Comparison of three different HCV genotyping methods: Core, NS5B sequence analysis and line probe assay
- Qingxian Cai
- Zhixin Zhao
- Ying Liu
- Xiaoqiong Shao
- Zhiliang Gao
- Corresponding author:
- Zhiliang Gao [ firstname.lastname@example.org ]
Published online on: Wednesday, December 12, 2012
- Pages: 347-352
- DOI: 10.3892/ijmm.2012.1209
The hepatitis C virus (HCV) is an enveloped virus with a single stranded, positive sense, nonsegmented RNA genome of approximately 9,500 nucleotides that encodes a polyprotein of approximately 3,000 amino acids (1,2). Analysis of the HCV genome has demonstrated extremely high heterogeneity in both structural and nonstructural coding regions and has identified at least six different genotypes (1 to 6) that have generally been divided into several subtypes (designated a, b, c) (3–5). These genotypes have distinct geographical distributions. Although HCV genotypes 1, 2 and 3 appear to have a worldwide distribution, their relative prevalence varies from one geographic area to another. HCV genotype 4 was found in the Middle East and North Africa, and genotypes 5 and 6 in South Africa and Asia, respectively (6–11).
Genotype identification is clinically important for prediction of responses to, and in determining the duration of, antiviral therapy (12). This is illustrated by the fact that genotypes 1 and 4 were more resistant to treatment with pegylated interferon-α and ribavirin than genotypes 2, 3 and 6 (13–15). Patients with chronic HCV genotype 1b infection showed more severe liver disease than patients infected with other genotypes (16). At present, most treatment protocols require genotype information for patients infected with HCV.
A variety of technologies have been developed for HCV genotype determination. The majority of these assays rely on the amplification of short HCV RNA regions from clinical specimens, followed by a type-specific assay, such as restriction fragment length polymorphism analysis (17), line probe reverse hybridization (18,19), or sequence analysis (20,21). Almost all available commercial assays target the 5′-untranslated region (5′-UTR), as the highly conserved sequences of this region are most suitable for reverse transcription-PCR (RT-PCR) amplification.
The Versant HCV genotype assay (LiPA) is one of the most widely used methods for HCV genotyping. In this assay, the 5′-UTR of HCV is amplified with biotinylated primers, after which the PCR product is hybridized to a membrane impregnated with genotype-specific probes and detected with streptavidin linked to a colorimetric detector (22). Despite the high conservation of the 5′-UTR, genotype determination of HCV based on the 5′-UTR is accurate for most genotypes (21,23,24). However, it has been noted that methods that are based on the use of the 5′-UTR falsely identify genotypes 6c to 6l from SEA as genotype 1, which is also the case in the Versant HCV genotype assay (22,25). Moreover, this assay is unable to distinguish genotype 1a from 1b in 5–10% of the cases (24,26). To improve the accuracy of distinguishing between genotypes 1a, 1b, and 6c to 6l, a new generation of the line probe assay (Versant HCV genotype 2.0 assay, LiPA 2.0) that uses core sequence information in addition to the 5′-UTR was developed.
Sequencing and phylogenetic analysis of the core/E1 or NS5B region were considered to be the gold standard for HCV genotyping since they accurately identified the subtype and were used to establish an epidemiological picture of circulating virus strains (27). Although such assay involves complicated and time-consuming experimental procedures of RNA extraction, reverse transcription, nested PCR, DNA sequencing and phylogram construction, genotyping HCV by sequence analysis has become increasingly more convenient than before with the continuous improvement of biochemical reagents and experimental techniques.
In the present study, three methods, i.e., core, NS5B sequence analysis and line probe assay (LiPA 2.0), were evaluated for their effectiveness in identifying HCV genotypes/subtypes in China.
Materials and methods
One hundred and ten patients from Guangdong, China were diagnosed with chronic hepatitis C in the Third Affiliated Hospital of Sun Yet-sen University. Prior to antiviral treatment, 5 ml serum samples were collected from each patient. All patients had signed the informed consent.
