RANTES Binding and Down-Regulation by a Novel Human Herpesvirus-6 β Chemokine Receptor

Cells and virus, preparation of infected cell DNA

HHV-6A strain U1102 (2) was cultured in the CD4⁺ T cell leukemic cell line JJhan in RPMI 1640 medium, supplemented with 10% (v/v) FCS and 2 mM glutamine, 50 IU/ml penicillin, and 50 μg/ml streptomycin. The human immortalized epithelial cell line, HaCaT (20), was maintained in the same medium. Vero cells were maintained in DMEM supplemented as described for RPMI. The human erythroleukemia (K562) and premonocytic (U937) cells were cultured in RPMI 1640 supplemented as described above. Transfected cell lines were maintained in 500 or 750 μg/ml of G418 for HaCaT or K562 lines, respectively.

HHV-6-infected JJhan DNA was prepared 5 days postinfection. Infected cells and uninfected controls were washed with PBS (Dulbecco’s A), taken into lysis buffer (10 mM Tris-HCl (pH 8.0), 10 mM EDTA, and 0.5% SDS), digested with proteinase K (0.25 mg/ml), extracted with phenol/chloroform/isoamyl alcohol (25:24:1), and then DNA precipitated from the aqueous phase with isopropanol.

Cloning of U51 gene into pCDNA3 and epitope tagging

The U51 gene was amplified using PCR with standard thermocycling procedures and Taq polymerase. The following primer set was used to amplify the coding region and to introduce BamHI restriction enzyme recognition sequences: 5′-AAAGGATCCTTACTTCGTTTATTC-3′ and 5′-TATGGATCCAAAGGATCCACTCCT-3′. The amplified product was digested with BamHI restriction enzyme, separated by agarose gel electrophoresis, and purified by phenol extraction followed by ethanol precipitation. The purified DNA fragment was ligated with BamHI-digested, phosphatased plasmid pCDNA3 (Invitrogen) and transformed into Escherichia coli strain JM109. The selected clone was checked by sequence analyses and designated pcDNA3-U51. This plasmid was then epitope tagged with the sequence encoding the 9-aa epitope from influenza hemagglutinin recognized by mAb 12CA5 (Babco, Richmond, CA; Boehringer Mannheim, Indianapolis, IN) as described previously (21). The following primer was used to introduce this sequence into the N-terminal end of the U51 gene, such that the epitope-encoding sequences were inserted after the initiating methionine and glutamic acid codons: 5′-CTTCGTTTCTTTAGCATAATCCGGCACATCGTATGGATACTCCATCCTG-3′. Site-directed mutagenesis was performed following the Kunkel method (22), as described previously (21). This plasmid was designated pcDNA3-U51D.

In vitro transcription/translation and transient expression

In vitro transcription and translation (IVTT) was conducted using the transcription and translation-coupled rabbit reticulocyte lysate system (Promega, Madison, WI) according to the manufacturer’s instructions. Briefly, 0.5 μg of plasmid was used in an IVTT mix containing 12.5 μl of rabbit reticulocyte lysate, 20 U of RNasin, 15 μCi of [³⁵S]methionine (SJ1515, Amersham, Arlington Heights, IL), and 7.5 U of T7 RNA polymerase. Reactions were conducted at 30°C for 90 min, then analyzed directly by SDS-PAGE or immunoprecipitated.

For transient expression, the T7 RNA polymerase-vaccinia system (23) was used as described previously (21). Vero cells (3 × 10⁵) were infected (input multiplicity 1 PFU/cell) with recombinant vaccinia virus vTF-3, expressing the T7 RNA polymerase (23). After 2.5-h incubation at 37°C, infected cells were transfected, using Lipofectin and OPTIMEM (Life Technologies, Gaithersburg, MD), with 7.5 μg of plasmid. After a further 4-h incubation, cells were washed with DMEM deficient in methionine and labeled at 37°C for 13 h with 2 μl (15 mCi/ml) of [³⁵S]methionine in methionine-deficient medium. Lysates were then prepared and immunoprecipitated.

