Human Heparanase

Heparan sulfate and heparan sulfate proteoglycans are present in the extracellular matrix as well as on the external cell surface. They bind various molecules such as growth factors and cytokines and modulate the biological functions of binding proteins. Heparan sulfate proteoglycans are also important structural components of the basement membrane. Heparanase is an endo-β-d-glucuronidase capable of cleaving heparan sulfate and has been implicated in inflammation and tumor angiogenesis and metastasis. In this study, we report the purification of a human heparanase from an SV40-transformed embryonic fibroblast cell line WI38/VA13 by four sequential column chromatographies. The activity was measured by high speed gel permeation chromatography of the degradation products of fluorescein isothiocyanate-labeled heparan sulfate. The enzyme was purified to homogeneity, yielding a peptide with an apparent molecular mass of 50 kDa when analyzed by SDS-polyacrylamide gel electrophoresis. Using the amino acid sequences of the N-terminal and internal heparanase peptides, a cDNA coding for human heparanase was cloned. NIH3T3 and COS-7 cells stably transfected with pBK-CMV expression vectors containing the heparanase cDNA showed high heparanase activities. The homology search revealed that no homologous protein had been reported.

Heparan sulfate and heparan sulfate proteoglycans are present in the extracellular matrix as well as on the external cell surface. They bind various molecules such as growth factors and cytokines and modulate the biological functions of binding proteins. Heparan sulfate proteoglycans are also important structural components of the basement membrane. Heparanase is an endo-␤-D-glucuronidase capable of cleaving heparan sulfate and has been implicated in inflammation and tumor angiogenesis and metastasis. In this study, we report the purification of a human heparanase from an SV40-transformed embryonic fibroblast cell line WI38/VA13 by four sequential column chromatographies. The activity was measured by high speed gel permeation chromatography of the degradation products of fluorescein isothiocyanate-labeled heparan sulfate. The enzyme was purified to homogeneity, yielding a peptide with an apparent molecular mass of 50 kDa when analyzed by SDS-polyacrylamide gel electrophoresis. Using the amino acid sequences of the N-terminal and internal heparanase peptides, a cDNA coding for human heparanase was cloned. NIH3T3 and COS-7 cells stably transfected with pBK-CMV expression vectors containing the heparanase cDNA showed high heparanase activities. The homology search revealed that no homologous protein had been reported.
Heparanase, which cleaves HS into characteristic large molecular weight fragments, was identified in murine metastatic melanoma cells by Nakajima et al. (14). Heparanase activity was correlated with the lung colonization potential of murine B16 melanoma sublines (14). Nakajima et al. (15) concluded that the enzyme responsible for HS degrading is an endoglucuronidase, cleaving the linkage between GlcUA and GlcNAc, and named it heparanase. Heparanase is a hydrolase distinct from flavobacterium heparitinase and heparinase, which are eliminases capable of specifically degrading heparan sulfate and heparin, respectively, into di-and tetrasaccharides (15)(16)(17).
Several groups have reported the purification of heparanase from different sources. Jin et al. (21) reported the purification of heparanase from murine melanoma cells. The molecular size of the purified heparanase was 97 kDa, with minor proteins of 205, 125, 76, and 52 kDa also detected (21). Freeman and Parish (22) have recently purified a human platelet heparanase by sequential chromatographies using ConA-Sepharose, Zn 2ϩchelating Sepharose/blue A-agarose, octylagarose, and Superose 12 columns, resulting in a single protein of 50 kDa. Ihrcke et al. (23) also purified a platelet heparanase with a molecular mass of 34 kDa. Gilat et al. (24) purified a 45-kDa heparanase from human placenta. The partial purification of heparanase from Chinese hamster ovary cells was also reported (25). The Chinese hamster ovary heparanase was reported as having a molecular mass greater than 30 kDa.
