Interaction between Pseudorabies Virus and Heparin/Heparan Sulfate PSEUDORABIES VIRUS MUTANTS DIFFER IN THEIR INTERACTION WITH HEPARIN/HEPARAN SULFATE WHEN ALTERED FOR SPECIFIC GLYCOPROTEIN C HEPARIN-BINDING DOMAIN*

,

been demonstrated to serve as an initial receptor for a number of herpesviruses, including herpes simplex virus type 1 and type 2 (1), pseudorabies virus (PrV) (2), bovine herpesvirus 1 (3,4), varicella zoster virus (5), human cytomegalovirus (6,7), and bovine herpesvirus 4 (8). Thus, evolutionary distant herpesviruses follow a common strategy to initiate infection of susceptible cells. An early study emphasized importance of the charge density of heparin/HS for interaction with herpes simplex virus (9), while recent competition experiments demonstrated that preparations of selectively modified heparin differentially inhibited attachment to cells of herpes simplex virus type 1 and 2 (10,11) and PrV (12). However, several facets of herpesviral interactions with heparin/HS remain obscure. These include involvement of specific structural features of heparin/HS, a possible association of virus binding to HS with further steps of herpesvirus entry into the cells, the importance of HS recognition for the specific cell and tissue tropism of different herpesviruses, and functional aspects of the HS-binding sites of herpesvirus glycoprotein C (gC).
According to one of the molecular concepts of glycosaminoglycan recognition by proteins, consensus sequences comprising clusters of basic and hydropathic amino acid residues have been regarded as primary heparin/HS binding units (13). HS/heparin-binding proteins frequently contain more than one functional cluster of basic amino acid residues (for review, see Ref. 14). Molecular models available for several such proteins show that these clusters are assembled in a specific region on a surface of folded protein to form either an elongated domain that could function in accommodation of linear structures of glycosaminoglycan chains (15,16) or a more complex domain that can bind with high specificity to relatively short stretches within heparin/HS (17,18). Since heparin/HS binding sequences of different proteins carry distinct compositions of amino acid residues that interspace or flank positively charged amino acids, specific clusters of basic residues may differentially contribute toward an overall heparin/HS binding activity or toward a regulation of protein-glycosaminoglycan interaction.
PrV, a pathogen closely related to herpes simplex virus, causes Aujeszky's disease in swine. In addition PrV can cause a deadly disease in many different animals except for equines and higher primates. It has been shown that gC of the virus envelope is the major component that interacts with cell surface HS, thus mediating efficient virus attachment to the cells (2,19,20). gC-null strains of PrV can still infect the cells, albeit with a significantly lower attachment efficiency than wild-type gC-positive virus. The HS-binding domain of PrV gC has been delimited to the amino-terminal one-third fragment of this glycoprotein (21). Mutational analysis of PrV gC revealed that the three clusters of basic amino acid residues at locations 76-RRKPPR-81, 96-HGRKR-100, and 133-RFYRRGRFR-141, respectively, are the major components of the HS-binding domain (22). Interestingly, these clusters seemed to function redundantly, since PrV mutants whose gC contains double amino acid substitutions in any two clusters, i.e. it retains intact only one specific cluster, could still efficiently attach to cell surface HS (22). In the present study we took advantage of this observation to examine the mode of interaction between heparin/HS and PrV mutants that contain HS-binding site of gC dominated by a single specific cluster of basic residues. Our competition experiments and direct binding assays revealed a number of differences in interaction between the respective PrV gC mutants and heparin/HS. Cells and Viruses-Rabbit kidney (RK-13) and Madin-Darby bovine kidney (MDBK) cells were cultivated in Eagle's minimum essential medium (EMEM) supplemented with 4% heat-inactivated fetal bovine serum. The wild-type PrV Becker (PrV-Be) strain (23) and its derivatives PrV580, PrV581, PrV582, and PrV583 (22) were used. Each of these mutants carry several substitutions for specific basic amino acid residues within the three major heparin/HS-binding subunits of the one-third amino-terminal portion of PrV gC (22) (Table I). To determine the virus particle/infectivity ratio, PrV mutants were propagated in RK-13 cells, harvested 24 h after infection, and purified by centrifugation through a three-step discontinuous sucrose gradient as described (19). The viral infectivity (plaque assay) was determined by titration of the purified virus preparations on RK-13 cells. After a 1-h adsorption period at 37°C, the unadsorbed virus inoculum was removed, and the cells were washed two times with EMEM and overlaid with 1% methylcellulose solution in EMEM. Based on the determination of DNA content in the purified virus preparations (20), the number of virus particles to plaque forming unit (pfu) ratios were 2.6 ϫ 10 2 for PrV-Be, 1.8 ϫ 10 3 for PrV580, 3.4 ϫ 10 2 for PrV581, 4.5 ϫ 10 3 for PrV582, and 8 ϫ 10 3 for PrV583.
