TEXT

Top Abstract Text References

Comparison of the nucleotide sequences of different hepatitis G virus (HGV) (also known as GB virus C [GBV-C]) isolates reveals a certain degree of genetic variability, similar to that found among hepatitis C virus strains (7, 11). Therefore, if specific consensus primers from highly conserved regions of the viral genome were designed, the chance of false-negative results could be reduced. Degenerate primers from the helicase gene were the first to be used in reverse transcription-PCR for detection of GBV-C/HGV RNA. However, due to the high level of nucleotide sequence divergence (17%) in that region, new oligonucleotide primers based on other, more conserved areas of GBV-C/HGV, such as the 5' untranslated region (5'-UTR), were constructed (4, 8, 11). Several sets of primers from the 5'-UTR have been used for single-round (10, 11) or nested (6, 13) PCR, and the authors claimed that there was an increased overall sensitivity for GBV-C and HGV detection. There are no published reports on the direct comparison of different sets of primers from the same genomic region and their specificities and sensitivities in PCR. We report here the systematic investigation of six sets of PCR primers: four from the 5'-UTR and two from the 3' end of the genome (the NS5a region). In addition, we have optimized the PCR protocol particularly for use with AmpliTaq Gold DNA polymerase, which has an increased functional activity (3).

For the present study, we selected 44 clinical samples that were previously found to be positive for GBV-C/HGV RNA by PCR with at least one of three sets of primers (no. 1, 2, and 5) (Fig. 1; Table 1) and 45 sera that were negative by PCR with primer set 5. To broaden the spectrum of primers, two additional sets of nested primers from the 5'-UTR and one set of single-round PCR primers from the NS5a region (Fig. 1; Table 1) were also included.

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Relative locations of GBV-C/HGV primers.

RNA was extracted from 0.1 ml of serum by the guanidinium thiocyanate-phenol-chloroform method (5) and reverse transcribed by using random hexamers [p(dN6); Pharmacia-LKB Biotechnology] and 200 U of Moloney murine leukemia virus reverse transcriptase (Bethesda Research Laboratories, Gaithersburg, Md.) at 42°C for 1 h. The cDNA from each sample was amplified in parallel with four sets of single-round PCR primers (Fig. 1, primer sets 1, 2, 5, and 6) and two sets of nested-PCR primers. The published sequences for all six sets of primers are shown in Table 1. For both first- and second-round PCRs, amplification with nested primers involved 25 cycles consisting of stepwise denaturation for 45 s at 94°C, annealing for 45 s at 55°C, and elongation for 90 s at 72°C with the commonly used, thermostable Taq DNA polymerase. The single-round PCR with primer sets 1, 2, 5, and 6 was carried out with the thermally activated AmpliTaq Gold DNA polymerase as follows, at the times and temperatures noted: an initial activation of the enzyme within the PCR mix at 95°C for 12 min followed by 60 cycles of amplification consisting of 94°C for 45 s, 55°C for 45 s, and 72°C for 45 s. Guidelines to avoid contamination (9) were strictly followed, with one positive and two negative controls being included for every 10 samples. Amplified products were separated by agarose gel electrophoresis and visualized under UV illumination. Subsequently, the separated products were blotted on Zetaprobe nylon membranes (Bio-Rad, Hercules, Calif.), immobilized by UV cross-linking, and hybridized with ³²P-end-labeled oligonucleotide probes (1). Samples that were found to be negative after SYBR Green 1 (FMC, Rockland, Maine) staining were scored as positive if, following Southern blotting and autoradiography, a hybridizable band of the expected size was observed.

AmpliTaq Gold, which is a modified Taq DNA polymerase, does not exhibit any activity at room temperature and requires a pre-PCR heating cycle that serves at the same time as a hot start. To optimize the PCR protocol with this enzyme, preliminary tests (data not shown) were run; they demonstrated an increased specific band yield at 60 cycles in comparison with the yield from 40 or 50 cycles. The specific band yield was as strong or stronger at 70 or 80 cycles; however, degradation of the PCR band could also be observed, which obscured the exact size of the "overamplified" product. The standard GeneAmp 10× PCR buffer (Perkin-Elmer) was used with 2.5 mM magnesium chloride because weaker specific bands were observed at 1.25 and 3.75 mM concentrations. The AmpliTaq Gold DNA polymerase was 10 times more sensitive in detection of HGV RNA than conventional Taq DNA polymerases, as demonstrated in a titration experiment using serial 10-fold dilutions of two viremic serum samples (data not shown). Another advantage of this enzyme was a significant reduction in the background despite the increased number of PCR cycles. All 45 samples that were determined to be negative for GBV-C/HGV RNA by PCR with primer set 5 were confirmed to be true negatives by using the other primer sets. A comparison of the six PCR protocols, four using primers from the 5'-UTR and two using primers from NS5a, is shown in Table 2.

