The chemokine receptor homologue encoded by US27 of human cytomegalovirus is heavily glycosylated and is present in infected human foreskin fibroblasts and enveloped virus particles

2.1. Plasmids

PCR amplification of US27 was done as described for UL33 (Margulies et al., 1996), using the HindIII B fragment of HCMV AD169 genomic DNA (provided by Marc Chee, Medical Research Council) and primers to the 5′ and 3′ end of the US27 ORF (Chee et al., 1990a). The resulting PCR product was cloned into the BamHI site of the vector pcDNA1 (Invitrogen, San Diego, CA) in the T7/CMV orientation to create BJM2. The insert in BJM2 was sequenced (Kraft et al., 1988; Sanger et al., 1977) and found to contain a single base change (C for T at base 867 of original PCR product) compared to the published sequence (Chee et al., 1990a,b). Because this change was not conservative (i.e., coding specificity was changed from isoleucine to threonine), it was corrected by replacing the Bsu36I–BspMI fragment of BJM2 with the corresponding fragment from the genomic HindIII B fragment. The corrected construct was called BJM2.1.

2.2. In vitro transcription–translation

Proteins were synthesized in vitro with a TNT rabbit reticulocyte kit (Promega, Madison, WI), using 1 μg of Qiagen column-purified (Qiagen, Chatsworth, CA) BJM2.1 (or appropriate control DNAs), T7 RNA polymerase, and 10 μCi [³⁵S]-methionine (New England Nuclear, Boston, MA, #NEG-009A) per reaction, according to manufacturers’ protocol.

2.3. Antisera and immunoprecipitation

Anti-27 is a polyclonal antipeptide antiserum prepared by immunizing one female New Zealand White rabbit with a synthetic peptide corresponding to the seventeen carboxy-terminal amino acids of the predicted protein product of US27 (i.e., NH₂-YDRKNAPMESGEEEFLL-CO₂H), conjugated to keyhole limpet hemocyanin via an EDC (1-ethyl-3-(dimethylaminopropyl) carbodiimide) linkage (Imject; Pierce, Rockford, IL), all essentially as described previously for the Anti-33 antiserum (Margulies et al., 1996). Anti-C1 is a polyclonal antipeptide antiserum specific for the carboxyl end of the CMV assembly protein precursor; its preparation and characterization were described elsewhere (Schenk et al., 1991; Welch et al., 1991b). Generation of the Anti-33 antiserum has been described (Margulies et al., 1996).

Immunoprecipitations were performed by combining 7 μl of each in vitro transcription–translation reaction with 18 μl of IP wash buffer, consisting of 0.4% CHAPS (3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulfonate; Sigma, St. Louis, MO) and 1 mM PMSF (phenylmethylsulfonyl fluoride; Sigma) in calcium and magnesium-free-PBS (CMF-PBS). One sample each of the diluted preparations was combined with 25 μL of Anti-27, and one of each was combined with 25 μL of an irrelevant antiserum (rabbit Anti-Neurothelin [Margulies, B.J., and Clements, J.E., unpublished data]). The samples were rocked overnight at 4 °C, combined with 50 μL Protein A-Sepharose beads (100 mg/mL in 0.02% NaN₃ and CMF-PBS; Sigma, #P3391), and rocked another 3 h at 4 °C. The beads were washed four times with 50 μL IP wash buffer and once with CMF-PBS, and the final pellet was analyzed by SDS-PAGE and fluorography (Section 2.4).

