Protocol for metagenomic virus detection in clinical specimens

Abstract

Sixty percent of emerging viruses have a zoonotic origin, making transmission from animals a major threat to public health. Prompt identification and analysis of these pathogens are indispensable to taking action toward prevention and protection of the affected population. We quantifiably compared classical and modern approaches of virus purification and enrichment in theory and experiments. Eventually, we established an unbiased protocol for detection of known and novel emerging viruses from organ tissues (tissue-based universal virus detection for viral metagenomics [TUViD-VM]). The final TUViD-VM protocol was extensively validated by using real-time PCR and next-generation sequencing. We could increase the amount of detectable virus nucleic acids and improved the detection of viruses <75,000-fold compared with other tested approaches. This TUViD-VM protocol can be used in metagenomic and virome studies to increase the likelihood of detecting viruses from any biological source.

Figure 1

Schematic description of tissue-based universal virus detection for viral metagenomics protocol. Estimated durations of each step are shown in parentheses. The protocol takes 12 h to complete.

Figure 2

Validation of test aliquots of infected mode used for development of tissue-based universal virus detection for viral metagenomics protocol. Every ninth aliquot was extracted, and viral copy numbers were determined by using a quantitative PCR. Standard deviations (error bars), medians (solid horizontal lines), and residual plots indicate homogeneity and mixture of test specimens. Ct, cycle threshold.

Figure 3

Comparison of tissue homogenization methods used for development of tissue-based universal virus detection for viral metagenomics protocol. Copy numbers were measured by quantitative PCR in duplicate. RQ, relative quantification: RQ (2 – ΔΔCt); (ΔΔCt = Δ purified – Δ unprocessed). Lower panel left y-axis indicates signal-to-noise ratio (RQ) for all viruses tested. The method with the highest score was used to establish the protocol and is shaded in yellow. Red stars indicate highest scores. Diagonally striped area indicates not significant. Ct, cycle threshold. Numbers along baseline indicate method used. 1, control; 2, Dounce homogenizer; 3, extended homogenization; 4, trypsin; 5, ultrasound; 6. QIAshredder (QIAGEN, Hilden, Germany).

Figure 4

Comparison of filtration methods used for development of tissue-based universal virus detection for viral metagenomics protocol. Copy numbers were measured by quantitative PCR in duplicate. RQ, relative quantification: RQ (2 – ΔΔCt); (ΔΔCt = Δ purified – Δ unprocessed). Lower panel left y-axis indicates signal-to-noise ratio (RQ) for all viruses tested. The method with the highest score was used to establish the protocol and is shaded in yellow. Red stars indicate highest scores. Diagonally striped area indicates not significant. Ct, cycle threshold. Numbers along baseline indicate method used. 1, control; 2, 0.22-μm filter; 3, 0.45-μm filter; 4, filter extraction 1; 5, filter extraction 2; 6, fractionated filtration; 7, filter tubes.

Figure 5

Comparison of enrichment methods used for development of tissue-based universal virus detection for viral metagenomics protocol. Copy numbers were measured by quantitative PCR in duplicate. RQ, relative quantification: RQ (2 – ΔΔCt); (ΔΔCt = Δ purified – Δ unprocessed). Lower panel left y-axis indicates signal-to-noise ratio (RQ) for all viruses tested. The method with the highest score was used to establish the protocol and is shaded in yellow. Red stars indicate highest scores. Diagonally striped area indicates not significant. Ct, cycle threshold. Numbers along baseline indicate method used. 1, control; 2, PEG-It (System Biosciences, Mountain View, CA, USA); 3, InRichment Virus Reagent Kit (Analytik Jena AC, Jena, Germany); 4, clearing centrifugation; 5, clearing centrifugation at 25,000 rpm for 2 h; 6, second clearing centrifugation after 20% sucrose centrifugation; 7, tissue enrichment; 8, Ribominus Eukaryote Kit (Life Technologies, Grand Island, NY, USA).

