Soluble LDL Minireceptors MINIMAL STRUCTURE REQUIREMENTS FOR RECOGNITION OF MINOR GROUP HUMAN RHINOVIRUS* human low density lipoprotein

Soluble minireceptors with less than seven ligand binding repeats (as are present in the native membrane receptor) were expressed in Sf9 insect cells with a hexa-His tag fused to the C terminus. The recombinant truncated proteins were affinity purified from the tissue culture supernatants by Ni-NTA column chromatography. Minireceptors with more than two repeats bound to rabbit b very low density lipoprotein and could thus be further purified by affinity chromatography. Binding and cell protection assays indicated that two ligand binding repeats are sufficient for attachment of minor group human rhinoviruses to im-mobilized receptors, whereas at least three ligand binding repeats are required to protect HeLa cells against viral infection. protocol Technologies). Bacmid DNAs encoding the various LDLR fragments were isolated and used for Lipofectin-mediated transfection of Sf9 insect cells. Recombinant bacu- loviruses were harvested, amplified, and used for expression of soluble minireceptors. His-tagged proteins in the cell culture supernatants (50 m l) were detected by electrophoretic transfer to nitrocellulose membranes (Schleicher and Schuell) from SDS-polyacrylamide gels run under reducing conditions. The membranes were blocked with 20 m M Tris-HCl (pH 7.5), 150 m M NaCl, 2 m M CaCl 2 (TBS-Ca) containing 2% Tween 20 (blocking buffer) for 1 h at 4 °C, washed twice with 0.1% Tween 20 in TBS-Ca (incubation buffer) for 10 min, and probed with 10 ml of incubation buffer containing HRP-conjugated Ni-NTA at a final dilution of 1:10 000 for 3 h at 4 °C. After washing twice with incubation buffer, HRP was revealed by chemiluminescence.

With 102 serotypes, human rhinoviruses (HRVs) 1 constitute the largest genus within the family Picornaviridae. They are composed of a small (ϳ35 nm) icosahedral shell containing 60 copies each of four viral capsid proteins (VP1 through VP4) and a positive (messenger sense) single strand RNA genome (for a recent review, see Ref. 1). Based on competition for cellular receptors, a major group (91 serotypes) and a minor group (10 serotypes) were defined (2,3). Major group HRVs attach to the intercellular adhesion molecule 1 (ICAM-1) (4 -6), and minor group viruses bind to several members of the low density lipoprotein receptor (7)(8)(9)(10); the only exception (HRV87) attaches to an unknown glycoprotein (3). Whereas the site of attachment of major group HRVs to ICAM-1 has been mapped to the canyon, a deep cleft encircling the 5-fold axes of icosahedral symmetry (11), the binding site of the low density lipoprotein receptor (LDLR) on minor group viruses remains elusive.
The most prominent common structural motive of all members of the LDLR family are various numbers of complement type-A repeats, about 40 amino acids in length, containing six cysteines each, which are involved in disulfide bridges; the C termini exhibit clusters of negatively charged amino acid residues. These repeats form the ligand binding domain and interact with a great number of structurally and functionally unre-lated ligands (for reviews, see Refs. 12 and 13). Whereas LDLR and the very low density lipoprotein receptor (VLDLR) contain seven and eight consecutive ligand binding repeats, respectively, LDLR-related protein (LRP) and megalin contain four clusters of various numbers of these repeats (for a discussion of the structural features of these proteins, see Ref. 14). The ligand binding domain is followed by (or interspersed with, in the case of LRP and megalin) regions with similarity to epidermal growth factor precursor; a region that is heavily O-glycosylated is present in some of the receptors and precedes the transmembrane segment. Finally, the short cytoplasmic domain contains typical internalization signals (NPXY) responsible for clustering of the receptors in coated pits.
