INTRODUCTION

Tat is a viral regulatory gene of the human immunodeficiency virus type 1 (HIV-1),¹ the etiologic agent of AIDS (1, 2). Tat is essential for viral replication since its protein product promotes transcription of the viral genome by interacting with the transactivation responsive element, located at the 5 end of viral mRNAs (3, 4). The HIV-1 Tat protein is a 86-102 amino acid polypeptide, depending on the virus strain, which is encoded by two exons and translated from multiply spliced 2-kilobase mRNAs (5). The amino acid sequence 1-72 encoded by the first exon of the Tat gene is endowed with a full transactivating activity (6-8), while the carboxyl-terminal region encoded by the second exon (amino acids 72 to 86) is not required for the transactivating activity of Tat nor for the replication of HIV-1 (6, 9).

Tat is released from HIV-1-infected cells (10). Extracellular Tat has the ability to enter the cell and nucleus in an active form where it may stimulate the transcriptional activity of HIV-LTR (11). Exogenous Tat taken up by uninfected cells can also transactivate endogenous genes, inducing the production of various cytokines and their receptors (12-23). It has been recently demonstrated that LTR transactivation and possibly cellular gene transcription depends on NF-B activation by Tat (24). Exogenous Tat is also able to exert an angiogenic activity in vitro (25) and in vivo (26-29) and stimulates cell proliferation (27, 28, 30), chemoinvasion (28), and activation (31, 32) of cultured endothelial cells.

The mechanisms of extracellular Tat uptake, intracellular transport to the nucleus, and biological activity exerted on uninfected cells are still unclear. Tat has been shown to bind to integrin receptors (33, 34) and to the vascular endothelial growth factor tyrosine kinase receptor Flk-1/KDR, in a manner analogous to vascular endothelial growth factor itself on endothelial cells (35). Also, the capacity of Tat to interact with heparin-heparan sulfate (HS) may play an important role in its interaction with the cell surface. Heparin inhibits the uptake of extracellular Tat and its HIV-LTR transactivating activity (11). Tat uptake is significantly reduced in Chinese hamster ovary cell mutants defective in glycosaminoglycan (GAG) synthesis (28). Similar to various heparin-binding angiogenic growth factors, the angiogenic activity exerted by Tat in vivo and its mitogenic and chemotactic activity for cultured endothelial cells are modulated by heparin (28). Finally, recombinant HIV-1 Tat protein binds to heparin-Sepharose columns and is eluted only at high salt concentrations (28). These observations suggest that cell-associated HS proteoglycans function as cell surface co-receptors for exogenous Tat.

Heparin consists largely of 2-O-sulfate L-iduronic acid  N-6-disulfate D-glucosamine disaccharide units. Other disaccharides containing unsulfated L-iduronic or D-glucuronic acid and N-sulfate or N-acetylate D-glucosamine are also present as minor components. This heterogeneity is more pronounced in HS, where the low-sulfated disaccharides are the most abundant (36). Recent observations have shown that heparin-HS interaction with various angiogenic growth factors, including basic fibroblast growth factor (bFGF) (37), hepatocyte growth factor (38), and vascular endothelial growth factor (39), depends on the molecular weight of the polysaccharide chain and on the degree and distribution of sulfate groups. Interestingly, distinct oligosaccharide sequences have been identified to retain bFGF- or hepatocyte growth factor-binding capacity (38, 41). In addition, binding studies involving chemically modified heparins or HS preparations have shown that 2-O- and N-sulfate groups are important for bFGF interaction while hepatocyte growth factor interacts mainly with 6-O-sulfate groups (40, 41). Taken together, these data suggest that distinct structural requirements are necessary for the interaction of heparin/HS with different growth factors.

Here we have investigated the molecular basis of Tat-heparin interactions. Different glycosaminoglycans, selectively desulfated heparins, and heparins with different molecular weight were evaluated for their capacity to interact with the Tat protein in cell-free systems. These compounds were also assessed for their capacity to modulate the LTR transactivating activity of extracellular Tat. The results indicate that, among the GAGs tested, only heparin and HS interact with Tat and inhibit its transactivating activity. The interaction requires at least some 2-O-, 6-O-, and N-sulfate groups and is significantly affected by the size of the polysaccharide chain.