Hoogewerf et al. (26) purified a heparanase-like enzyme from human platelets. Their purified protein was identified as connective tissue activating peptide III, having a glucosaminidase activity but not a glucuronidase activity and cleaving HS into disaccharides. However, the relationships among HS-degrading endoglycosidases have not yet been elucidated due to a lack of sufficient information regarding their structures and enzymatic characteristics.
In the present study, we purified human heparanase to homogeneity by sequential column chromatographies and cloned a cDNA encoding human heparanase. We report here the enzymatic characteristics of the purified native heparanase, the * 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.
‡ To whom correspondence should be addressed. cloning of the cDNA coding for the enzyme, and the expression of the enzyme in NIH3T3 and COS-7 cells.

Labeling of Heparan Sulfate
Heparan sulfate (sodium salt) from bovine kidney was labeled with fluorescein isothiocyanate (FITC). Five milligrams of heparan sulfate and 5 mg of FITC were dissolved in 1 ml of 0.1 M sodium carbonate buffer, pH 9.5, and stirred at 4°C for overnight in darkness. The solution was then loaded on a PD-10 desalting column to isolate FITClabeled heparan sulfate (FITC-HS) from unreacted FITC. The FITC-HS was further subjected to chromatography through Sephacryl S-300 HR equilibrated with column buffer 1 (25 mM Tris-HCl, 150 mM NaCl, pH 7.5) to fractionate high molecular weight products (M r Ͼ30,000). The fractionated materials were concentrated with a Microcon® 10 concentrator (Amicon). The quantity of FITC-HS was measured using the Blyscan proteoglycan and glycosaminoglycan assay (Biocolor Ltd., Belfast, Ireland).

Heparanase Assay
Human heparanase activity was determined toward FITC-HS. The reaction was carried out in 100 l of 50 mM sodium acetate, pH 4.2, containing 0.5-1 g of FITC-HS. The protein material containing heparanase was added to the reaction mixture without exceeding the salt concentration of 0.25 M and incubated at 37°C for an appropriate period. The reaction was then stopped by the addition of 100 g of heparin and subsequent heating at 100°C for 5 min. The reaction mixture was centrifuged at 15,000 rpm for 5 min to precipitate the insoluble material. The products of FITC-HS yielded by this reaction were analyzed by high speed gel permeation chromatography. Briefly, a 20-l aliquot of the supernatant was injected into a TSKgel G3000SWXL column (7.8-mm inner diameter ϫ 30 cm) equilibrated with column buffer 1 and run at 0.5 ml/min. The fluorescence in the eluent was measured by a F-1050 fluorescence spectrophotometer (Hitachi, Tokyo, Japan). The activity was determined from HPLC chromatograms by measuring a forward half area of the peak of the intact FITC-HS. The decrease of this area following heparanase treatment was measured by using an integrator, and the amount of the degraded FITC-HS was calculated from the decrease of fluorescence intensity (FI). The degradation of FITC-HS was 0 -30% of the substrate added to obtain kinetic parameters using this calculation method.

Purification of Human Heparanase
All of the procedures were performed in a total of 5 days at 4°C Cell Extraction-Cell lysate was prepared by homogenizing cells in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5% Triton X-100, 0.2 mM AEBSF, 10 g/ml leupeptin, 10 g/ml pepstatin A, and 1 g/ml aprotinin). The lysis was carried out at 4°C for 30 min and centrifuged at 9000 rpm. The supernatant was then subjected to heparin-Sepharose affinity chromatography. The concentration of protein was measured by BCA protein assay (Pierce).
Heparin-Sepharose Affinity Chromatography-The cell lysate obtained as above was loaded on to a heparin-Sepharose column (5.0 ϫ 10 cm) at a flow rate of 2.5 ml/min. The column was washed with 10ϫ gel volume of the column buffer and then with washing buffer containing 0.5 M NaCl in column buffer. Heparanase was eluted in column buffer containing 1.0 M NaCl. The eluate was then diluted with an equal volume of dilution buffer 1 (50 mM sodium acetate, pH 6.0, 0.4% CHAPS). The concentration of protein was measured by BCA protein assay (Pierce).
