Heparin and its derivatives bind to HIV-1 recombinant envelope glycoproteins, rather than to recombinant HIV-1 receptor, CD4

Abstract

We have employed a direct radiolabel binding assay to investigate the interaction between ³H-heparin and recombinant envelope glycoproteins, rgp120s, derived from several different isolates of HIV-1. Comparable dose-dependent binding is exhibited by rgp120s from isolates IIIB, GB8, MN and SF-2. Under identical experimental conditions the binding of ³H-heparin to a recombinant soluble form of the cellular receptor for gp120, CD4, is negligible. The binding of ³H-heparin to rgp120 is competed for by excess unlabeled heparin and certain other, but not all, glycosaminoglycan and chemically modified heparins. Of a range of such polysaccharides tested, ability to compete with ³H-heparin for binding was strictly correlated with inhibition of HIV-1 replication in vitro. Those possessing potent anti-HIV-1 activity were effective competitors, whereas those having no or little anti-HIV-1 activity were poor competitors. Scatchard analysis indicates that the K_d of the interaction between heparin and rgp120 is 10 nM. Binding studies conducted in increasing salt concentrations confirm that the interaction is ionic in nature. Synthetic 33–35 amino acid peptides based on the sequence of the V3 loop of gp120 also bind to heparin with high affinity. V3 loop peptides that are cyclized due to terminal cysteine residues show more selective binding than their uncyclized counterparts. Overall, these data demonstrate further that heparin exerts its anti-HIV-1 activity by binding to the envelope glycoprotein of HIV-1, rather than its cellular receptor, CD4. This study confirms that the V3 loop of gp120 is the site at which heparin exerts its anti-HIV-1 activity. Moreover, it reveals that high affinity binding to heparin is shared by all four rgp120s examined, despite amino acid substitutions within the V3 loop.

Introduction

The heparin/heparan sulfate family of glycosaminoglycans are highly charged, linear polysaccharide components of the extra-cellular matrix and cell surfaces. Although these polysaccharides share a common disaccharide repeat structure of alternating uronic acid and N-acetyl glucosamine, extensive and variable postincorporation modifications result in considerable diversity in sulfated oligosaccharide sequence (for review, see Gallagher et al., 1992). It is now emerging that among their multiple biological roles, heparin and heparan sulfate participate in, or influence, the mechanisms by which certain envelope viruses bind to and infect their target cells. In a well defined example, cell surface heparan sulfate is the cellular receptor for herpes simplex virus type 1 and is essential for virion-plasma membrane fusion (Shieh and Spear, 1994).

The in vitro replication of a range of enveloped viruses is inhibited by both heparin and the synthetic sulfated polysaccharide, dextran sulfate (Baba et al., 1988a). A prominent virus inhibited in this way is the human immunodeficiency virus type 1 (HIV-1). Although HIV-1 replication is inhibited by various sulfated polysaccharides, including curdlan sulfate, dextrin sulfate, fucoidan, and pentosan polysulfate (Baba et al., 1988b,c; Anand et al., 1990; Aoki et al., 1992; McClure et al., 1992), heparin and dextran sulfate are the two best studied sulfated polysaccharides in this regard. Heparin and dextran sulfate are both highly potent inhibitors of HIV-1 replication in vitro with EC50 values around 1-5 µg/ml (Baba et al., 1988c; Bagasra and Lischner, 1988; Ueno and Kuno, 1988).

The mode of action by which sulfated polysaccharides inhibit HIV-1 replication n vitro is not entirely clear. Indeed two different mechanisms have been proposed. One of these, which is particularly well established in the case of dextran sulfate, is the inhibition of the high affinity binding of the envelope glycoprotein of HIV-1, gp120, to its primary cell surface receptor, CD4. This mode of action is evidenced by the ability of dextran sulfate to inhibit HIV-1 virion binding to CD4⁺ cells (Baba et al., 1988c), and also to inhibit the formation of syncitia between infected cells, thereby gp120-expressing, and uninfected CD4⁺ cells (Baba et al., 1990). In solid phase ELISA studies, it has been shown that dextran sulfate inhibits binding of rgp120 to rsCD4 (Lederman et al., 1989). These workers also presented evidence that heparin had the same inhibitory activity. However in a previous study we found that dextran sulfate alone was able to inhibit the binding of rgp120 to rsCD4, whereas by contrast heparin was inactive in this regard (Harrop al.,1994).