Isolation of HS Chains-Freshly confluent monolayers of MDBK cells were metabolically labeled for 48 h with 50 Ci/ml Na 2 35 SO 4 in sulfate-free EMEM supplemented with 10% fetal calf serum and antibiotics. Cell-associated HS chains were isolated according to the procedure described by Lyon et al. (31).
Purification of PrV gC-Confluent, dense monolayers of 4-day-old RK-13 cells were infected with PrV, and when the cytopathic effect was advanced, the cells were spun down by low speed centrifugation and stored at Ϫ70°C. The virus was pelleted from the medium by centrifugation at 160,000 ϫ g for 1 h. Both virus and infected cell pellets were combined and resuspended in cold 0.02 M Tris-HCl buffer, pH 7.5, containing 1% sodium deoxycholate, 2% Nonidet P-40, 2 mM EDTA, and 2 mM 4-(2-aminoethyl)benzenesulfonyl fluoride. The mixture was homogenized with several strokes of a Dounce homogenizer and left on ice for 1 h. The unsolubilized material was pelleted by centrifugation at 130,000 ϫ g for 1 h, and the supernatant was circulated for 1 h through an immunosorbent column containing either F2E3.1E8 or F3E3.1C4 anti-PrV gC monoclonal antibody (32). Subsequently the column was washed with 0.02 M Tris-HCl, pH 7.5, containing 0.1% Nonidet P-40, 0.5 M NaCl, and 2 mM EDTA. The adsorbed material was eluted with 0.1 M glycine HCl, pH 2.4, and immediately neutralized with 1 M Tris-HCl, pH 8.0. To exchange the buffer into PBS (137 mM NaCl, 2.7 mM KCl, 8.1 mM Na 2 HPO 4 , 1.5 mM KH 2 PO 4 ), the eluted material was centrifuged over 10,000-Da microcentrifugal concentrator (Palfiltron). Protein concentration was determined according to the standard Lowry method (DC protein assay kit, Bio-Rad).
PrV Attachment in the Presence of Modified Heparin-All experiments were carried out at 4°C (cold room). Serial 5-fold dilutions of the indicated compound in 2 ml of Hanks' balanced salt solution were mixed with 200 l of the same medium containing ϳ200 pfu of PrV and incubated for 15 min. Confluent, dense monolayers of MDBK (3-dayold) cells in six-well plates, precooled for 20 min at room temperature and for 20 min at 4°C, were washed with 2 ml of Hanks' medium and then 1-ml portions of the virus/heparin mixture were added. Following an adsorption for 1 h, the cells were washed twice with 2 ml of Hanks' medium and overlaid with 4 ml of EMEM containing 1% methylcellulose, 2% fetal bovine serum, and antibiotics. The plaques were stained with crystal violet solution after 3 days of incubation at 37°C.
PrV Attachment to Heparinase-or Heparitinase-treated Cells-Confluent monolayers of 2-day-old MDBK cells in six-well plates were washed two times with 2 ml of Hanks' medium, and 1-ml portions of the same medium containing the indicated number of heparinase or heparitinase units were added. Following incubation at 37°C for 1 h and at 4°C for 20 min, the cells were washed two times with 2 ml of cold Hanks' medium, and ϳ200 pfu of PrV in 1 ml of Hanks' medium were added. Following virus adsorption for 1 h at 4°C, the cells were washed twice with 2 ml of Hanks' medium and overlaid with 4 ml of methylcellulose solution. The plaques were stained with crystal violet solution after 3 days of incubation at 37°C.