ConA-agarose Affinity Chromatography-The heparin-Sepharose eluted fractions were diluted and applied to a ConA-agarose column (2.5 ϫ 10 cm) at a flow rate of 1 ml/min. The column was washed with ConA buffer (50 mM sodium acetate, pH 6.0, 0.5 M NaCl, 0.2% CHAPS), and heparanase was eluted with ConA buffer containing 0.7 M ␣-methylmannose. The eluate was concentrated in a Centriprep® 10 concentrator (Amicon) and diluted with 5 volumes of dilution buffer 2 (50 mM sodium acetate, pH 6.0, 0.2% CHAPS) to adjust the NaCl concentration to 0.1 M. The concentration of protein was measured by the Dye assay (Bio-Rad).
Carboxymethyl-Sepharose Cation Exchange Chromatography-The ConA-agarose eluted fractions were diluted and applied to a carboxymethyl-Sepharose column (2.5 ϫ 10 cm) at a flow rate of 1 ml/min. The column was washed with CM buffer A (50 mM sodium acetate, pH 6.5, 0.1 M NaCl, 0.2% CHAPS) and eluted with a step gradient of NaCl from 0.1 to 0.5 M. Fractions containing heparanase were collected, concentrated in a Centriprep® 10 concentrator (Amicon), and mixed with an equal volume of 5 M NaCl solution. The concentration of protein was measured by Micro BCA protein assay (Pierce).
Phenyl-Sepharose Hydrophobic Exchange Chromatography-The purified fractions from carboxymethyl-Sepharose were applied to a phenyl-Sepharose column (2.5 ϫ 10 cm) equilibrated with phenyl buffer A (50 mM Tris-HCl, 2.5 M NaCl, pH 7.3) and washed at a flow rate of 0.5 ml/min from 0 to 40 min. The heparanase was eluted with a gradient of phenyl buffer B (50 mM Tris-HCl, 0.1% CHAPS, pH 7.3) from 0 to 100% over 100 min. The heparanase-containing fractions were collected and concentrated with a Centricon® 10 concentrator (Amicon). The protein quantity was estimated by comparing the staining of various known amounts of bovine serum albumin in a silver-stained SDS-polyacrylamide gel after running according to the method of Laemmli (27). glutaric acid, 50 mM Tris, 50 mM 2-amino-2-methyl-1,3-propanediol, pH adjusted with HCl and NaOH) at various pH values and incubated for 2 h at 37°C. The heparanase activity was calculated as described for the heparanase assay method.

Amino Acid Sequence of Heparanase Protein and Derived Peptides
The purified heparanase from phenyl-Sepharose chromatography was subjected to SDS-polyacrylamide gel electrophoresis (27) and then electrotransferred onto a ProBlott membrane (Applied Biosystems, Inc). The transferred protein was visualized with Coomassie Brilliant Blue. The 50-kDa band, which had heparanase activity, was excised and used for N-terminal amino acid sequencing with the HP G1005A protein sequencing system (Hewlett Packard). The transferred protein band prepared as above was digested in situ with Lys-C endoproteinase. The digests were subjected to reverse-phase chromatography and eluted with a gradient of 0 -100% acetonitrile. The amino acid sequences of the separated peptides were analyzed using the HP G1005A protein sequencing system.