EXPERIMENTAL PROCEDURES

Type I HS (42) was from Opocrin (Corlo, Italy). K5 polysaccharide was prepared as described (43). Chondroitin sulfates A and C, dermatan sulfate, and hyaluronic acid were a gift of Dr. M. Del Rosso, University of Florence, Italy.

Conventional heparin (13.6 kDa) was obtained from a commercial batch preparation of unfractionated sodium heparin from beef mucosa (1131/900 from Laboratori Derivati Organici S.p.A., Milan, Italy) which was purified from contaminants according to described methodologies (44). Purity was higher than 95% as assessed by electrophoresis in acidic buffer (45), uronic acid quantitative determination (46), and high performance liquid chromatograph analysis (44). The ¹³C NMR spectrum performed according to Casu et al. (47) showed 78% N-sulfate glucosamine, 80% 6-O-sulfate glucosamine, and 59% 2-O-sulfate iduronic acid. Very low molecular weight beef mucosal heparins (3.0 and 2.1 kDa) were obtained by controlled nitrous acid degradation of unfractionated heparin as described elsewhere (47, 48).

2-O-Desulfated beef mucosal heparin was obtained by selective 2-O-desulfation of the starting material under alkaline conditions (49). The ¹³C NMR spectrum of the product showed 95% 2-O-desulfation and absence of 6-O-desulfation. 6-O-Desulfated mucosal heparin was obtained by preferential 6-O-desulfation (treatment with dimethyl sulfoxide, 10% methanol at 80 °C for 6 h (50)) along with a partial 2-O-desulfation and N-desulfation of the starting material followed by re-N-sulfation. The ¹³C NMR spectrum of the product showed 100% 6-O-desulfation, 15% 2-O-desulfation, and a complete re-N-sulfation. Chemical N-desulfation/N-acetylation of heparin was carried out as described (51). Totally O-desulfated heparin was prepared according to Ogamo et al. (52). The ¹³C NMR spectrum of the product showed 100% 6-O and 2-O-desulfation, and a complete re-N-sulfation. Selectively desulfated heparins were a generous gift of Dr. B. Casu, Ronzoni Institute, Milan. The molecular weight determination and the sulfate/carboxyl ratio analysis of the different GAGs tested were performed according to Harenberg and De Vries (53) and to Casu and Gennaro (54), respectively. The average molecular weight and the SO₃/COO ratio of the GAGs described above are shown in Table I.

Table I. Characterization of the GAGs utilized in the present study For each compound, the molecular mass (kDa) and the number of sulfate groups per disaccharide unit (SO3/COO ratio) were evaluated as described under "Experimental Procedures." GAG kDA SO3/COO1 Heparin   Conventional 13.6 2.14   Very low molecular weight 3.0 2.10   Very low molecular weight 2.1 1.97   2-O-Desulfated 9.6 1.50   6-O-Desulfated 10.3 1.50   Totally-O-desulfated 9.3 1.00   N-Des./N-acetylated 12.3 1.59 Heparan sulfate 18.0 0.93 K5 polysaccharide 5.0 0.00 Chondroitin sulfate A and C 30.0 1.00 Dermatan sulfate 28.0 1.22 Hyaluronic acid 100.0 0.00

Recombinant HIV-1 Tat (86 amino acid) and the one-exon form Tat-1e, that lacks of the amino acid sequence encoded by the second exon, were expressed in Escherichia coli as a glutathione S-transferase (GST) fusion protein as described previously (24). Briefly, the plasmids pGST-Tat-1e and pGST-Tat were constructed by cloning the first exon only or both exons of HIV-1 Tat, respectively, in the commercial vector pGEX2T (Pharmacia, Uppsala, Sweden). GST-chimeric proteins were purified to homogenity by glutathione-agarose affinity chromatography (Sigma) according to the manufacturer's instructions. When required, Tat proteins were cleaved from the GST moiety of the chimera by digestion with thrombin. In detail, 100 µg of GST-Tat were incubated for 3 h at 30 °C with 400 µl of 0.15 M NaCl in Tris-HCl, pH 7.5 (TBS), containing 25 mM CaCl₂ and 2 µg of thrombin (Sigma). Tat protein was then stored at 80 °C until use.