Cell lysis, immunoprecipitation, and gel electrophoresis

Labeled cells from transient expression experiments were washed three times with ice-cold PBS, then lysed for 30 min into 0.5 ml of ice-cold RIPA buffer (21). Lysates were sonicated (1 min on setting 6; Heat Systems (Farmingdale, NY) XL2020 sonicator with cup horn), cleared by microcentrifugation (13,000 rpm, 30 min, 4°C), and stored at −20°C. After preclearing with 50 μl of protein A-Sepharose (50% swollen volume in RIPA buffer), equal amounts of lysate (quantitated by scintillation counting of TCA precipitates) were incubated at 4°C overnight with Abs (1 μl of anti-influenza hemagglutinin (HA) type 1 mAb 12CA5 (Boehringer Mannheim)) recognizing the epitope tag. Immune complexes were precipitated with protein A-Sepharose, washed three times with RIPA buffer, and eluted in sample buffer.

For direct analysis, IVTT reaction products were diluted with an equal volume of 2× SDS sample buffer (without DTT) and resolved by SDS-PAGE. For immunoprecipitation, IVTT reactions were diluted to 0.5 ml with RIPA buffer.

Except where noted in the text, to avoid aggregation of the hydrophobic U51 protein, all samples were sonicated, rather than heated, before separation by electrophoresis. SDS-PAGE was performed in 12% precast gels (Novex, San Diego, CA) as described previously (21).

Generation of cell lines expressing U51

The plasmids pcDNA3-U51 (wild-type U51), pcDNA3-U51D (epitope tagged), and (pcDNA3) vector-only control were transfected into either HaCaT epithelial cells using Lipofectin (Life Technologies) according to the manufacturer’s protocol or K562 cells using electroporation (electroporator with capacitance extender, Bio-Rad, Hercules, CA). HaCaT-expressing colonies were selected and grown using RPMI culture medium (10% FCS) supplemented with 500 μg/ml geneticin (G418, Life Technologies). Epithelial cell colonies were Perspex ring cloned and analyzed by DNA PCR using the SP6 and T7 primers to amplify across the multiple cloning site in pcDNA3. The U51 clones were screened for expression by reaction with HHV-6 human sera in immunofluorescence assays as well as by RT-PCR as described below. Two each of independently derived vector (V)-only, U51 wild-type (U51), and U51tag (U51D) cell lines were used for further study and were labeled V-2, V-5, U51-3, U51-6, U51D-3, and U51D-5. The K562 cells were similarly selected using G418 at 750 μg/ml followed by cloning from the resistant population using limiting dilution. Four clonal lines of U51-expressing cells were selected. Cells were stored in liquid nitrogen before use, and only low passage clones (less then passage 5) were used in these studies.

RT-PCR

Confluent HaCaT or dense K562 cell lines cultured in 25-cm² flasks were washed twice with PBS, then lysed in 1 ml of cold RNA Isolator reagent (Genosys, The Woodlands, TX) and incubated at room temperature for 5 min. Chloroform (0.2 ml) was added, and the mixture was shaken until emulsified, then incubated for 15 min. After microcentrifugation (13,000 rpm for 30 min at 4°C), RNA was isopropanol precipitated from the aqueous phase, then pelleted by microcentrifugation. The RNA pellet was washed with 75% (v/v) ethanol (in water) and air-dried. Contaminating DNA was removed using DNase (Promega; 0.08 U/ml, 2 h at 37°C). DNase-treated RNA was extracted with phenol/chloroform/isoamyl alcohol (25:24:1), precipitated with ethanol, washed with 70% (v/v) ethanol (in water), and resuspended in water to 150 ng/ml.

Complementary DNA was synthesized from 2 μg of template RNA using 8 U/ml Moloney murine leukemia virus reverse transcriptase (Promega) in a 60-μl reaction containing 0.5 mM dNTP, 100 U of RNasin, 10 ng/ml random hexamers, 1× RT buffer (50 mM Tris-Cl (pH 8.3), 75 mM KCl, 3 mM MgCl₂, and 10 mM DTT). RNA was heated to 65°C for 10 min before cDNA synthesis, chilled on ice for 5 min, then added to the prepared reaction mix. Parallel reactions, excluding RT, were prepared for each RNA template to confirm the absence of contaminating DNA. Reactions were incubated at 37°C for 2 h, then diluted with an equal volume of water and stored at −20°C.