Identification of the Human Heparanase cDNA Using Expressed Sequence Tag (EST) Sequences
The peptide sequences identified from the purified human heparanase were used to search for homologues in the gene data bases stored at the NCBI by using the tBLASTn sequence alignment program. Translation of EST sequences (yw97a02.r1 Homo sapiens cDNA clone 260138 5Ј, GenBank TM accession no. N45367 and yw70a03.s1 Homo sapiens cDNA clone 257548 3Ј, GenBank TM accession no. N30824) 2 in one frame presented a significant sequence similarity with three internal peptide sequences of the human heparanase. These sequences were originally obtained from a human placenta 8 -9-week cDNA library primed with oligo(dT) primer. Aligning overlapping regions, the sequences of these EST clones were connected to make one DNA fragment 968 bp in length. A 731-bp DNA fragment corresponding to the connected DNA of these EST clones was then amplified by reverse transcriptase-PCR using WI38/VA13 cDNA. For this purpose, a total RNA was prepared from the cultured WI38/VA13 cells using ISOGEN (Nippon Gene, Corp. Japan), and poly(A) RNAs were purified with an mRNA purification kit (Amersham Pharmacia Biotech). poly(A) RNA was reverse-transcribed with the avian myeloblastosis virus reverse transcriptase (Takara, Japan) according to the manufacturer's protocol and used for PCR with a sense oligonucleotide primer (5Ј-CTTCTAA-GAAAGTCCACCTTC-3Ј) and an antisense oligonucleotide primer (5Ј-AAACTATATGAGAAAGCTGGC-3Ј). PCR was performed using the LA PCR TM kit (Takara). PCR conditions were as follows: 98°C for 20 s and 68°C for 5 min. The DNA fragment was then subcloned into pCR TM 2.1 cloning vector using the TA Cloning® kit (Invitrogen). The clone was sequenced to confirm its length and identity with the EST sequence by using an automatic sequencer (Applied Biosystems model 377).

Construction of ZAP Express TM Library and Cloning of Human
Heparanase cDNA from WI38/VA13 Cells Synthesis of the cDNA and ligation into EcoRI-and XhoI-digested ZAP Express TM were carried out using the ZAP-cDNA synthesis kit (Stratagene). Poly(A) RNA was prepared from WI38/VA13 cells as described above, and oligo(dT) primer was used for the reverse transcriptase reaction. The ligated DNA was packaged in vitro using the ZAP-cDNA Gigapack III Gold cloning kit (Stratagene). The library was screened by 32 P-labeled DNA fragments containing a connected EST fragment that was expected to code a heparanase cDNA, and the obtained cDNA clones were sequenced using an automatic sequencer (Applied Biosystems model 377).

Expression of Human Heparanase in NIH3T3 and COS-7 Cells
The 3726-bp-long cDNA coding for a human heparanase was inserted into an expression vector, pBK-CMV (Stratagene), directly used for transfection of NIH3T3 and COS-7 cells by LipofectAmine TM reagent (Life Technologies, Inc.). Cells (1 ϫ 10 5 ) were transfected with 1 g of plasmid DNA in the presence of 10 l of LipofectAmine TM reagent in a well of six-well tissue culture plate containing OPTI-MEM® reduced serum medium (Life Technologies, Inc.). After a 5-h incubation, the medium was replaced with fresh RPMI medium supplemented with 10% FBS, 2 mM L-glutamine, penicillin (100 units/ml), and streptomycin (100 g/ml). Cells were further cultured for 72 h and then collected, washed by phosphate-buffered saline, and lysed in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5% Triton X-100, 0.2 mM AEBSF, 10 g/ml leupeptin, 10 g/ml pepstatin A, and 1 g/ml aprotinin). A 20-g aliquot of the cell lysate was used for heparanase activity analysis as described above.

