Discovery of a Novel Human Pegivirus in Blood Associated with Hepatitis C Virus Co-Infection

Hepatitis C virus (HCV) and human pegivirus (HPgV), formerly GBV-C, are the only known human viruses in the Hepacivirus and Pegivirus genera, respectively, of the family Flaviviridae. We present the discovery of a second pegivirus, provisionally designated human pegivirus 2 (HPgV-2), by next-generation sequencing of plasma from an HCV-infected patient with multiple bloodborne exposures who died from sepsis of unknown etiology. HPgV-2 is highly divergent, situated on a deep phylogenetic branch in a clade that includes rodent and bat pegiviruses, with which it shares <32% amino acid identity. Molecular and serological tools were developed and validated for high-throughput screening of plasma samples, and a panel of 3 independent serological markers strongly correlated antibody responses with viral RNA positivity (99.9% negative predictive value). Discovery of 11 additional RNA-positive samples from a total of 2440 screened (0.45%) revealed 93–94% nucleotide identity between HPgV-2 strains. All 12 HPgV-2 RNA-positive cases were identified in individuals also testing positive for HCV RNA (12 of 983; 1.22%), including 2 samples co-infected with HIV, but HPgV-2 RNA was not detected in non-HCV-infected individuals (p<0.0001), including those singly infected by HIV (p = 0.0075) or HBV (p = 0.0077), nor in volunteer blood donors (p = 0.0082). Nine of the 12 (75%) HPgV-2 RNA positive samples were reactive for antibodies to viral serologic markers, whereas only 28 of 2,429 (1.15%) HPgV-2 RNA negative samples were seropositive. Longitudinal sampling in two individuals revealed that active HPgV-2 infection can persist in blood for at least 7 weeks, despite the presence of virus-specific antibodies. One individual harboring both HPgV-2 and HCV RNA was found to be seronegative for both viruses, suggesting a high likelihood of simultaneous acquisition of HCV and HPgV-2 infection from an acute co-transmission event. Taken together, our results indicate that HPgV-2 is a novel bloodborne infectious virus of humans and likely transmitted via the parenteral route.


Introduction
The only members of the Hepacivirus and Pegivirus genera that are known to infect humans are Hepatitis C virus (HCV) and human pegivirus (HPgV, formerly GB virus C/GBV-C).HCV infects approximately 3% of the world's population and can cause liver damage resulting in cirrhosis, hepatocellular carcinoma, and the need for transplantation [1].In contrast, HPgV infection is considered non-pathogenic, although prevalence can exceed 40% in populations at high risk for exposure to blood-borne agents [2,3].While controversial, several groups have also reported that co-infection of HPgV and human immunodeficiency virus (HIV) can delay progression to AIDS, presumably by decreasing HIV replication or perturbing the host immune response [3][4][5].More recently, co-infection with HPgV in patients with Ebola virus disease has been reported to be associated with improved survival [6].
Novel animal hepaciviruses and pegiviruses have previously been found in multiple animal reservoirs, including bats, rodents, cows, and horses [7][8][9][10][11][12][13]. While hepaciviruses such as HCV in humans and GBV-B are established hepatitis agents [14], traditional criteria for classification of viruses as pegiviruses have included phylogenetic relatedness, persistent infection in the host and, importantly, apparent lack of pathogenicity [15,16].However, the reported discovery of Theiler's disease-associated virus (EPgV-TDAV) [7], a novel pegivirus associated with acute hepatitis outbreaks in horses, belies this general classification and suggests that at least one member of the Pegivirus genus is able to cause hepatitis in its animal host.
Unbiased metagenomic next-generation sequencing (NGS) is an established approach for the detection and discovery of novel pathogens that can be highly divergent and thus undetectable using conventional molecular methods [17].Here we employed NGS and the sequencebased ultra-rapid pathogen identification (SURPI) bioinformatics pipeline for pathogen detection [18] to identify a novel pegivirus, provisionally named HPgV-2, in an HCV-infected patient who died from sepsis and multi-organ failure of unknown etiology.Subsequent phylogenetic, PCR, and serological analyses confirm that HPgV-2 is a novel blood-borne virus infectious to humans.

