Background
BK virus (BKV), along with JC virus (JCV) and Simian virus 40 (SV40), are members of the family
Polyomaviridae. BKV and JCV are ubiquitous in human populations worldwide with a seroprevalence in adults of 70%-80% [
1‐
4]. They establish persistent, latent infections and are capable of reactivating in immunosuppressed hosts [
5‐
7]. BKV in particular is recognized as a significant cause of allograft failure in renal transplant recipients [
8]. In addition, these viruses may also be associated with renal dysfunction in nonrenal transplant recipients [
9,
10]. Prospective monitoring of patients at risk for BKV-associated nephropathy (BKVAN) or BKV associated morbidity may identify those patients with active infection before renal function deteriorates [
11‐
13]. Early identification of active BK infection in transplant recipients is advantageous for controlling BKV replication and preventing BKVAN via reduction of immunosuppression or use of cidofovir antiviral therapy [
14,
15].
BKV screening protocols and quantitative BKV testing are increasingly performed in molecular virology laboratories. Guidelines for quantitative cutoffs for nucleic acid tests for BK viruria and BK viremia that indicate the need for additional clinical testing were proposed in 2005 [
5]. However, the usefulness of these cutoffs is hampered by the lack of standardized assays, uniform external viral standard, the existence of viral subtypes and the presence of viral polymorphisms. Primers and probes developed for quantitative BKV testing based on limited available sequence data from few viral isolates suffer reduced performance in detection of viral isolates with sequence variations in these regions. This in turn can lead to inconsistent detection of virus, inaccurate quantitation of viral load and difficulty comparing results between assays.
We describe an approach for the development of a stable nucleic acid assay when limited nucleotide sequence information is available. We report that interspecies amino acid and nucleotide sequence analysis, in conjunction with intraspecies nucleotide sequence alignment, can elucidate genomic regions that may be under potential functional constraints and that these regions can be targeted for primer and probe design to improve assay performance.
Discussion
There have been previous efforts to develop real-time PCR based BKV quantification assays [
27,
28]. However, many of these assays were developed when the number of available DNA sequences in public databases was quite limited. Although multiple sequence alignments are routinely used to identify stable target regions for assay development, sequence variants have been observed in these regions as more strains and sequences become available. Therefore, when the number of isolates and sequences is limited, merely selecting a target region for assay development based on intraspecies nucleotide sequence alignments alone may not be sufficiently predictive of interstrain sequence conservation to result in a sensitive and specific assay.
We proposed that identifying areas of the genome that are conserved within and between species could be used to supplement limited nucleotide sequence information when designing assays. Target regions identified via standard intraspecies sequence alignment of full genome sequences from available isolates were further evaluated by interspecies amino acid analysis to select a single target region. The interspecies amino acid alignments assisted in narrowing down potential target regions by identifying areas containing functionally and/or structurally constrained sequences. Conservation in amino acid sequence among closely related viruses does not necessarily preclude the possibility of non-synonymous substitutions. Functional and structural constraints, however, do suggest that there would likely be fewer number of amino acid residues that can reside at that particular position than if the same region was under less stringent functional or structural constraint.
However, mutations in the nucleotide sequences giving rise to synonymous substitutions can potentially have equally adverse effects on the performance of a PCR assay. Possible mechanisms for constraints limiting synonymous mutations may include overlapping reading frames, especially if those overlapping regions are functionally or structurally important. Identifying potential cold spots in the genome that are less likely to undergo synonymous substitutions is a significant challenge and a major obstacle to developing a robust PCR-based viral assay. Intra-species nucleotide alignment of available BKV sequences was followed by detailed intra- and interspecies nucleotide analysis of the target region to aid in designing assay primers and probe.
In the present work, the utilization of the intraspecies nucleotide alignment of BKV with the intraspecies nucleotide alignment of a closely related virus, JCV, for which a significant number of sequences were available, proved to be useful in substantiating that the Polyoma_4.2(f) and the BKV-MGB probe binding regions within BKV was less likely to experience nucleotide sequence variation. The lack of polymorphisms at the corresponding positions in both BKV and JCV isolates provides stronger evidence for sequence conservation than does intra-species nucleotide analysis of BKV isolates alone. Conversely, the forward primer BKV_2.1(f) is reserved for use only as a secondary backup to Polyoma_4.2(f) due to the extent of polymorphisms seen among JCV isolates along the nucleotide stretch corresponding to the BKV_2.1(f) binding site.
Since the assay was first developed using ~160 BKV sequences, the number of available BKV complete genome sequences increased to 271 (Additional file
1). As presented in the results section, the increase in BKV sequence data did not demonstrate increased sequence variation in the region targeted by this analysis, validating the approach of incorporating the additional analyses discussed herein when designing a real-time PCR assay.
The amino acid and nucleotide sequence analyses that were performed identified the VP2 C-terminus as a probable functionally conserved region, consistent with the presence of an alpha-helix [
18] and nuclear localization signal [
19‐
22]. Targeting the VP2 gene for assay development is in contrast to the location of primers/probes for most existing BKV assays [
27]. Therefore this approach was successful in identifying a region different from those targeted by other assays that have become limited with the recognition of additional sequence variations.