Binding of Glycosaminoglycans to PrV gC-Purified gC (1.5 g) in 0.2 ml of PBS supplemented with 0.05% bovine serum albumin (BSA) was mixed with ϳ5000 cpm of MDBK cell-specific [ 35 S]HS and incubated for 2 h at room temperature. Bound glycosaminoglycan was trapped on nitrocellulose filters as described (33). For heparin oligosaccharide binding, purified gC (1.5 g) in 0.2 ml of PBS-BSA was mixed with ϳ5000 cpm of [ 3 H]heparin oligosaccharide composed of 4 through 18 monosaccharide units. Following incubation at 4°C for 90 min, the mixtures were transferred to microcentrifugal concentrators (Microsep) with 10,000/30,000-Da cutoff and centrifuged in a horizontal rotor at 3400 ϫ g for 30 min at 4°C. Subsequently using the same centrifugation conditions the filters were washed two times with 0.2 ml of PBS, and the bound heparin oligosaccharides were eluted with 0.2 ml of 2 M NaCl. The eluted fractions were diluted in 1 ml of water and transferred to liquid scintillation vials for quantification of radioactivity.

PrV gC Attachment Mutants Differ in Their Sensitivity to Modified Heparin Compounds-Several
PrV mutants in which the HS-binding site of gC is dominated by a specific cluster of basic amino acid residues (22) were selected for competition experiments (Table I). In PrV580 C-III , PrV581 C-II , and PrV582 C-I two out of the three clusters of basic residues that compose the HS/heparin-binding domain were disrupted by simultaneous replacement of two pairs of basic amino acid residues with neutral amino acids. Consequently, PrV580 C-III carries the 133-RFYRRGRFR-141 cluster (Cluster-III) intact, PrV581 C-II the 96-HGRKR-100 cluster, and PrV582 C-I the 76-RRKPPR-81 cluster as a dominant HS-binding sequence of gC. Of note, these mutants showed reactivity with three different anti-gC monoclonal antibodies (22) that recognize conformationdependent epitopes and that have been reported (20) to be able to inhibit PrV binding to cells. The retention of all three epitopes by the gC mutants reduces the likelihood that amino acid substitutions introduced significant alterations in the folding of gC. Since more profound gC changes such as deletion of the entire 133-RFYRRGRFR-141 sequence or the 96-HGRKR-100 cluster abolished gC reactivity with some monoclonal antibodies (22), such mutants were not included in the present study. PrV583, whose gC contains dysfunctional HS-binding domain due to double substitutions for arginine and/or lysine residue in each of the three clusters of basic amino acids (Table  I), was used as control. As shown earlier on pig kidney cells (22) and repeated in the present study on MDBK cells (Fig. 1) parental PrV-Be as well as PrV580 C-III , PrV581 C-II , and PrV582 C-I were attachment-proficient, whereas the control PrV583 strain demonstrated significantly lower attachment efficiency, manifested by the fact that ϳ65% of input virus could be removed from cell surface by simple washing the cells with Hanks' medium ( Fig. 1). In this respect PrV583 resembles gC-null strains of PrV.
The pattern of inhibition of PrV-Be attachment to MDBK cells by desulfated heparins (Fig. 2; for IC 50 values, see Table  II) resembles that observed previously on RK-13 cells, except that N-desulfated heparin was somewhat more active on the latter cells (12). Thus, singular removal of only one type of sulfate groups affected the competition capacity of heparin only slightly with 6-O-desulfated heparin being the least active competitor followed by 2-O-and N-desulfated heparin ( Fig. 2; compare samples 2, 3, and 4). This weak PrV preference for the specific sulfate groups, i.e. 6-O-sulfates Ͼ 2-O-sulfates Ͼ Nsulfates is further confirmed by the observation that modified heparin sample 6, which contains 6-O-sulfate groups as a major sulfate component left, was more active than sample 5, which retains 2-O-sulfates as predominant groups, and the preparation carrying predominantly N-sulfates (#7). Native heparin poorly inhibited attachment of control PrV583 strain. In the presence of the highest concentration of heparin tested (25 g/ml), an average number of PrV583 plaques expressed as a percentage of plaques formed in the absence of heparin was 89 (Ϯ27) (three experiments).