HCV viral load tests
The viral loads were determined with Amplicor HCV monitor version 2.0 (Roche, Meylan, France). According to the supplier’s instructions, serum samples were not prediluted except as otherwise stated. RNA extraction, reverse transcription, amplification, detection, and calculation of the number of HCV copies per ml were performed according to the manufacturer’s protocol. Briefly, 100 μl of serum was added to 400 μl of lysis buffer containing the internal quantitative standard (IQS). After incubation for 10 min at 60°C and precipitation, the pellet was diluted in 1 ml of specimen diluent. For PCR, 50 μl of diluted extract was transferred into reaction tubes containing 50 μl of the PCR master mixture. Reverse transcription was performed at 60°C for 30 min, and amplification was performed for 2 cycles at 95°C for 15 sec and 60°C for 20 sec and then 33 cycles at 90°C for 15 sec and 60°C for 20 sec in GeneAmp® PCR System 9700 (ABI, Alabaster, AL, USA). Following amplification, 100 μl of denaturing reagent was transferred to each reaction tube, and the tubes were incubated for at least 10 min. Detection was performed on microtiter plates coated with capture oligonucleotides specific for HCV sequences and containing 100 μl of hybridization buffer/well by pipetting 25 μl of the amplification product into the wells of the first row and then performing 1/5, 1/25, 1/125 and 1/625 dilutions in the following rows. Similarly, a 1/1, 1/5 and 1/25 dilution series was performed in wells coated with capture oligonucleotides specific for the IQS. Following incubation, washing, and the color reaction, the OD of each well was measured in an ELISA reader RT-6000 (Rayto, Shenzhen, China) at A450. The HCV concentration was calculated by multiplying the OD value of HCV and IQS of the highest dilution that gave an OD of between 0.5 and 2.0 by the respective dilution factor and dividing the value for HCV by the value for the IQS. The result was then multiplied by 100 for the IQS copies and a sample dilution factor of 200, resulting in the number of HCV copies/ml of serum. For each sample, the OD for the IQS was above 0.5 in at least one of the wells and the ODs of HCV and IQS were subtracted by the background value according to the supplier’s instructions.
RNA extraction and reverse transcription
RNA was isolated from the first RNA-positive serum sample obtained from each patient using 500 μl serum and a RNAiso™ Plus extraction kit (Takara, Dalian, China). HCV RNA was eluted in 10 μl of Tris-EDTA (TE) buffer and was subsequently transcribed into cDNA using the ReverTra Ace-α-reverse transcription kit (Toyobo, Shanghai, China). This cDNA was used as the input for two separate PCR assays targeting the HCV core and NS5B regions.
Amplifying and sequencing core and NS5B fragments
The core and NS5B regions were amplified using a nested polymerase chain reaction (nPCR). Primers used for amplifying the core region were the same as previously reported Lole et al (28). The core outer primers were: forward, 5′-ACTGCCTG ATAGGGTGCTTGC-3′ and reverse, 5′-ATGTACCCCAT GAGGTCGGC-3′; the inner primers were: forward, 5′-AGG TCTCGTAGACCGTGCA-3′ and reverse, 5′-CATGTGAG GGTATCGATGAC-3′. The primers used for amplifying the NS5B region derived from Laperche et al (29). The NS5B outer primers were: forward, 5′-CNTAYGGITTCCA RTACTCICC-3′ and reverse, 5′-GAG GARCAIGATGTT ATIARCTC-3′; the inner primers were: forward, 5′-TATGA YACCCGCTGYTTTGACTC-3′ and reverse, 5′-GCNGAR TAYCTVGTCATAGCCTC-3′.
Outer PCR (30 μl): 10X PCR buffer 3 μl 2.5 mM dNTP 2 μl, dH2O 17.6 μl, primers (10 pmol/μl) 1.5 μl, Taq enzyme (2.5 U/μl) 0.4 μl, template cDNA 4 μl. Inner PCR system (30 μl): 10X PCR buffer 3 μl, 2.5 mM dNTP 2 μl, dH2O 19.6 μl, primers (10 pmol/μl) 1.5 μl of each, Taq enzyme (2.5 U/μl) 0.4 μl, template cDNA 2 μl. PCR conditions were: 94°C for 5 min; 94°C for 30 sec, 55°C for 1 min, 72°C for 40 sec, 30 cycles; 72°C for 10 min.