PCR was conducted in 30-μl reactions containing, as standard, 200 μM of each dNTP, 250 ng of each primer, and 1.5 U of Taq polymerase (Promega; storage buffer B) in 1× Taq reaction buffer. RANTES reactions contained 0.7 mM MgCl₂; all others contained 1.5 mM. Primer sequences were as follows (expected product sizes in base pairs are given in parentheses): β-actin: actin-F, GATGGAGTTGAAGGTAGTTT; actin-B, TGCTATCCAGGCTGTGCTAT (445 bp); HHV-6 U51: U51-F, TCGGTCGAGAATACGCTGTG; U51-B, AGATACGTAGTCACGGTCGA (493 bp); IL-8: IL-8-1, CTTCCTGATTTCTGCAGCTCTGTG; IL-8-2, CAAAAACTTCTCCACAACCCTCTG (245 bp); α₂ integrin: AL2-1, CCCTCTGGACAGCTTCTAGAG; AL2-2, GAAATCCCCGCTTACCTTGAC (189 bp); and RANTES: RAN-1, TCGCTGTCATCCTCATTGCTACTG; RAN-2, CATCTCCAAAGAGTTGATGTACTC (248 bp). Thermal cycling parameters for β-actin and U51 reactions were 94°C for 1 min, 58°C for 1 min, and 72°C for 2 min, for 40 cycles; for IL-8 reactions the parameters were 94°C for 30 s, 55°C for 30 s, and 72°C for 2 min, for 40 cycles; for RANTES and α₂ integrin reactions the parameters were 95°C for 30 s, 49°C for 30 s, and 72°C for 2 min for 40 cycles. PCR product identities were confirmed by sequencing using the ABI Prism 377 automated sequencer (Applied Biosystems, Foster City, CA) according to the manufacturer’s instructions.

Chemokine binding assay

Binding assays were conducted as described previously (24). Log phase cells (K562-U51 or K562-vector) were washed twice in unsupplemented RPMI, then resuspended at 2.5 × 10⁷ cells/ml in binding medium (RPMI, 0.1% BSA, and 20 mM HEPES, pH 7.4) on ice. Where appropriate, U937 cells were used as positive controls. Assays, in duplicate or triplicate, contained 2.5 × 10⁶ cells, 166 pM radiolabeled chemokine, and varying concentrations of unlabeled competitor chemokine. After 2-h incubation on ice with regular mixing, the cells were separated from the unbound chemokine by microcentrifugation through a pthlalate oil cushion (1.5 parts dibutyl phthalate to 1 part bis-(2-ethylhexyl)phthalate). Bound radioactivity was then counted using a gamma counter. Data analysis was conducted using the EBDA and LIGAND programs (version 2, 1985) (25). ¹²⁵I-labeled RANTES (sp. act., 2000 Ci/mmol), IL-8, and MIP-1α were obtained from Amersham, unlabeled chemokines were obtained from PeproTech (Rocky Hill, NJ), and RANTES was obtained from R&D Systems (Minneapolis, MN).

Quantification of RANTES secretion

A RANTES enzyme immunoassay was used according to the manufacturer’s instructions (R&D Systems, DRN00). Each sample of culture supernatant was tested, in duplicate, undiluted as well as diluted 1/10 and 1/100. A series of standards was also tested in each assay, and the results were used to plot a standard curve from which the test results were determined. The range of sensitivity for the assay was 31.25–2000 pg/ml, with usually two of the three sample dilutions giving concentrations within this range. Unused culture medium without G418, processed in parallel to exclude any effect of the drug on results, gave readings equivalent to background. All within-assay range results for a sample were used to calculate a weighted mean concentration.

Definition of a betaherpesvirus-specific GCR gene

U51 defines a new herpesvirus GCR family with only limited similarity to chemokine receptors. Although the degree of overall sequence identity between members of the U51 family is low, they have numerous common features and a conserved genomic location. The U51 family genes are located within block IV of the seven conserved herpesvirus gene blocks (2). Although the flanking genes within this block, notably those for glycoprotein H (HHV-6 U48) and the viral protease/assemblin (HHV-6 U53), are present in herpesviruses of all three subfamilies, the U51 homologues are found only in the betaherpesviruses. Using FASTA together with multiple alignment analyses, U51 was found to be a member of the class A family of GCRs that includes rhodopsin, adrenergic receptors, and chemokine receptors (26).