The effects of modified heparin samples on PrV580 C-III , PrV581 C-II , and PrV582 C-I attachment to MDBK cells are shown in Fig. 2. Although an overall PrV preference for different sulfate groups of heparin, i.e. 6-O-Ͼ 2-O-Ͼ N-sulfates remained similar with all mutant strains, these viruses exhibited different patterns of sensitivity to heparin samples, especially evident when the competition capacities of native heparin versus desulfated heparin preparations are compared. PrV581 C-II , PrV582 C-I , and to lesser extent, PrV580 C-III , were less efficiently inhibited by native heparin than parental PrV-Be strain. However, PrV580 C-III exhibited higher sensitivity than PrV-Be to all the desulfated samples of heparin (compare samples 2-7). In fact both N-and 2-O-desulfated samples appeared to be slightly better inhibitors than native heparin of PrV580 C-III binding. Quite the opposite pattern was observed in PrV581 C-II . Selectively N-, 2-O-, and 6-O-desulfated heparin compounds were ϳ2 orders of magnitude less active competitors of the binding of this mutant than PrV580 C-III . Thus, our results that selectively desulfated heparin samples were poor inhibitors of PrV581 C-II but efficient inhibitors of PrV580 C-III binding suggest that PrV581 C-II may show preference for more extensively sulfated stretches within the HS chain than those required for PrV580 C-III . PrV582 C-I exhibited intermediate sen-sitivity to modified heparins. While selectively N-desulfated and 2-O-desulfated heparin preparations exhibited comparable competition capacities when tested with PrV580 C-III , PrV581 C-II, and the parental strain, there was a 5-fold difference between relative potencies of these samples with PrV582 C-I ( Fig.  2; Table II). This suggests that an overall binding preference for the specific sulfate groups might be somewhat changed in this mutant.
PrV gC Attachment Mutants Differ in Their Abilities to Bind Heparin Oligosaccharides of Different Size-Differential susceptibility of PrV mutants to desulfated heparin samples suggested that specific subunits of the HS-binding domain of gC might show preference for different structural features of heparin/HS. To verify this interpretation, gCs purified by immunoaffinity chromatography from the respective PrV mutants were tested for ability to bind 3 H-labeled size-fractionated heparin oligosaccharides composed of 4 through 18 monosaccharide units. Mutated gCs showed different patterns of binding heparin fragments (Fig. 3). PrV-Be gC bound significant proportions of 8-mer and near-maximal levels of 10-mer and larger fragments. Of all the mutants, only PrV581 C-II demonstrated a similar binding pattern as the wild-type protein, although slightly larger fragments were required for efficient binding. The remaining mutants exhibited a clearly lower capacity to bind to short heparin fragments and were demonstrating a linear dependence of affinity and fragment length.
PrV583 gC bound little but definite amounts of some heparin fragments. In a separate experiment we noticed that gC from PrV583 weakly interacted with heparin oligosaccharides, since repeated cycles of washing the gC-oligosaccharide complexes that had been retained on filters continuously removed some   (22). Indicated are double amino acid substitutions introduced within the specific clusters of basic amino acid residues of the HS-binding domain of viral gC. Underlined are clusters of basic amino acids that dominate in gC of the respective mutant.
amounts of labeled oligosaccharide (data not shown). This is in line with the observation that a significant proportion of input PrV583 virions can be removed from cell surfaces by simple washing the cells with PBS (Fig. 1).
PrV gC Attachment Mutants Differ in Their Binding to He-paritinase-or Heparinase-treated cells-Knowing that each of specific clusters of basic amino acid residues could separately support efficient PrV attachment to cell, we were interested in a purpose for the presence of redundant HS-binding subunits in wild-type PrV gC. To this end gCs purified from the respective mutant and wild-type strain were tested for their ability to bind HS isolated from MDBK cells. Under conditions described under "Experimental Procedures," PrV-Be gC bound 32% (Ϯ4%), PrV580 C-III gC 7% (Ϯ1%), PrV581 C-II gC 10% (Ϯ1%), and PrV582 C-I gC bound 6% (Ϯ2%) of input HS (three experiments). Thus, the presence of all three clusters of basic residues endows parental gC with high HS binding capacity. To examine whether and under which conditions the high HSbinding ability of parental gC is beneficial for PrV interaction with cell surface HS, we compared the binding of wild-type and mutant strains to cells pretreated with relatively low amounts of HS-degrading enzymes (Fig. 4). PrV gC mutants differed in their abilities to attach to and to infect of enzyme-treated cells. Similar tendencies were observed with both heparinase and heparitinase. The binding of PrV582 C-I was most affected, then followed PrV580 C-III , and the binding of PrV581 C-II was only slightly reduced as compared with parental strain PrV-Be. Pretreatment of cells with heparinase and heparitinase had little and moderate effects, respectively, on binding of the control PrV583 strain. Note that the two PrV mutants that have been the most affected by heparinase/heparitinase, i.e. PrV582 C-I and PrV580 C-III required relatively long heparin oligosaccharides for efficient binding (Fig. 3).