2.4. Gel electrophoresis, Western immunoassay, fluorography, and autoradiography

Proteins were separated by electrophoresis in 12% polyacrylamide gels containing sodium dodecyl sulfate (SDS-PAGE), essentially as described by Laemmli (1970); specific modifications and optimizations, including not heating samples prior to electrophoresis, were implemented as before (Gibson, 1981; Margulies et al., 1996). Urea was added to some samples to a final concentration of 4 M before electrophoresis (see Section 3.9) (Soulie et al., 1996). Western immunoassays with the Anti-33 (Margulies et al., 1996), Anti-C1 antisera (Welch et al., 1993), and Anti-27 antisera (Section 2.3) were done by the general procedure of Towbin et al. (1979), with modifications as described previously (Margulies et al., 1996; Welch et al., 1993). Fluorograms were obtained from the processed blots with XAR-5 film (Eastman-Kodak, Rochester, NY) or Biomax MS film (Eastman-Kodak) and a calcium–tungstate intensifying screen at −80 °C (Laskey and Mills, 1977). Gels with [³⁵S]-methionine-labeled in vitro-synthesized samples were processed for sodium salicylate-enhanced fluorography (Chamberlain, 1979) and an image was recorded similarly. The autoradiogram in Fig. 2B was obtained from direct exposure of Biomax MR film (Eastman-Kodak) at room temperature for 1 week.

2.5. Cells and virus

Human foreskin fibroblast (HFF) cultures were prepared, maintained, and infected with the AD169 strain (ATCC VR-538) of human cytomegalovirus (HCMV) as described before (Gibson, 1981; Weiner and Gibson, 1983). The three types of extracellular enveloped virus particles, non-infectious enveloped particles (NIEPs), virions, and dense bodies were effectively purified, separated from one another, and concentrated by banding three times in potassium tartrate gradients as described (Irmiere and Gibson, 1983; Margulies et al., 1996; Talbot and Almeida, 1977). The inhibitory chemicals cycloheximide (CHX; supplied in DMSO and used at a final concentration of 50 μg/mL, Sigma, St. Louis, MO) (Jeang and Gibson, 1980) or phosphonoformic acid (PFA; 200 μg/mL, Sigma) (Welch et al., 1991a) were added to some cultures to inhibit translation or virus DNA replication, respectively (Honess and Roizman, 1974) (see also Section 2.7).

2.6. Membrane preparation and peptide N-glycosidase F (PNGase F) treatment

Membrane fractions were prepared from HCMV-infected and non-infected HFF 8 days after infection as described before (Margulies et al., 1996; Nathans et al., 1989; Parker et al., 1993; Sung et al., 1991). Membrane preparations and virus particle preparations were treated or not treated overnight with PNGase F (200 mU per sample) as outlined previously (Haltiwanger and Hart, 1993; Margulies et al., 1996; Tarentino and Plummer, 1994).

2.7. RNA isolation

Tissue culture plates (six-well) were seeded with HFFs and infected with HCMV AD169, clarified by low-speed centrifguation, the next day. One hour prior to infection, cells were treated with the corresponding levels of CHX, PFA (Section 2.5), or DMSO (at 0.05%); these treatments were left in the culture medium throughout the course of infection. Total RNA from the cells in a single well (approximately 10⁶ cells) was isolated at 6, 24, and 168 h post-infection with the Qiagen RNEasy system, supplemented with an on-column RNAse-free DNAse step (Qiagen), all according to manufacturer’s protocols. Total RNA was isolated from extracellular virus particles (100 μL of clarified inoculum), which were first pretreated with DNAse-free RNAse (0.5 μg, Roche Diagnostics Corporation, Indianapolis, IN) at 37 °C for 1 h to remove any residual RNA outside any virus particles, via a QIAamp Viral RNA Mini Kit (Qiagen). All RNAs isolated from both cellular and virion sources were also subsequently treated with a second round of RNAse-free DNase (Amplification Grade DNAse I; Invitrogen) at 25 °C for 30 min to remove any remaining genomic DNA prior to RT-PCR. One-twentieth of each eluted RNA was used per amplification reaction.