Following verification by agarose gel electrophoresis, the PCR products were sent to the Beijing Genomics Institute at Guangzhou for sequencing.
Basic local alignment search tool (BLAST) and phylogenetic analysis were used to identify HCV genotypes. First, the nucleotide sequences from the core and NS5B regions were analyzed by HCV BLAST in the Los Alamos HCV sequence database (http://hcv.lanl.gov/content/index). Then, using the ClustalW 1.8 software package, the sequences of HCV strains were then aligned with a reference panel of sequences representative of each subtype provided by the Los Alamos National Laboratory (30). Pairwise distances were generated using the Jukes-Cantor corrected distance algorithm of the program MEGA 4.0 (31). Phylogenetic analysis was performed using the neighbor-joining method for tree drawing. The reliability of phylogenetic classification was evaluated by a 1,000-cycle bootstrap test.
The LiPA 2.0 was a reverse hybridization line probe assay in which biotinylated DNA PCR products were hybridized to immobilized oligonucleotide probes that were specific for the 5′-UTRs and core regions of the six HCV genotypes. The probes were bound to a nitrocellulose strip by a poly (T) tail. After hybridization of the biotinylated targets to the probes, unhybridized PCR products were washed from the strips, and alkaline phosphatase-labeled streptavidin (conjugate) was bound to the biotinylated hybrid. After washing the strips, 5-bromo-4-chloro-3-indolylphosphate (BCIP)-nitroblue tetrazolium chromogen (substrate) reacted with the conjugate, forming a purple/brown precipitate, which resulted in a visible line pattern on the strip that was specific for each genotype. Each strip had three control lines and 22 parallel DNA probe lines containing sequences specific for HCV genotypes 1–6. The conjugate control at line 1 monitored the color development reaction and gave a positive result if the strip was processed correctly. The amplification control at line 2 (AMPL CTRL 1) contained universal probes that hybridized to PCR products from the 5′-UTR. AMPL CTRL 2 was located at line 23 and contained universal probes that hybridize to PCR products from the core region. HCV genotypes were determined by aligning the strips with a reading card and comparing the line patterns from the strip with the patterns on the interpretation chart.
Statistical analysis was performed with the SPSS 17.0 software package. Statistical significance was defined as a 2-sided P-value of ≤0.05. The differences in the distribution of categorical variables were assessed by Pearson’s Chi-squared test.
Determination of HCV genotype by sequence analysis
Of the 110 HCV samples, 8 failed to be amplified in any region, 57 were amplified in both regions, 40 were amplified in the core region only and 5 were amplified in the NS5B region only (Tables I and II). The amplification rate of the core region (92.7%) was significantly higher (P<0.001) than that of the NS5B region (56.4%). Correlation analysis revealed that amplification rate of both regions was correlated with viral load (Table I). When the viral load was ≥1.E+03 IU/ml, the amplification rate of the core region was satisfactorily high (87.5–100%). However, it decreased markedly (P<0.001) to 41.7% when the viral load was <1.E+03 IU/ml. A similar phenomenon was observed in amplifying the NS5B region. When the viral load was ≥1.E+04 IU/ml, the amplification rate ranged from 62.1 to 73.9%, whereas when the viral load was <1.E+04 IU/ml, it decreased significantly (P<0.001) to only 26.7% at 1.E+03–1.E+04 IU/ml and 16.7% at <1.E+03 IU/ml.