DISCUSSION
It has been demonstrated that the HS-binding domain of PrV gC is composed of three discrete clusters of basic amino acid residues and that each of these clusters could independently support PrV attachment to cells (22). In this work we attempted to define whether this design of the HS-binding domain would provide PrV gC with an ability to interact with different structural features of heparin/HS. To this end, we have used PrV mutants (22) in which gC contains intact only a single specific cluster. Two remaining clusters were disrupted by double substitutions of neutral amino acids for basic amino acid residues. Due to the fact that only two out of several basic amino acids in each cluster had been altered, a residual binding activity of the mutated cluster cannot be excluded. However, we restricted our analysis to these mutants because (i) amino acid substitutions, in contrast to the entire cluster deletion, introduced little or no changes in the conformation of gC as detected with monoclonal antibodies, and (ii) mutations caused significant functional impairments of the clusters, since the mutant virus in which all three clusters of the HS-binding domain of gC were disrupted poorly attached to cells. In addition we have used modified heparin compounds to relate specific amino acid sequence of the respective cluster with binding preferences for specific sulfate groups and chain length of heparin.
Our results suggest that PrV mutants whose gC contains a HS-binding site dominated by a single specific cluster of basic residues may preferentially interact with different structural features of heparin/HS. This stems from the observations that mutant viruses exhibited different patterns of sensitivity to selectively desulfated heparin compounds in cell attachment assays and that gC, purified from the respective mutants, differed in their abilities to bind size-fractionated heparin oligosaccharides. Inhibition of attachment of the PrV581 C-II mutant to cells was strongly dependent on the presence of all, i.e. N-, 2-O-, and 6-O-sulfate groups in the heparin preparation. When either of the sulfate group was removed, a drastic reduction in the potency of the inhibitor was observed, suggesting  Table II. that the C-II cluster of the HS-binding domain of gC may require a highly sulfated stretch within HS for an efficient interaction to occur. gC isolated from PrV581 C-II bound heparin oligosaccharides almost as efficiently as wild-type gC. Heparin 8 -10-mer fragments were the shortest oligosaccharides binding and a near-maximal binding capacity reached with oligosaccharides composed of 12 monosaccharide units.
In a sharp contrast to PrV581 C-II , selective N-, 2-O-, or 6-Odesulfation of did not decrease heparin ability to inhibit attachment to cells of PrV580 C-III . Interestingly, two heparin derivatives, i.e. N-and 2-O-desulfated preparations, were even somewhat more efficient than native heparin at inhibition of this mutant. This suggests that a high density of sulfate groups within glycosaminoglycan chain is not required for interaction with PrV580 C-III . In addition gC, isolated from PrV580 C-III , required longer heparin oligosaccharides than gC from PrV581 C-II for effective binding. Efficient inhibition of PrV580 C-III by desulfated heparin compounds may correlate with the presence of clusters of acidic amino acid residues ETFEV and SPDADPE flanking the C-III cluster (133-RFYR-RGRFR-141; see Table I). Since HS chains are known to be composed of blocks of virtually nonsulfated disaccharide units that alternate with stretches of moderately/extensively sulfated disaccharides, it is likely that acidic residues may direct, due a low charge repulsion with weakly sulfated blocks of HS, the 133-RFYRRGRFR-141 cluster toward interaction with somewhat more restricted sequences within HS chain. For the same reason PrV580 gC may, to some extent, be repulsed by native heparin in which sulfate groups are known to be evenly distributed over the chain. It should be noted that in contrast to PrV580 C-III , the clusters of negatively charged amino acids are not present near the dominant HS-binding sequences in PrV581 C-II and PrV582 C-I . It has been recently suggested that the ability of some chemokines to interact with subfractions of heparin correlated with the presence of acidic residues within a putative heparin/HS-binding domain (34).
Yet another pattern of sensitivity to desulfated heparin samples was observed with remaining mutant PrV582 C-I . While selective N-desulfation of heparin had no effect, 2-O-or 6-Odesulfation decreased an ability of heparin to inhibit PrV582 C-I attachment to cells. This suggests that the binding of PrV582 C-1 is more dependent on the O-sulfates than N-sulfates. In addition gC purified from PrV582 C-1 less efficiently than gC of the other mutants bound heparin oligosaccharides;  an extensive binding was not demonstrated, even with the longest oligosaccharide fragments.