Results
Discovery and whole-genome sequencing of a novel pegivirus in a hepatitis patient Chronic liver disease patients enrolled at the University of Chicago Medical Center (n = 169) were screened by unbiased metagenomic NGS to identify potential novel etiologies.Metagenomic NGS and SURPI analysis done at University of California, San Francisco identified 3 reads in a plasma sample from one patient that were assembled into two contiguous sequences (contigs) sharing 60% amino acid identity to simian pegivirus A (SPgV-A/GBV-A) (Fig 1A).An additional 13.4 million raw metagenomic sequences were then generated from the sample, and 537 reads corresponding to the novel pegivirus (identified by BLASTx translated nucleotide alignment to proteins from SPgV-A, GBV-B, HPgV, and bovine pegivirus / BPgV) were de novo assembled using PRICE [19] into 10 contigs spanning >40% of the putative viral genome (Fig 1B).The NGS data were then remapped to these contigs (Fig 1C), followed by PCR and Sanger sequencing to close remaining gaps and recover the nearly complete draft genome (Fig 1D).
The index case, UC0125.US, was a 70 y/o female with a history of sickle-cell disease (SCD), hypertension, HCV infection, and chronic renal insufficiency who was admitted to University of Chicago Medical Center in October of 2011 with abdominal pain.The patient rapidly decompensated after a few days in the hospital and underwent urgent surgical exploration of the abdomen, showing a fulminant colitis secondary to mesenteric ischemia versus gastrointestinal infection.She died approximately 2½ weeks after admission after developing sepsis and resultant multi-organ failure.The autopsy report revealed extensive hemorrhagic necrosis of the abdomen and stage 3 fibrosis of the liver consistent with chronic HCV infection.
The patient worked in the textile industry and had a prior history of alcohol and illegal intravenous drug use (IVDU).She was initially diagnosed with HCV genotype 1a infection in 1998, and failed treatment with interferon-α in 1999, resulting in chronic aminotransferase elevation from 2000-2009, with HCV viral loads ranging from 226,000 to 12,685,000 copies per mL (S1 Fig) .A liver biopsy performed in April of 2003 revealed mild portal inflammation and stage 1 fibrosis secondary to chronic HCV infection, and repeat biopsies in 2006 and 2009 showed progression of fibrosis.Notably, the patient had undergone multiple exchange and blood transfusions throughout her lifetime due to repeated vaso-occlusive crises as a result of her SCD.The patient's plasma sample from which pegivirus sequences were originally identified was drawn 8/28/08.
To confirm the presence of a novel pegivirus and exclude the possibility of contamination, a separate aliquot of plasma from the index case collected on the same day (8/28/08) was independently processed at Abbott Laboratories.Metagenomic NGS yielded 249,693 reads out of 16,306,796 (1.53%) mapping to the draft genome.Reads covered 98.4% of the predicted complete genome with an average depth of 3,314X ± 426 reads per nucleotide (Fig 1E).This independently-assembled consensus genome shared 99.73% identity to the draft genome, with every mismatch either conserved or resolving an ambiguous base, and the 5' end was extended by an additional 306 nucleotides.Due to its structural and phylogenetic similarity to pegiviruses, we provisionally named the novel virus human pegivirus 2 (HPgV-2, strain UC0125.US), in accordance with recently established criteria [16].
At the amino terminus of the polyprotein, six potential initiator (ATG) codons were located in-frame with the predicted ORF.The highest identity of the HPgV-2 5'UTR to rat (RPgV-cc61, 65.9%) and bat (BPgV-34.1,61.1%) pegivirus sequences, respectively, occurred with the initiator codon positioned at the fifth ATG (nt 326).Beyond this ATG, homology dropped off sharply to 42.8%, suggesting that the beginning of the polyprotein coding sequence was at the fifth ATG and corresponding to a predicted core protein size of 79 aa.