HL3T1 cells are derived from HeLa cells and contain integrated copies of pL3CAT, a plasmid where the bacterial gene for chloramphenicol acetyltransferase (CAT) is directed by the HIV-1 LTR (55). They were grown in Dulbecco's modified Eagle's medium (Life Technologies, Inc., Grand Island, NY) with 10% fetal calf serum (Life Technologies, Inc.). For LTR/CAT transactivating assay, HL3T1 cells were seeded in 24-well dishes at the density of 20,000 cells/cm² in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. After 24 h, cell cultures were washed twice with phosphate-buffered saline and incubated for a further 24 h in fresh medium containing 10% fetal calf serum and 200 ng/ml GST-Tat in the presence of the GAG under test. 100 µM Chloroquine was routinely added to cell cultures to prevent lysosomal degradation of cell-internalized Tat (56). Then, conditioned medium was removed and cell cultures were incubated for a further 24 h in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. At the end of incubation, cells were extracted with MOPS-buffered saline lysis buffer (Boehringer, Mannheim, Germany). The amount of CAT present in the cell extracts was determined by ELISA using the CAT ELISA kit (Boehringer, Mannheim, Germany) according to the manufacturer's instructions. Protein concentration was evaluated by the Bio-Rad protein assay (Bio-Rad Ltd., Brussels, Belgium) according to manufacturer's instructions.

1,6-Diaminoexyl-derivatized Sepharose gel was prepared from CNBr-activated Sepharose 6B gel (Pharmacia, Uppsala, Sweden) according to the manufacturer's instructions. One-ml aliquots of the gel were then suspended in 3 ml of distilled water, pH 4.5, containing 1 mg of conventional or selectively desulfated heparin and 10 mg of N-(3-dimethylaminopropyl)-N-ethylcarbodiimide HCl. The suspension was kept overnight at 4 °C under gentle mixing. Gels were then sequentially washed with 10 ml of H₂O, pH 4.5, 10 ml of 2.0 M NaCl, and 20 ml of H₂O. To evaluate their capacity to bind Tat, 3-µg aliquots of native GST-Tat were loaded onto the different heparin-Sepharose columns (5 × 60 mm) which were then eluted with a 0.15-3.0 M NaCl gradient. Tat immunoreactivity in the various fractions was evaluated by dot-immunoblot analysis with an anti-Tat antibody (MRC AIDS Reagent Project, National Institute for Biological Standards & Control, Potters Bar, Herts, United Kingdom) and quantified by soft-laser scanning of the nitrocellulose membrane.

Heparin was ³H-labeled as described previously (57) with minor modifications. Briefly, 3 mg of conventional heparin were dissolved in 600 µl of 0.1 M Tris-HCl, pH 8.0, containing 2.5 mCi of NaB[³H]₄ (1Ci/mmol) (DuPont New England Nuclear, Wilmington DE) and incubated for 3 h at room temperature under gentle shaking. 18 mg of glucose was then added to the suspension and incubated for a further 3 h. At the end of incubation the sample was dialyzed (cut off: 1,000 Da) against distilled water and kept at 20 °C until use. The specific radioactivity of the ³H-labeled heparin was approximately 3,000 cpm/nmol.

To evaluate their relative affinity for Tat protein, GAGs and heparin molecules utilized in the present study were tested for their capacity to compete for the binding of ³H-labeled heparin to immobilized GST-Tat. 400 µl of glutathione-agarose gel were mixed with 250 µg of recombinant GST-Tat fusion protein for 6 h at 4 °C. After extensive washing, the supernatant was removed and resin beads were resuspended in TBS and stored at 4 °C until use. Under these conditions, up to 90% of originally added GST-Tat remained bound to the resin.

For competition binding assay, 80-µl GST-Tat-glutathione-agarose columns were loaded with 300-µl samples containing 50 µg of ³H-labeled heparin (20,000 cpm) and increasing concentrations of the unlabeled GAG. Columns were washed with TBS and eluted with 3.0 M NaCl. Radioactivity in the eluate was measured in a liquid scintillation counter.

100 µg of 3.0-kDa heparin were loaded onto a GST-Tat-glutathione-agarose column (200 µl). The column was then extensively washed with TBS and eluted with 500 µl of 50 mM Tris-HCl, pH 7.5, containing 3.0 M NaCl. Heparin present in the flow-through of the column and in the 3.0 M NaCl eluate was quantified by the carbazole reaction (58) and analyzed by polyacrylamide gel electrophoresis according to Hilborn et al. (59).