Unlike other herpesvirus GCRs characterized, U51 shares only limited similarity to the chemokine receptors. The closest homologues show only borderline overall similarity, with 20–24% identity to the functional CXC receptor from herpesvirus saimiri, ORF 74, as well as the cellular receptor, CCR7 (EBI1), inducible by both HHV-6 and EBV (16, 27, 28). However, although only distantly related, U51 does have a similar length and overall structure as the chemokine receptors. Other similar features to chemokine receptors are shown by amino acid sequence alignment using CXCR2 (IL-8β receptor) and CCR1 (MIP-1α and RANTES receptors) as prototypes for CXC and CC receptors, respectively (Fig. 1⇓). U51 shares some features with the human chemokine receptors and GCRs (4, 29). It has the characteristic seven-transmembrane predicted structure (Fig. 1⇓) together with basic residues in the first intracellular loop as well as conserved tyrosines in the first and fifth transmembrane domains; prolines in the second, fifth, and sixth transmembrane domains; tryptophan in the first extracellular loop; and cysteine in the second extracellular domain. It also has conserved cysteine, proline, and histidines in the seventh transmembrane domain, although these cannot be exactly aligned. It has potential sites for chemokine ligand binding at the acidic residues in the extracellular N-terminal domain and charged residues in the third extracellular domain, whereas regions in the transmembrane and intracellular domains have potential for mediating intracellular signaling. Here, like other GCRs, it has conserved prolines in the sixth and seventh transmembrane domains, and in common with the CXC chemokine receptors, U51 has a predicted N-glycosylation site in the second extracellular domain. However, like herpesvirus saimiri ORF 74, it does not have the conserved aspartate in the second transmembrane domain and has only a few clusters of serines in the C-terminal intracellular loop.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Multiple alignment of U51 family compared with U12 and prototypes of the chemokine receptor family. CCR1 represents the CC chemokine receptor family (binds RANTES); CXCR2 represents the CXC family (IL-8β, binds IL-8). CLUSTALV (Higgens algorithm) was used for the multiple alignment followed by manual adjustment to maximize conserved positions. The predicted seven-transmembrane regions (TM) found in the GCR family are marked I–VII (26 ). Potential extracellular loops are before TMI and between TM II and III, IV and V, and VI and VII. N-linked glycosylation sites in these exposed loops are indicated as well as two conserved cysteines, marked in bold, in loops between TM II and III, and IV and V. Consensus residues (identity of more than three, using HHV-6 strains as one count) are marked in small letters; capital letters mark sites where the majority (more than six) or all are conserved. HHV6A is strain U1102, HHV6B is strain Z29, and HHV7J is strain JI. Asterisks mark conserved residues specific for the U51 family.

U51 also significantly diverges from other chemokine receptors in an unusually short N-terminal extracellular domain with acidic residues; although it includes characteristic acidic residues, it lacks the conserved cysteine and predicted N-glycosylation site common in these receptors (4, 29). This cysteine residue in other chemokine receptors is predicted to be disulfide-bonded with a conserved cysteine in the third extracellular domain, which is also absent from HHV-6 U51, although the U51 family features more conserved cysteines at distinct positions that may play a compensatory role (Fig. 1⇑).

Overall, the U51 family appears to define a new betaherpesvirus-specific family of G protein-coupled receptor genes, with only distant similarity to the chemokine receptors. Amino acid residues specific to this family are marked in Fig. 1⇑.

Biochemical characterization of U51

For a preliminary investigation of the properties of U51, the HHV-6A strain U1102 U51 gene and a modified version (termed U51D) containing the HA epitope tag at the N-terminus were each inserted into the eukaryotic expression vector pcDNA3, putting them under the control of the T7 promoter and the human CMV major immediate-early promoter.

IVTT of the U51 constructs generated products with properties typical of a seven-transmembrane-spanning protein. Wild-type U51 showed a discrete band of 28 kDa, consistent with the U51 monomer, although smaller than the predicted M_r of 34.7 kDa, and a broad range of higher M_r species centered around 180 kDa (Fig. 2⇓A). IVTT of U51D gave rise to a similar band profile, with the monomer slightly increased in size to 30 kDa, most likely due to a conformational effect of the 9-aa epitope tag (Fig. 2⇓A). When samples were heated to 95°C before electrophoresis, the 28-kDa (U51) or 30-kDa (U51D) band disappeared (Fig. 2⇓A), consistent with aggregation of a multiply hydrophobic protein. The ladder of bands seen with IVTT U51 and U51D (Fig. 2⇓A, lanes 1 and 2) increase in size by ∼8-kDa increments, consistent with the size of ubiquitin, which may be added to misfolded hydrophobic proteins synthesized by IVTT (30). Inclusion of canine pancreatic microsomal membranes in the IVTT mix caused an increase in the size of the discrete band, consistent with use of the single predicted N-glycosylation site in the second extracellular domain of U51 (Fig. 2⇓B). As with unglycosylated U51, heating the glycosylated form caused increased aggregates of high M_r oligomers at the top of the gel (Fig. 2⇓, A and B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