Using functionally constrained regions for assay development might be expected to result in lack of specificity. However, sufficient nucleotide differences existed within the stable region to design an assay specific for BKV. The three nucleotide difference between the BKV and JCV at the probe binding site was sufficient to meet assay specificity requirements by utilizing a Taqman®-MGB probe. By having the probe bind to the anti-sense strand, the respective positions of the inter-species nucleotide differences between the BKV and JCV were such that the minor groove binder of the Taqman®-MGB probe could be ideally positioned to provide a high degree of specificity. This configuration also provided the added benefit of having the probe bind to the same strand (anti-sense) as the BK virus specific reverse primer rather than on the sense strand that binds the Polyoma_4.2(f) primer, which amplifies both JCV and BKV. It should be noted that the primer combination BKV_5.1(r) and Polyoma_4.2(f) is the only primer combination that has been designed and tested to perform under a BKV and JCV co-infection scenario.
Preliminary assay performance results indicate the BKV assay has a 6-log dynamic range, a lower detection limit of 1.5 × 101 copies/reaction and an intra-assay CV of ≤ 5.0%. Good precision is important for distinguishing between assay variation and assay results that are associated with clinical treatment or indicative of the need for treatment. A wide dynamic range is significant for clinical application of the assay since the clinical need is for identification of patients with high viral load who may be at immediate risk for developing BKVAN and also for identification of low viral loads in patients being monitored for BKV reactivation. The assay also does not detect JCV and the presence of JCV does not inhibit detection of BKV at the concentrations tested, suggesting that accurate detection and quantitative results for BKV will be obtained for BKV clinical samples that are co-infected with JCV.
The relative lack of sequence variants in the VP2 gene suggests that an assay targeting this region will permit accurate detection of BKV and estimates of viral load regardless of subtype (Additional file
2). Such information is important in identifying active BKV infections in transplant recipients in time to initiate appropriate treatment. Further testing of this BKV assay is needed to fully characterize assay performance characteristics, clinical performance and to evaluate the impact of sequence variations.
Methods
Sequence alignments: Amino Acid & DNA
A total of 185 amino acid sequences for the BKV VP2 gene were obtained from the NCBI database. The amino acid sequences of the VP2 genes for the other members of the family
Polyomaviridae were collected for species in which the information was available. Multiple sequence alignment of the amino acid sequence of the VP2 gene was performed using ClustalW2
http://www.ebi.ac.uk.
At the time of preliminary assay development, a total of 160 nucleotide sequences of the Agno protein, 159 nucleotide sequences of the large-T antigen, 157 nucleotide sequences of the small-T antigen, and 159 nucleotide sequences of the three structural genes, VP1, VP2, and VP3 were obtained from the NCBI database
http://www.ncbi.nlm.nih.gov. Multiple alignments for the sequences of the six respective genes were performed using ClustalW2
http://www.ebi.ac.uk and MUSCLE
http://www.ebi.ac.uk.
The entire amplicon sequence was input into the BLAST software
http://www.ncbi.nlm.nih.gov to conduct species-specific searches and to analyze intra-species polymorphisms as well as interspecies sequence variation amongst closely related, clinically significant viruses. The BLAST search and results formatting were adjusted according to the number of available sequences within the database for each specific species. The results were formatted with inquiry sequence anchored with stars for identity to generate a multiple sequence alignment of the BLAST results.
Oligonucleotides
The theoretical Tm calculations as well as secondary structure and primer-dimer analysis for oligonucleotides were performed using OligoCalc [
29] & Primer3 [
30]. The Tm under the actual reaction conditions was confirmed empirically. The need for a probe was confirmed by empirically testing for primer-dimer formation using the LightCycler
®480 SYBR Green I Master reagent as well as post-amplification gel-electrophoresis.
Primers were purchased from Invitrogen Inc. (Carlsbad, CA). Two forward and two reverse primers were designed to amplify the C-terminus region of the VP2/3 gene. The BKV_5.1(r) comprised the primary reverse primer of the assay while the BKV_4.1(r) was utilized as an alternate. The reverse primer denoted as BKV_5.1(r) comprises the nucleotide sequence 5'-CTG CCC CTG GAC ACT CTC-3'. The reverse primer denoted as BKV_4.1(r) has the sequence 5'-TCA (I)AT GCT TCA AGA GCA GGT GT-3', wherein the nucleotide (I) represents deoxyinosine. The nucleotide sequence of the two forward primers denoted as Polyoma_4.2(f) and BKV_2.1(f) are 5'-GCT CCT CAA TGG ATG TTG CCT-3' and 5'-CCC AGG AGG TGC TAA TCA AAG A-3', respectively. The Polyoma_4.2(f) comprised the primary forward primer of the assay while the BKV_2.1(f) was utilized as an alternate.