RESULTS

Recombinant HIV-1 Tat and the one-exon form Tat-1e, which lacks the amino acid sequence encoded by the second Tat gene exon, were expressed in E. coli as GST-Tat fusion proteins. GST-chimeric proteins were then purified from the cell extracts by glutathione-agarose affinity chromatography (Fig. 1A). Tat proteins were cleaved from the GST moiety of the chimera by digestion with thrombin and compared with intact GST-Tat proteins for the capacity to transactivate HIV-1 LTR. The recombinant proteins were added to the culture medium of HL3T1 cells that contain integrated copies of pL3CAT in which the bacterial CAT gene is directed by HIV-1 LTR. After incubation, the level of HIV-1 LTR transactivation by Tat was determined by measurement of CAT concentration by ELISA of the cell extracts. As shown in Fig. 1B, the one-exon and two-exon forms of Tat exert the same transactivating activity both as GST-Tat fusion proteins and as thrombin-cleaved Tat proteins. These data indicate that the GST moiety does not interfere with the capacity of extracellular Tat to be uptaken by HL3T1 cells and to exert its transactivating activity after internalization. These data also confirm previous findings indicating that the amino acid sequence of the second exon of Tat is dispensable for the transactivating activity of Tat protein (6, 9).

Heparin has been previously reported to inhibit the transactivating activity of extracellular Tat (11). To further validate the use of GST-Tat fusion proteins for studying Tat-heparin interactions, we assessed the capacity of conventional heparin to inhibit the transactivating activity of both wild-type Tat and Tat-1e when fused to the GST moiety. As shown in Fig. 1C, heparin inhibits the HIV-1 LTR transactivating activity of both GST-Tat fusion proteins with a similar potency (ID₅₀ equal to 1-5 nM). Similar results were obtained when Tat proteins were assessed after thrombin cleavage of the GST moiety (data not shown). Again, the data indicate that the GST moiety does not interfere with Tat-heparin interactions and suggest that the amino acid sequence coded by the second Tat exon is not involved in this interaction. On this basis, an uncleaved, full-length GST-Tat fusion protein was utilized in subsequent experiments.

The indication that the GST moiety does not affect the capacity of Tat to interact with heparin prompted us to establish an experimental protocol in which we took advantage of the ability of GST to bind to glutathione-agarose beads, permitting immobilization of GST-Tat onto a solid matrix. The capacity of immobilized GST-Tat to interact with heparin was then investigated. Conventional heparin binds to a GST-Tat-glutathione-agarose column and can be eluted with a 3.0 M NaCl wash (Fig. 2A). Specificity of this interaction was demonstrated by the inability of heparin to interact with glutathione-agarose columns in which no protein or a recombinant GST protein devoid of the Tat moiety were bound to the resin beads. Determination of the maximum amount of heparin retained by the GST-Tat resin, evaluated by detection of the GAG in the flow-through of GST-Tat columns loaded with increasing amounts of heparin, indicated that 1 ml of GST-Tat-glutathione-agarose beads, corresponding to 400 µg of Tat protein, bound 650 µg of conventional heparin.

We then examined the capacity of unlabeled conventional heparin to compete for the binding of ³H-labeled heparin to immobilized GST-Tat. A series of 80-µl GST-Tat-glutathione-agarose columns were loaded with 300-µl samples containing 50 µg of ³H-labeled heparin (20,000 cpm) and increasing concentrations of unlabeled heparin. Columns were washed extensively and radioactivity in the 3.0 M NaCl eluates was measured. As shown in Fig. 2B, unlabeled heparin competes for the binding of ³H-labeled heparin in a dose-dependent manner, with an ID₅₀ equal to 10 µM. Specificity of the competition is shown by the lack of competition by a non-sulfated glycosaminoglycan, the E. coli capsular K5 polysaccharide (Fig. 2B).

The capacity of different GAGs to compete with ³H-labeled heparin for binding to immobilized GST-Tat was then evaluated in the same assay. As shown in Fig. 3A, only heparin, and to a lesser extent HS, were able to compete for Tat binding, while equimolar concentrations of dermatan sulfate, chondroitin sulfates A and C, hyaluronic acid, and the K5 polysaccharide were ineffective. Accordingly, 0.5 µg/ml heparin or HS, but not the other GAGs tested, were able to significantly inhibit the LTR transactivating activity exerted by the GST-Tat fusion protein in HL3T1 cells (Fig. 3B).

