heparan sulfate may either stimulate or inhibit cell growth depending on the cell type. ^{43 44} With primary cultures of rat intestinal epithelial cells, heparin stimulated proliferation very significantly, whereas heparan sulfate did not. ⁴⁵  

All the proteins in the HDGF family of proteins have been purified by heparin-Sepharose chromatography. This family includes hepatoma-derived growth factor (HDGF), ⁴⁶ LEDGF, ^{1 2 3} hepatoma-related proteins 1 and 2 (HRP-1, HRP-2), ⁴⁷ and p52. ⁵ The protein p52 is an alternative splicing product of the LEDGF gene. ³ Its N-terminal 325 amino acids are identical with LEDGF; only the eight residues at the C terminus are different. All these proteins share a common N-terminal region, and we can speculate that it is this so-called HATH region that binds to heparin. Many growth factors are highly charged (e.g., fibroblast, epidermal, vascular endothelial, and platelet-derived growth factors), and they bind to and are stabilized by highly sulfated, negatively charged glycosaminoglycans, such as heparin. ^{22 29 48} The binding between LEDGF and heparin may occur in the extracellular matrix. The results in Figure 10 suggest that heparin binds to full-size LEDGF and facilitates transport of this complex through the cytoplasm into the nucleus. Heparin or a heparin-like GAG may facilitate this transportation. The role of GAGs in the various functions of LEDGF in the lens and other ocular tissues will be the focus of our continuing research. 

 Supported by National Institutes of Health Grants EY-12015, EY-10958, and EY-10824; Shojin Research Associates; and The Massachusetts Lions Eye Research Fund.

 Submitted for publication May 14, 1999; revised August 17 and October 4, 1999, January 21 and April 8, 2000; and accepted April 13, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Lanes 1 through 4 are from a Coomassie blue–stained SDS-PAGE gel. Lanes 5 and 6 are immunostained nitrocellulose blots of the SDS-PAGE gel. Lane 1: molecular weight markers in kilodaltons; lane 2: supernatant from the mixture of purified GST-LEDGF (eluted from glutathione-Sepharose beads) and heparin-Sepharose beads. Lane 3: GST-LEDGF forms released from heparin-Sepharose. The uppermost band is the 89-kDa GST-LEDGF fusion protein (MW of GST, 29 kDa and LEDGF, 60 kDa), and the others are degraded GST-LEDGF. Lane 4: full-size and degraded forms of purified GST-LEDGF; lane 5: similar to lane 4 but blotted on a nitrocellulose filter and immunostained with an anti-GST pAb (diluted 1:1000). Multiple bands with sizes ranging from 89 to 25 kDa were found. The lower molecular weight forms are degraded forms of GST-LEDGF. In lane 6 the gel is stained with a pAb to a small peptide (residues 419–437) in the C terminus, and only full-size GST-LEDGF, two of the larger degradation products of LEDGF, and a small peptide fragment (presumably containing the epitope) are present.

Figure 2.

 

(A) Heparin protected recombinant GST-LEDGF from degradation. Heparin was added in LB-ampicillin medium at final concentrations of 7.1 and 71 mg/ml. GST-LEDGF in the Coomassie blue–stained SDS-PAGE gel showed less degradation when heparin was added at the beginning of the bacterial culture. Lane 1: molecular weight markers in kilodaltons; lane 2: GST-LEDGF without heparin; lanes 3 and 4: heparin present at final concentrations of 7.1 and 71 mg/ml, respectively. Densitometric scanning of dried gel revealed that the amount of full-size GST-LEDGF (arrow) as a percentage of total protein in the lane increased from 32% (lane 2, without heparin) to 43% (lane 3, heparin 7.1 mg/ml) to 56% (lane 4, heparin 71 mg/ml). (B) Samples from an experiment identical with that described in (A) except that heparin (7.1 mg/ml) was added three times (at each stage of the prokaryotic production process), and samples were taken at different stages in the culture (when the OD_600nm reached 0.20 (lanes 2 and 3), 0.40 (lanes 4 and 5), 0.60 (lanes 6 and 7), and 0.80 (lanes 8 and 9). Lane 1: molecular weight markers (as in A). Lanes 2, 4, 6, and 8: GST-LEDGF without heparin; lanes 3, 5, 7, and 9: GST-LEDGF from system in which heparin was added at each of three stages in the synthetic process. Proteolytic degradation of GST-LEDGF decreased and accumulation of full-size ST-LEDGF increased when heparin was present.

Figure 3.

 

Effect of heparin on prokaryotic production and accumulation of GST-LEDGF. Equal volumes of washed glutathione-Sepharose 4B beads were boiled in sample buffer, and samples of the supernatant were loaded on an SDS-PAGE gel. The y-axis represents the area under the curve of a densitometric scan of the full-size GST-LEDGF band in a dried SDS gel. The first bar is from the culture without IPTG; the second is the positive control (only IPTG was added). The third, fourth, and fifth bars are from cultures to which heparin, but not IPTG was added. Heparin at all concentrations without IPTG increased the amount of full-size GST-LEDGF protein. Error bars, SDs. Replicates, three.

Figure 4.