Development of an Assay for Analysis of Heparanase
Activity-We first developed a highly sensitive, quantitative, and reliable assay system to detect heparanase-mediated degradation of heparan sulfate by utilizing a fluorescence-labeled substrate. First, heparan sulfate from bovine kidney was labeled with FITC as described under "Experimental Procedures." The FITC reacted with the free amino residues of glucosamines present in heparan sulfate. The reaction was carried out at pH 9.5, since exceeding this pH would result in alkaline hydrolysis of heparan sulfate. The amount of FITC linked to heparan sulfate was calculated from its absorbance at 494 nm at pH 8.0 using the absorption coefficient (⑀) of 72,000 cm Ϫ1 M Ϫ1 . The amount of heparan sulfate was measured by a Blyscan proteoglycan and glycosaminoglycan assay with a standard curve drawn using unlabeled heparan sulfate. When the molar ratio of FITC to HS was greater than 1.0, the FITC-HS became less susceptible to heparanase and even worked as an inhibitor of enzyme activity (data not shown). The FITC-HS was further fractionated by molecular size to obtain a sharp single peak when analyzed by high speed gel permeation chromatography. The substrate prepared as described above was homogeneous at least in molecular weight.
The reaction of heparanase was carried out at pH 4 -5 and

TABLE I
Purification of human heparanase from WI38/VA13 cells Enzyme activity against FITC-HS was measured as described under "Experimental Procedures." Each sample was assayed at points where the incubation time and the amount of degraded FITC-HS showed a linear correlation. Heparanase activity yielded in heparin-Sepharose affinity chromatography fractions was the highest. Therefore, we assumed that the total activity in heparin-Sepharose affinity fractions represented the total heparanase activity in the whole cell lysate. Thus, the purification level was also calculated as a ratio relative to the total activity in heparin-Sepharose fractions (shown in parentheses). stopped by adding an excess of heparin followed by boiling of the reaction mixture. The heparin was added before boiling to prevent FITC-HS from precipitating with the protein. The degraded products were analyzed by HPLC equipped with an automated injector. A 20-l aliquot of the reaction mixture was injected into a G3000 SWXL column to analyze the products by molecular size. Fig. 1a shows a chromatographic profile of the intact FITC-HS with a sharp peak and retention time of 15.99 min. When the products of FITC-HS incubated with the cell extract prepared from SV40-transformed human embryonic fibroblast WI38/VA13 cells were analyzed, a very broad peak with decreased FI appeared in the lower molecular weight range as compared with the intact FITC-HS (Fig. 1b). From the retention time, the average molecular mass of the degradation products was determined to be 20 -30 kDa. The FITC-HS was also tested for degradation following a treatment with 1 milliunit of heparitinase of bacterial origin (Fig. 1c). The peaks at retention times of 29.98 and 31.94 min represent tetra-and disaccharides produced by heparitinase treatment, respectively. Purification of Human Heparanase-A human embryonic lung fibroblast cell line WI38 did not show any heparanase activity toward FITC-HS when analyzed by HPLC (data not shown), while the SV40-transformed clone WI38/VA13 showed very high heparanase activity. Therefore, WI38/VA13 cells were used as a source of the human heparanase in this purification. The cell extract containing 880 mg of protein was first subjected to heparin-Sepharose affinity column chromatography at pH 7.5. The amount of protein eluted from this column was 10.2 mg, which was approximately 1% of the original cell extract (Table I). Then ConA-agarose affinity chromatography was employed. This procedure again resulted in significant enrichment of the heparanase in the eluate, but the recovery of the enzyme activity was less than 50%. The third step was carboxymethyl-Sepharose cation exchange chromatography, which did not result in any reduction in the activity recovered. The final step was performed utilizing hydrophobic interaction chromatography. Among various hydrophobic interaction matrices tested, phenyl-Sepharose showed the highest binding to heparanase. Fig. 2A shows the chromatographic profile of phenyl-Sepharose chromatography and the catalytic activity toward FITC-labeled heparan sulfate. The heparanase was detected in fractions 62-78. Although the recovery of the enzyme activity was less than 5%, the degree of purification was ele- vated to 740-fold. Fractions 51-58, 59 -66, 67-74, 75-82, and 83-90 were combined, respectively, and concentrated. The samples were subjected to SDS-polyacrylamide gel electrophoresis followed by silver staining (Fig. 2B). Only a single band of approximately 50 kDa was detected in fractions 67-74 and was at a high level compared with fractions 59 -66 or 75-82, which coincides with the level of activity toward FITC-HS analyzed by HPLC. In summary, approximately 500 ng of the human heparanase with a molecular mass of 50 kDa was purified from 880 mg of protein from the WI38/VA13 cell extract by four sequential chromatographic procedures with 740-fold purification and 3.6% recovery.