2.8. Polymerase chain reaction (PCR) and reverse-transcription-PCR (RT-PCR)

Each RNA (Section 2.7) was amplified by PCR alone, using only Platinum Taq DNA polymerase (Invitrogen), or by RT-PCR, using a Superscript II One-Step RT-PCR kit (Invitrogen). Each RNA was tested for the presence of transcripts for the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH), with the deoxyoligonucleotide primers 5′-ACCACCATGGAGAAGGCTGG-3′ and 5-CTC-AGTGTAGCCCAGGATGC-3′ (Ercolani et al., 1988), for transcripts from the US27 gene, with primers 5′-GCGTGCGA-AATAACAAGCGG-3′ and 5′-GCAACAGGCGAATACCC-TTAGG-3′ (Chee et al., 1990b), and for the RNA encoded by the HCMV UL109 gene, with the primers 5′-CGTTAC-GATGAACTGCGG-3′ and 5′-AGCTTATTGAGCGCAGCC-3′ (Bresnahan and Shenk, 2000) (all oligonucleotides synthesized by MWG Biotech, High Point, NC). Every PCR and RT-PCR reaction contained 1/2 volume of 2× reaction mix (Superscript II One-Step RT-PCR kit; contains buffer, MgCl₂, and dNTPs; Invitrogen), 200 μM each primer, 5 U. recombinant RNasin (Promega, Madison, WI), and 1 μL of enzyme (either Platinum Taq alone or the Superscript II/Platinum Taq mix); these reaction conditions corresponded to manufacturer’s specifications for the One-Step RT-PCR kit. Both PCR and RT-PCR reactions for GAPDH were amplified under the same thermal cycling parameters: 50 °C for 30 min (the RT step), followed by 30 cycles of 94 °C for 15 s, the T_m of the primers (60 °C) for 30 s, and 72 °C for 1 min; PCR and RT-PCR amplification for US27 and UL109 was performed with the same times as those used for GAPDH in each step, except the initial reverse transcription was conducted at 52 °C (for US27) or 45 °C (for UL109), cycling was continued 30 or 60 times (as noted in Sections 3.6 and 3.7), and the T_m during each cycle was 62 °C (for US27) or 55 °C (for UL109). Approximately 50 ng of a green fluorescent protein-HCMV (AD169) recombinant bacterial artificial chromosome (Marschall et al., 2000) was also employed as a positive control for each set of reactions. PCR products were resolved on 2% agarose gels in SB buffer (Brody and Kern, 2004), stained with ethidium bromide, and images of the gels were captured on a Bio-Rad Molecular Imager Gel Doc XR system (Bio-Rad, Hercules, CA). Digital images were then compiled with Adobe Photoshop and Illustrator.

3.1. In vitro transcription–translation and immunoprecipitation

These experiments were initiated by PCR amplifying (Scharf, 1990) the HCMV US27 ORF (Chee et al., 1990a,b) from a genomic fragment and cloning it into a vector, pcDNA1, to produce the plasmid BJM2.1, as described in Section 2.1. The protein encoded by BJM2.1 was then synthesized in a T_NT coupled in vitro transcription–translation system, along with the luciferase protein, encoded on a plasmid included in the T_NT kit as a positive control. Small portions of the reaction mixtures were subjected to SDS-PAGE in a 12% polyacrylamide gel.

An image prepared from the resulting gel showed that the luciferase control plasmid predominantly gave rise to the expected ≈61-kDa luciferase protein (Fig. 1, lane 1) and US27 in BJM2.1 gave rise to a broad, ≈36–38-kDa band (pUS27) (Fig. 1, lane 4). Neither of these proteins was present in the control with no DNA added (Fig. 1, lane 2). The apparent size of this pUS27 band is smaller than expected from the predicted amino acid sequence of the US27 ORF (i.e., ≈41 kDa). When heated in solubilization buffer before SDS-PAGE, pUS27 aggregated and did not appreciably enter the resolving gel (Fig. 1, lane 3). The anomolously small size and heat sensitivity of pUS27 are attributed to strong irreversible intra- and inter-molecular bonds, resulting from non-covalent interactions between the proteins’ multiple hydrophobic transmembrane domains. Both the size discrepancy and heat sensitivity in SDS-PAGE are frequently observed with membrane-bound proteins (Maeshima, 1991; Sazuka et al., 2004; Schubert et al., 1992; Semenza et al., 1990), especially cellular (Keinanen et al., 1987) and viral (Chiou et al., 2002; Zhen et al., 2005) GPCRs, including pUL33 (Margulies et al., 1996).

