A Novel Field-Deployable Method for Sequencing and Analyses of Henipavirus Genomes From Complex Samples on the MinION Platform

Abstract

Viruses in the genus Henipavirus encompass 2 highly pathogenic emerging zoonotic pathogens, Hendra virus (HeV) and Nipah virus (NiV). Despite the impact on human health, there is currently limited full-genome sequence information available for henipaviruses. This lack of full-length genomes hampers our ability to understand the molecular drivers of henipavirus emergence. Furthermore, rapidly deployable viral genome sequencing can be an integral part of outbreak response and epidemiological investigations to study transmission chains. In this study, we describe the development of a reverse-transcription, long-range polymerase chain reaction (LRPCR) assay for efficient genome amplification of NiV, HeV, and a related non-pathogenic henipavirus, Cedar virus (CedPV). We then demonstrated the utility of our method by amplifying partial viral genomes from 6 HeV-infected tissue samples from Syrian hamsters and 4 tissue samples from a NiV-infected African green monkey with viral loads as low as 52 genome copies/mg. We subsequently sequenced the amplified genomes on the portable Oxford Nanopore MinION platform and analyzed the data using a newly developed field-deployable bioinformatic pipeline. Our LRPCR assay allows amplification and sequencing of 2 or 4 amplicons in semi-nested reactions. Coupled with an easy-to-use bioinformatics pipeline, this method is particularly useful in the field during outbreaks in resource-poor environments.

The genus Henipavirus within the family Paramyxoviridae, subfamily Paramyxovirinae, are non-segmented, negative-strand ribonucleic acid viruses, including the highly pathogenic Hendra virus (HeV) and Nipah virus (NiV) and the non-pathogenic Cedar virus (CedPV) [1, 2]. The recently identified Ghanaian bat virus and Mojiang virus species are also members of this genus, but only viral genomes and not viral isolates were recovered, and these novel viruses are currently of unknown zoonotic potential [3, 4]. HeV and NiV can cause respiratory distress and fatal encephalitis with case fatality rates ranging from 40% to 100% upon spillover into humans [5]. HeV outbreaks are currently limited to Australia, where humans contract the virus indirectly from fruit bats through horses. To date, there have been 53 spillover incidents involving over 70 horses and 7 human cases with 4 fatalities. All HeV spillover events occurred in the northeastern coastal region of Australia [6]. NiV is transmitted to humans through the consumption of contaminated food sources or indirectly through livestock, with almost yearly outbreaks in Bangladesh since 1999 [7, 8]. Recent surveillance studies in bats have found evidence for henipavirus circulation as far as West Africa and Brazil [3, 9, 10], suggesting a potential risk of zoonotic spillover across a broader geographic distribution than previously thought.

Despite the threat henipaviruses pose to human health, comparatively little full-genome data has been available from wildlife reservoirs, hindering a more comprehensive understanding of the evolution and ecology of these viruses. Current deep-sequencing methods to recover full viral genomes are both costly and inefficient without prior virus isolation. Direct sequencing from complex sample sources, including blood, swabs, tissue, or urine, would therefore allow us to conduct detailed molecular studies on henipavirus variation in the natural reservoir. In addition, these molecular tools would allow us to perform rapid sequencing during outbreaks to study molecular epidemiology and transmission chains; this was shown to be crucial in our understanding of outbreaks of Ebola virus and Lassa virus [11, 12]. However, rapid analysis of sequences generated in resource-poor settings is currently hindered by slow, unreliable, or unavailable internet connections. Stand-alone bioinformatics pipelines for both offline and online analyses of next-generation sequencing (NGS) data are therefore highly desirable. In this study, we describe the development of a long-range polymerase chain reaction (LRPCR) assay for efficient genome amplification of NiV, HeV, and CedPV. We developed both 2- and 4-amplicon LRPCR to amplify henipavirus full henipavirus genomes. The amplicons were subsequently sequenced on the portable Oxford Nanopore MinION platform and analyzed on a newly developed field-deployable bioinformatics pipeline for analyzing sequence data from Oxford Nanopore MinION.