HPgV-2 Strain Diversity and Phylogeny
Comparative alignment of 8 HPgV-2 polyproteins showed 93.8% overall identity at the nucleotide level and 94.7% at the amino acid level,with no apparent hypervariable region observed in the genome.The E1, NS2, NS3, and NS5B genes were the most conserved between strains (92-96% nt identity), whereas more diversity was present in the E2 (91-97%), X (89-94%), NS4A (89-97%), and NS5A (90-94%) proteins.Pairwise identity plots showed that HPgV-2 is substantially divergent from other representative pegiviruses in the National Center for Biotechnology Information (NCBI) nucleotide (nt) database (Fig 2A ), including the two most closely related strains, BPgV-34.1 and RPgV-CC61, with which it shared less than 32% overall amino acid identity.Phylogenetic analysis of the NS3, NS5B, and full polyprotein revealed that HPgV-2 strains cluster tightly together on a very deep branch within a clade that includes one rodent and two bat pegiviruses (Fig 2B and S3 Fig).
Newly developed serologic and molecular diagnostic tools uncover additional strains of HPgV-2 Previous studies have revealed frequent detection of pegivirus GBV-C (HPgV, referred herein as HPgV-1) in HCV-infected individuals [3,15].Since HPgV-2 was discovered in an HCV coinfected individual (UC0125.US), initial screening of 200 plasmapheresis donors positive for HCV by NAT and Ab testing (HCV NAT+/Ab+) was performed for HPgV-2 using a real-time RT-PCR directed against a region spanning the NS2 and NS3 genes (S2A Fig).Three samples were found to be positive for HPgV-2 RNA (ABT0070P.US, ABT0096P.US, and ABT0188P.US).Metagenomic NGS of these HPgV-2 isolates revealed regions of conservation in the 5'UTR and the NS2-NS3 junction, for which a multiplex assay for detection of HPgV-2 RNA was then developed (S2B Fig).
The multiplex assay was also supplemented with primers/probes situated in the 5'UTR of HPgV-1, enabling simultaneous detection of active infection with both HPgV-1 and HPgV-2.
Initial serologic testing of 200 HCV-positive plasmapheresis donors for HPgV-2 was performed using an automated, indirect immunoassay utilizing 16 peptides selected on the basis of predicted HPgV-2 antigenic regions and minimal sequence identity with HCV and HPgV-1 (S4A and S4B Fig) .Three of the 16 peptides, located in NS3 (P4), NS5A (P9) and NS4A/B (P16), were predictive of the presence of HPgV-2 RNA and thus were judged to be useful as serological markers (S5 Fig) .The remaining 13 peptides were excluded from further testing due to absent or low frequency of reactivity and lack of correlation with RT-PCR.Importantly, the index case was found to be reactive to all three peptides (P4, P9, and P16).A pool of these 3 peptides was used for screening of additional cohorts; samples reactive with the peptide pool were subsequently tested with individual peptides.Reactivity to at least one peptide was considered positive, with reactivity to two or more peptides providing increased confidence (Table 1).