In vitro-synthesized pUS27 was then used in an immuno-precipitation experiment to validate the reactivity of Anti-27, an antiserum prepared to identify the US27 protein in HCMV-infected cells (Section 2.3). Portions (7 μL) of the BJM2.1 and luciferase in vitro-translation reactions described above were subjected to immunoprecipitation with either Anti-27 or Anti-Neurothelin (irrelevant antiserum control), as described in Section 2.3. Anti-27 immunoprecipitated in vitro-synthesized pUS27 (Fig. 1, lane 6), but not luciferase (Fig. 1, lane 8), and neither pUS27 nor luciferase was immunoprecipitated by the irrelevant antiserum (Fig. 1, lanes 5 and 7). As was observed with immunoprecipitation of the in vitro-synthesized pUL33 with the Anti-33 antiserum (Margulies, 1996) and another GPCR (Rozell et al., 1998), immunoprecipitation with Anti-27 promoted a significant amount of aggregation of pUS27 (Fig. 1, lane 6, see top of lane).

3.2. The US27 gene product is present in HCMV-infected cell membranes and is N-glycosylated

The qualified Anti-27 serum was then used as a probe to identify the US27-encoded GPCR produced in HCMV-infected HFF cells; we called this protein GCR27 to differentiate it from the recombinant protein, pUS27 (Fig. 1), and for consistency with the designation GCR33, given the UL33-encoded protein in HCMV-infected cells (Margulies et al., 1996). Membrane fractions, prepared 8 days after infection with HCMV, were used as the starting material because earlier work showed that GCR33 was enriched and more easily identified there (Margulies et al., 1996). These membrane preparations from HCMV-infected and non-infected cells were subjected to SDS-PAGE in a 12% polyacrylamide gel, followed by Western immunoassay with the Anti-27 antiserum, as described in Sections 2.3–2.6.

An image prepared from the Western blot showed a broad, high-molecular-weight band of Anti-27-reactive material (≈ 80–180 kDa) in the membrane preparation from infected or cells (Fig. 2A, lane 3, left-margin brackets), but not in that from non-infected cells (Fig. 2A, lane 1). The high-molecular-weight material was also discernable as two discrete regions, from approximately 80 to 130 kDa and from about 150 to 180 kDa. The size of this material, which was larger than predicted based on the translated nucleotide sequence of the US27 ORF (Chee et al., 1990a,b), and its heterogeneity suggested that GCR27, like GCR33, is extensively N-glycosylated (Margulies et al., 1996). To test this hypothesis, equal portions of the same membrane preparations were treated with PNGase F (Tarentino and Plummer, 1994), to remove N-linked oligosaccharides, and analyzed in parallel. PNGase F treatment resulted in the appearance of two smaller bands (≈41 and ≈37 kDa, Fig. 2A, lane 4, circles) that were not present in either the nontreated, infected cell (Fig. 2A, lane 3) or the PNGase F-treated, non-infected (Fig. 2A, lane 2) preparations. Deglycosylation did not appear to be complete, though, as there was residual high-molecular-weight, Anti-27 reactive material still present after PNGase F treatment (Fig. 2A, lane 4). The finding that GCR27 was detected as a pair of bands following analysis of material treated with PNGase F is in contrast to findings with GCR33, which was detected as a single band in similar assays (Margulies et al., 1996); however, the deglycosylated GCR27 doublet (Fig. 2A, lane 4) is consistent with the two separate electrophoretic mobility groups of glycosylated GCR27 (Fig. 2A, lane 3) and with the appearance of two proteins from in vitro-translated BJM2.1 (Fig. 1, lane 4).