METHODS

Ethics Statement

Approval of animal experiments was obtained from the Institutional Animal Care and Use Committee of the Rocky Mountain Laboratories. Experiments were performed following the guidelines of the Association for Assessment and Accreditation of Laboratory Animal Care, International (AAALAC) by certified staff in an AAALAC-approved facility, following the guidelines and basic principles in the US Public Health Service Policy on Humane Care and Use of Laboratory Animals and the Guide for the Care and Use of Laboratory Animals. Work with infectious NiV and HeV under biosafety level 4 conditions was approved by the Institutional Biosafety Committee (IBC). Inactivation and removal of samples from high containment was performed per IBC-approved standard operating procedures.

Viruses and RNA

Henipavirus isolates were obtained from the Special Pathogens Branch of the Centers for Disease Control and Prevention (Atlanta, GA), Public Health Agency (Winnipeg, Canada), and the Australian Animal Health Laboratory (Victoria, Australia). Nipah virus Bangladesh (NiV-B) (GenBank accession no. AY988601) and HeV (GenBank accession no. AF017149) were passaged 3 times on Vero-E6 cells, whereas CedPV (GenBank accession no. JQ001776) was passaged 2 times in PaKi cells and 8 times on Vero-E6 cells. RNA was extracted from virus isolates using the QiaAmp Viral RNA extraction kit (QIAGEN) according to manufacturer’s instructions.

Hendra virus RNA from infected animals was obtained using lungs and brain tissues collected from Syrian hamsters (SHs) inoculated intraperitoneally with 6 × 10³ 50% tissue culture infective dose (TCID₅₀) HeV virus (AF017149) and euthanized on 4 days post infection (dpi) [13]. Nipah virus RNA was obtained from blood, axillary lymph node, brain, lung, and urinary bladder tissues obtained from an infected African green monkey inoculated with 10⁵ TCID₅₀ NiV-B via the intranasal and intratracheal route and euthanized at 8 dpi. Tissues were homogenized in RLT buffer, and RNA was extracted using the RNeasy extraction kit (QIAGEN). Viral loads of stock virus and all tissues used in this study are listed in Supplementary Table 1.

To determine the limit of amplification for LRPCR, 10-fold serial dilutions were performed on NiV, HeV, and CedPV stocks from concentrations of 10⁸ or 10⁶ TCID₅₀/mL through 10^–1 or 10^–3 TCID₅/mL (for NiV and HeV, and CedPV, respectively). RNA was then extracted for each dilution with the QIAamp viral RNA Mini Kit (QIAGEN). The RNA extractions were performed using modifications necessary for inactivation of high-containment pathogens [14]. Complementary DNA (cDNA) was synthesized from 10 µL of each RNA extract using the Superscript IV First-Strand Synthesis system with random hexamers according to the manufacturer’s instructions (Thermo Fisher Scientific).

Long-Range Polymerase Chain Reaction Assay and MinION Sequencing

Primers were designed targeting regions of the NiV, HeV, and CedPV genomes that are broadly conserved across all available genomes and that meet the specifications for the high-fidelity PrimeSTAR GXL DNA Polymerase (Takara Bio USA), and the LRPCR amplification was performed as described in Seifert et al [15]. In brief, after cDNA synthesis, the henipavirus genomes were amplified in 2 (~10 kilobase [kb]) or 4 (~6 kb) amplicons with a semi-nested LRPCR, in which the PCR product from the first round of PCR was used as template for the second round of PCR. Each 50-µL LRPCR master mix contained 0.2 µM of each primer (list of primers in Supplementary Table 2), 1 × PrimeSTAR GXL Buffer, 200 µM each deoxyribonucleotide triphosphate, 5 µL cDNA template, and 1.25 or 2.5 units of PrimeSTAR GXL DNA Polymerase (Takara Bio USA). The LRPCR mixture was incubated at 98°C for 2 minutes for the initial denaturation, followed by 4 cycles at 98°C for 10 seconds, 68°C for 15 seconds (−2°C per cycle), and 72°C for 10 minutes before an additional 26 cycles of 98°C for 10 seconds, 56°C (NiV and CedPV) or 60°C (HeV) for 15 seconds, and 72°C for 10 minutes. PCR products were visualized on a 0.75% agarose gel with 0.5 µg/mL ethidium bromide (Supplementary Figure 1). The LRPCR products amplified from NiV-infected African green monkey and HeV-infected SHs tissue samples were purified with 0.37 ratio of AMPure XP beads (Beckman Coulter) to exclude any non-specific fragments. The concentration of each LRPCR product was determined with a Qubit 3.0 as directed by the dsDNA HS Assay Kit (Thermo Fisher Scientific). The LRPCR products were pooled at equal molar ratios to get a final total concentration of 1 μg in 49 μL. The sequencing libraries were prepared using the 1D Native barcoding genomic DNA (with EXP-NBD104, EXP-NBD114, and SQK-LSK109) protocol (Oxford Nanopore Technologies). Sequencing was done with a MinION Mk1B on an R9.4.1 Flow Cell for approximately for 24 hours, and live base calling was performed using the MinKNOW v18.12.9 software (Oxford Nanopore Technologies).