In vitro transcription and translation of U51 and U51D. A, IVTT of U51 (lanes 1 and 4), U51D (lanes 2 and 5), and pcDNA3 vector (lanes 3 and 6). Samples were sonicated (lanes 1–3) or were heated to 95°C (lanes 4–6) before gel electrophoresis. (M, M_r markers shown in K_d). B, IVTT with canine pancreatic microsomal membranes (CMM). The pcDNA3 vector with CMM (lane 1) and without CMM (lane 2) (the white area is from gel tearage); U51D with CMM (lanes 4 and 6) or without CMM (lanes 3 and 5) is shown. Samples were sonicated (lanes 1, 2, 5, and 6) or heated to 95°C (lanes 3 and 4) before gel electrophoresis. M, M_r markers shown in kilodaltons. Discrete U51 bands, glycosylated and unglycosylated U51D, and the oligomeric forms are indicated.

Both the discrete band and the high M_r aggregate were immunoprecipitated by the anti-HA mAb 12CA5 from IVTT reactions with U51D (Fig. 3⇓A). Transient expression of U51D in Vero cells using the T7 RNA polymerase vaccinia system followed by immunoprecipitation gave rise to a similar pattern of bands (Fig. 3⇓B). The discrete band comigrated with the glycosylated IVTT U51D (data not shown), suggesting that the glycosylation site was used in vivo, as had been shown in the in vitro expression system. An additional band, which had not been seen in IVTT experiments, was prominent in the immunoprecipitations of the cellularly expressed U51D (Fig. 3⇓B), and its size of 58 kDa was consistent with that expected for a U51 dimer. Similar results for Vero cells were found for transient expression in epithelial, HaCaT, cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Immunoprecipitation of U51D. A, IVTT (without CMM) of U51D (lane 1), U51 (lane 2), or pcDNA3 vector (lane 3), immunoprecipitated with the anti-HA mAb 12CA5. M, M_r markers shown in kilodaltons. B, Transient cellular expression of U51D (lane 1) or pcDNA3 vector (lane 2) transfected into Vero cells amplified by vaccinia T7 recombinant superinfection and then immunoprecipitated with anti-HA mAb 12CA5. Discrete U51 bands, dimers, and oligomers are indicated. Identical results were shown with transfection and superinfection of HaCaT cells.

These results showed that the U51 protein has characteristics of a protein containing multiple hydrophobic domains, that the conserved N-glycosylation site in U51 can be used, and that a proportion of the protein might exist as a dimer.

Stable expression of U51 in epithelial cells

HHV-6 in vivo can infect and be secreted from epithelial cells, which are important for host-to-host transmission in saliva and also as a barrier in systemic infections (31, 32, 33). However, the virus then permissively infects circulating cell types. Virus chemokines and their receptors may modulate interactions between these cell types. To investigate the effect of U51 expression on an epithelial cell, HaCaT cells were used as a model system, because they can undergo normal differentiation (20), and epithelial in vitro culture systems that are fully permissive for HHV-6 replication have not been defined. Furthermore these cell lines provide an unlimited standardized system for characterization. Transient and stably transfected cell lines were established with the same plasmid clones used to demonstrate protein expression in the Vero-vaccinia T7 transient expression assays (pcDNA3-U51, pcDNA3-U51D, or vector alone). U51D could easily be detected in the transiently transfected, vaccinia T7-infected HaCaT cells by immunoprecipitation with the tagging mAb, 12CA5, with results similar to those shown above for Vero cells. However, in the absence of amplification of gene expression by the T7 polymerase, the U51 gene product could not be detected by immunoprecipitation in transient or stably transfected cell lines, although the U51 transcript was detected by RT-PCR (Fig. 4⇓, upper panels). Thus, this level is lower than that detectable using the tagging mAb, and possibly higher levels are not consistent with cell survival, as shown for other multiple membrane proteins. The U51 cell lines were further screened by immunofluorescence for reactivity with HHV-6-seropositive sera, and three of seven sera tested showed positive fluorescence with no reaction to the vector-only cell lines.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Expression of U51D in HaCaT cells. RT-PCR for U51 and β-actin in HaCaT cell lines V7 (lane 1), U51D-5 (lane 2), V2 (lane 3), and U51D-3 (lane 4). Light micrographs demonstrate the morphology of HaCaT cell lines V7 (panel 1), U51D-5 (panel 2), V2 (panel 3), and U51D-3 (panel 4). Original magnification, ×200.