Characterization of the Catalytic Activity of Human Heparanase-Heparanase showed the highest activity toward FITC-HS at pH 4.2 (Fig. 3). No significant activity was observed below pH 3.5 and above pH 7.0. The same results were obtained when other buffers such as those containing sodium citrate and sodium acetate were used (data not shown).
Amino Acid Sequences of Human Heparanase-Human heparanase purified from WI38/VA13 cell extracts was immobilized on a polyvinylidene difluoride membrane after SDSpolyacrylamide gel electrophoresis. The N-terminal amino acid sequence of the 50-kDa protein is shown in Table II. We then treated the immobilized protein with Lys-C endoproteinase in situ. The digests were then separated on a reverse-phase column by HPLC. The N-terminal amino acid sequences were further determined from three independent peptides (Table II).
Molecular Cloning of Human Heparanase-Protein sequences of human heparanase were used to search for homologous sequences in gene data bases by using the tBLASTn sequence alignment program (28). As a result, we identified two overlapping EST clones of 553 and 587 bp that were originally obtained from a human placenta library. The deduced amino acid sequence of ESTs had high homology with the three internal amino acid sequences of human heparanase. We postulated that these ESTs were partial copies of an mRNA coding for the human heparanase. Two primers were then designed from the EST clones and used in PCR amplification to screen for the presence of human heparanase in SV40-transformed WI38/VA13 human lung fibroblast cells. The cDNA from WI38/ VA13 cells produced an expected 731-bp PCR fragment. This fragment was cloned and sequenced, confirming its identity with two EST sequences. This PCR fragment was then used as a probe to screen full-length cDNA from the WI38/VA13 cDNA library. We obtained five positive clones from 1 ϫ 10 6 plaques. An independent clone was isolated and subcloned. The nucleotide sequence of the clone contained a 3726-bp insert (Fig. 4).
The cDNA encoded the entire open reading frame of human heparanase of 543 amino acids as shown in Fig. 4. It had six potential N-linked glycosylation sites. It also contained the N-terminal amino acid sequence of human heparanase obtained previously starting from Lys 158 . The hydropathy plot using the Kyte and Doolittle program suggested that human heparanase has no hydrophobic domain (data not shown). Two in-frame ATG codons were found in the amino-terminal sequence of the human heparanase cDNA.
The 5Ј-noncoding region of the clone was found to be GC-rich, which, although it does not match the consensus eukaryotic translation initiation sequence, is a typical feature of housekeeping gene promoters (29). The 3Ј-noncoding region of the clone was found to contain a L1.t1.L1 repetitive element when we searched for related sequences in gene data bases as described previously.
Expression of Heparanase and Characterization-To test whether the human heparanase cDNAs really encode a catalytically active heparanase, the clone containing respective full-length cDNA in pBK-CMV was directly used for expression in NIH3T3. The lysates prepared from NIH3T3 cells and mock (pBK-CMV alone) transfected cells did not show any degradative activity against FITC-HS when analyzed by high speed gel permeation column chromatography (Fig. 5, a and b). In contrast, the cell extracts prepared from the transfectant clones showed very high heparanase activity toward FITC-HS (Fig. 5c).