Although each of the specific HS-binding clusters of gC could separately support efficient PrV attachment to cells, we found that in comparison with wild-type PrV gC, PrV581 C-II gC bound ϳthree times less and gCs of PrV580 C-III and PrV582 C-I bound ϳfour to five times less of HS isolated from MDBK cells. Thus, it is likely that the individual HS binding capacities of the respective clusters could potentiate through an additive or synergistic effects an overall HS binding capacity of wild-type gC. Further insight into functional significance of specific HSbinding cluster of PrV gC was obtained in the experiments of PrV attachment to cells digested with HS-degrading enzymes. Both PrV582 C-I and PrV580 C-III poorly attached to cells pretreated with certain, relatively low, enzyme amounts, whereas under the same conditions the binding of PrV581 C-II and wildtype PrV was unaffected. Since similar tendencies were observed with both heparinase that cleaves HS chain in regions of high sulfation, i.e. containing GlcNSO 3 (Ϯ6-OSO 3 )-Ido(2-SO 3 ) disaccharides and heparitinase, which attacks HS stretches of low sulfation at hexosaminidic linkages adjacent to glucuronic acid residue, it is likely that truncated chains of HS that remained on proteoglycan core following enzyme treatment of cells poorly interacted with PrV582 C-I and PrV580 C-III . It is perhaps interesting to note that these two mutants required relatively long heparin fragments for efficient binding. These results also suggest that the HS-binding subunits of gC, although seemingly redundant with respect to PrV binding to in vitro cultured cells, may facilitate PrV infection of cells that express low amounts or undersulfated HS. It is known that HS expressed on the surfaces of different cells may differ with respect to the overall quantity produced, chain length, and its fine structure (35,36). PrV infection in vivo frequently concerns polarized cells. It has been shown that the specific sorting mechanisms in certain polarized cells may favor a low expression of HS on the apical surface (37).
The molecular model of some heparin/HS-binding proteins revealed that discrete clusters of basic residues, or the same clusters originating from separate monomers of a particular protein, frequently juxtapose each other to form an elongated domain that could favorably accommodate linear blocks of HS/ heparin chain (15,16,38). If this is the case also with PrV gC, then mutant viruses that retain only singular cluster might preferably accommodate shorter fragments of heparin oligosaccharides than wild-type gC that contains three distinct clusters. However direct binding experiments revealed that wildtype gC bound much shorter oligosaccharides than PrV582 C-I and PrV580 C-III and only slightly shorter fragments of heparin than PrV581 C-II gC. Thus, either spatial arrangement of all three clusters endows wild-type gC with flexibility to adopt relatively short heparin fragments or the C-II cluster plays a dominant role. In this respect it is noteworthy that PrV581 C-II attached to heparinase/heparitinase-treated cells almost as efficiently as wild-type virus. This suggests that some important attachment functions related to efficient binding to HS are, at least in part, retained in PrV581 C-II gC. However, competition experiments suggested that the binding of this mutant to cells may, in contrast to wild-type virus, require more densely sulfated stretches of HS. Based on this we postulate that the C-II cluster may represent the main HS-binding site of PrV gC, supporting virus attachment to short, heparin-like stretches within HS chains. The likely function of the C-I and C-III clusters, which exhibited clearly different binding preferences, would be to increase gC ability to bind to different structural features or regions within the HS chain including those of low/moderate sulfation. The molecular model of lipoprotein lipase revealed that several clusters of basic residues were congregated on one face of the molecule (16). Based on this as well as on the results of mutational analysis of lipoprotein lipase (39,40), it was proposed (16) that the ensemble of clusters may be involved in vivo in interaction with HS and that the domination of a specific cluster may depend on the nature of available HS chain. For the virus it would be of benefit to adopt to different types of host cell HS by having suitable adapter molecules for different HS motifs. Finally, several abilities of different PrV mutants, such as attachment to heparitinase/ heparinase-treated cells, binding to heparin oligosaccharides, and binding to isolated HS correlated with the values of their specific infectivity. This indicates that the specific interaction of PrV with HS is important for an efficient infection of cells by the virus.