Amplicons of the core and NS5B regions were sequenced directly. Then, the sequences were submitted to GenBank (accession no. JN572940 to JN572983) and were aligned with all genotyped sequences in the Los Alamos HCV sequence database using HCV BLAST (http://hcv.lanl.gov/content/sequence/BASIC_BLAST/basic_blast.html). All the core sequences hit highly similar sequences in the database with scores >450 bit and identities ≥97%. Similarly, all the NS5B sequences matched sequences from the database with scores >450 and identities ≥93%. Since the sequence similarities in the core and NS5B regions between tested sample and strains from the HCV database were >93 and 87.8% (which were the thresholds between variants of the same genotypes for the core and NS5B regions, respectively) (27,38), genotypes of sequences at the top BLAST outputs were considered as identical to the tested samples. Based on the BLAST result, genotypes identified by core and NS5B sequence analysis were compared. The result showed that genotypes assigned by sequence analysis of both regions were identical. In summary, of the 102 HCV samples amplified in either or both regions, 63 (61.8%) were identified as subtypes 1b, 10 (9.8%) as 2a, 4 (3.9%) as 3a, 4 (3.9%) as 4a, and 21 (20.6%) as 6a (Table II).
To further verify genotypes of HCV, phylograms were constructed with the reference stains for the core and NS5B regions (Fig. 1). The results showed that the core sequences of 97 samples and the reference strains of genotypes 1b, 2a, 3a, 3b and 6a were grouped into five clusters. The bootstrap values of each cluster exceeded 80%, indicating that the topology of core sequence was highly reliable. The genotypes identified by phylogenetic analysis were identical to those assigned by BLAST analysis. Similarly, the NS5B sequences were grouped into five clusters with high bootstrap values (74–96%), and the results of phylogenetic analysis were consistent with those of BLAST analysis. Comparison of the phylograms of the core and NS5B regions showed that the topology of both regions was similar, but the Jukes-Cantor molecule genetic distance of NS5B was larger than that of core, which was consistent with the fact that the NS5B region is more variable than the core region.
Phylogenetic tree of HCV sequences. Left, phylogenetic tree constructed based on core sequences. Right, phylogenetic tree constructed based on NS5B sequences. Subtypes 1a, 2a, 3a, 3b and 6a are shown with 5 different colors in the phylogenetic tree. The reference sequences of HCV variants were cited from Kuiken et al (30). Accession nos: M62321 (1a); D90208 (1b); D00944 (2a); D10988 (2b); D50409 (2c); AB031663 (2k); D17763 (3a); D49374 (3b); D63821 (3k); Y11604 (4a); Y13184 (5a); Y12083 (6a); D84262 (6b); D84263 (6d); D63822 (6g); D84265 (6h); and D84264 (6k).
Genotyping by LiPA 2.0
Table III shows that except for genotype 6, the other genotypes were distinguished correctly by LiPA 2.0 at the genotype level. However, at the subtype level, only 1b and 3b were distinguished accurately.
The accuracy for genotyping 1b was 100% with LiPA 2.0. Interpreting the results obtained from all genotype 1 samples with LiPA 2.0 according to the amplified region, 44 (69.8%) were correctly genotyped by taking into account the 5′-UTR alone, whereas 63 (100%) were genotyped correctly using the additional information on the core region. In this study, 8 (80%) and 2 (20%) of the subtype 2a samples were incompletely classified into subtype 2a/2c and genotype 2, respectively. LiPA 2.0 was unable to completely distinguish 2a from 2c, due to the lack of probes specific to 2a and 2c. Despite all 8 of the genotype 3 samples (4 were 3a, and 4 were 3b) correctly identified at the genotype level, there were 37.5% misidentified at the subtype level (i.e., 3 subtype 3a samples were misidentified as 3b). Compared with the other genotype, the accuracy of identifying genotype 6 by LiPA 2.0 was low. Ten of the 21 genotype 6a samples were incompletely classified as 6a/6b, and the others (52.4%) were misclassified into 1b. Considering the whole panel, the overall rate of concordance (correct genotype and correct subtype) was 66.7% (68 samples) for LiPA 2.0. The percentage of incomplete results (indistinguishable or not identified subtype) was 19.6% (20 samples). Misclassifications were observed for 13.7% (14) of the tested samples.