 

Relative potency of GST-LEDGF and heparin-GST-LEDGF as enhancers of mouse LEC growth. Five thousand cells were added to each well of a 96-well culture plate. Each pair of wells comprised one with GST-LEDGF and the other with the heparin-GST-LEDGF complex. In each pair the amount of GST-LEDGF was the same. The control cells had no GST-LEDGF or heparin-GST-LEDGF. MTS assays were performed after 96 hours of incubation. At each concentration, the growth-enhancing effect was higher in the GST-LEDGF than control (P < 0.001) and even higher in the heparin-containing GST-LEDGF preparation (P < 0.001). Error bars, SDs. Replicates, three.

Figure 5.

 

Effect of varying concentrations of heparin with a constant concentration of heparin-GST-LEDGF on mouse LEC growth. Five thousand cells in DMEM (without serum) were added to each well of a 96-well culture plate, and the plates were incubated in a 6% CO₂ atmosphere for 96 hours at 37°C. LECs without serum, heparin, or heparin-GST-LEDGF served as controls. In each pair of wells, one contained heparin alone, and the other contained 100 ng/ml heparin-GST-LEDGF. MTS assays were performed after 96 hours of incubation; the y-axis is the OD_490nm. Heparin alone doubled the number of viable cells at the end of the 96-hour incubation, but not in a dose-dependent manner. GST-LEDGF approximately tripled the number of viable cells. Heparin-GST-LEDGF increased the number of viable cells sixfold (P < 0.001), and the greatest potentiation was with 12.5–50 μg /ml heparin. Error bars, SDs. Replicates, three.

Figure 6.

 

Protection of GST-LEDGF by heparin. y-Axis: OD_{490 nm} in the MTS assay. Heparin at its lowest concentration (12.5μ g/ml) protected GST-LEDGF from heat inactivation at all three temperatures. At higher concentrations of heparin (up to 50 μg/ml), the protective effect was similar. Error bars, SDs.

Figure 7.

 

Heparin preserved the growth-enhancing potency of GST-LEDGF in acidic and basic conditions. y-Axis: OD_490nm in the MTS assay. The control samples had no GST-LEDGF or heparin-GST-LEDGF. Error bars, SDs. Replicates, three.

Figure 8.

 

Heparin protected GST-LEDGF against trypsin (T) and chymotrypsin (CT) digestion. T or CT was added to the GST-LEDGF or the heparin-GST-LEDGF samples: lanes 3 and 4 (1 U T/mg), lanes 5 and 6 (10 U T/mg), lanes 7 and 8 (1 U CT/mg), and lanes 9 and 10 (10 U CT/mg). The mixtures were incubated for an additional hour at 37°C. Lane 1: molecular size markers; lane 2: GST-LEDGF without proteinase; lanes 4, 6, 8, and 10: heparin-GST-LEDGF treated with T (lanes 4 and 6) or CT (lanes 8 and 10); lanes 3, 5, 7 and 9: heparin-free GST-LEDGF treated with T (lanes 3 and 5) or CT (lanes 7 and 9). Marked reduction in GST-LEDGF degradation occurred in all samples in which heparin was present.

Figure 9.

 

Localization of FITC-heparin (A) and FITC-heparin-GST-LEDGF (B) in mouse LECs. FITC-heparin or the FITC-heparin-GST-LEDGF complex was added to LECs grown on coverslips. FITC-heparin entered the cells after 4 to 5 hours but remained in the cytoplasm and around the nucleus for 1 to 3 days (A). The FITC-heparin-GST-LEDGF complex, although first localized in the cytoplasm, was found in the nucleus after 1 day (B). The conditions for fluorescent photography in (A) and (B) were identical. Magnification, ×400.

Figure 10.

 

(A) Mouse LECs cultured in the absence of any form of LEDGF and immunostained with a polyclonal anti-GST antibody (diluted 1:500). Absence of staining was noted in both nucleus and cytoplasm. (B) Mouse LECs cultured with GST-LEDGF and immunostained with polyclonal anti-GST antibody (diluted 1:500) showed weak staining in the nucleus and cytoplasm in almost all cells. The positive staining reflects the location of exogenous GST-LEDGF. (C) Mouse LECs cultured with heparin-GST-LEDGF and stained with polyclonal anti-GST (diluted 1:500) showed strong staining of nuclei and cytoplasm in almost all cells, indicating the presence of high nuclear and cytoplasmic levels of exogenous GST-LEDGF. (D) Mouse LECs incubated in the absence of any form of LEDGF and immunostained with polyclonal anti-(C-terminal) LEDGF antibody (diluted 1:200) showed moderate intensity of nuclear and cytoplasmic staining in all cells, indicating the presence of endogenous LEDGF. (E) Mouse LECs incubated with GST-LEDGF and immunostained with polyclonal anti-(C-terminal) LEDGF antibody (diluted 1:200) showed variable intensity of staining of both nucleus and cytoplasm in most cells. The positive staining reflects some cytoplasmic and nuclear uptake of exogenous GST-LEDGF and the location of endogenous LEDGF. (F) Mouse LECs incubated with heparin-GST-LEDGF and stained with polyclonal anti-(C-terminal) LEDGF antibody (diluted 1:200) showed intense staining of cytoplasm and nucleus, indicating marked heparin enhancement of nuclear and cytoplasmic uptake of LEDGF. Bars, (A, B, and D) 30 μm; (C, E, and F) 20 μm.