Previously, we expressed the entire ligand binding domain of the low density lipoprotein receptor in a soluble form and demonstrated that this region is sufficient for recognition of minor group HRVs (10). We have also shown that a recombinant fragment of human very low density lipoprotein receptor, which contains eight complement type repeats, binds minor group HRVs as well (9). In order to identify the smallest structure element of the LDLR molecule required for viral recognition, we expressed soluble LDLR fragments encompassing various numbers of the complement type repeats in the baculovirus system and tested their binding to HRV2, a representative of the minor receptor group. In addition, the potential of these soluble minireceptors with respect to protection of HeLa cells against viral infection was compared. In the present communication, we show that fragments with two or more complement type repeats are recognized by minor group HRVs when immobilized on PVDF membranes and that the presence of more than two repeats is required to inhibit viral infection of HeLa cells.

EXPERIMENTAL PROCEDURES
Materials-All chemicals were purchased from Sigma, unless otherwise specified. The Bac-to-Bac expression system and Lipofectin were from Life Technologies; restriction enzymes and T4-DNA ligase were from New England Biotechnology; Pfu polymerase was from Stratagene; and human rhinovirus serotype 2 (HRV2) was originally obtained from ATCC and plaque-purified before use. Virus was metabolically labeled with [ 35 S]methionine and [ 35 S]cysteine (ARS110, American Radiolabeled Chemicals) as described (15). Alkaline phosphatase-conjugated goat anti-rabbit antibodies were from Southern Biotechnology; nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate were used as substrates. Ni-NTA beads and HRP-Ni-NTA conjugate were from Qiagen; for detection, the chemiluminescence system from Pierce (Supersignal) was employed. CNBr-activated Sepharose was from Amersham Pharmacia Biotech.
Generation of the Transposition Vector pTM1-A derivative of pFast-Bac1 (Life Technologies) was constructed by replacing the SnaBI-KpnI fragment for the EcoRV-KpnI fragment excised from pVT_Bac_His2 (10). The resulting vector (pTM1) harbors the polyhedrin promoter, which directs expression of the melittin signal peptide. A multiple cloning site allows for insertion of the gene of interest between the melittin sequence and a C-terminal collagenase site followed by a hexa-His tag. As a pUC based vector, it is easily amplified under ampicillin selection; the presence of the transposon sites makes it compatible with the Bac-to-Bac expression system. * This work was supported by Grant P12189-MOB from the Austrian Science Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The pTM1 plasmids were used for transposition in DH10Bac cells according to the manufacturer's protocol (Life Technologies). Bacmid DNAs encoding the various LDLR fragments were isolated and used for Lipofectin-mediated transfection of Sf9 insect cells. Recombinant baculoviruses were harvested, amplified, and used for expression of soluble minireceptors. His-tagged proteins in the cell culture supernatants (50 l) were detected by electrophoretic transfer to nitrocellulose membranes (Schleicher and Schuell) from SDS-polyacrylamide gels run under reducing conditions. The membranes were blocked with 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 2 mM CaCl 2 (TBS-Ca) containing 2% Tween 20 (blocking buffer) for 1 h at 4°C, washed twice with 0.1% Tween 20 in TBS-Ca (incubation buffer) for 10 min, and probed with 10 ml of incubation buffer containing HRP-conjugated Ni-NTA at a final dilution of 1:10 000 for 3 h at 4°C. After washing twice with incubation buffer, HRP was revealed by chemiluminescence.
Ni-NTA Purification of Soluble Low Density Lipoprotein Minireceptors-Expression and purification of recombinant proteins was performed as described previously (10). Briefly, Sf9 insect cells grown in 0.4-liter suspension cultures were infected with recombinant baculoviruses at a multiplicity of infection of 5. At 90 h postinfection, the cell culture supernatants were cleared by low speed centrifugation and passed over Ni-NTA affinity columns (3 ml). Bound proteins were eluted with 15 ml of 250 mM imidazole in TBS-Ca. Amino-terminal sequencing of proteins electrophoretically transferred onto PVDF membranes was carried out with an instrument from Applied Biosystems.
Ligand Binding Properties of Soluble LDL Minireceptors-Recombinant proteins were separated on SDS-polyacrylamide gels under nonreducing conditions and electrophoretically transferred to PVDF membranes (Millipore). The membranes were incubated in blocking buffer for 1 h at 4°C, washed twice with incubation buffer for 10 min, and probed with 10 ml of incubation buffer containing 2 ϫ 10 4 cpm/ml of 35 S-labeled HRV2 for 1.5 h at 4°C. The membranes were again washed for 10 min with incubation buffer, dried, and exposed to Kodak Biomax film for 24 h.