These data suggest that the capacity of GAGs to interact with the Tat protein is related, at least in part, to differences in their structure and degree of sulfation. Indeed, heparin, which shows the highest number of sulfate groups per disaccharide unit (SO₃/COO = 2.14) is more effective than HS, while no significant interaction was observed with the non-sulfated hyaluronic acid and K5 polysaccharide. On the other hand, HS is much more effective than dermatan sulfate or chondroitin sulfates, even though they all share a similar SO₃/COO value of approximately 1.0. In HS, sulfate groups are arranged in clusters at high charge density intercalated with low charge density regions (60). Thus, the backbone structure of the polysaccharide and/or the distribution of the sulfate groups along the chain may play a role in mediating its interaction with Tat protein.

To assess the role of the specific sulfate groups of heparin in Tat interaction, selectively desulfated unlabeled heparins were evaluated for the capacity to compete with ³H-labeled heparin for the binding to GST-Tat immobilized onto glutathione-agarose beads. As shown in Fig. 4A, selective 2-O-desulfated heparin, 6-O-desulfated heparin, and N-desulfated/N-acetylated heparin show a significant reduction in their capacity to compete for the binding to GST-Tat (ID₅₀  330 µM) when compared with conventional heparin (ID₅₀ = 10 µM). Moreover, total O-desulfation completely abolished the capacity of heparin to interact with Tat. Finally, a mixture of 2-O-desulfated, 6-O-desulfated, and N-desulfated/N-acetylated heparins, each at 110 µM concentration (total heparin concentration equal to 330 µM), inhibits the binding of ³H-labeled heparin to immobilized GST-Tat by 30% only. Taken together, the data indicate that at least some 2-O-, 6-O-, and N-sulfate groups must be organized on the same heparin chain to allow an optimal interaction with immobilized GST-Tat.

To confirm the requirement for 2-O-, 6-O-, and N-sulfate groups for optimal Tat-heparin interaction, conventional and selectively desulfated heparins were conjugated to Sepharose gel. The capacity of GST-Tat fusion protein to bind to the different heparin-Sepharose columns was then evaluated (see "Experimental Procedures" for further details). As shown in Fig. 5A, GST-Tat binds to immobilized conventional heparin from where it elutes at 1.5 M NaCl. A similar elution profile was obtained when Tat was loaded onto the column after thrombin cleavage of the GST-Tat fusion protein (28). It is interesting to note that under the same experimental conditions, the prototypic heparin-binding factor bFGF eluted from the column at 2.2 M NaCl (Fig. 5B). In contrast, recombinant GST lacking Tat (Fig. 5A) and a heat-denaturated GST-Tat fusion protein (Fig. 5B) did not bind to heparin-Sepharose beads and eluted with the flow-through of the column. These data indicate that the high affinity interaction of GST-Tat with immobilized heparin reflects the heparin-binding capacity of the Tat moiety and occurs only when the protein is present in the proper native conformation.

In agreement with the competition binding studies, GST-Tat fusion protein did not bind to immobilized totally-O-desulfated, 2-O-desulfated, 6-O-desulfated, and N-desulfated/N-acetylated heparins when loaded at 0.15 M NaCl (Fig. 5, C-F). In contrast, bFGF still bound to the different desulfated heparins, even though, in agreement with previous observations in different experimental systems (40, 41), with a reduced affinity, being eluted at NaCl concentrations ranging from 0.9 to 1.5 M (Fig. 5, C-F). In conclusion, the data confirm the requirement for 2-O-, 6-O-, and N-sulfate groups for optimal, high affinity heparin-Tat interaction.

The selectively desulfated heparins were then evaluated for their capacity to inhibit the LTR transactivating activity exerted by GST-Tat fusion protein in HL3T1 cells (Fig. 4B). As expected from the competition binding studies, totally-O-desulfated heparin shows a very limited capacity (ID₅₀ = 10 µM) to inhibit the transactivating activity of GST, when compared with conventional heparin (ID₅₀ = 0.8 nM). A significant decrease of the antagonist activity of heparin was caused also by selective 2-O-desulfation (ID₅₀ = 100 nM). Unexpectedly, 6-O-desulfated or N-desulfated/N-acetylated heparins significantly inhibited the transactivating activity of GST-Tat, in apparent contrast with their reduced activity in the competition binding studies.