A second mode of action is binding to the third variable, V3, loop of gp120. This sequence delineated by invariant cysteines 303 and 338 is an important determinant of cellular tropism (Hwang et al., 1991). Since antibodies to the V3 loop neutralize HIV-1 albeit in a strain specific manner, the V3 loop is termed the principal neutralizing domain. The V3 loop is not involved in the initial binding of CD4, but appears to be involved in subsequent cell surface events of plasma membrane-viral envelope fusion (reviewed by Nara et al., 1991). Although initially viewed as a variable sequence, the V3 loop has a number of conserved features, including an array of at least five basically charged amino acids (LaRosa et al., 1990; Rider et al., 1994).

Evidence that the V3 loop of gp120 is a binding site for dextran sulfate and heparin has arisen from several studies. Dextran sulfate was found to inhibit the binding of V3 loop-specific monoclonal antibody (Mab) to HIV-1 infected cell surfaces or to rgp120 in ELISA studies (Schols et al., 1990; Callahan et al., 1991). Moreover Batinic and Robey (1992) demonstrated that ³H-heparin binds to recombinant gp120, and also to beads coated with short, basically charged synthetic peptides derived from the V3 sequence of several HIV-1 isolates. In our laboratory, we found that both heparin and dextran sulfate selectively block the binding of V3-loop specific Mabs to immobilized gp120 in solid phase ELISA (Harrop et al., 1994). Our conclusion was therefore that heparin is more specific in its anti-HIV-1 activity than dextran sulfate, in that it acts only at the V3 loop of gp120. This hypothesis is of some significance, since if heparin and its derivatives were indeed to act only via the surface glycoprotein of the pathogen, and not via CD4, the host cell surface receptor, then such compounds would have a clear therapeutic advantage over dextran sulfate which interacts strongly with CD4. We have therefore sought to explore further the anti-HIV-1 activity of heparin and its derivatives with the view to developing novel therapeutic agents.

Aside from its potential for therapeutic exploitation, the anti-HIV-1 activity in vitro of heparin also raises the question of whether cell surface heparin-like molecules have a physiologically important role in the cellular infection of HIV-1 in vivo. Heparan sulfate is synthesized and secreted by many cell types including T cells, and CD4⁺ve cell lines (Wilson and Rider, 1991). Might such sulfated polysaccharides provide some protection from cellular HIV-1 infection, or alternatively is the inhibitory activity of soluble heparin a reflection of some role of cell-surface heparan sulfate in facilitating cell surface virion-plasma membrane fusion events? In the latter context it has been postulated that heparan sulfate may function as a coreceptor, being involved in the formation of a trimolecular complex with the HIV-1 envelope glycoprotein gp120 and its cell surface receptor glycoprotein CD4 (Patel et al., 1993; Rodriquez et al., 1995).

Because of the potential importance of understanding fully the anti-HIV activity of heparin we have studied its interaction with soluble rgp120 in a direct radiolabel binding assay. This assay enables us to compare the affinities of heparin for various rgp120 isolated from different HIV-1 isolates, and to confirm that the binding site for heparin is on this viral envelope glycoprotein and not its primary receptor on the cell surface, CD4.

Results

Radiolabeled heparin binding to rgp120

We have previously shown, using solid phase ELISA, that heparin can compete with and block the binding of anti-V3 loop Mabs to immobilized rgp120. However heparin has no effect on the binding of rsCD4 to either immobilized gp120 or anti-CD4 Mabs (Harrop et al.,1994). Although the ELISA approach we employed in these studies enabled the sensitive detection of the interaction of heparin with the V3 loop of gp120, it suffers from the disadvantage that such binding is only measurable indirectly, through the displacement of Mab binding. One consequence of this is that the binding of heparin to rgp120s from different HIV-1 isolates cannot be directly compared. Similarly, the relative affinity of heparin for rgp120 and rsCD4 could not be examined. We have therefore examined the binding of rgp120 and rsCD4 using to ³H-heparin in a direct binding assay.

