This study involving human subjects was approved by the Tzu-Chi University, National Science Committee and the National Blood Center or Taichung Veteran Hospital Review Boards and approved by the Committee of Ethics in Tzu-Chi University 15. A selected portion of patients' sera were removed from this study subsequently due to restriction from Institutional Review Boards. All subjects in this study gave their informed consents. Patients were classified based on the classification criteria of the American College of Rheumatology as SLE (n = 61), rheumatoid arthritis (RA, n = 50), Sjögren's syndrome (SS, n = 13) and systemic sclerosis (SSc, n = 20). Normal sera (n = 45) were collected from qualified, sex- and age-matched adult blood donors.

Normal six- to eight-week-old female BALB/c mice were purchased from the National Laboratory Animal Center (NLAC), Taipei, Taiwan. Animals were housed in a pathogen-free facility with an independent ventilation cage system at the Laboratory Animal Center of Tzu-Chi University, Hualien, Taiwan. All animal experiments were approved by Tzu-Chi University Animal Experimental Ethics Committee (reference number 94-A-06).

The pp65_1-167, pp65_167-336 and pp65_336-561 sequences are amplified using the following primer pair sequences, respectively listed in Table 1. The sequences were designed using the published nucleotide sequence of pp65 (strain: AD-169, GenBank: FJ527563). The fragments of pp65_1-167, pp65_167-336, and pp65_336-561 were prepared from PCR and digested by restrictive enzymes, and then ligated into pET30. The pp65_336-379, pp65_379-455 and pp65_455-561 fragments were digested from pp65_336-561 to form 132-bp (BamHI/HindIII), 231-bp (HindIII/NotI) and 321-bp (NotI/XhoI) fragments, respectively. The pp65_336-422, pp65_336-439 and pp65_336-448 encoding sequences were amplified from a pp65_336-561 clone using both upstream and downstream primers (Table 1). The pp65 sub-fragments mentioned above were cloned and inserted into pET30. The murine C3d encoding sequence (GenBank: DQ408205) was PCR amplified with C3d primers and ligated into pET32 (Table 1). For yeast two-hybrid analysis, PCR product of pp65_336-439 was cloned into pAS-1 plasmid to create a pAS-1-pp65_336-439-binding domain (BD) plasmid.

Recombinant proteins were over-expressed in E. coli with 1 mM isopropyl β-D-thiogalactoside (IPTG, Sigma-Aldrich, St. Louis, MO, USA) induction and purified by nickel affinity column. The C3d biotinylation and streptavidin (SA) conjugation (Pierce, Thermo Scientific, Rockford, IL, USA) were performed by the manufacturers' instructions. In brief, maleimide-activated streptavidin (Pierce) was conjugated with proteins containing reduced disulfide bonds from a disulfide reducing gel (Pierce) and mixed with biotinylated C3d to form the protein-SA-C3d tetramer, including pp65_1-167, pp65_336-439 and SA-C3d only. Tetramers were generated and prepared for immunization within four hours.

A total of 35 six- to eight-week-old female BALB/c mice were randomly separated into groups of pp65_1-167-C3d (n = 11), pp65_336-439-C3d (n = 17), SA-C3d (n = 5) and PBS (n = 2). Mice were inoculated intraperitoneally with 50 μg pp65_336-439-C3d or pp65_1-167-C3d, or SA-C3d in complete Freund's adjuvant (Complete Freund's Adjuvant, Sigma-Aldrich) or phosphate-buffered saline (PBS, 3.2 mM Na₂HPO₄, 0.5 mM KH₂PO₄, 1.3 mM KCl, 135 mM NaCl, pH 7.4). Boosting was performed with antigens in incomplete Freund's adjuvant (Incomplete Freund's Adjuvant, Sigma-Aldrich) three times in three weeks. Mice were bled via the retro-orbital vein one day prior to each assay and at two-week intervals. Unused sera were stored at -20°C and the diluted sera for use were kept at 4°C.

Immunoblotting was performed as previously described 15. In brief, 1 × 10⁸ cultured HeLa cells or 2 μg purified HCMV were prepared, homogenized and separated by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE/slab gel format). Separated proteins were transferred to nitrocellulose paper and blocked by 5% skim milk then analyzed with mice or human sera at dilutions of 1:500 or 1:1,000 in PBS. The antibody reactivity was detected by horseradish peroxidase (HRP) conjugated secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) and chemiluminescent detection agents (Perkin Elmer, Norwalk, CT, USA).

