Effect of Epstein-Barr Virus Sequence Variation on Performance of the Roche Real-Time LightCycler PCR Assay

Epstein-Barr virus (EBV), one of the most common human viruses, is a member of the Herpesviridae family. EBV infects more than 90% of the world's population and remains latent in B-cell lymphocytes. EBV is associated with the development of certain malignancies, including lymphoma and posttransplant lymphoproliferative disease (PTLD). The monitoring of EBV viral load in peripheral blood is important for guiding empirical treatment of transplant recipients to prevent the development of PTLD. Recently, Roche Diagnostics developed an analyte-specific reagent for real-time LightCycler PCR to detect EBV. The assay amplifies the EBV latent membrane protein 2a gene (LMP2a) by utilizing a pair of PCR primers and a set of hybridization probes labeled with LightCycler Red 640 and fluorescein for amplicon detection.

From 1 November 2005 through 31 July 2007, a total of 644 patients were tested in our laboratory. Multiple samples were received from the same patient for monitoring purposes. In addition, 197 of the 644 patients were EBV positive at some point during the study (limit of detection = 12 EBV copies/reaction). Samples received from six patients and one external exchange specimen exhibited 7°C peak shifts in melting curve (55°C versus the expected 62°C) (Table 1 and Fig. 1A and B). To determine the cause of the shifted melting temperature (T_m), additional PCRs using primers specific to the EBV LMP2a gene (nucleotides 166704 to 166985) were performed. Electrophoresis of PCR products from samples with the expected (62°C) and shifted (55°C) melting curves revealed the same band size (Fig. 1C). Further, sequencing of PCR products by use of a fluorescence chain terminator method revealed a common variation (G→T) at the 166905 position of the EBV genome (EBV strain B95-8; GenBank accession no. V01555) compared to the DNAs without T_m shifts (Table 1). To confirm this sequence variation, a T_m analysis was performed by adding fluorescent probes (primer/hybridization probes; Roche LightCycler EBV assay) to the purified PCR DNA in the presence of reaction buffer (Roche LightCycler FastStart DNA Master HybProbe, vial 1b). The results confirmed that the G→T nucleotide change at position 166905 caused a 7°C shift of the melting curve (Fig. 1D).

TABLE 1.

Sequencing analysis of 13 EBV-positive samples^a

Group or sample no.	Tm (°C)	Assay date	Nucleotide at indicated position
			166750	166796	166805	166810	166896	166905
Without shifted melting curve
Cell line (B95-8)	62		T	C	A	C	C	G
1	62	7/19/2006	T	C	A	C	C	G
2	62	7/31/2006	G	A	C	T	C	G
3	62	9/15/2006	T	C	A	C	T	G
4	62	10/18/2006	T	C	A	C	C	G
5	62	10/16/2006	T	C	A	C	T	G
6	62	11/22/2006	T	C	A	C	T	G
With shifted melting curve
Sample 7
7A	55	12/27/2005	G	A	C	T	C	T*
7B	55	3/8/2006	G	A	C	T	C	T*
7C	55	5/17/2006	G	A	C	T	C	T*
Sample 8
8A	55	3/1/2006	G	A	C	T	C	T*
8B	55	7/7/2006	G	A	C	T	C	T*
8C	55	7/14/2006	G	A	C	T	C	T*
9	55	7/31/2006	G	A	C	T	C	T*
10	55	7/31/2006	G	A	C	T	C	T*
11	55	9/15/2006	G	A	C	T	C	T*
12	55	11/22/2006	G	A	C	T	C	T*
13	55	7/11/2007	G	A	C	T	C	T*

^aA T→G nucleotide change at position 166750 represents a Y→D amino acid change, A→C at 166796 represents N→T, C→A at 166805 represents P→Q, T→C at 166810 represents L→L (no amino acid change), C→T at 166896 represents Y→Y, and T→G at 166905 represents A→A. *, a change that alters the melting curve.

FIG. 1.

Melting curve analysis of EBV PCR products. (A and B) The amplification curve of a sample (sample B) with a shifted melting curve is compared with that of the EBV standard (STD). Although the threshold cycle (CP) value (28.87), copy number (343.8), and sigmoid amplification curve are consistent with a positive result for sample B (A), the melting curve analysis revealed a 7°C left shift for sample B compared to the level for the standard (channel F2/back F1) (B). (C) PCR primers specific for LMP2 were used on specimens with and without shifted melting curves (samples 2, 7C, 8B, and 8C in Table 1). Gel electrophoresis of the PCR products demonstrated the same band size. (D) After purification of PCR products, the fluorescent probe (primer/hybridization probes; Roche LightCycler EBV assay) as well as buffer (Roche LightCycler FastStart DNA Master HybProbe, vial 1b) was added to the PCR products and the melting curve analysis was performed using programs 1, 3, and 4 (95°C for 10 min, 40°C for 60 s, and 40°C for 30 s) on the LightCycler. The results showed that the samples with shifted melting curves had T_ms of 55°C, but the samples without shifted curves had T_ms of 62°C. These results confirmed that the sequence variation (G→T) at position 166905 was caused by a shifted melting curve. n/a, not applicable; N, negative EBV patient sample; W, water blank control.

Within the 282-nucleotide fragment, we identified five additional sequence variations at different locations (Table 1). It is interesting to note that the five additional sequence variations occurred only in samples without shift of the T_m. Some variations altered the amino acid sequence, while others did not (Table 1). However, these changes may not alter the viral replication or virulence (1). Based on the presented data, the specific sequence variation observed in a particular patient remains stable over time. For example, three samples, obtained on 27 December 2005, 8 March 2006, and 17 May 2006, demonstrated the same G→T change at position 16905. This finding is consistent with a previous report indicating that in chronically EBV-infected patients, the dominant EBV strains remained unchanged over several years (3).

EBV sequence variation is a recognized phenomenon (2-4) and has been used for EBV genotyping (4). Sequencing of the carboxyl-terminal region of the LMP1 gene showed that multiple variants (one to seven clones) may coexist in an individual (3). In addition, active EBV infection may be caused by endogenous EBV reactivation in posttransplant recipients (2).

It is important for a clinical laboratory performing EBV viral load testing to recognize the G→T nucleotide change at position 166905. The incidence of this particular variant represented 3% of our patient population (7 of 197 patients). Therefore, encountering this phenomenon is not a rare diagnostic circumstance. Although the precise fluorescent probe binding site is proprietary information, it has become evident that the unique EBV sequence variation interferes with one of the fluorescent hybridization probes during EBV target binding, resulting in a shifted melting curve. Therefore, carefully analyzing PCR amplification and melting curves is paramount for avoiding false-negative and false-positive data interpretation. Furthermore, samples with shifted melting curves of 62°C ± 2°C should not be hastily considered negative. Reflex testing for sequencing analysis should be performed to identify EBV sequence variants and exclude nonspecific amplification or background fluorescence.