3.3. Direct comparison of in vitro-translated pUS27 and deglycosylated GCR27 by SDS-PAGE

The appearance of a protein doublet for both pUS27 (Fig. 1, lane 4) and GCR27 (Fig. 2A, lane 3) led us to examine the relationship between these polypeptides. Samples of deglycosylated GCR27 and in vitro-synthesized pUS27 were compared side-by-side on the same gel to see if the respective bands of each doublet co-migrated. In vitro-synthesized pUS27 was radiolabeled during synthesis with [³⁵S]-methionine (Section 2.2) while deglycosylated GCR27 was detected by Western immunoassay with Anti-27 and [¹²⁵I]-labeled protein A (Sections 2.3–2.6).

As was shown in Fig. 2A, no material was detected in non-infected cell membranes (Fig. 2B, lanes 5 and 6); HCMV-infected cell membranes that were not deglycosylated contained high-molecular-weight GCR27 that was too large to be displayed in the selected window for Fig. 2B, but is displayed for other preparations (Figs. 2A and C, 3A, 4C, and 5). Deglycosylated GCR27 appeared as a partially resolved doublet of approximately 37 and 41 kDa (Fig. 2B, lane 8), as previously observed (Fig. 2A, lane 4). An in vitro translation reaction containing no DNA did not give rise to any [³⁵S]-labeled polypeptides that resolved in the selected range (Fig. 2B, lane 9), while pUS27 synthesized from BJM2.1 gave rise to a poorly resolved ≈36–38-kDa doublet (Fig. 2B, lane 10), as was previously observed (Fig. 1, lane 4). Therefore, although both deglycosylated GCR27 and pUS27 can be resolved as doublets in SDS-PAGE, their respective apparent molecular weights do not match exactly. This characteristic is common for membrane-bound proteins, in which recombinant polypeptide products are misfolded when compared to the bona fide polypeptides, and these misfolded species migrate anomalously in SDS-PAGE (Chiou et al., 2002; Keinanen et al., 1987; Maeshima, 1991; Margulies et al., 1996; Sazuka et al., 2004; Schubert et al., 1992; Semenza et al., 1990; Zhen et al., 2005).

3.4. Anti-27 detects GCR27 uniquely and specifically

To further prove the specificity of Anti-27, we compared the reactivity of this antiserum against a previously characterized antiserum, Anti-33, which reacts specifically with another HCMV GPCR encoded by UL33 (Margulies et al., 1996). A single gel (Fig. 2C) was run with mock-infected (Fig. 2C, lanes 11, 12, 15, and 16) and HCMV-infected (Fig. 2C, lanes 13, 14, 17, and 18) cell membranes that were either not treated (Fig. 2C, odd numbered lanes) or treated (Fig. 2C, even-numbered lanes) with PNGase F. One-half of the gel was processed for a Western blot with Anti-27 (Fig. 2C, lanes 11–14); the other was processed with Anti-33 (Fig. 2C, lanes 15–18). As shown previously, Anti-27 reacted with only HCMV-infected cell material (Fig. 1C, lanes 13 and 14), and deglycosylated GCR27 appeared as a doublet of closely spaced bands (Fig. 2C, lane 14). Anti-33 reacted with only infected cell material (Fig. 2C, lanes 17 and 18); however, deglycosylated GCR33 was specifically detected as a ≈43-kDa band (Fig. 2C, lane 18), a slightly higher apparent molecular weight than the upper band of deglycosylated GCR27 (Fig. 2C, lane 14). Therefore, the Anti-27 antiserum is not cross-reactive with GCR33, and the Anti-33 antiserum is not cross-reactive with GCR27.