ELISA was performed as previously described 15. In brief, for the anti-dsDNA antibody assay, 1 μg/well of purified calf thymus dsDNA (Sigma-Aldrich) in ddH₂O was coated to a microtiter plate (Corning, Lowell, MA, USA). After blocking with 5% skim milk, mice or human sera at 1:100 and 1:1,000 dilutions in PBS, respectively, were added and incubated at room temperature (RT) for two hours. At the end of incubation, the plate was washed and bound antibodies were detected by HRP conjugated secondary antibodies at dilutions of 1:10,000 (for anti-dsDNA IgG) or 1:2,000 (for anti-dsDNA IgG subtypes, Bethyl Laboratories, Montgomery, TX, USA, purified HCMV or 1 μg/well of HeLa lysate in PBS were coated on a microtiter plate at 4°C overnight 15. After the plate was skim-milk-blocked, mouse or human sera were added at dilutions of 1:500 or 1:1,000, respectively, in PBS and incubated at 4°C for two hours. The bound antibodies were detected by a secondary antibody at dilutions of 1:10,000 at 4°C for two hours. The o-phenylenediamine dihydrochloride (OPD, Sigma-Aldrich) was used as the substrate and HRP activity was read at 450 nm with a micro-ELISA reader (DYNEX MRX II).

For the detection of protein-to-protein interaction between pp65_336-439 and HeLa proteins, whole HeLa extract (1 × 10⁸ cultured HeLa cells) was separated by 12% SDS-PAGE/slab-gel and transferred onto nitrocellulose paper. Before the experiment, the blot was cut into strips, skim-milk-blocked and then incubated with either pp65_336-439 or pp65_1-167 His-tag fusion protein at concentrations of 20, 10, 5, 2.5, 1.25, 0.625 mg/ml or at concentrations of 20, 10, 5, 2.5 mg/ml, respectively for one hour. The pp65_336-439 or pp65_1-167 bound HeLa proteins were detected by 10,000X diluted HRP-conjugated mouse anti-His tag IgG (Serotec, Raleigh, NC, USA) after one hour of incubation. The reactions were visualized by chemiluminescent detection agents.

Mouse sera were tested for anti-nuclear antibodies (ANAs) at a dilution of 1:100 in PBS by a standard anti-nuclear antibody (ANA) test (Binding site). The reactivity of anti-dsDNA antibody was tested by immunofluorescent stain using the Crithidia luciliae test (binding site) with mice sera at a dilution of 1:40 in PBS, as suggested by the manufacturer. In brief, 25 μl of diluted mice sera were incubated with slide-coated HEp-2 or Crithidia luciliae for 20 minutes in humid chamber at RT. HEp-2 or Crithidia luciliae slides were washed three times in PBS at RT for 10 minutes each. The bound antibodies were detected by 100X diluted Fluorescein isothiocyanate (FITC)-conjugated anti-mouse IgG (Jackson ImmunoResearch Laboratories) for 20 minutes in a humid chamber at RT in the dark. For nuclear visualization, HEp-2 slide was incubated in 25 μl of DAPI (0.5 μl/ml, Sigma-Aldrich) at RT for two minutes in the dark. At the end of staining, slides were washed (PBS) and mounted (mounting medium) for investigation by Nikon E800 fluorescence microscopy (Nikon, Tokyo, JP).

For an immunofluorescent stain on glomerulus, kidneys were removed from mice, immediately placed in the OCT gel and frozen at -80°C for 24 hours. The 5-mm-thick frozen sections were stained with FITC-conjugated anti-mouse IgG at a 1:100 dilution in PBS for 20 minutes in a humid chamber at RT in the dark. After PBS washing, coverslips with mounting medium on tissue slides were prepared for investigation by Nikon E800 fluorescence microscopy.