Cell Protection Assay-Protection of HeLa cells from viral infection with HRV2 was assayed as in (10) with minor modifications. Cell culture supernatants (20 l) from recombinant baculovirus infected Sf9 cells were mixed with 80 l of infection medium (minimal essential medium containing 2% fetal calf serum and 30 mM MgCl 2 ), and serial 2-fold dilutions were made in the same medium. To each dilution (50 l), 100 TCID 50 of HRV2 in 50 l of infection medium was added, and the mixtures were incubated for 1.5 h at 34°C. They were then transferred onto subconfluent Rhino HeLa cells (Flow Laboratories) in 96well plates (containing 100 l of infection medium) and incubated for 3 days at 34°C. Cells remaining attached to the plastic were stained with Amido Black (0.1% in acetic acid/methanol/water, 10:40:50, v/v). Controls used were cell culture supernatants from mock-infected and from wild type baculovirus-infected Sf9 cells and tissue culture media (In-sectXpress and HeLa infection medium).
Purification of Recombinant Soluble LDLR Minireceptors-To specifically enrich for minireceptors active in ligand binding, Ni-NTA affinity purified recombinant proteins were subjected to chromatography on ␤-VLDL columns. ␤-VLDL was prepared from 150 ml of blood col-lected from three bleedings of a rabbit fed with a high cholesterol diet. Fifteen ml of 0.15 M NaCl, 0.1 M EDTA (pH 8.0) were added, and cells were removed by centrifugation in a Heraeus Megafuge 1.0 at 3000 rpm for 30 min. The plasma was transferred into quick seal tubes (Beckman) and centrifuged in a Type 45Ti rotor at 35000 rpm for 16 h. ␤-VLDL, concentrated as a floating cushion on top of the tube, was saved and gently resuspended in 15 ml of phosphate-buffered saline (pH 7.5). ␤-VLDL (30 mg, as determined with the SDS-Lowry assay (17) using bovine serum albumin as standard) was coupled to 8 ml of CNBractivated Sepharose gel according the manufacturer's protocol. Ni-NTA-purified receptor fragments (from 100 ml of Sf9 suspension culture) were passed over 1 ml of ␤-VLDL-Sepharose and washed with 0.5ϫ TBS-Ca, and bound proteins were eluted with 1 M NH 3 in 0.5ϫ TBS-Ca. Fractions (1 ml) were then concentrated to 0.5 ml, and NH 3 was removed in a Speedvac concentrator.
Determination of the Minimal Inhibitory Concentration (MIC 50 ) with Respect to HRV2 of Soluble LDL Minireceptors-Forty l of fraction 2 (purified rLDLR 1-7 h, rLDLR 1-5 h, rLDLR 3-5 h, and rLDLR 3-7 h) and fraction 9 (used as a control) from the ␤-VLDL columns that had been loaded with the various minireceptors (as exemplified in Fig. 6 for rLDLR 1-5 h), were mixed with 60 l of infection medium, and 2-fold serial dilutions were made in a 96-well plate. Fifty l of infection medium containing 100 TCID 50 HRV2 were added to each well and incubated for 1.5 h at 34°C. The mixture was transferred onto HeLa cell monolayers in 96-well plates and cell lysis was determined after 5 days of incubation at 34°C. Cells remaining attached to the plastic were stained with Amido Black and dissolved in 1 M NaOH, and A 560 was determined in a microplate reader. Complete cell lysis was set to 0% protection, and cells were grown in the absence of virus to 100% protection. The concentration of the respective recombinant minireceptor affording 50% cell protection corresponds to the MIC 50 (18). Protein concentrations were determined with the BCA assay (Pierce) using bovine serum albumin as a standard.