Six independent clonal HaCaT lines were studied, U51-3 and U51-6 expressing wild-type U51; U51D-3 and U51D-5, expressing the epitope-tagged U51D; and V2 and V7, containing the vector alone. RT-PCR analysis of the cell lines with primers specific for a sequence within the U51 gene and standardized with RT-PCR for β-actin showed expression of U51 in the U51 cell lines but not in the vector controls (Fig. 4⇑). Representative results are shown for the tagged lines (Fig. 4⇑), and similar results were found for the wild-type lines. The cell lines showed altered cellular morphology compared with the two vector cell lines. In Fig. 4⇑ differences in cell spreading or flattening are shown in the U51-expressing cell lines. The morphological differences were not accounted for by growth differences or viability, as cell yields were similar (see below and Fig. 6⇓).

Down-regulation of RANTES expression

Chemokine expression has been shown to alter cellular morphology (34), and infection with HHV-6 or HCMV can alter both chemokine expression (35, 36) as well as cellular morphology (1). In HCMV infections, RANTES can be sequestered via HCMV US28 chemokine receptor (36). It was therefore of interest to investigate the effect of U51 on chemokine expression and possible autologous feedback mechanisms. To this end, RT-PCR was used to test for expression of RANTES and IL-8, as representative CC and CXC chemokines, with β-actin used as a control. α₂ integrin was used as an additional control, with indistinguishable results from actin (not shown).

Amplification from equal amounts of cDNA template generated from the HaCaT cell lines U51-3, U51-6, U51D-3, U51D-5, V2, and V7 showed that, compared with the control cell lines V2 and V7, the level of RANTES RT-PCR product was reduced in all U51 and U51D lines by at least 10-fold (Fig. 5⇓).Titrations of cDNA template in serial dilutions followed by separate PCR amplifications confirmed that the reduction was ∼10-fold, whereas β-actin, α₂ integrin, and IL-8 RT-PCR product yields were similar in all cell lines (Fig. 5⇓). On the basis of these results, it seemed that U51 was specifically causing regulation of RANTES expression.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Down-regulation of RANTES RNA in U51-expressing cell lines. RT-PCR analysis of IL-8, RANTES, and β-actin expression in U51- and U51D-expressing cell lines. Randomly primed cDNA samples generated from equal amounts of total RNA. Products corresponding to the linear range of amplification are shown. A, 10⁻² and 10⁻³ dilutions are shown. B, Products from two separate flasks of each cell line were further tested to show reproducibility (1/50 dilution of cDNA for the RANTES PCR and 1/1000 dilution of cDNA for the β-actin PCR). Lanes are: HaCaT cell line V-2 (lanes 1 and 2), V-7 (lanes 3 and 4), 51D-3 (lanes 5 and 6), 51D-5 (lanes 7 and 8), 51-3 (lanes 9 and 10), and 51-6 (lanes 11 and 12).