Moderated Cyanogen bromide (CnBr) powder (Sigma-Aldrich) was activated as described by the manufacturer. In brief, purified and sonicator-homogenized HCMV virions were dissolved in a coupling buffer (0.1 M NaHCO₃, 0.5 M NaCl, pH 8.3) with activated CnBr gel at 4°C overnight. The free active groups on CnBr were deactivated by 0.1 M Tris-HCl (pH 8.0) at RT for two hours. After deactivation, CnBr gel was washed with alternating buffer (0.1 M NaAc, 0.5 M NaCl, pH 4.0 and 0.1 M Tris-HCl, 0.5 M NaCl, pH 8.0) twice and washed with 10 ml PBS once. For purification, 200 μl of pooled pp65_336-439 or pp65_1-167 mouse sera in 10 ml PBS were added to HCMV-CnBr gel and rolled at 4°C overnight. The unbound portion of sera, flow through, was collected and concentrated as a negative control, while bound antibodies were eluted by 1 ml of 0.1 M glycine (pH 2.0) 19. The eluted samples were neutralized immediately with a 30 μl of neutralizing buffer (1 M Tris-HCl, 2 M NaCl, pH 8.8).

The Matchmaker yeast two-hybrid screening system (Clontech. Mountain View, CA, USA) was used to identify the proteins that were able to interact with pp65_336-439 peptide. In this system, yeast two-hybrid library screening using yeast mating was performed as modified from the manufacturer's manual. The DNA fragment encoding pp65_336-439 was cloned into the Gal4BD (DNA-binding domain of the transcription factor Gal4) vector pGBT-7 and the resulting plasmid was designated as BD-pp65_336-439. The BD-pp65_336-439 plasmid was transformed into the yeast strain AH109 (MATa) for screening the yeast library strain (Clontech Co.), which is the yeast strain Y187 (Matα) transformed with the AD-cDNA plasmid, HeLa cDNA cloned into the AD (activation domain of Gal4) vector pGAD-T7. The AH109 cells bearing BD-pp65_336-439 were cultured in the synthetic dextrose medium lacking tryptophan at 30°C until O.D.₆₀₀ was approximately 0.8. The AH109 cells bearing BD-pp65_336-439 were then collected and mated with the yeast library strain in 2X YPDA medium (1% Bacto yeast extract, 2% Bacto peptone, 2% Dextrose, 4% Adenine hemisulfate) at 30°C. After mating, cells were screened on the synthetic dextrose solid medium lacking leucine, tryptophan and histidine (SD/-L/-W/-H) to assay expression of reporter gene HIS3 at 30°C. The screened colonies were further screened on the synthetic dextrose solid medium lacking leucine, tryptophan, histidine and adenine (SD/-L/-W/-H/-A) to assay expression of reporter genes HIS3 and ADE2 at 30°C. The AD-cDNA plasmid was isolated from the screened colony grown on the SD/-L/-W/-H/-A solid medium and transformed into E. coli for amplification. To further confirm the interaction between pp65_336-439 peptide and the cDNA-encoding protein, both BD-pp65_336-439 plasmid and the purified AD-cDNA plasmid were transformed into yeast strain YRG2 (Stratagene, La Jolla, CA, USA) and tested on the SD/-L/-W/-H solid medium containing 15 mM 3-aminotriazole to assay for the expression of the reporter gene HIS3 at 30°C. The purified AD-cDNA plasmid was then sequenced after confirmation.

Statistical methodology for differences of titer and prevalence in test results was analyzed by GraphPad Prism (GraphPad Software Inc. La Jolla, CA, USA) and using the Student t-test and Fisher's two-tailed exact test, respectively. Results with a P-value of < 0.05 were considered to be significant.

To verify the existence of B-cell epitope(s), HCMVpp65 tegument protein (pp65) was cloned, truncated and expressed as his-tagged fragments (pp65_1-167, pp65_167-336 and pp65_336-561) that covered the entire antigen (Figure 1). Results showed that HCMV-seropositive SLE patients responded strongly to pp65_336-561 (61%, 37/61) compared to pp65_1-167 (7%, 4/61) or pp65_167-336 (20%, 12/61). The elevated positive rate to pp65_336-561 by SLE patients' sera was not found on either healthy or other disease controls (Table 2). In order to reveal the dominant epitope(s) within pp65_336-561, pp65_336-561 was sub-cloned into three fragments, expressed and re-screened (Figure 1). Of the original 37 pp65_336-561-positive sera, 7 were removed from subsequent tests due to various availability issues. Of the rest of the 30 pp65_336-561-positive sera, 22 were positive to pp65_379-455 (73%, 22/30), 3 were positive to pp65_455-561 (10%, 3/30) and 0 was positive to pp65_336-379 (Table 2). Subsequently, pp65_336-448, pp65_336-439 and pp65_336-422 fragments were created by partial deletion from the C-terminus of pp65_336-561 (Figure 1). Of 22 pp65_379-455-positive sera, 17 were positive to pp65_336-448 (77%, 17/22), 16 were positive to pp65_336-439 (73%, 16/22) and 9 were positive to pp65_336-422 (41%, 9/22). The sero-reactivity to three fragments (pp65_336-422, pp65_336-439 or pp65_336-448) by pp65_379-455 positive sera was listed in Table 3.