TaqMan®-MGB probes were supplied by Applied Biosystems Inc. (Foster City, CA). The BKV probe denoted as BKV-MGB probe (5'-6FAM-CCG TAC AGG CCT AGA A-MGB-NFQ-3') was tagged with 6FAM™ at the 5'-end and tagged with a MGBNFQ (Minor Groove Binder/Non-fluorescent quencher) at the 3'-end. The JCV probe denoted as JCV-MGB probe (5'-VIC-CGT ACA ACC CTA AAA G-MGB-NFQ-3') was tagged with VIC® at the 5'-end and MGBNFQ (Minor Groove Binder/Non-fluorescent quencher) at the 3'-end.
Viral DNA Standards
The BKV (strain MM) and JCV (strain MAD1) DNA standards were obtained from Advanced Biotechnologies Inc. (Columbia, MD), supplied at 1-2 × 105 DNA copies/μl. Viral standard at 1.5 × 106 DNA copies/μl were produced from clinical BKV isolates using TOPO® TA Cloning kit (Invitrogen Inc., CA, U.S.A.). Cloned viral DNA was quantified on the Roche LightCycler®480 using the commercially purchased standard.
Quantitative real-time PCR
Quantitative real-time PCR was performed on a Roche LightCycler®480 instrument. PCR amplification was run in a final volume of 40 μl per reaction comprising 30 μl of PCR master mix and 10 μl of sample, standard or PCR-grade water. PCR master mix for one reaction comprised of: 0.42 μM forward primer, 0.42 μM of reverse primer, 0.33 μM of TaqMan®-MGB probe, and 20 μl of LightCycler® 480 Probes Master (2×).
Thermal cycling conditions when using the reverse primer BKV_4.1 were as follows: pre-incubation at 95°C for 10 minutes; 45 cycles of a three-step PCR wherein one cycle comprising of 95°C for 10 seconds (ramp rate 4.4°C/s), 60°C for 15 seconds (ramp rate 2.2°C/s) and 72°C for 1 second (ramp rate 4.4°C/s); and ending with a cool down step at 40°C for 30 seconds. Fluorescence was measured at each cycle following the third step at 72°C.
Thermal cycling conditions when using the reverse primer BKV_5.1 were as follows: pre-incubation at 95°C for 10 minutes; 45 cycles of a three-step PCR wherein one cycle comprised 95°C for 10 seconds (ramp rate 4.4°C/s), 58°C for 20 seconds (ramp rate 2.0°C/s) and 65°C for 1 second (ramp rate 4.4°C/s); and ending with a cool down step at 40°C for 30 seconds. Fluorescence was measured at each cycle following the third step at 65°C.
BKV external standards at seven different concentrations (106 copies/μl, 105 copies/μl, 104 copies/μl, 103 copies/μl, 102 copies/μl, 10 copies/μl and <10 copies/μl) were included with each assay in order to generate the standard curve. Linear regression of the external standards and quantification of unknowns were performed using the LightCycler® 480 software.
Specificity testing. Specificity data were generated by testing analytical samples containing BKV, JCV and mixed samples containing both BKV and JCV. Serial dilutions of BKV to result in from 1.5 × 101 copies/reaction to 1.5 × 107 copies/reaction were tested individually (2 replicates of each concentration) and in the presence of 7 × 103 copies/reaction of JCV (10 replicates of each concentration). JCV at 1.4 × 104 copies/reaction (10 replicates) was also tested individually. A total of 94 specificity samples were tested. Five replicates of a negative control were also tested.
Analytical sensitivity evaluation. Analytical sensitivity samples comprised of BKV at concentrations of 150, 75, 15 and 1.5 copies/reaction were used to evaluate assay sensitivity. The BKV samples were prepared by serial dilution of a BKV standard. A total of 120 samples, 10 replicates of each concentration with 3 primer combinations, were tested in the assay. Five replicates of a negative control were also tested.
Assay Precision. The intra-assay precision of the BKV assay was evaluated. The samples were constructed by serial dilution of BKV standard and contained from 1.5 × 101 copies/reaction to 1.5 × 107 copies/reaction. Multiple replicates (10) of the seven different concentrations of the BKV standard were tested with 3 primer combinations in one run and the variance in the Cp (crossing point) calculated to determine the reproducibility of the assay. A total of 210 samples were tested. Five replicates of a negative control were also tested.
Competing interests
The primary author is listed as an inventor in a published International Patent Application filed by Metic Immunogenetic Consultant, Inc., entitled "Detection of Polyomavirus".
Authors' contributions
KKI carried out the assay design, sequence alignment and assay performance studies. SHQ, YM, DZ, MEM and JG-G assisted with the assay optimization and cloning. MS and YQ participated in acquiring the samples. MT participated in assay validation. KM, YQ, and YI participated in the conception and design of the studies. AK reviewed patient charts and confirmed clinical diagnoses. WTS reviewed histopathological findings to confirm clinical diagnoses. DDI reviewed test results and conducted further blind testing using this diagnosing system. All authors read and approved the final manuscript.