Discussion
Here we report the identification of HPgV-2, a second human pegivirus, and recovery of 8 whole and 4 partial genome sequences from 12 strains, all in HCV-infected patients.HPgV-2 was found to be highly divergent, sharing <32% amino acid identity with other pegiviruses.Notably, no cases of active HPgV-2 infection were found in 494 singly HIV-infected individuals, 488 HBV-infected individuals, or 476 volunteer blood donors, as compared to detection of HPgV-2 in 12 of 983 HCV-infected individuals (p<0.0001).The observed overall nucleotide diversity (93.0%-94.4% identity between strains), together with detection of antibodies directed against multiple proteins, establish that HPgV-2 is a bona fide infectious agent of humans.Utilizing the prototype serological assay described herein, 9 of 12 HPgV-2 RNA positive samples were seroreactive, with the majority of these samples (67%; 6 of 9, including the index case) reactive to two or more peptides (Table 3).HPgV-2 seronegative results in three RNA positive individuals may be explained by (1) recent infection, as evidenced by higher viral loads relative to the other HPgV-2 samples (Table 3), (2) host-related factors, such as an immunocompromised state (e.g.HIV infection), resulting in a lack of antibody production, and/or (3) viral amino acid polymorphisms that may impact assay performance (S5 Fig) .Interestingly, among the three HPgV-2 RNA+/Ab-samples, one was found in an HCV RNA+/Ab-individual (ABT0239AN.US), suggesting an acute HCV infection occurring within the pre-seroconversion "window" and raising the possibility of a likely HCV/HPgV-2 co-transmission event.
HPgV-1, originally designated GBV-C, was discovered in 1995-1996 independently by two groups in sera from patients with non-A, non-B hepatitis [21,22].Although initially suspected to be a cause of chronic hepatitis, HPgV-1 was later found to be lymphotropic (and not hepatotropic), and to date has not been associated with hepatitis or any other acute clinical illness in humans, despite its capacity for persistent viremia [15].Nevertheless, there is convincing evidence that at least one pegivirus, EPgV-TDAV, is an etiologic agent of hepatitis in horses [7].Given co-infection with HCV, we are unable to ascertain whether HPgV-2 played a role in the chronic liver disease of our index patient (UC0125.US), or indeed, whether the virus is hepatotropic.However, longitudinal sampling of one individual in the current study did reveal the capacity of HPgV-2 to produce sustained viremia of at least 7 weeks duration.
Initial serological screening suggests that there are significant differences in the markers and epidemiology of HPgV-2 infection as compared to HPgV-1 infection.First, HPgV-2 antibodies were frequently detected for HPgV-2 viremic samples (9 of 12 cases) in the current study, whereas antibodies are uncommonly detected in actively infected HPgV-1 individuals [2].Second, unlike the case for HPgV-1, where there is frequent association with HIV infection [23], the only HPgV-2 viremic cases for HIV-infected persons were found among those also coinfected with HCV.Third, HPgV-2 viremia was not as prevalent as HPgV-1 viremia in all of the groups tested (Table 1, S1 Table ).We found HPgV-1 in 9.5% of HCV-infected individuals and in 4.2% of volunteer blood donors, comparable with other reports [24][25][26][27][28][29], while HPgV-2 viremia was 1.1% in HCV-infected individuals (p < 0.0001 versus HPgV-1 viremia) and absent in volunteer donors (p < 0.0001).This difference in prevalence may reflect differences in the mode or efficiency of HPgV-2 transmission relative to HCV and HPgV-1.Given the index patient's multiple bloodborne exposures, the established association of IVDU with HCV transmission, tight co-detection of HPgV-2 with HCV, and the identification of a likely co-transmission event from an individual dually infected with HPgV-2 and HCV but seronegative for both viruses, it is probable that the mode of transmission for HPgV-2 is also via the parenteral route.
Importantly, all HPgV-2 cases identified in the current study were dually infected with HCV.The association with HCV was statistically significant (p<0.0001),but may not be exclusive given the observed low frequency of HPgV-2 seroreactivity in non-HCV infected individuals.The presumptive clearance of HPgV-2 infection in seroreactive HCV negative individuals also raises the possibility that that HCV co-infection may facilitate persistence of HPgV-2.Interestingly, pegivirus and hepacivirus co-infections have also been reported in horses [30].Whether HPgV-2 co-infection in humans influences the natural history or course of HCV infection will be of interest, although establishing the pathogenic potential of HPgV-2 is likely to be confounded by its apparent association with HCV.The HCV-positive samples tested to date all originated from the United States; ongoing studies will be needed to determine the global prevalence of HPgV-2 viremia and the full extent of its sequence diversity.Since serology can detect past as well as current infection, future refinement of serological diagnostic assays for HPgV-2 will be critical in establishing the true prevalence of HPgV-2 infection in humans.
Studies to date have suggested that pegivirus evolutionary lineages corresponding to each mammalian host tend to cluster together, consistent with a model of co-divergence of the virus with its host [11,13,31,32].Our report of at least 12 strains of HPgV-2 infecting humans is consistent this pattern, with each strain sharing ~94% identity to one another.Over 80 novel hepaciviruses and pegviruses have now been reported in bats [13], as well as multiple novel hepaciviruses in rodents [9], cattle [8] and horses [7,10,12], underscoring the extraordinary diversity of these viruses in animal reservoirs.Thus, whether HPgV-2 arose early in the evolutionary history of pegiviruses or is the result of a more recent zoonotic spillover event from a mammalian reservoir remains to be established.Although the long phylogenetic branch lengths corresponding to HPgV-2 would indicate early acquisition of this virus by an ancestral mammalian lineage, the high sequence identity between individual strains (94-96%) as compared to the high diversity of HCV genotypes [33,34] suggests either geographical sampling bias or a recent introduction of HPgV-2 into the human population.Further epidemiological and genomic studies will be needed to pinpoint the precise evolutionary origins of HPgV-2.