Incubation, as described in Materials and methods, of 6 µg ³H-heparin in the absence of protein, followed by filtration through nitrocellulose filters resulted in the retention of counts between 100–400 c.p.m. per 80 µl of incubation mixture. This represents less than 1% of radiolabel binding nonspecifically to the membrane. However, as shown in Figure 1, in the presence of rgp120s from various isolates of HIV-1 there is a concentration dependent retention of labeled heparin. In the case of the rgp120 from isolates IIIB and GB8, which because of availability were used at the highest concentration of 240 nM, the binding appears saturable. The maximal binding achieved is around 25% of added ³H-heparin. Over the concentration range tested, the rgp120s from the viral isolates IIIB, GB8, MN and SF-2 bind to ³H-heparin in a similar concentration-dependent manner. Because of observed experiment to experiment variation in the amount of binding achieved it is not possible to discern a rank order of heparin binding for these rgp120s.

In contrast to these results with rgp120, very low binding was observed between rsCD4 and ³H-heparin under the same experimental conditions (Figure 1). The slope of the binding curve obtained with rsCD4 as seen in Figure 1 remained unchanged even up to rsCD4 concentrations of 900 nM (data not shown). At around 250 nM protein, at which concentration saturable binding of rgp120 is observed, the rsCD4 binds only around 10% of the radiolabel bound by rgp120. These data clearly demonstrate that compared to strong binding to rgp120, heparin binds poorly to rsCD4.

Displacement of ³H-heparin binding to rgp120 by chemically modified heparin and heparan sulfate

The ability of unlabeled heparin, heparin derivatives and heparan sulfate to compete with and displace the binding of ³H-heparin to rgp120 was examined. As may be seen in Figure 2A, both unmodified heparin and Enoxaparin, a clinical low M_r heparin, are able to completely displace the binding of labeled heparin to rgp120 from MN strain. Dextran sulfate, used at the same concentration, is also able to completely inhibit ³H-heparin binding. By contrast chondroitin sulfate is completely inactive. The two preparations of heparan sulfate from porcine intestinal mucosa employed show contrasting activity. Heparan sulfate II, which is a highly sulfated preparation, gives some 80% inhibition of binding whereas HSI, a low sulfated preparation from the same source, is a weak inhibitor.

The inhibitory activity of chemically modified heparins is shown in Figure 2B. The parent unmodified bovine lung heparin, like pig intestinal mucosal heparin, completely inhibits ³H-heparin binding. N-Desulfation almost completely removes inhibitory activity, however subsequent N-acetylation partially restores activity. N-desulfated heparin was almost inactive, and acetoacetylated and totally desulfated heparins are completely inactive. By contrast, decarboxylated heparin is active, giving 80% inhibition of binding. This pattern of activity in the series of chemically modified heparins is very similar to that previously found for both the V3 loop-binding ELISA and the inhibition of HIV-1 infectivity in vitro (Rider et al., 1994). Throughout all of these studies, the same rank order of activity is seen: unmodified heparin > carboxyl-reduced = acetylated, N-desulfated > N-desulfated > totally desulfated = acetoacetylated N-desulfate; the latter two preparations were essentially inactive.

Scatchard analysis of ³H-heparin binding to rgp120

The binding of ³H-heparin to rgp120 was further examined, by incubating increasing concentrations of this glycosaminoglycan with a fixed concentration of gp120. As may be seen in Figure 3, with MN rgp120, binding is dose-dependent with saturation at around 30 nM heparin. The maximal binding of radiolabel represents 40% of ³H-heparin present. Scatchard analysis of this data is shown in Figure 3 inset. From the slope of the line a value for affinity binding constant 10 nM is calculated. The intercept on the abscissa indicates that the total number of binding sites for the rgp120 (50 nM) is present on 10 nM heparin. Assuming that all heparin chains participate equally in the binding, this would give an average of five binding sites per chain.

Binding of ³H-heparin to rgp120 and rsCD4 at varying salt concentrations

As previously shown in Figure 1, ³H-heparin binds to rgp120 to a much greater extent than to equivalent concentrations of rsCD4 when incubated in a low conductivity buffer of 25 mM Tris-HCl. In order to investigate the effects of increasing ionic strength on these interactions we examined the binding at various NaCl concentrations. Rgp120 IIIB and rsCD4 were used at a concentration of 166 nM, which in the case of the rgp120 gives near saturable binding at low ionic strength (Figure 1). On increasing salt concentrations up to that of physiological saline (16 milliSieverts), the binding of ³H-heparin to the rgp120 is decreased in an apparent linear fashion by approximately 50% (see Figure 4). The low level of heparin binding to rsCD4 seen in the absence of salt decreases by approximately 30% with the addition of 30 mM NaCl (4 milliSieverts). Thereafter further increases in ionic strength give no detectable decreases in binding. Thus, at 16 milliSieverts, equivalent to physiological saline, the degree of binding to rgp120 remains some four-fold greater than that observed with rsCD4. Bovine serum albumen (BSA) at the same molar concentration, used as a control to measure nonspecific binding of ³H-heparin to protein, produces only very low background retention of ³H-heparin on the filter throughout the range of ionic strengths tested.