The influence of size of the heparin chain on its capacity to interact with immobilized GST-Tat was investigated by competition binding assays. Unlabeled very low molecular mass heparins (3.0 and 2.1 kDa) were compared with conventional heparin (13.6 kDa) for their capacity to compete with ³H-labeled heparin for binding to GST-Tat-glutathione-agarose. When the different heparin molecules were compared on a molar basis (Fig. 6A), the capacity of very low molecular mass heparins to interact with immobilized GST-Tat was dramatically reduced (ID₅₀ equal to 10, 200, and 330 µM for 13.6, 3.0, and 2.1 kDa heparin, respectively). In agreement with these competition binding studies, 3.0- and 2.1-kDa heparins showed a reduced capacity to inhibit the LTR transactivating activity exerted by GST-Tat fusion protein in HL3T1 cells (ID₅₀ equal to 50-100 nM compared with 1.0 nM for conventional 13.6-kDa heparin) (Fig. 6B).

To select high affinity components present in the very low molecular mass heparin preparation, the 3.0-kDa heparin was fractionated by GST-Tat affinity chromatography. A 200-µl GST-Tat-glutathione-agarose column was loaded with 100 µg of sample, washed extensively, and eluted with a 3.0 M NaCl wash. 15% of the sample loaded onto the column was recovered in the 3.0 M NaCl fraction (Fig. 7A). Polyacrylamide gel electrophoresis (59) confirmed that the Tat-bound heparin fraction had an average molecular mass of approximately 3.0 kDa (data not shown). When bound and unbound fractions were evaluated for their capacity to inhibit the transactivating activity of GST-Tat, the bound fraction was approximately 20 times more potent than the parent compound and 1,000 times more effective than the unbound fraction (Fig. 7B). Thus, the inhibitory activity of the Tat-bound fraction appeared to be even higher than that anticipated on the basis of its relative concentration in parent heparin, suggesting that unbound molecules in the original preparation may interfere with Tat-heparin interaction and/or with LTR transactivation. Further experiments are necessary to clarify this point. In conclusion, the data demonstrate that GST-Tat affinity chromatography can be used to isolate very low molecular weight heparin oligosaccharide chains with a potent Tat antagonist activity.

DISCUSSION

This is the first biochemical characterization of HIV-Tat protein interactions with heparin/HS. Disaccharide composition, degree of sulfation, and arrangement of the charges along the polysaccharide chain are all important in determining the capacity of a glycosaminoglycan to bind Tat protein and inhibit its transactivating activity. Compounds with low charge density, such as K5 polysaccharide and hyaluronic acid, do not interact with Tat. Also, a different binding capacity is shown by low sulfated GAGs with a similar SO₃/COO ratio but which differ in their backbone conformation and charge distribution (HS versus dermatan sulfate and chondroitin sulfates). Finally, selective 2-O-, 6-O-, or N-desulfation all prevent Tat-heparin interaction in vitro, indicating that 2-O-, 6-O-, and N-sulfate groups in heparin/HS are required for high affinity binding to Tat. Disaccharide units containing 2-O-, 6-O-, and N-sulfate groups are predominantly present in heparin and to a lesser extent in HS. This appears to be in keeping with the slightly reduced capacity of HS to bind immobilized GST-Tat in competition binding assays. The heparin-binding angiogenic factor bFGF binds avidly to a pentasaccharide region composed of such disaccharide repeating units (40) even though only N-sulfate groups and a single 2-O-sulfate group are essential for bFGF interaction while the remaining 2-O-sulfate groups and the 6-O-sulfate groups in the pentasaccharide structure are redundant (61). In contrast, 6-O-sulfate groups are mainly involved in the interaction with hepatocyte growth factor (38). These findings suggest that different heparin/HS-binding cytokines may bind sulfated GAGs in a distinct, possibly specific manner. Even though specific factor binding sequences may be hidden in heparin due to its high degree of sulfation, the high heterogeneity in HS structure allows a more refined tailoring of selective binding regions that may influence the biological activity and bioavailability of heparin/HS-binding growth factors, including the Tat protein. This possibility is exemplified by the shift in cell-surface HS proteoglycan properties from a bFGF- to an acidic FGF-binding phenotype in murine neuronal cells during embryonic development (62).