Expression of Soluble Low Density Lipoprotein Receptor
Fragments-DNA fragments encoding various numbers of complement type repeats of human LDLR (as shown in Fig. 1) were amplified by polymerase chain reaction from the plasmid pTZ1, a derivative of pLDLR2 containing the entire LDLR cDNA sequence (16,19). Primers (summarized in Table I) were synthesized as to contain BamHI or XmaI restriction sites to allow for subsequent cloning into the transfer plasmid pTM1. Recombinant plasmids were selected, and DNA was isolated and used for transposition into the bacmid DNA by transfection into DH10Bac. White colonies were selected, and correct insertion was verified by restriction analysis. The recombinant bacmids were used for Lipofectin-mediated transfection of Sf9 insect cells. Recombinant baculoviruses were used for infection of Sf9 cells grown in 0.4-liter suspension cultures as described previously (10). In accordance with earlier observations with the recombinant LDLR fragment rLDLR 1-7 h, most of the virus binding activity was recovered in the tissue culture supernatant; therefore, only the supernatants were saved at 90 h postinfection. Samples (20 l) of the cleared supernatants were subjected to SDS-polyacrylamide gel electrophoresis followed by electrophoretic transfer to a PVDF membrane. The presence of His-tagged proteins in the tissue culture supernatants was then verified with HRP-conjugated Ni-NTA followed by chemiluminescence detection. As seen in Fig. 2, infection of the insect cells with the various recombinant baculoviruses resulted in the expression of soluble His-tagged polypeptides with mobilities roughly reflecting the number of complement type repeats expected to be present in the fusion proteins. However, as observed earlier (12,20,21), the apparent molecular masses substantially deviated from those calculated from the amino acid sequences (Fig. 1). For example, human LDL receptor (calculated molecular mass, 95,375 Da) migrates with an apparent molecular mass of 164 kDa (see Ref. 20). Possible Nglycosylation at the sites indicated in Fig. 1 (9). We thus asked whether the various soluble LDL minireceptors (Fig. 1) secreted into the cell supernatant of recombinant baculovirusinfected Sf9 cells (Fig. 2) were also capable of protecting HeLa cells from infection with a minor receptor group rhinovirus (HRV2). HRV2 at 100 TCID 50 was incubated with dilutions of the Sf9 cell supernatants for 1.5 h at 34°C, and HeLa cells grown in 96-well plates were challenged with the mixtures. Upon incubation at 34°C for 72 h, cells remaining attached to the plastic surface were stained with Amido Black (10). As a control, HRV2 was preincubated with tissue culture supernatant obtained from Sf9 cells that had been infected with wild type baculovirus. In order to exclude any effect of the media on the viability of the HeLa cells, control incubations were also carried out in the absence of virus. As seen in Fig. 3, all tissue culture supernatants containing minireceptors encompassing more than two complement type repeats protected the cells against viral infection in a concentration-dependent manner, although to different degrees. Whereas rLDLR 1-5 h afforded the highest and rLDLR 3-5 h the lowest protection, the other minireceptors were intermediate under these experimental conditions. However, this experiment does not allow for a quantitative estimate of the protection potency because it does not take into account differences in expression yield and the fraction of denatured protein present in the various minireceptor preparations. In order to allow for a quantitative comparison of the cell-protecting activity, purification was attempted.
Purification of Recombinant Soluble LDL Minireceptors-The hexa-His tag fused to the C terminus of the various recombinant minireceptors was then used for purification on Ni-NTA columns as described under "Experimental Procedures." Proteins eluted from the columns were analyzed by SDS-polyacrylamide gel electrophoresis under reducing (Fig. 4A) and nonreducing (Fig. 4B) conditions followed by Coomassie Blue staining. It can be inferred from the gel run under reducing conditions that all minireceptors were obtained in comparable yields and migrated as single, well resolved bands reflecting the number of the respective repeats; however, as pointed out above, the apparent molecular masses substantially deviated from those calculated from the amino acid compositions. A completely different picture was obtained upon analysis under nonreducing conditions (Fig. 4B). Most of the proteins migrated as a smear or as badly resolved bands, indicating substantial heterogeneity, aggregation, and misfolding.
For N-terminal sequencing, proteins were separated under reducing conditions and electrophoretically transferred to a

FIG. 1. Primary structure of recombinant soluble human low density lipoprotein minireceptors as expressed in Sf9 insect cells.