Furthermore, another cell line COS-7 of monkey kidney epithelial origin was used for the same transfection study. COS-7 cells themselves had a low level of heparanase activity (Fig.  5d). Although the mock transfectant showed a similar low level of heparanase activity (Fig. 5e), the heparanase transfectant showed a significantly elevated activity compared with the parental and mock control cells (Fig. 5f). DISCUSSION A highly sensitive heparanase assay was developed by labeling heparan sulfate with FITC and using high speed gel permeation chromatography. The FI of FITC-HS was high enough to detect a very low level of heparanase activity. Most of all, the advantage of using fluorescence for detection was that the sensitivity was almost equivalent to that achieved by use of radioisotopes, but experiments could be performed safely and easily. The degraded products were analyzed by gel permeation column chromatography. Continuous routine analyses with a 7-min interval were performed in our assay system. Calculating the reduction of FI in the forward half area of the intact FITC-HS peak enabled us to perform quantitative analysis. We employed this calculation procedure rather than measuring retention time shift or the area of the fully cleaved heparan sulfate because several heparanase cleavage sites have been reported to exist in the heparan sulfate sequence (14,15). The cleavage of only one susceptible site by heparanase was enough to cause a reduction of the forward area of the intact FITC-HS peak (15,18,19).
In the present study, human heparanase was purified from SV40 virus-transformed WI38/VA13 cells. The parental human embryonic lung fibroblast WI38 cells did not show significant heparanase activity, suggesting that heparanase expression was activated by SV40 transformation in WI38/VA13 cells. Similar results were reported on hyaluronidase expression in SV40-transformed 3T3 cells by Orkin et al. (30). Thus, the expression of these endoglycosidases may be controlled by the same transcriptional regulation mechanism and closely linked to malignant transformation. Alternatively, the production of TABLE II N-terminal amino acid sequences of the intact purified human heparanase and peptides derived from Lys-C endoproteinase digestion The peptides derived from Lys-C endoproteinase digestion of the purified human heparanase were separated by reverse phase chromatography. All amino acid sequences were obtained as described under "Experimental Procedures." X represents an unidentified amino acid residue.

Lys-Lys-Phe-Lys-X-Ser-Thr-Tyr-Ser-Arg-Ser
Peptide 45 Val-Phe-Gln-Val-Val-Glu-Ser-Thr-Arg-Pro-Gly-Lys Peptide 58 Glu-Gly-Asp-Leu-Thr-Leu-Tyr-Ala-Ile-Asn-Leu-His-Asn-Val-Thr-Lys Peptide 61 Tyr-Leu-Leu-Arg-Pro-Leu-Gly-Pro-His-Gly-Leu-Leu-Ser-Lys endogenous inhibitors may be lost by SV40 transformation. It is thus of great interest that the heparanase activity of various malignant tumor cells correlates well with their metastatic potential as reviewed by Nakajima et al. (31). High expression of such enzymes is probably necessary for malignant cells to invade basement membranes and cause angiogenesis (31,32) In our purification, following the initial use of the heparin-Sepharose column, total enzyme activity rose from 5.1 to 82.5 FI/h⅐10 Ϫ8 . This result suggests that in the cellular extract, there is an endogenous inhibitor(s) of heparanase. The possible existence of such an endogenous inhibitor(s) has been implied by Ihrcke et al. (23) working on heparanase purification from platelets. The high recovery of heparanase activity after heparin-Sepharose chromatography cannot be explained solely by the removal of HS/HSPG present in the cell extract. In fact, Keren et al. (33) have partially purified an endogenous heparanase inhibitor(s) of M r 2,000 -10,000 with an isoelectric point of pH 5.6 -5.8 from murine melanoma cells. There might also be endogenous small proteins other than HS/HSPG that strongly inhibit heparanase activity in WI38/VA13 cells.
The second purification step we applied was ConA affinity chromatography. Heparanase bound strongly to ConA-agarose, suggesting that it had several high mannose-type oligosaccharides linked to asparagine residues. Incubation of heparanase bound to ConA-agarose in elution buffer even overnight yielded less than 50% recovery of the heparanase activity, suggesting not only that there is a high affinity of heparanase oligosaccharides for ConA but also that strong hydrophobic interactions exist among these proteins. The third step of purification was carboxymethyl-Sepharose cationic exchange chromatography in which heparanase was eluted at a very low concentration of NaCl.
