To express TSV viral protein 1 (VP1) as a glutathione-S-transferase (GST) fusion protein, we created a pGEX4t3-TSV VP1.tag plasmid. For cloning of TSV VP1, sense (5′-GGATCCGGATCCGCCCCCAAAAGAAAAGG-3′) and antisense (5′-GTCGACGTCGACATAAAGCCGGGCGGGGAAG-3′) primers (BamHI and SalI restriction sites are underlined) were generated (Eurogentec, Cologne, Germany). Using these primers, we performed a PCR on the pUC19-TSV plasmid (9). A 2-step AmpliTaq gold PCR program was performed as described (9). TOPO TA cloning (Invitrogen, Carlsbad, CA, USA) of the amplified PCR product resulted in a construct used for cloning TSV VP1 into the pGEX4t3-BKV VP1.tag plasmid (18) after removal of the BKV VP1 sequence. The pGEX4t3-TSV VP1.tag construct was verified by sequencing using the BigDye Terminator Kit (Applied Biosystems, Foster City, CA, USA) and analyzed on an ABI Prism 3130 Genetic Analyzer (Applied Biosystems).

We sequenced VP1 from MCV isolate 344 and verified amino acid residues aspartic acid (D) and arginine (R) at positions 288 and 316 as found in MCV isolates 339 and 162. These residues are likely involved in proper folding of the VP1 for conformation-dependent epitope recognition (21). A pGEX4t3-tag plasmid was included to express tagged GST alone, which is necessary for serologic background determinations. In every construct, the tag sequence included codes for the 11-aa KPPTPPPEPET epitope of simian virus 40 (SV40) large T-antigen (31,32). GST and GST-fusion proteins of TSV, BKV, and MCPyV VP1 were expressed in the Bl21 Rosetta Escherichia coli strain as described (21,31–33). Expression of the GST-fusion proteins was analyzed by using Western blotting.

A total of 528 serum samples from a population-based serum bank of healthy persons from the Province of Utrecht in the Netherlands were analyzed. This serum bank was set up in 1994 as a pilot study, the prePienter study, for a nationwide serum bank that would be used to evaluate long-term seroepidemiologic changes of diseases included in the Dutch National Immunization Program (34) (M.A. Conyn van Spaendonk et al., pilot study for Pienter project, logistical evaluation (part 1), RIVM-report no. 213675001/1995). Approval of the prePienter study was obtained from the Medical Ethical Committee of the Dutch Organization for Applied Scientific Research (TNO) (Leiden, the Netherlands), and every participant provided written informed consent. The age distribution within the population was <1–9 years, n = 79; 10–19 years, n = 66; 20–29 years, n = 51; 30–39 years, n = 64; 40–49 years, n = 76; 50–59 years, n = 54; 60–69 years, n = 79; and 70–79 years, n = 56.

We also tested 80 serum samples obtained in 1995 from immunocompromised renal transplant patients who came to a specialized dermatologic outpatient clinic at Leiden University Medical Center. These samples were obtained after informed oral consent was obtained from the patients, which was documented in patient files. The Medical Ethics Committee of Leiden University Medical Center reviewed and approved this study. The average age of the patients was 45 years (range 26–64 years).

A serum sample was also obtained from a 16-year-old immunocompromised heart transplant patient with TS. A detailed description of this patient was reported by van der Meijden et al. (9). The TS patient and his mother provided oral informed consent for the patient to provide serum for detection of antibodies against TSV, which was recorded in the patient’s medical file. The Medical Ethics Committee of the Leiden University Medical Center declared in writing that no formal ethical approval was needed to analyze this sample for viral diagnosis.

Samples were analyzed for polyomavirus seroreactivity by using the multiplex antibody-binding assay developed and described by Waterboer et al. (33). Briefly, glutathione–casein (GC) coupled Bio-Plex polystyrene beads (Bio-Rad Laboratories, Hercules, CA, USA) containing a combination of fluorescent dyes were coupled to either GST-TSV VP1.tag, GST-BKV VP1.tag, GST-MCV VP1.tag, or GST.tag. For each antigen, 3,000 GC-coupled beads per sample were loaded with crude bacterial lysates containing relevant GST fusion protein. Samples were preincubated with GST.tag containing bacterial crude lysates (2 mg/mL) in blocking buffer to reduce nonspecific GST binding. For cross-reactivity studies, samples were preincubated with GST-TSV VP1.tag, GST-MCV VP1.tag, or GST-BKV VP1.tag. After preincubation, antigen-coated bead mixtures were incubated with samples diluted 1:100. For detection of bound serum antibodies, beads were incubated with goat anti-human total immunoglobulin G–biotin (1:1,000 dilution; Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA), streptavidin R–phycoerythrin (1:1,000 dilution; Invitrogen), and washed. Beads were analyzed in a Bio-Plex 100 analyzer (Bio-Rad Laboratories). Results are presented as median fluorescent intensity (MFI) units. For each sample, antigen-specific binding was obtained by subtracting the MFI for beads coated with GST alone from those of beads coated with GST VP1.