Although pp65 antigen immunized BALB/c animals possessed anti-pp65 antibodies and enhanced autoantibody activities, the titers were reduced in a few weeks after immunization 15. To improve the immunogenicity of pp65, C3d was used as an adjuvant. The immunization results showed that pp65_336-439-immunized mice gradually increased developed anti-HCMVpp65 IgG reactivity started at four weeks and continued to the end of the experiment (12 weeks post-immunization, Figure 2a, b). In contrast, the titer of anti-HCMVpp65 IgG was significantly less for pp65_1-167-immunized mice (Figure 2a). The anti-HCMVpp65 IgG was not detected from either SA-C3d or PBS challenged mice (Figure 2b). Quasi-quantitative analysis showed that the titers of anti-HCMVpp65 IgG from pp65_336-439 immunization was twice as much as sera from either pp65_1-167 (pp65_336-439 vs. pp65_1-167, 0.78 ± 0.02 vs. 0.44 ± 0.05, P < 0.0001) or SA-C3d (pp65_336-439 vs. pp65_1-167, 0.78 ± 0.02 vs. 0.43 ± 0.02, P < 0.0001) immunized animals at eight weeks post-immunization (Figure 2a). The IgG reactivity to HCMV of pp65_1-167 and SA-C3d was statistically insignificant.

In order to demonstrate that the immunization of pp65_336-439 could lead to the development of cross-reactive autoantibodies, total HeLa lysate was prepared as the substrate for the detection of anti-HeLa antibodies (Figure 2c). Although immunization of pp65_336-439 and pp65_1-167 induced anti-HeLa IgG at 4 weeks and continued to 12 weeks post-immunization, pp65_336-439 immunization exhibited significantly higher anti-HeLa IgG activity than pp65_1-167 immunization (pp65_336-439 vs. pp65_1-167, 0.50 ± 0.03 vs. 0.38 ± 0.02, P = 0.0191) at 8 weeks post-immunization. To exclude the possibility of HCMV contamination, HeLa lysate were immunoblotted by pp65 sub-fragment immunized sera (Figure 2d). The results showed that of eight anti-pp65 positive sera, only one strongly and another weakly react to HeLa antigens at 65 kD position.

To determine if pp65_336-439 immunization could induce antibodies against nuclear components from HeLa cells, the anti-nuclear antibody (ANA) test was performed. The results showed that pp65_336-439 immunization induced multiple ANA staining patterns, including speckled (5/17, Figure 3a, i), nucleosome (4/17, Figure 3b, j), chromatin (4/17, Figure 3c, k), mitotic spindle type I (MSA I, 4/17, Figure 3d, l), mitotic spindle type II (MSA II, 10/17, Figure 3e, m) centriole (6/17, Figure 3f, n) and nucleolar (14/17, Figure 3g, o) stains at 1:100 dilution at 8 weeks and continued to 12 weeks post-immunization. In several occasions, ANA patterns were detected at dilution as much as 500-fold. Nuclear stains, however, were not detected from either SA-C3d or PBS-immunized animals (0/5, 0/2, Figure 3h, p). Four pp65_1-167-immunized mice developed weak anti-nucleolar reactivity (4/11, Figure 3g, o) detectable at 1:40 dilution. Taken together, pp65_336-439 immunization could induce cross-reactive antibodies to multiple nucleus components (Table 4).