Figure 6

Comparison of extraction methods used for development of tissue-based universal virus detection for viral metagenomics protocol. Copy numbers were measured by quantitative PCR in duplicate. RQ, relative quantification: RQ (2 – ΔΔCt); (ΔΔCt = Δ purified – Δ unprocessed). Lower panel left y-axis indicates signal-to-noise ratio (RQ) for all viruses tested. The method with the highest score was used to establish the protocol and is shaded in yellow. Red stars indicate highest scores. Diagonally striped area indicates not significant. Ct, cycle threshold. Numbers along baseline indicate method used. 1, Nucleospin RNA II (Macherey Nagel, Dueren, Germany); 2, Nucleospin DNA (Macherey Nagel); 3, RTP DNA/RNA Virus Ultra Sense (Invitek, Berlin Germany); 4, RTP DNA/RNA Virus Mini Kit (Invitek); 5, QIAmp UltraSens Virus Kit (QIAGEN, Hilden, Germany); 6, Viral Mini Kit (QIAGEN); 7, QIAmp MinElute Virus Spin Kit (QIAGEN); 8, PureLink Viral RNA/DNA (Invitrogen Life Technologies, Grand Island, NY, USA); 9, TRIzol LS; 10, phenol chloroform.

Figure 7

Comparison of primers and random amplification methods used for development of tissue-based universal virus detection for viral metagenomics protocol. Copy numbers were measured by quantitative PCR in duplicate. RQ, relative quantification: RQ (2 – ΔΔCt); (ΔΔCt = Δ purified – Δ unprocessed). Lower panel left y-axis indicates signal-to-noise ratio (RQ) for all viruses tested. The method with the highest score was used to establish the protocol and is shaded in yellow. Red stars indicate highest scores. Diagonally striped area indicates not significant. Ct, cycle threshold. Numbers along baseline indicate method used. 1, control; 2, K primer; 3, 3′ locked primer; 4, N12 primer; 5, N primer; 6, whole transcriptome amplification (QIAGEN, Hilden, Germany); 7, whole genome amplification (QIAGEN).

Figure 8

Results of comparative next-generation sequencing used for development of tissue tissue-based universal virus detection for viral metagenomics (TUViD-VM) protocol. A) Sample preparation flowchart to generate 4 next-generation sequencing approaches. B) Results obtained for model tissue (chicken) infected with 4 viruses: vaccinia virus (poxvirus) Sendai virus (paramyxovirus), influenza virus (A/PR8/1934), or reovirus (T3/Bat/G/342/08). The x-axis is log-scaled, and normalized read numbers are indicated. C) Results of marmoset sample proof of principle, Sendai virus–infected lung tissue. The baseline is log-scaled, and normalized read numbers are indicated.

Figure 9

Changes in virus-to-host nucleic acid signal-to-noise ratio during development of tissue-based universal virus detection for viral metagenomics (TUViD-VM) protocol. Next-generation sequencing results for virus-infected chicken tissue comparatively sequenced were obtained by using 4 approaches: standard RNA library preparation (Chicken RNA), standard DNA library preparation (Chicken DNA), DNA library from random-amplified chicken tissue (Chicken random), and DNA library from random-amplified chicken tissue prepared by using TUViD-VM (Chicken purified random). The y-axis is log-scaled.

Figure 10

Changes in virus-to-host nucleic acid signal-to-noise ratio during development of tissue-based universal virus detection for viral metagenomics (TUViD-VM) protocol. Next-generation sequencing results for virus-infected marmoset tissue comparatively sequenced were obtained by using 4 approaches: standard RNA library preparation (Monkey RNA), standard DNA library preparation (Monkey DNA), DNA library from random-amplified marmoset tissue (Monkey random), and DNA library from random-amplified marmoset tissue prepared by using TUViD-VM (Monkey purified random). The y-axis is log-scaled.