3.5. Presence of GCR27 in HCMV enveloped virus particles

GCR27 was also detected in preparations of glycerol/tartrate gradient-purified (Talbot and Almeida, 1977) enveloped virus particles. Virions, non-infectious enveloped particles (NIEPs), and dense bodies, the three forms of extracellular enveloped particles produced by HCMV (Irmiere and Gibson, 1983), were treated with PNGase F or not treated as described (Margulies et al., 1996) and subjected to SDS-PAGE and Western immunoassay with the Anti-27 antiserum. The antiserum Anti-C1, specific for the capsid assembly protein precursor (pAP) that is present in NIEPs, but not virions or dense bodies (Irmiere and Gibson, 1983; Schenk et al., 1991; Welch et al., 1991b), was mixed with Anti-27 to demonstrate the resolution of these extracellular particles, to provide an internal size marker for the GCR27 bands, and to help visualize and distinguish the NIEP patterns. Cell membrane preparations were included as controls.

Results obtained from the cell membrane controls were essentially the same as before (i.e., Fig. 2A); Anti-27 reacted with high-molecular-weight, heterogeneous material in HCMV-infected cell membrane preparations (Fig. 3A, lane 3), and PNGase F treatment of these membranes resulted in the appearance of ≈37- and ≈41-kDa bands (Fig. 3A, lane 4, circles), neither of which was seen in the non-treated membrane preparation (Fig. 3A, lane 3) or in non-infected cells (Fig. 3A, lanes 1 and 2). The ≈45–50-kDa pAP doublet, detected by Anti-C1 in this experiment, is not known to be glycosylated, consistent with its unchanged size in the PNGase F-treated preparations (Fig. 3A, lanes 3–6). Furthermore, the similar intensity of the pAP band in lanes 4 versus 3, and 6 versus 5, indicates that the PNGase F treatment did not grossly alter the amount of material present.

Results similar to those seen with infected cell membranes were obtained with the NIEP (Fig. 3A, lanes 5 and 6), virion (Fig. 3A, lanes 7 and 8), and dense body (Fig. 3A, lanes 9 and 10) preparations. High-molecular-weight, Anti-27-reactive material was present in enveloped particles that were not treated with PNGase F (Fig. 3A, lanes 5, 7, and 9, brackets); this material resolved as a pair of glycosylated forms of GCR27 of approximately 80–130 and 150–180 kDa, consistent with the electrophoretic migration of glycosylated GCR27 found in infected cell membranes (Fig. 2A, lane 3). Bands of ≈37- and ≈41-kDa were observed in all three preparations of enveloped virus particles following PNGase F treatment (Fig. 3A, lanes 6, 8, and 10, circles). Characteristically only NIEPs contained detectable pAP (Fig. 3A, lanes 5 and 6). Consistent with the membrane preparations (Fig. 2A, lane 4 and Fig. 3A, lane 4), deglycosylation of GCR27 in particles was incomplete, as evidenced by residual higher-molecular-weight material that reacted with Anti-27 (Fig. 3A, lanes 8 and 10, in particular). The fact that GCR27 presumably contains seven membrane-spanning domains, is homologous to other GPCR proteins that are characteristically plasma membrane constituents, and is present in HCMV dense bodies, which are composed of only a viral envelope surrounding an aggregation of predominantly the lower matrix protein (Gibson, 1983; Irmiere and Gibson, 1983; Varnum et al., 2004), provides compelling evidence that this protein is an envelope constituent, like other herpesvirus GPCR homologues (Fraile-Ramos et al., 2002; Margulies et al., 1996; Oliveira and Shenk, 2001; Penfold et al., 2003).

3.6. US27 RNA is not present in HCMV virions

Several reports have described the presence of HCMV-specific RNAs in mature virions, usually associated with the tegument (Bresnahan and Shenk, 2000; Greijer et al., 2000; Sarcinella et al., 2004; Terhune et al., 2004). If the US27 RNA were also incorporated into virions, it would provide a means for GCR27 expression prior to de novo RNA synthesis, as part of the infectious material delivered by incoming particles. Therefore, to investigate this possibility, RNA was isolated from the clarified inoculum that was used for the previous experiments (Figs. 2 and 3A).