Anti-dsDNA antibody is a feature and a disease indicator for SLE patients 202122. ELISA assays showed that pp65_336-439-immunized sera exhibited significantly enhanced anti-dsDNA antibody activity compared to animals immunized with pp65_1-167 (pp65_336-439 vs. pp65_1-167, 0.66 ± 0.02 vs. 0.48 ± 0.03, P < 0.0001), or SA-C3d (pp65_336-439 vs. SA-C3d, 0.66 ± 0.02 vs.0.42 ± 0.02, P < 0.0001) at 8 weeks and continued to 12 weeks post-immunization (Figure 4a). The differences of anti-dsDNA antibody between pp65_1-167 and SA-C3d immunized mice were insignificant. The IgG2a to dsDNA is the dominant isotype to SLE nephritis 23. ELISA-based assays showed that 13 of 17 pp65_336-439 immunized mice were positive to dsDNA. Isotyping showed that enhanced IgG1 (dsDNA (+) IgG1 vs. dsDNA (-) IgG1, 0.50 ± 0.02 vs. 0.35 ± 0.03, P = 0.0029) and IgG2a isotypes (dsDNA (+) IgG2a vs. dsDNA (-) IgG2a, 0.33 ± 0.02 vs. 0.22 ± 0.02, P = 0.0134) were the contributors of anti-dsDNA activity (Figure 4b). To confirm the ELISA-based anti-dsDNA analysis, the Crithidia luciliae stains were performed. Of 17 pp65_336-439-immunized animals, 11 were positive for anti-dsDNA antibody (1:40 dilution) at 8 weeks and continued to 12 weeks post-immunization, compared to 2 of 11 pp65_1-167-immunized mice (Figure 4c). All Crithidia luciliae-positive sera were positive at ELISA tests. Only one pp65_1-167-immunized mouse was positive for Crithidia luciliae at 12 weeks post-immunization (Figure 4d).

To elucidate the relation between pp65_336-439 immunization and anti-nuclear antibody found in animals, antibodies to either pp65_336-439 or pp65_1-167 were affinity purified from pooled pp65_336-439 or pp65_1-167 immunized mouse sera. The results showed that affinity-purified pp65_336-439-specific IgG exhibiting significantly enhanced anti-HCMV activity compare to pp65_1-167 specific IgG (pp65_336-439 vs. pp65_1-167, 1.08 ± 0.05 vs. 0.27 ± 0.01, P < 0.0001, Figure 5a). Unbound fractions (flow through) from purification processes remain anti-HCMV positive, but the titer reduced significantly (pp65_336-439 vs. flow through, 1.08 ± 0.05 vs. 0.41 ± 0.02, P = 0.0003, Figure 5a, b). As immunofluorescent stains performed in Figure 3, affinity-purified anti-pp65_336-439 antibodies reproduced all ANA stains found in direct serum-staining (Figure 3), including speckled (Figure 5c1,c7), chromatin (Figure 5c2,c8), centriole (Figure 5c3,c9) or MSA II (Figure 5c4,c10) stains. Antibodies purified from flow through or anti-pp65_1-167 immunized sera, however, did not produce noticeable nuclear staining patterns (Figure 5c5,c11 and Figure 5c6,c12). In addition to nuclear stain, affinity-purified anti-pp65_336-439 antibody also possessed reactivity to dsDNA as ELISA and Crithidia luciliae slides demonstrated (0.49 ± 0.02, Figure 6a, b). The difference of anti-dsDNA activity between purified anti-pp65_1-167 antibody and flow through were insignificant (pp65_1-167 vs. flow through, 0.12 ± 0.01 vs. 0.15 ± 0.01).

Deposition of immunoglobulin or immune complexes on glomeruli is a characteristic of early nephritis and is often found in SLE patients 24. Immunofluorescent stains with anti-mouse IgG in renal section showed that pp65_336-439-immunized mice developed signs of deposition of immune complex on glomeruli (Figure 6c1). A total of 6 of 17 pp65_336-439-immunized mice showed IgG deposition on glomeruli, but such stains were not found in pp65_1-167 (0/11, Figure 6c2), SA-C3d (0/5, Figure 6c3) or PBS (0/2, Figure 6c4) treated animals. We did not observe clinical symptoms, such as proteinuria or lesions on the kidneys of immunofluorescent-positive mice. Nevertheless, it is noteworthy that the levels of immunoglobulin deposition showed positive correlation to the titers of the anti-dsDNA antibody (data not shown).

Immediately after infection, HCMVpp65 was transported into the nucleus and migrated to nucleolus 25. We hypothesized that the pp65_336-439 fragment may complex to intracellular antigens during infection. Immunoblotting and yeast two-hybrid verified this hypothesis and showed the association between HeLa proteins and pp65_336-439 (Figure 7, Table 5). Such binding was not found on either pp65_1-167 or protein-free tests. Among those pp65_336-439 binding colonies from yeast two-hybrid, 10 were verified as nucleic acid binding proteins, heat-shock proteins and apoptosis-related proteins (Table 5).