Previous studies have shown that the bFGF-heparin interaction depends on the molecular weight of the polysaccharide and that a pentasaccharide sequence represents the minimal binding site for bFGF (see above). Here, by using heparin preparations with different average molecular weight, we have observed that the affinity of heparin for Tat protein decreases with size reduction of the polysaccharide. Consequently, a reduced capacity to inhibit the transactivating activity of extracellular Tat was also found. Even though no attempts were done to determine the minimal size required to bind Tat, we have demonstrated that fractionation of very low molecular weight heparin by Tat affinity chromatography isolates a subpopulation of oligosaccharide chains that retain the capacity to bind Tat and to inhibit its transactivating activity with high potency. Further studies are in progress in our laboratory to characterize this subpopulation and to determine the minimal binding structure.

The 6-O-desulfated and N-desulfated/N-acetylated heparins retained the capacity to inhibit the transactivating activity of Tat even though they lost the ability to bind Tat with high affinity. This suggests that inhibition of LTR transactivation may require only partial interactions between Tat and heparin/HS. Alternatively, multiple interactions with the cell surface may be involved in heparin inhibition of the HIV-LTR transactivating activity of Tat. Extracellular Tat binds integrins (33, 34, 63, 64) and Flk-1/KDR (35) on cell surfaces and heparin has been reported to affect the capacity of ligands to bind to these receptors (33, 39). Similar observations have been made for bFGF-heparin interactions. In this system, polysaccharide chains longer than the minimal bFGF-binding sequence are required for the modulation of the biological activity of the growth factor (65). In addition, the 6-O-sulfate groups play a role in the formation of the bFGF-heparin-FGF receptor ternary complex even though they are not involved in bFGF interaction (61, 66).

Our studies on the biochemical properties of GST-Tat chimeras demonstrate that the two-exon and one-exon forms of Tat exert the same transactivating activity and are similarly inhibited by heparin. They also bind to heparin/HS with the same affinity when immobilized onto glutathione-agarose column (data not shown). In contrast to the apparently dispensable second exon, several observations point to a role for amino acid residues 49-57 of the first Tat exon in Tat-heparin interaction. This highly charged, basic domain is critical in both transactivating activity (67) and stability (68) of the protein and it has been implicated in the interaction of Tat with _v5 integrin (33). Peptides encompassing the Tat basic domain induce growth and migration of cultured endothelial cells (28), potentiate Tat uptake and transactivation (11), and activate tyrosine kinase receptors (35). Alignment of the domain with other heparin-binding factors gives a rough consensus sequence for heparin interaction with clustering of Arg and Lys residues (28). Site-directed mutagenesis studies of the Tat protein will be required to validate this hypothesis.

Tat is a potent transactivator essential for viral replication. It can be released from infected cells and enter new cells in an active form, where it stimulates the transcriptional activity of the HIV-LTR. This mechanism may explain the "burst" replication associated with early phases of HIV infection, where synchronized virion replication takes place (69). Tat also appears to have a number of biological effects apart from LTR transactivation which may be responsible for some AIDS-associated syndromes. Expression of the HIV1-Tat gene in transgenic mice results in angiogenic skin lesions and increased incidence of adenocarcinomas, lymphomas, and hepatocarcinomas (70). Tat is a growth factor for Kaposi's sarcoma-derived cells and for endothelial cells (27, 28, 30, 34) and exerts angiogenic activity in vivo (26-29). Tat may play an important role also in tumor metastasis, possibly by modulating protease production by transformed cells (29, 71, 72). Finally, Tat appears to mediate the neurological affections that often occur in AIDS (73) possibly as a consequence of a direct toxic effect onto neurons (74, 75) or abnormal activation of endothelium in the central nervous system (32).

These observations indicate that pharmacological interference of Tat-HS interactions may be a target for blocking AIDS-associated pathologies, including Kaposi's sarcoma, and possibly HIV replication itself. To this respect, it is interesting to note that sulfated polysaccharides, including heparin and dextran sulfate, have been reported to be effective in vitro inhibitors of HIV infection (76, 77). Sulfated polysaccharides may act at the level of viral binding and penetration into the host cell (78) and prevent gp120-CD4 interaction (79). Our data rise the possibility that sulfated polysaccharides may inhibit HIV replication also by inhibiting Tat activity. Tat has been recently proposed as a specific target for AIDS vaccine (80). This observation strongly supports the potential use of Tat antagonists in HIV infection. Thus, the elucidation of the molecular bases of Tat-heparin (HS) interaction will help to synthesize specifically tailored saccharide analogs designed for pharmacological therapies.