Binding of ³H-heparin to V3 loop peptides

Since our previous data (Harrop et al., 1994) supports the view that the V3 loop is the major binding site for heparin on gp120, we have used this binding assay to examine the structural requirements of ³H-heparin to V3 loop peptides. In particular we sought to investigate the importance of the peptide being in disulfide looped conformation compared an open state.

As may be seen in Figure 5, ³H-heparin incubated with increasing concentrations of V3 loop-derived peptides is retained on the filter in a dose-dependent manner and saturable manner. The linear 33-mer peptide 1840, lacking terminal cysteines, reaches saturable binding at 20 µg/ml peptide. Under these conditions of maximal binding, approximately 40% of the added labeled heparin is retained on the filter. Likewise the cyclic peptide 1841, a 35-mer of sequence identical to 1840 but with the addition of a cysteine residue at each terminus, also binds to ³H-heparin in a dose dependent manner. However, maximal binding is only approximately half that obtained with the linear peptide. Moreover with the cyclic peptide saturable binding is reached at a peptide concentration of only 5 µg/ml. Under the same conditions the peptide EVA 7019, of identical sequence to 1841 but from an independent source, gave an essentially identical binding curve, with maximal binding of around 7000 c.p.m. attained by 4 µg/ml peptide (data not shown). Similarly the cyclic V3 loop peptide EVA 7026 gave the same level of maximal binding, but required 8 µg/ml to attain saturation. These data clearly indicate that V3 loop peptides cyclized by virtue of possessing terminal cysteines show reduced levels of maximal binding to radiolabeled heparin.

Discussion

Previous investigations of the anti-HIV activities of the sulfated polysaccharides heparin and dextran sulfate have produced somewhat conflicting findings. An early study which employed solid phase binding assays suggested that both dextran sulfate and heparin bind to the CD4, the primary cell surface receptor for HIV, and inhibit the subsequent binding of gp120 (Lederman et al., 1989). The existence of a polyanion binding site in the amino-terminal domain of CD4, close to the binding site for gp120, was later proposed (Parish et al.,1990). In our previous work using recombinant proteins in solid phase binding assays, we confirmed that dextran sulfate indeed binds to the amino-terminal domain of CD4, and that this blocks the subsequent binding of gp120 (Harrop et al., 1994). Overall it would appear that dextran sulfate can occupy the polyanion binding site on CD4, and in so doing blocks the adjacent binding site for the viral envelope glycoprotein. However by contrast we found that heparin, despite possessing a similar inhibitory effect on HIV-1 replication in vitro, was ineffective at blocking the binding to CD4 of either recombinant gp120, or anti-CD4 monoclonal antibodies specific for the amino-terminal domain (Harrop et al., 1994). These findings concur with the observation that heparin binds poorly to the polyanion binding site on CD4 (Parish et al., 1990).

A second site at which sulfated polysaccharides are likely to exert their anti-HIV-1 activity is the V3 loop of the viral envelope glycoprotein gp120. This is a disulfide-bridged sequence delineated by cysteines 303 and 338, and is referred to as the principle neutralizing domain of gp120, since antibodies which bind to the V3 loop neutralize HIV-1 replication by blocking cell surface membrane fusion events (for review, see Nara et al., 1991). Moreover, it has recently been shown that the V3 loop is a critical determinant of susceptibility to inhibition by the C-C chemokines, RANTES, MIP-1α and MIP-1β, consistent with the V3 loop binding to the CCR5 chemokine receptor (Cocchi et al., 1996). Several groups have shown that the V3 loop is a binding site for both dextran sulfate and heparin (Callahan et al., 1991; Batinic and Robey, 1992; Lederman et al., 1992). In our own studies and those of our collaborators we have shown that heparin blocks the binding of a total of 11 V3-specific Mabs to recombinant gp120 derived from HIV-1 strains MN, SF-2 and IIIB (Harrop et al., 1994; Hammond and McKeating, personal communication). By contrast heparin does not block the binding to rgp120 of 12 Mabs specific for epitopes outside the V3 loop (Harrop et al., 1994; Harrop et al., unpublished observations; Hammond and McKeating, personal communication). The V3 loop has the highest density of positively charged amino acids of any region of gp120 (Callahan et al., 1991). Moreover despite sequence variations between different strains of HIV-1, there is considerable conservation of basic residues at five particular positions within the loop (Rider et al., 1994). Thus the V3 loop would appear to be a major site at which sulfated polysaccharides exert anti-HIV-1 activity.