To determine whether the transcriptional down-regulation of RANTES was reflected in reduced levels of RANTES secretion, culture supernatants from the six cell lines were tested using a RANTES-enzyme immunoassay. Cells were seeded at equivalent density, and assays were conducted on supernatants 4 days postinfection of now confluent monolayers. RANTES levels in control supernatants were consistently high, averaging 2000–3000 pg/ml for the vector-only controls, V2 and V7, whereas the U51 and U51D cell line supernatants contained much lower levels of RANTES, averaging 30–300 pg/ml (Fig. 6⇓A). This down-regulation was not related to decreased cell numbers, as there was no significant difference between the final cell densities of the cell lines (Fig. 6⇓B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Down-regulation of chemokine RANTES secretion in epithelial cell lines stably expressing U51 and U51D. Cleared supernatants from confluent monolayers of HaCaT cell lines were assayed for RANTES levels using a commercial ELISA. Cells were seeded at 8 × 10⁵ cells/flask. Supernatants were taken, and cells were counted after 4 days. A, RANTES titers per milliliter of supernatant for two vector lines, two U51 lines, and two U51D lines. Expt. A and B represent data from two sets of flasks processed in parallel. B, Cell counts from Expt. A and B and from a third identical experiment, Expt. C. Data are the means of two counts for each cell line. The results show that RANTES levels are significantly reduced in the U51- and U51D-expressing cell lines, with no significant variation in cell number.

Ligand binding characteristics of HHV-6 U51

Having identified U51 as a chemokine receptor-like protein and shown that its expression is associated with down-regulation of RANTES, but not that of IL-8, it was important to determine ligand binding. To investigate this, U51 was stably expressed in K562 cells, which have been used previously for similar assays, for example HCMV US28 and HHV-6 U12, and do not express the MIP-1α/RANTES receptor or show specific competable binding of chemokines (11, 13). Binding assays using ¹²⁵I-labeled chemokines showed clearly that U51 binds RANTES, but not IL-8 or MIP-1α (Fig. 7⇓A). The positive control cell lines used were U937 cells that express functional receptors (Fig. 7⇓A).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Chemokine binding profile of U51. A, ¹²⁵I-labeled RANTES, IL-8, and MIP-1α binding was tested in the presence and the absence of cold competitor (1,000-fold molar excess for IL-8 and MIP-1α, 250-fold molar excess for RANTES). Assays using U51-8 (U51 clone 8) and K562 vector cells were performed in triplicate; control assays with U937 cells were performed in duplicate. Mean bound counts are shown, with error bars representing ±1 SD for U51-8 and K562 vector cells and ±1 range for U937 cells. Background binding was not subtracted from these data. B, Binding of ¹²⁵I-labeled RANTES to K562 cells stably transfected with pcDNA3-U51. Displacement of ¹²⁵I-labeled RANTES binding by unlabeled RANTES. Each point represents the mean ±1 SD for determinations performed in triplicate. The mean total binding shown is 20,516 ± 729 cpm. Nonspecific binding was 3,710 ± 64 cpm (vector control). The parameters for binding are shown in the lower corner derived from the Scatchard analyses shown in the upper corner, which gave a high affinity binding site with a K_d of 0.83 ± 0.13 nM.

Titration of unlabeled RANTES into the binding assay showed that ¹²⁵I-labeled RANTES binding by U51 was saturable and competable, indicating that the binding was specific. The titration was repeated in three separate experiments with similar results, and a representative titration is shown in Fig. 7⇑B. Scatchard analysis of the competition curve was performed using EBDA and LIGAND programs (25) (Fig. 7⇑B). The results identified a high affinity binding site, with a K_d of 0.83 ± 0.13 nM (Fig. 7⇑B). The number of receptors per cell assayed, as calculated from the r value, was 9375 ± 750. Thus, as well as regulating the expression of RANTES, U51 binds RANTES in a saturable and competable manner. U51 is therefore the most divergent member of the chemokine receptor family to have been shown to bind ligand to date.

To compare the ligand binding profile of U51 with those of other receptors, we tested a panel of unlabeled chemokines for competition of RANTES binding to U51. Of those tested, MCP-1, MCP-3, MCP-4, eotaxin, and vMIPII did compete binding, while, consistent with the direct binding assays described above (Fig. 8⇓) MIP-1α and IL-8 did not. Thus, U51 shows a unique ligand binding specificity compared with data on known chemokine receptors.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

Competition for RANTES binding to U51 by heterologous chemokines. Unlabeled chemokines were added at a 625-fold molar excess (125-fold molar excess for RANTES) over ¹²⁵I-labeled RANTES. The amount of ¹²⁵I-labeled RANTES bound in the presence of each competitor is shown as a percentage of the amount bound in the absence of competitor. Background binding (to vector K562 cells) was subtracted in each case. The mean background count was 4119 cpm, and mean uncompeted binding was 9420 cpm. Each chemokine was tested in triplicate. The mean is shown with error bars representing ±1 SD.