The entire LDL receptor with the boundaries between the various domains and the respective complement type repeats, as given in the Swissprot data bank (AC P01130), is schematically shown at the top. The melittin signal sequence (m) (not present in the mature proteins); the respective complement type repeats; cysteine-rich epidermal growth factor repeats A, B, and C; putative glycosylation sites within the ligand binding domain (arrowhead); and the hexa-His tag are indicated. Molecular mass calculations are based on N-terminal sequence determination of the mature proteins and thus exclude the melittin signal sequence but include the His tag (see also Ref. 10). a g was missing in the oligonucleotide used for amplification due to a synthesis error; therefore, the correct reading frame was regenerated by digestion with BamHI followed by filling in with Klenow enzyme and religation (for explanation, see text).

FIG. 2. His-tagged polypeptides are released into the supernatant of recombinant baculovirus-infected Sf9 insect cells.
Fifty-l samples of cleared cell culture supernatants harvested from Sf9 cells infected with various recombinant baculoviruses harboring LDL minireceptor sequences at 90 h postinfection were run on a 12% (A) or 15% (B) SDS-polyacrylamide gel under reducing conditions. Proteins were electrotransferred onto a nitrocellulose membrane. Soluble Histagged proteins were detected with HRP-conjugated Ni-NTA (diluted 1:10 000) followed by chemiluminescence substrate and exposure for 3 min to x-ray film. Complement type repeats present in the minireceptors are indicated at the top of the lanes. The positions of marker proteins run on the same gel are also shown. PVDF membrane. As already shown for rLDLR 1-7 h (10) signal peptidase cleavage in rLDLR 3-5 h, rLDLR 2-3 h, rLDLR 3-4 h, and rLDLR 4-5 h was found to occur at ISYIYA2DR. For rLDLR 1-5 h, the cleavage site was not determined but was assumed to be identical to that previously found for rLDLR 1-7 h (10).
We have previously shown that HRV2 attaches to LDLR, LRP (7), vitellogenin receptor (8), recombinant soluble human VLDLR (9), and recombinant soluble rLDLR 1-7 h (10) in ligand blots from SDS-polyacrylamide gels run under nonreducing conditions. This well established technique, which has also been extensively used for the detection of ligand binding to LDL receptor and to other members of this protein family (see, for example, Refs. 22 and 23) was thus employed to investigate whether the smaller LDL minireceptors would also bind HRV2. Moreover, we asked whether all conformationally different forms present in the preparation of one given minireceptor (see Fig. 4B) would be recognized by rhinoviruses. Ni-NTA-purified proteins were thus separated on a SDS-polyacrylamide gel under nonreducing conditions as in Fig. 4B and electrophoretically transferred to a PVDF membrane. The membrane was incubated with 35 S-labeled HRV2 followed by autoradiography. As seen in Fig. 5, the virus only bound to distinct molecular forms present within each minireceptor preparation; however, there was no apparent correlation between these virus-binding proteins and the major proteins stained with Coomassie Brilliant Blue (see Fig. 4B). The material active in virus binding thus represents only a minor constituent in the Ni-NTA-purified minireceptor preparation.