To measure seroreactivity against TSV, an immunoassay was developed with TSV VP1 antigen expressed as a GST-fusion protein in E. coli. The TSV VP1 immunoassay was developed according to the Luminex-based assay described by Waterboer et al. for simultaneous measurement of seroresponses against different human papillomavirus types (33). We first analyzed the reproducibility of the new assay with 80 serum samples from renal transplant patients. These samples were tested 3 months apart by using GC-coated beads coupled independently to the same crude TSV VP1 bacterial extract. This comparison showed reproducible results with a correlation coefficient of r² 0.89 (Figure 1).

Although not expected on the basis of amino acid sequence comparison (RefSeq TSV NC_014361 and MCV: NC_010277) for a randomly selected subset of 30 renal transplant serum samples, we investigated a possible association between seroreactivity against VP1 of TSV and that of MCPyV because MCPyV is phylogenetically the closest related human polyomavirus to TSV (9). No association between TSV and MCPyV VP1 seroresponses was observed (r² 0.036; Figure 2, panel A). Similar findings were obtained when TSV VP1 seroresponses were compared with those against the more distantly related BKV VP1 (r² 0.065; Figure 2, panel B).

We also evaluated potential cross-reactivity between TSV VP1 and MCPyV VP1 in detail for 2 TSV- and MCPyV-reactive serum samples. These samples were titrated and preincubated with soluble GST, GST-TSV VP1, or GST-MCV VP1. Subsequently, TSV VP1 and MCPyV VP1 seroresponses were measured. In both samples tested, TSV VP1 reactivity was inhibited by preincubation with TSV VP1 only and not with MCPyV VP1, whereas MCPyV VP1 reactivity could be inhibited by preincubation with MCPyV VP1 only (Figure 3). Similar results were obtained in a TSV VP1 and BKV VP1 competition experiment with TSV-reactive and BKV-reactive samples (Figure 4).

TSV seroreactivity was determined in a TS patient previously reported (9). A serum sample was obtained 6 months after detection of TSV and daily facial treatment with cidofovir-containing cream had started. At the time the sample was obtained, treated lesions had resolved but untreated skin (e.g., of the legs) still had typical spicules indicative of active TSV infection.

Serial dilutions of the TS serum sample were tested by using the TSV VP1 assay. High reactivities were observed (Figure 5, panel A). This response could be exceeded by using soluble GST-TSV VP1 but not with GST-BKV VP1. Conversely, the BKV seroresponse observed in this patient was exceeded only by using soluble recombinant GST-BKV VP1 and not by using GST-TSV VP1 (Figure 5, panel B). No seroresponse against MCV VP1 was detected for this patient.

TSV VP1 seroreactivity was determined for 528 healthy persons and 80 immunosuppressed renal transplant patients. BKV VP1 was included in the analyses as a positive control because of the known high BKV seroprevalence in the general population (19–21). In every experiment, a panel of 3 reference serum pools was also included, which showed little variance over time. Results for TSV VP1 and BKV VP1 are shown in Figure 6.

To investigate age-specific TSV seroreactivity and to calculate a cutoff value to determine TSV seropositivity, we subdivided the healthy population into different age groups (Figure 7, panel A). For persons <1–9 years of age, a clear distinction could be made between patients who were seronegative for TSV (MFI ≈0) and children with TSV seroreactivities of 4,000–12,000 MFI units. To calculate TSV seropositivity, a cutoff value of 877 MFI units was calculated on the basis of mean seroreactivity of the TSV serononresponders from the lowest age group + 3 SD. Although the distinction between seronegative persons and seropositive persons in the first age group was less clear for BKV (Figure 7, panel B), a similar strategy was used for BKV and resulted in a cutoff value of 1,051 MFI units.

Using the calculated cutoff values, we determined the age-specific seroprevalence for TSV in each age group of the immunocompetent population. In the children <10 years of age, the seroprevalence for TSV was 41% (Figure 7, panel C). This percentage increased to ≈75% at 30 years of age, and remained stable for higher age groups. BKV seroprevalences were calculated and showed values between 75% and 97% (Figure 7, panel C). Analyses of the youngest age group showed an increasing trend for TSV and BKV seropositivity starting at 10% for TSV in children 1–2 years of age (Figure 7, panel D).

On the basis of these calculations, overall seroprevalence for TSV VP1 was 70% for the healthy population and 89% for the immunocompromised population (Figure 6, panel A). For BKV VP1, the overall seroprevalence for both groups was somewhat higher (85% and 99%, respectively) (Figure 6, panel B).