Using 30 cycles of RT-PCR and PCR, we were unable to detect the presence of RNA for UL109, previously shown to be incorporated into virions (Bresnahan and Shenk, 2000), in our HCMV inoculum (data not shown). Improving the sensitivity of the RT-PCR and PCR reactions by increasing to 60 cycles of amplification revealed the presence of the expected UL109 RNA in virions (Fig. 3B, lane 17). Under the same PCR conditions, and using the same starting material, RNA for US27 was never detected (Fig. 3B, lane 11), indicating either its absence from virions or an abundance below the limit of detection.

3.7. US27 is expressed as a γ1 RNA in the presence of cycloheximide or phosphonoformic acid

Previous reports showed that blocking HCMV infection with phosphonoformic acid prevented expression of the RNA for US27, placing this open reading frame into the true late gene (γ2) category (Honess and Roizman, 1974; Mocarski and Courcelle, 2001; Roizman and Pellett, 2001; Welch et al., 1991a). More recent work suggested that the US27 gene is transcribed at earlier times during infection (Chambers et al., 1999), so we attempted to resolve this discrepancy by using a comparatively more sensitive RT-PCR assay to investigate the gene’s time course of expression. HCMV-infected cells were blocked at various times after infection with either CHX, to prevent translation of immediate-early proteins and abate the transition to early stage infection (Fioretti et al., 1973; Honess and Roizman, 1974; Jeang and Gibson, 1980; Michelson-Fiske et al., 1977; Mocarski and Courcelle, 2001), or PFA, to diminish DNA synthesis and the complete transition to late stage infection (Honess and Roizman, 1974; Mocarski and Courcelle, 2001; Welch et al., 1991a). A parallel set of mock- and HCMV-infected cells were also treated with only 0.05% DMSO, the solvent for CHX, as a control for this solvent in the CHX-treated cells.

RT-PCR amplification of these samples (Fig. 4A) showed that RNA for the GAPDH housekeeping gene was present in every RNA sample recovered (Fig. 4A, lanes 1–12), as evidenced by the presence of an approximately 500-bp PCR product that was not present in RT-PCR reactions without RNA (Fig. 4A, lane marked with minus sign). To control for the possible presence of residual genomic DNA in the samples, the same RNAs were subjected to the same amplification conditions without a reverse transcriptase (i.e., with Platinum Taq DNA polymerase alone) (Fig. 4A, lanes 13–24); no PCR product was synthesized in any of these experiments. Hence there was either no DNA contamination present in the samples, to serve as a false template for amplification in the RT-PCR reactions (Fig. 4A, lanes 1–12), or the amount of DNA was below the limit of detection for the assay. In either instance, the lack of a PCR product, yet the qualitative presence of an RT-PCR product, for GAPDH showed that the RNA samples were intact and of good quality for the subsequent detection of US27 transcripts.

RT-PCR of these qualified RNAs under similar conditions (see Sections 2.7 and 2.8) to detect US27 transcripts resulted in amplification products of 527 base pairs (Fig. 4B, lanes 35 and 36) that were not present in samples without RNA (Fig. 4B, lane 38), or RNA from non-infected cells (Fig. 4B, lanes 25–30). These US27-derived PCR products were present in only those HCMV-infected cell samples from 168 h post-infection, whether treated with PFA (Fig. 4B, lane 36) or left untreated (Fig. 4B, lane 35), while the same RT-PCR product was not present in RNAs isolated from any other times post-infection and treated with either CHX (in DMSO) or DMSO alone (Fig. 4B, lanes 31–34). As with GAPDH, PCR amplification alone with only Platinum Taq DNA polymerase (Fig. 4B, lanes 40–51) generated no product, proving that the RT-PCR products in Fig. 4B, lanes 35 and 36 arose from only RNA.