In our previous work binding of sulfated polysaccharides to the V3 loop was assessed by the inhibition of subsequent binding to recombinant gp120 of V3 loop specific Mabs (Harrop et et al., 1994; Rider et et al., 1994). Since this is a competitive approach, it is not possible to derive a dissociation constant for the interaction of the sulfated polysaccharide and rgp120. Moreover, it is not possible to compare that binding of sulfated polysaccharide to the differing rgp120s derived from various HIV-1 strains, because as the V3 loop sequence changes, the affinity of any Mab employed will be altered and indeed cross-reactivity may be completely lost. Thus, different Mabs are necessary for binding to particular rgp120s. We have therefore studied the interaction of heparin with rgp120 using a direct binding assay.

This study demonstrates that the recombinant envelope glycoprotein from four different strains of HIV-1 binds with high affinity to ³H-heparin in a very similar concentration-dependent manner. Considerable sequences differences exist between the gp120s, with variation occurring throughout all regions of the protein including the V3 loop. In the case of rgp120 derived from HIV-1 isolate MN, we determined the affinity binding constant of this interaction to be 10 nM. This dissociation constant compares with a value of 30 nM for the binding of heparin to platelet factor 4 measured by the same technique (Stringer and Gallagher, 1997). For the heparin binding of antithrombin III and basic fibroblast growth factor, the Kd have been estimated respectively as 50 nM (Watton et al., 1993) and 6 nM (Lee and Lander, 1991). Thus, the affinity for heparin of gp120 is in a similar range to that of other proteins which are generally recognized to have a high affinity for heparin. Despite the considerable variation in V3 loop sequence between the HIV-1 isolates, the binding curves are similar. It would therefore appear that high affinity binding to heparin is a widespread property of gp120 from diverse HIV-1 isolates. This suggests that inhibition of HIV-1 replication by heparin will occur with a high proportion of isolates. The ability of heparin to bind to rgp120s of isolates SF-2 and MN is of particular significance in the development of heparin derivatives as potential anti-HIV therapeutic agents, since the V3 loop sequences of these isolates are close to the North American consensus sequence for syncytial-inducing isolates. Our experimental results support a recent modeling study which suggested that the V3 loop of syncytial-inducing HIV-1 variants is folded to bring adjacent positive charges into close proximity so as to facilitate binding of sulfated sugars. It was predicted that this would not occur in non-syncytial-forming viral isolates (Bhattacharyya et al., 1996). An important limitation of our studies is that we have used soluble rgp120 which is presumed to exist in monomeric form in solution. On the viral surface, gp120 occurs as tetramers formed in association with the transmembrane protein gp41. Our present studies do not address whether the affinity for heparin would be altered in this associated state.

The direct binding assay used in this study clearly demonstrates that heparin binds poorly to CD4. Our data show that as may be expected for an ionic interaction, the binding of heparin to rgp120 is reduced by increasing salt concentrations. However, even at physiological salt concentrations, the binding of heparin to rsCD4 is considerably lower than that to rgp120. This direct measurement of heparin binding confirms the previous conclusion that heparin does not readily occupy the polyanion binding site adjacent to the gp120 binding site on the amino terminal domain of CD4 (Parish et al., 1990; Harrop et al., 1994). Thus inhibition of gp120-CD4 binding is not the primary mode of anti-HIV-1 activity of heparin.