The best performance in terms of PCR sensitivity was achieved with primer sets 1, 3, and 4, all from the 5'-UTR. However, another set of primers (no. 2) from the same region performed poorly, missing 6 of 44 HGV-positive serum samples. All single-round PCR sets required radiolabeled hybridization to further detect GBV-C/HGV-positive sera (from 2.3% for set 6 to 22.7% for set 5). Primer sets 5 and 6, located in the NS5a region, missed three positive samples (no. 677 was missed by both sets, no. 671 was missed by set 5, and no. 64 was missed by set 6). These three samples were positive with 5'-UTR primer sets 1, 3, and 4. Thus, in terms of sensitivity, we may assume that the two sets from the NS5a region performed equally well; however, amplification with set 6 seems to be more efficient, since 93.2% of the samples identified as positives did not require subsequent hybridization with radiolabeled oligonucleotides, compared to 72.7% with set 5.

Primer sets 1 and 3 each missed one (no. 682) of the 44 GBV-C-/HGV-positive serum samples. The same sample was found to be positive with both primer sets from the NS5a region and with sets 2 and 4 from the 5'-UTR. The availability of other PCR primer sets from the 5'-UTR allowed us to discover which of the primers of these two sets was responsible for the failure to detect HGV viremia in this patient. Thus, we used two different combinations of primers from sets 1 and 2: S1 forward plus B2 reverse, and B1 forward plus A2 reverse (Fig. 1 and Table 2). A positive result was scored with the combination B1-A2, but not with S1-B2, which suggests a possible mismatch of primer S1 with the cDNA template of this particular HGV isolate (no. 682). Sequencing of this primer binding region was hampered by the extreme 5' location of primer S1 and the preceding short pyrimidine tract, which do not allow the designing of a proper upstream primer. The same strategy was applied for nested-primer set 3, which also had not identified this sample as a positive. A combination of sets 2 and 3 included the forward B1 primer (set 2) and both reverse primers from set 3, B1-531 (outer) and B1-476 (inner) (Fig. 1). An amplified product of the expected size (377 bp) was observed with the combination B1-476 (inner), but not with B1-531 (outer), after a single round of PCR and gel electrophoresis. Five microliters of the PCR mix from the first round of PCR with each of these two combinations was subjected to further amplification with the inner primers from set 3 (134 plus 476), and a strong specific band of 343 bp was observed only with the combination B1-476, not with B1-531. Thus, a mismatch for the reverse primer 531 and the cDNA template of this particular HGV isolate was suspected. An attempt to amplify this region by utilizing the reverse primer GBVCE1wb (5'-CAGGGCGCAACAGTTTGGGAG-3') (11), located downstream of primer 531 (nucleotides 638 to 659), failed, and thus sequences for the primer 531 binding region could not be obtained for this HGV isolate.

Primer set 4 detected 100% of all GBV-C/HGV-positive samples, although 6 of the originally selected 44 samples were not included because there was not enough serum available. Primer set 2 performed poorly in comparison with all other sets due to the inadequacy of reverse primer B2. When replaced with A2 (i.e., when the combination B1-A2 was used), all six samples that were missed with this set were scored as positives after a single round of PCR. Sequencing of the region from which this primer was derived in North American GBV-C/ HGV isolates (2, 10) showed up to five mismatches, which explains its inefficiency in PCR. This is one of two commercially available sets sold under the name HGV Primer and Capture Probe Set, which is used with the Enzymun-Test DNA Detection System (Boehringer Mannheim) for HGV diagnostics (11). Replacement of primer B2 with any of the reverse primers from the 5'-UTR (A2, 531, 476, G131, or G75) would greatly improve this assay, which otherwise allows chemiluminescence detection instead of hybridization with radiolabeled oligonucleotides and has a user-friendly microtiter plate format. The reverse transcription-nested PCR also does not require radioactive labels, and assuming that cross-contamination can always be prevented, it may be the method of choice for detection of GBV-C/HGV RNA. However, our results indicate that GBV-C/HGV yields may be higher than those in hepatitis C virus infection, since the majority of the samples tested were positive after only one round of PCR and then staining with SYBR Green 1 (Table 2). Thus, the relatively high GBV-C/HGV concentrations may facilitate cross-contamination in nested-PCR protocols.

In conclusion, primer sets 1, 3, and 4 from the 5'-UTR are almost equally effective for detection of GBV-C/HGV RNA. Our data suggest that although the 5'-UTR is highly conserved, genetic heterogeneity may still cause false-negative results in a very small proportion of samples (2%). Routine testing using one primer set from the 5'-UTR in parallel with a second set from the NS5a region may further improve the detection of this virus and, at the same time, confirm the findings of the 5'-UTR-based PCR.