Bioinformatics Analysis

After sequencing and base calling, the raw reads were demultiplexed and barcodes were removed using porechop [16]. Variant calling and consensus derivation was done with the pipeline https://github.com/niaid/viral_variant_pipeline. In this pipeline, all derived FASTQ files were filtered and trimmed with Nanofilt (-q 8 -l 500 --headcrop 50) [17]. Reads were then aligned to the reference genome using Minimap2 [18] and recommended parameters for MinION reads (-ax map-ont). Variants were then called using both samtools [19] and bcftools [20] and filtered for Allelic depths (AL >30) and Phred-scaled genotype likelihoods (PL >60). Coverage plots were created using R from a bedgraph file generated using BEDtools [21] and plotted with R ggplot2 package [22]. A consensus sequence was generated for each sample incorporating variants called. All consensus sequences were aligned together with the reference sequence using MAFFT [23] with default parameters. To aid in convenience and reproducibility, an image containing the source code and dependencies was created and is available at https://cloud.docker.com/u/philipmac/repository/docker/philipmac/vir_pipe. The analysis was run in a Docker container created from this image.

RESULTS

We developed a novel, seminested LRPCR with primers targeting conserved regions for all available NiV, HeV, and CedPV sequences in GenBank. These primers would be predicted to amplify 99.0%, 99.3%, and 99.7% of NiV, HeV, or CedPV genomes, respectively, in 2 or 4 overlapping amplicons with semi-nested LRPCR (Figure 1). To test the sensitivity of the assay, virus stocks were serially diluted from 1 × 10⁵, 1.5 × 10⁵, or 1.2 × 10⁵ genome copies/μL of NiV, HeV, or CedPV stocks virus, respectively. For 2 amplicon LRPCR, primers targeting NiV and CedPV genomes could amplify full genomes down to titers of approximately 1 genome copies/μL, whereas primers against HeV genome could go as low as 2 genome copies/μL. The 4-amplicon assay was slightly more sensitive with titers as low as <1 genome copies/μL for NiV and HeV and 1 genome copies/μL for CedPV. The 2-amplicon assay is a less expensive and more useful assay for samples with higher viral titers. The 4-amplicon assay, although twice as expensive, is more sensitive and therefore more useful to amplify genomes in samples with lower viral titers.

Figure 1.

A schematic diagram showing the primer binding sites across the henipavirus genome for the 2-apmlicon (A) and 4-amplicon (B) seminested long-range polymerase chain reaction (LRPCR) protocol.

To test the LRPCR method on samples reflective of ecologically or outbreak derived samples, we amplified NiV and HeV genomes from various tissues samples from an infected African green monkey and SHs, respectively (Supplementary Figure 1B and C). To further evaluate the assay, we amplified NiV genome from NiV-B-positive African green monkey blood (Supplementary Figure 1D). Sequencing on the portable MinION yielded approximately 72 000 reads, and at least 86% of these reads were mapped to the reference genomes with an average coverage of more than 6000 reads. However, in some cases, there was a disparity in depth of coverage for each individual amplicon per sample especially for HeV sequences from SH tissues, despite pooling the equal molar amount of each amplicon per sample (Supplementary Figure 2). We found a single-nucleotide polymorphism (SNP) between the reference HeV sequence and all HeV genomes sequenced from SHs at position 132 from C to A (Figure 2). SH brain sample No. 4 had 3 additional SNPs, 1 of which was shared in the viral genome sequence amplified from a lung tissue sample for SH No. 3. For NiV, 2 SNPs were found between the reference NiV sequence of the input virus and the NiV genomes sequenced from African green monkey urinary bladder after productive infection. After Sanger sequencing, only the SNP at position 132 in HeV SH tissues were determined to be a true SNP. This nucleotide change leads to an amino acid change from alanine to glutamate within the nucleoprotein. The remainder of the SNPs are possibly as a result of error during a LRPCR cycle, sequencing error, or insufficient MinION sequencing depth. If the latter is true, with a greater depth of coverage, the error rate is low, and hence this method can be successfully used for variant calling for within and between population genetics comparison.