We show here that the binding of ³H-heparin to rgp120 is completely displaceable with excess unlabeled heparin. Displacement also occurs with excess amounts of clinical low molecular weight heparin (Enoxaparin), dextran sulfate, and HSII, a highly sulfated heparan sulfate preparation. These preparations all possess potent anti-HIV-1 activity and are able to displace V3 loop-specific Mabs in our ELISA (Harrop et al., 1994; Coombe et al., 1995). By contrast, chondroitin sulfate and HSI, a low sulfated heparan sulfate from the same source as HSII, fail to compete with heparin, and both lack HIV-1 inhibitory activity in vitro. This correlation between anti-HIV-1 activity and gp120 binding is further illustrated by chemically modified heparins. We have previously shown that certain chemical modifications selectively affect the ability of heparin to block V3 loop Mabs binding to immobilized rgp120 in solid phase ELISA assays (Rider et al., 1994). In cellular HIV-1 infectivity assays these chemically modified heparins have a range of activities from inactive to near full activity. For these modified heparins there is a strong correlation between HIV-1 inhibition in vitro, and V3 loop blocking activity in the ELISA assay (Rider et al., 1994). We now show that there is a complete correlation between binding to rgp120, displacement of V3 loop Mabs, and antiHIV-1 activity in vitro. Taken overall, this is further evidence that the primary mode of anti-HIV-1 activity of heparin and its derivatives is exerted by binding to the V3 loop of the HIV envelope glycoprotein.

A second conclusion that arises from the study of the chemically modified heparins is that binding to gp120 involves molecular specificity of glycosaminoglycan structure. The decrease on gp120 binding observed with increasing ionic strength shown here clearly indicates that the interaction, as might be anticipated from the polysulfated nature of heparin, is largely ionic. The failure of totally desulfated heparin to compete for binding is also consistent with this view. N-desulfation of heparin, which removes some 25% of the total sulfate (Rider et al., 1994) almost completely abolishes binding, but subsequent N-acetylation partially restores binding activity. This latter modification will neutralize the positive charges as well as reintroduce a bulky subsistuent on to the free amino groups of the glucosamine residues. The markedly different binding properties of N-desulfated and N-desulfated-N-reacetylated heparins may arise from either the charge differences, or the differing secondary structures possessed by these two polysaccharides. These data therefore suggest the interaction of heparin with gp120 is not a simple ionic interaction, but instead has requirements for the charged groups on the heparin organized into favorable three-dimensional arrays.

The molecular specificity of in the interaction between heparin and gp120 is also indicated by studies involving V3 loop derived peptides. Cyclic peptides with V3 loop sequences were found to bind at saturation only around half the radiolabeled heparin bound by a noncyclic counterpart. NMR studies of the solution structure of the cyclic peptide 1841 indicate that it is able to adopt a loop-turn-helix conformation consistent with structural predictions whereas the linear counterpart is disordered in both C- and N-terminal regions. Moreover, cyclic peptide 1841 bound to V3-loop Mabs better than the linear peptide 1840 in solid phase ELISA, further supporting the view that cyclization constrains the conformation which is otherwise disordered (Catasti et al., 1995). The heparin binding data presented here show that the more ordered cyclic structures are more selective in binding to heparin, but do so with a higher affinity than the linear V3 loop peptide.

Taken overall, our findings strongly support the hypothesis that the V3 loop of the HIV-1 envelope glycoprotein is the target site at which heparin and its derivatives exert anti-HIV-1 activity. Compared to the high affinity of heparin for the V3 loop, binding to CD4 is weak. The interaction between heparin and the V3 loop is primarily ionic, however the three-dimensional conformation of the V3 loop, presumably by orientating the basic amino acid side chains relative to each other, confers a structural specificity on the interaction. Although the V3 loop sequence varies from one HIV-1 isolate to another, the conserved pattern of basic residues results in high affinity for heparin being a common property of diverse synytial-inducing isolates of HIV-1. This conservation of basic residues within the V3 loop is an intriguing phenomenon implying a functionally essential role. One possibility is that cell surface heparan sulfate plays an important part in the mechanisms of cellular infection by HIV-1 (Patel et al., 1993).