Additional enrichment of minireceptors in a native conformation was obtained with rabbit ␤-VLDL covalently attached to Sepharose. This natural ligand of LDLR, VLDLR, and LRP was previously used for the purification of vitellogenin receptor (24). Ni-NTA-purified minireceptors were thus applied onto ␤-VLDL Sepharose, and the retained material was eluted with 1 M NH 3 in 0.5ϫ TBS-Ca. The purification procedure was monitored by SDS-PAGE under reducing and nonreducing conditions followed by Coomassie Blue staining and by ligand blotting with HRV2; it is exemplified for rLDLR 1-5 h in Fig. 6. Analysis of the original sample (S), the flow-through (F), and the pool of eluted fractions 1-9 (E) by SDS-polyacrylamide gel electrophoresis under reducing conditions (Fig. 6A) showed that most of the protein was not retained on the column and was thus recovered in the flow-through. Although the amount of protein in the pool of the eluted fractions (E) was below the detection limit of Coomassie Blue staining, a faint band could be seen in fractions 2 and 3 upon separate analysis. When the same samples were analyzed under nonreducing conditions (Fig. 6B), several bands were seen in the original sample (S) and the flow-through (F) (compare also to Fig. 4B). Again, no staining was apparent in the pool of the eluate (E), but a faint band now appeared in fraction 2. This indicates that within the many differently folded and aggregated conformers, only a small fraction exhibited the correct native structure enabling binding to ␤-VLDL. Proteins in the samples were then sepa-   Fig. 4B and electrophoretically transferred to a PVDF membrane. Virus binding activity was revealed with 35 S-labeled HRV2 followed by autoradiography. Positions of marker proteins run on the same gels are indicated. rated under the same conditions as in the experiment shown in Fig. 6B and transferred electrophoretically onto a PVDF membrane. Virus binding was detected by incubation with HRV2 followed by rabbit anti-HRV2 antiserum and alkaline phosphatase staining. As seen in Fig. 6C, virus binding to a protein migrating between the 29-and 45-kDa makers was about equal in the original sample (S) and in the pool of the eluted fractions (E), with no binding activity being found in the flow-through (F). Activity was strongly enriched in fractions 2 and 3, indicating a substantial increase in specific virus binding activity.
Determination of the MIC 50 of the Various Recombinant Soluble Minireceptors toward HRV2-To determine the cell protection capacity of the various minireceptors in a quantitative manner, Ni-NTA and ␤-VLDL Sepharose-purified rLDLR 1-7 h, rLDLR 1-5 h, rLDLR 3-5 h, and rLDLR [3][4][5][6][7] h were incubated at decreasing concentrations with HRV2 at 100 TCID 50 as described above for the tissue culture supernatants (see Fig. 3). HeLa cells in 96-well plates were challenged with the mixtures and incubated for 5 days at 34°C. Cells were stained with Amido Black, which was then dissolved in 1 M NaOH, and the color intensity was measured with a microplate reader (10, 25). As seen in Fig. 7, rLDLR 3-7 h afforded the highest cell protection efficiency reducing cell damage to 50% of the control at a concentration of 0.3 g/ml, whereas the lowest protection was seen for rLDLR 3-5 h, with a value of 3 g/ml. The other recombinant proteins revealed intermediate values. Because the tissue culture supernatants containing minireceptors with only two repeats had not shown any cell protective effect at the lowest dilution (Fig. 3) and the Ni-NTA enriched material could not be further purified due to lack of binding to ␤-VLDL, they were excluded from this analysis. DISCUSSION We have previously shown that representatives of the minor group of human rhinoviruses bind LDLR, LRP, VLDLR, and vitellogenin receptor (7)(8)(9)(10). By expression of soluble human LDLR lacking the epidermal growth factor precursor homology domain, we then demonstrated that the ligand binding domain was sufficient for virus recognition and inhibition of infection (10). In the present work, we analyzed the LDL receptor in more detail and investigated how many complement type repeats were necessary for the structural integrity of the virus binding site.
Using the Bac-to-Bac system we expressed LDL minireceptors with several combinations and numbers of ligand binding repeats in Sf9 insect cells and purified them from the cell culture supernatant by affinity chromatography on Ni-NTA columns. Analysis of the recombinant proteins by reducing SDS-PAGE revealed a high electrophoretic purity and homogeneity (Fig. 4A). However, under nonreducing conditions, most of the fragments migrated as broad weakly resolved bands, indicating extensive heterogeneity and aggregation. Apparently, correct assembly of the large number of disulfide bridges in these truncated receptors constitutes a major problem for the protein folding machinery in these cells. The heterogeneity is mostly due to incorrect disulfide bridges because boiling in SDS in the absence of reducing agents did not change FIG. 6. Retention of recombinant rLDLR 1-5 h on a rabbit ␤-VLDL Sepharose column. Recombinant protein purified by Ni-NTA column chromatography was applied onto a rabbit ␤-VLDL Sepharose column. Equivalent aliquots from the original sample (S), from the flow-through (F), and from the eluted material (pool of fractions 1-9) (E) were then analyzed by 12% SDS-PAGE under reducing (A) and nonreducing (B) conditions. Individual fractions from the eluate were also analyzed. Either gels were stained with Coomassie Brilliant Blue (A and B) or the proteins were transferred to a PVDF membrane (C), and virus binding activity was revealed by incubation of the blot with a crude virus preparation (5 ϫ 10 7 TCID 50 in total) followed by immunodetection of bound virus as described in Ref. 10. Maker proteins transferred to the PVDF membrane were stained with Amido Black.