The most accurate method for genotyping is sequencing the entire genome of HCV. However, this is time-consuming and difficult to apply on a large scale. Despite the sequence diversity of HCV, all genotypes share an identical component of co-linear genes of similar or identical size in the large open reading frame, and the genetic inter-relationships of HCV variants are remarkably consistent throughout the genome (4). This has enabled many of the recognized variants of HCV to be provisionally classified, based on partial sequences from subgenomic regions (27). Core/E1 or NS5B are considered the most reliable regions for genotyping HCV, as sufficient genetic diversity is presented in these regions. Although genotyping results in these regions were consistent, amplification rates of these regions appear different. In the present study, the amplification rate of the core region was higher than that of NS5B with the same viral load, which indicated that the amplification rate of subgenomic regions was not only associated with viral load, but also with sequence conservation. Since the core region was more conservative than NS5B, to ensure a satisfactory amplification rate and high genotyping accuracy, amplification of the core region was recommended.
The accuracy of genotyping HCV by BLAST analysis depends on the number of genotyped sequences in the HCV database. Recently, with genotyped sequences being continuously submitted, genotyping HCV by BLAST analysis has become increasingly more reliable. In the present study, all the sequences of tested samples hit highly similar genotyped sequence in the HCV database. These results highly support BLAST analysis. Moreover, compared with phylogenetic analysis, BLAST analysis was simpler and more timesaving.
The accuracy of identifying genotype 6, subtypes c to l, was improved greatly in LiPA 2.0 using sequence motifs from the core region in addition to the 5′-UTR (32). In addition, the accuracy of identifying 1a and 1b was also improved greatly in LiPA 2.0 (Ross et al) (33). In the present study, only 69.8% of genotype 1b would have been characterized correctly using 5′-UTR information alone, whereas 100% of them have been correctly genotyped 1b with core information. The results confirmed the benefit of including the core region in LiPA 2.0. However, the capacity to identify the other genotypes/subtypes was not improved in LiPA 2.0, since only 5′-UTR information was used. Despite several studies showing that most genotypes, excluding subtypes 1a, 1b and 6c-l, could be distinguished correctly by 5′-UTR (21,24,34), it is not always the case. In some countries or regions, the HCV genotypes were not well distinguished by 5′-UTR. For example, Stuyver et al (22) used LiPA to analyze 506 HCV-infected sera from different geographical regions, representing a multitude of subtypes, and found that only 11% of HCV samples from Western Africa could be identified at the subtype level by 5′-UTR. The present study showed that only 66.7% of HCV samples were genotyped accurately, 52.4% of genotype 6a samples were misclassified in 1b. Based on this result, the suitability of LiPA 2.0 for genotyping HCV from South China should be reconsidered. A recent study showed that subtype 1b remained the most prevalent and widely distributed genotype in China (35). However, subtype 6a was more prevalent in the southern provinces of China (36,37). Clearly, with the increasing prevalence of 6a in South China, using LiPA 2.0 to identify HCV strains will lead to biased diagnostic and research results. Therefore, LiPA 2.0 was not suitable for identifying HCV genotypes in South China.
In addition, the LiPA 2.0 kit is relatively expensive. In China, the reagent cost was about 200 dollars/sample by LiPA 2.0, but less than 5 dollars by sequencing analysis. Genotyping HCV by sequencing analysis is therefore more cost-effective. Moreover, since the performance of commercial reverse transcriptase enzymes and DNA polymerases are improving continuously, genotyping HCV by sequencing analysis is not as difficult as previously. Currently, a reliable genotyping result can be obtained in 48 h with direct sequencing and BLAST analysis by a skilled technician. Therefore, direct sequencing of HCV core fragment plus BLAST analysis represents an ideal genotyping approach for laboratories with skilled technicians.
hepatitis C virus;
internal quantitative standard;
basic local alignment search tool
This study was supported by the National Science and Technology Key Project during the 11th Five-Year Plan Period (no. 2008ZX10002-013). Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the Ministry of Health of China.