Figure 2.

A schematic diagram showing the location of single-nucleotide polymorphisms detected between reference Hendra virus (HeV) (A) and Nipah virus (NiV) (B) sequence and sequences obtained from various henipavirus-infected tissues from Syrian hamsters (SH) and African green monkey, respectively. Nucleotide discrepancies between reference and animal-derived sequences are indicated by black vertical lines.

DISCUSSION

Henipaviruses are on the World Health Organization blueprint list of diseases with potentially high pandemic risk [24, 25]. Therefore, a better understanding of their evolution and molecular determinants of zoonotic and cross-species transmissions is urgently needed. Currently, only sparse full henipavirus genome sequence information is available from outbreaks and reservoir hosts. In this study, we describe the development of a LRPCR to amplify full genomes of HeV, NiV, and CedPV, and we have demonstrated that this assay effectively amplifies viral genomes from experimentally challenged animal tissues and blood with viral titers as low as 52 genome copies/mg, which is well within range of NiV and HeV loads observed in organs of various experimentally inoculated animals [26, 27]. This indicates that the sensitivity of our assay is sufficient to amplify henipavirus genomes from infected wildlife if titers are at least 52 genome copies/mg or 2 genome copies/μL. To increase the sensitivity, we also designed a 4-amplicon assay that would therefore be more useful with titers lower than 2 genome copies/μL (or 52 genome copies/mg), such as typically observed in samples obtained from the bat reservoir [28, 29]. The detection limit of the 2- and 4-amplicon assays are comparable to unbiased sequencing methods and a positive selection system such as VirCapSeq-VERT, which demonstrated a detection limit of ~1.2 genome copies/μL [30]. For both unbiased sequencing and positive enrichment systems, detection does not necessarily indicate coverage of the complete genome. These methods typically yield lower proportion of reads per target genome relative to the host background when virus titers are low, making it either costly or impossible to obtain desirable depth of coverage [31]. Although preamplified samples yield a high proportion of reads per target, these methodologies require the a priori knowledge of the virus’s genomic sequences, which, in the case of newly emerging pathogens, are likely not directly available [15].

CONCLUSIONS

Next-generation sequencing has revolutionized studies into the evolution and molecular epidemiology of viral pathogens in the past decade. However, many developing countries have not taken part of this revolution primarily because of lack of computational expertise, cost, or required infrastructure (e.g., fast internet access) [32, 33]. The direct need for within-country NGS has clearly been shown in the recent and ongoing Ebola outbreaks [34–36], in which transmission chains and effectiveness of medical countermeasures were assessed almost in real time using in-country NGS. There is a clear need for the development of fast and affordable viral genomic sequencing methods accessible to researchers in the developing world where henipaviruses such as NiV is endemic. Next-generation sequencing platforms, such as MinION, minimize the infrastructural requirements and cost. However, data analyses remain a major challenge due to reliance on cloud-based analysis. In addition, many bioinformatics programs that require the development of a pipeline are available only on Linux-based operating systems (OS), but many researchers in less-developed countries have Windows OS. These information technology (IT)-related issues hinder data analysis in the field and force researchers to bring the data to places with stable IT infrastructure, thereby delaying time from data collection to analysis. Therefore, in addition to establishing a rapid and sensitive full-genome method for henipavirus sequencing, we also developed a field-deployable and easy-to-use bioinformatics pipeline requiring less computational power and expertise. This field-deployable pipeline contains all of the necessary software in Docker container, which enables cross-platform performance. Once the image is installed, analysis can be done without internet connection in the field. These advances will make MinION sequencing and data analysis more applicable to usage in resource-limited settings.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Financial support. This work was funded by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health and the DARPA PREEMPT Program Cooperative Agreement (No. D18AC00031).

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.

Presented in part: American Society of Virology Conference 2019, Minneapolis, Minnesota, USA.