FIG. 7. Minimal inhibitory concentration of various soluble recombinant human LDL minireceptors toward HRV2.
Ni-NTA/ ␤-VLDL-purified minireceptors (rLDLR 1-7 h, rLDLR 1-5 h, rLDLR 3-7 h, and rLDLR 3-5 h) were incubated with 100 TCID 50 of HRV2 for 1.5 h at 34°C at the final concentrations shown as determined with the BCA assay using bovine serum albumin as a standard. HeLa cell monolayers in 96-well plates were then challenged with the mixtures and incubated for 5 days at 34°C. The supernatants together with floating cells were removed, and firmly attached cells were stained with Amido Black. The stain was solubilized in 1 M NaOH; A 560 was determined with a microplate reader and was taken as a measure for the integrity of the monolayer and hence for cell protection. The value determined for cells challenged with virus in the absence of any minireceptor was taken as 0% protection and that from noninfected cells was taken as 100% protection. MIC 50 is defined at that concentration of the respective minireceptor affording 50% protection of the cells against viral damage (31). the migrational behavior of the recombinant proteins upon SDS-PAGE (data not shown, but see also Ref. 10). In order to enrich for conformations active in ligand binding, we therefore further purified the material by rabbit ␤-VLDL column chromatography. This method substantially enhanced the homogeneity of the proteins, as seen upon analysis by nonreducing SDS-PAGE (Fig. 6B), and also increased the specific binding activity considerably (Fig. 6C). Fragments containing only two complement type repeats were not retained on the column. This indicates a higher stringency for the recognition of ␤-VLDL than for recognition of HRV2 because ligand blotting experiments demonstrated that even two repeats are sufficient for viral recognition (Fig. 5). However, protection of HeLa cells against infection with HRV2 was only observed with receptor fragments encompassing at least three ligand binding repeats (Fig. 3). This is not entirely unexpected because the avidity of the receptor is greatly increased upon immobilization, leading to high local concentrations and consequently to multivalent interaction, which apparently allows for binding to the otherwise weakly binding two-repeat minireceptors. Considerable differences in ligand interaction between membrane-bound receptors and soluble receptors have been reported previously (see, for example, Ref. 26).
Minireceptors with more than two complement type repeats sequentially purified on Ni-NTA and ␤-VLDL columns were then used for the determination of the minimal inhibitory concentration (which is defined as that concentration that affords protection of 50% of the cells against viral damage). All minireceptors showed MIC 50 values between 0.3 and 3 g/ml under these experimental conditions. It is interesting to note that rLDLR 3-7 h, with only five complement type repeats, exhibited a stronger protection potential than rLDLR 1-7 h, which encompasses the entire ligand binding domain. The high proportion of dimers present in rLDLR [3][4][5][6][7] h (that are also retained on the ␤-VLDL column) might be the reason for the high cell protection potency (Fig. 5A).
Currently, a bacterially expressed two-domain ICAM-1-IgA chimera is being evaluated as antiviral agent against major group HRVs (27). The recent determination of the three-dimensional structure of a virus-binding fragment of ICAM-1 (28,29) might now aid the design of lower molecular mass antiviral agents. Bacterial expression of human LDLR and refolding into a native conformation has recently been achieved (30). However, recovery was low, and use of this large polypeptide as antiviral appears impractical. The aim of our current study was to define the minimal structure requirements for inhibition of a minor group human rhinovirus. We thus expressed various minireceptors using the baculovirus system. Without the necessity for in vitro refolding, native proteins were recovered in amounts sufficient for cell protection assays. The experiments showed that the presence of at least three ligand binding re-peats is required for inhibition of viral infection. Experiments aimed at engineering such small fragments with increased virus neutralization efficiency are currently under way.