Although strictly not the RNA expression pattern of a typical truly late gene (Honess and Roizman, 1974; Mocarski and Courcelle, 2001; Roizman and Pellett, 2001), the US27 gene may be characterized as γ1, “leaky” or “early-” late gene; that is, the US27 RNA is expressed regardless of the presence of PFA (Boriskin and Butcher, 2001; Honess and Roizman, 1974; Mocarski and Courcelle, 2001), yet is not transcribed in the presence of CHX (Honess and Roizman, 1974; Mocarski and Courcelle, 2001).

3.8. Temporal regulation of GCR27 expression

We also examined the temporal control of expression of the glycoprotein encoded by US27, GCR27. Cells were treated with either CHX, DMSO, or PFA, as above (Sections 2.5, 2.7 and 3.7); membrane preparations were collected and were either treated or not treated with PNGase F (Section 2.6), then subjected to SDS-PAGE and Western blot with the Anti-US27 antiserum (Section 2.4). A single aliquot (50 μL) of inoculum was also prepared in the same fashion, to assay for the presence of GCR27 as a component of entering virions.

The GCR27 glycoprotein was detected in only samples that were left untreated through 168 h after infection (Fig. 4C, lanes 67 and 68). As was shown previously (Figs. 2 and 3A), the heavily glycosylated forms of GCR27 (Fig. 4C, lane 67) could be resolved as two closely spaced products of about 37 and 41 kDa when samples were first treated with PNGase F (Fig. 4C, lane 68). No Anti-27 reactive material was found in mock-infected cell membrane preparations (Fig. 4C, lanes 55, 56, 61, 62, 64, and 65), and also none could be detected when HCMV-infected cells were treated with CHX for 24 h (Fig. 4C, lanes 57 and 58), DMSO for 24 h (Fig. 4C, lanes 59 and 60), or PFA for 168 h (Fig. 4C, lanes 63 and 64), perhaps because the amount produced in this last case was below our limit of detection. Although extremely light, a very small amount of deglycosylated GCR27 is present in the inoculum at approximately 37 and 41 kDa (Fig. 4C, lane 70), showing the extremely low quantities of GCR27 that arrive in the infected cell with virus particles. Therefore, although the US27 RNA is present as a γ1 transcript (Section 3.7), the glycoprotein expressed from that RNA appears to be expressed most abundantly with truly late kinetics, corroborating the first report describing US27 RNA expression (Welch et al., 1991a).

3.9. Aggregation of GCR27 by heating in the presence of SDS can be decreased with the addition of urea

To determine whether the heat-induced aggregation of GCR27 could be abrogated, samples of deglycosylated GCR27 were analyzed for their electrophoretic mobility once SDS was added and the samples were boiled. Dense bodies, enriched in GCR27, were isolated by positive density-negative viscosity gradients (Section 2.5) and deglycosylated with PNGase F (Section 2.6), then analyzed by SDS-PAGE and Western blot with Anti-27 (Section 2.4). As seen with pUS27 expressed in vitro from BJM2.1 (Fig. 1, lane 3), heating the samples containing SDS before electrophoresis resulted in significant aggregation of GCR27 (Fig. 5, lane 3, asterisk) that was not observed when the samples were not heated (Fig. 5, lane 1). Addition of urea to a final concentration of 4 M did not alter the mobility of the polypeptides when left unheated (Fig. 5, lane 2). Most importantly, adding urea significantly reduced the heat-induced aggregation of deglycosylated GCR27 (Fig. 5, lane 4, circles), which migrated with the same electrophoretic mobility as deglycosylated GCR27 that was not heated (Fig. 5, lanes 1 and 2). Therefore, as has been observed with other membrane proteins (Schubert et al., 1992; Soulie et al., 1996), addition of urea to SDS-PAGE sample buffer appears to diminish the non-denaturable intra- and inter-molecular interactions that cause aggregation of these polypeptides when heated with SDS.