Materials and methods

Materials

Porcine intestinal mucosal heparin, sodium salt, was purchased from Sigma Chemical Co., Poole, Dorset, UK. ³H-Heparin (0.55 mCi/mg), prepared by the reduction of porcine mucosal heparin with tritiated sodium borohydride, was purchased from Du Pont NEN, Stevenage, Herts, UK. Chemically modified bovine lung heparins, previously assayed for both anti-HIV-1 activity in vitro and inhibiting the binding of V3 loop-specific Mabs to rgp120 (Rider et al., 1994; Series B) were prepared as described previously (Mulloy et al.,1994). The sulfate content of these preparations has been presented previously (Rider et al.,1994). Two preparations of heparan sulfate, HSI and HSII originally purified by Johnson (1984) were kindly donated by Dr. B. Mulloy, NIBSC, Hertfordshire, UK. HSI, mean Mr 20 kDa, has a sulfate/carboxylate ratio of 1, and an N-acetyl/carboxylate ratio of 0.6, and possesses low anticoagulant activities; whereas HSII, mean Mr 8 kDa, is more heparin-like with sulfate/carboxylate and N-acetyl/carboxylate ratios of 1.9 and 0.2, respectively. The low Mr (4-5 kDa) heparin, Enoxaparin, was purchased from Rhone Poulenc Rorer, Dagenham, UK. Dextran sulfate, Mr 8 kDa was donated by Dextran Products Ltd., Scarborough, Ontario, Canada.

Rgp120 from the HTLV-IIIB isolate of HIV-1 was produced by American Biotechnologies Inc., Cambridge, MA. CHO cell-derived rgp120 from the SF-2 strain of HIV-1 was prepared by the Chiron Corp., Emeryville, CA, USA. Rgp120 from the GB8 strain of HIV-1 was prepared by Dr. R. Daniels, NIMR, Mill Hill, London. Baculovirus-derived rgp120 from the MN strain of HIV-1 was produced by Agmed Corp., Bedford, MA. Human rsCD4, also produced by a baculovirus expression system, was from SmithKline Beecham Pharmaceuticals, King of Prussia, PA. These recombinant proteins were all provided by the UK Medical Research Council AIDS Directed Programme. The sequences of the V3 loops of these proteins is shown in Table I.

Two peptides derived from the V3 loop of the MN isolate of HIV-1 were provided by the AIDS Research and Reference Reagent Program of the National Institutes of Health, USA. As shown in Table I, the 35-mer peptide, catalog number 1841, has an identical V3 loop sequence to rgp120 from HIV-1 isolate MN. The terminal cysteines confer cyclization in solution. The 33-mer peptide, catalog number 1840, has an identical sequence except that it lacks the two terminal cysteines, and thus is not cyclized. The cyclic peptides EVA 7019, which has an identical sequence to 1841 but is from an independent source, and EVA 7026 (see Table I), were obtained from the Medical Research Council AIDS Directed Program.

Assay of binding of radiolabeled heparin to rgp120

The binding of proteins to heparin was assayed by filtration through nitrocellulose as described by Maccarana and Lindahl (1993) but adapted to a 96 well microtiter plate format. Radiolabeled heparin, typically 2 × 10⁵ c.p.m.,∼6 µg, was incubated with rgp120, typically 10 µg in 25 mM Tris–HCl buffer, pH 7.5, in a total volume of 400 µl in Eppendorf tubes. Following incubation for 18–20 h at room temperature on a rotating platform, five 80 µl aliquots of each incubation mixture were rapidly filtered through a reinforced nitrocellulose membrane filter (0.45 µ pore size; Optitran BA-S 85, Schleicher and Schuell, Dassel, Germany) using a 96 well micromanifold (Anderman and Co., Kingston, UK). The wells were washed twice with Tris–HCl buffer, and the filtered areas of the membrane were then cut out. The filter pieces, ∼1cm², were placed in scintillation vials containing 200 µl of 2 M NaCl. The vials were then vortexed and shaken for 1 h before the addition of 4 ml of scintillation fluid (Ecoscint, Packard, Meriden, CT). ³H-radioactivity was then determined on a Beckman LS 6500 scintillation counter. In some experiments, the ability of unlabeled heparin or heparin analogs to displace the binding of labeled heparin was determined. These competitors were incubated in Eppendorf tubes with rgp120 for 1 h with shaking before the addition of ³H-heparin. The procedure described above was then followed.

Acknowledgments

This work was supported by grants from the Medical Research Council AIDS Directed Programme.

Abbreviations

Abbreviations
 
• HIV-1

  human immunodeficiency virus type 1

 
• gp120

  glycoprotein 120kD

 
• rgp120

  recombinant glycoprotein 120kD

 
• Mab

  monoclonal antibody

 
• rsCD4

  recombinant soluble CD4

References