The final and very efficient purification step was phenyl-Sepharose hydrophobic interaction chromatography. This step was chosen after many failed attempts with various column chromatographies. Although the recovery of the enzyme was 3.6% of that recovered in the eluate from the heparin-Sepharose column, we detected only a single band of molecular size 50 kDa in the purified fraction from phenyl-Sepharose chromatography. The molecular mass of the purified heparanase reported so far varies from 10 up to 98 kDa. There might be a post-translation modification and/or an activation of heparanase enzymes in an organ-specific manner, resulting in the different sizes of heparanase observed. Alternatively, they might be truly different enzymes.
The optimal pH for heparanase enzymatic activity was 4.2. This result coincides with previous reports from several groups. From the optimal pH, it is suggested that heparanase is localized in lysosomes. Mollinedo et al. (34) have reported that heparanase is localized in the tertiary granules of human neutrophils. Since there have been extensive reports concerning glycosaminoglycans in the nucleus modulating cell growth, heparanase might also be localized in the nucleus as a regulator of the cell cycle (35). In relation to this, there is an interesting report that the activity of topoisomerase I was inhibited by HS of approximately 15 kDa (11), which is equivalent to the average size of HS degradation products following heparanase treatment.
Heparanase was reported to act as an adhesion molecule above pH 7.0 (24). It would be very interesting to know whether the enzyme we have purified could also act as adhesion molecules on the cell surface. We know that at least at neutral pH, the enzyme still possesses a strong affinity toward HS and heparin. Heparanase may have the ability to adhere to HSPG expressed on the cell surface or HS existing in the extracellular matrix at physiological pH. We are currently attempting to produce a very specific antibody against heparanase. When this has been obtained, further studies regarding the localization as well as the function of heparanase on the cell surface could be carried out.
We report here, for the first time, amino acid sequences of human heparanase peptides. In a search against the Swiss-Prot protein data base, the sequences showed no homology to any known protein sequences. Thus, we performed molecular cloning of human heparanase cDNA from the WI38/VA13 cDNA library by using the obtained internal amino acid sequences of human heparanase, and one cDNA coding a fulllength heparanase was cloned. This cDNA had no homology with connective tissue activating peptide III, which was reported to have heparan sulfate-degrading endoglucosaminidase activity (26). There were six potential N-glycosylation sites in the open reading frame, which coincided with our findings that the purified heparanase had high mannose-type N-linked oligosaccharide chains.
The expression study clearly demonstrated that the clone is the cDNA encoding for human heparanase. When the fulllength heparanase was expressed in mammalian cell lines, NIH3T3 and COS-7, both transfectants showed high levels of heparanase activity against FITC-HS.
In summary, our studies reveal for the first time the cDNA encoding human heparanase. Bame and Robson (36) proposed that there might be at least two different kinds of heparan sulfate-degrading enzyme. In their study, the products of heparan sulfate degraded by heparanase were grouped into two distinct classes, one of each recognizing the substrate by differences in sulfate content between modified and unmodified regions. Their work suggests that there might be another homologous heparanase, or alternatively the heparanase coded by the same cDNA may express different substrate specificities when modified by proteinases post-translation. In the latter model, differences in sizes of the heparanase reported so far would provide a supportive observation. It would be very interesting to elucidate the actual site of heparan sulfate that the heparanase catalytically cleaves. Once we identify concrete structures of the cleavage sites, we might be able to carry out the kinetic study using a substrate molecule having a single heparanase cleavage site.
Further studies will be needed to clarify the roles of heparanase in regulating biological activities of heparan sulfate and heparan sulfate proteoglycans. The expression of a recombinant human heparanase as well as the production of antibodies against it has been well advanced. These materials will be useful in studying heparanase functions (cellular localization, substrate specificity, etc.) in relation to diseases including cancer, inflammation, wound healing, and central nervous diseases.