Methods
initial 64 serum samples were pooled, filtered with 0.22 μm filter spin columns (Millipore) and treated with a cocktail of Baseline DNase and Turbo DNase (Ambion) at room temperature for 2.5 hours as previously described [35].Following nuclease treatment, samples were extracted for viral nucleic acids using the QIAamp Viral RNA Mini Kit (Qiagen).Libraries were prepared for unbiased metagenomic NGS as previously described [36].The remaining 105 samples were pooled into groups of 6 (50 μl each) and similarly treated with DNase, followed by automated bead-based extraction on the EZ1 instrument using the EZ1 Viral Mini Kit 2.0 (Qiagen).Extracted pools were combined in pairs prior to library preparation and barcode indexing using a modified Illumina TruSeq protocol [36].Libraries were sequenced as 100 nt paired-end reads on 4 lanes of an Illumina HiSeq2000 instrument, generating a total of 754 million raw reads with an average of 11 million paired-end reads per barcode.Reads were processed using the SURPI bioinformatics pipeline [18].Briefly, using the SNAP nucleotide aligner, human reads were first computationally subtracted, followed by alignment of remaining reads to all sequences in the comprehensive NCBI nt database.Viral and unmatched reads were then de novo assembled, translated in all 6 reading frames, and aligned to all viral protein sequences using RAPSearch [37].We identified 3 paired-end reads corresponding to a single barcode with remote homology to SPgV-A proteins (Fig 1A).Primers targeting 1 of the 3 reads were used to pinpoint the individual sample harboring HPgV-2.
To recover additional HPgV-2 sequence, additional libraries were prepared from the pool containing the HPgV-2 sample using two independent library methods, TruSeq (Illumina) and ThruPLEX (Rubicon Genomics).NGS libraries were barcoded and sequenced on an Illumina MiSeq, generating 13.4 million raw reads.When aligned to existing pegivirus amino acid sequences in the reference NCBI non-redundant (nr) protein database using BLASTx [38] at an e-value of 10 -5 , 1,080 candidate reads were identified, with 537 of these reads assembling into 9 longer contiguous sequences (contigs) (

Selection of samples for screening
Retrospective screening was performed on various sets of de-identified samples.Two-hundred samples from first-time plasmapheresis donors, testing positive both by an antibody test for HCV (Abbott Laboratories, Abbott Park IL 60064) and a HCV RNA test (Bayer Versant HCV RNA 3.0 assay) were purchased from ProMedDx (Norton, MA).The majority of samples were obtained from the Midwestern USA.In addition, 742 HCV RNA+/ Ab+ samples and 240 HCV RNA+/Ab-samples were purchased from the American Red Cross (Gaithersburg, MD, USA), while 498 HIV RNA+/Ab+ samples (collected from the East Coast, USA) and 488 HBV DNA +/Ab+ samples (collected from the West Coast, USA) were purchased from ProMedDx.Volunteer blood donor samples pre-screened for HCV, HBV, and HIV were purchased from Gulf Coast, Houston TX and were collected mainly from the Southwestern USA.

Design and synthesis of HPgV-2 peptides
Peptides were designed to target amino acid sequences found in the HPgV-2 polyprotein of the index case.Predicted surface exposure (hydrophilicity profile and surface probability) and antigenic index scores were computed using Protean (DNAStar, Madison WI, USA), and amino acid stretches from S, E1, E2, NS3, NS3, NS4A, NS4B, NS5A, and NS5B were chosen to be synthesized as peptides.Hydropathy profiles for each ORF were assessed for areas with potential exposure in aqueous solution (e.g.hydrophilicity).Surface probability was determined using the method of Emini et al. [39].Antigenic index was determined using the Jameson-Wolf algorithm, where surface accessibility of residues is combined with predicted backbone flexibility and secondary structure [40] (S4 Fig) .Peptides were generated with an amino terminal biotinylation modification (Genscript).

Serological screening
Antibody detection with biotinylated peptides was performed on an ARCHITECT instrument (Abbott Laboratories).Human serum or plasma was added to a reaction vessel along with sample diluent buffer (e.g.buffering salts and detergents) containing streptavidin beads (Dynabeads M-270, Life Technologies) and a pool of biotinylated peptides (800 ng/ml each).The sample was incubated for 18 minutes to permit solid phase capture of both the biotinylated peptide and the immune complex (antibody complexed to the peptide).Following incubation, unreacted sample was removed, and a chemiluminescent signal-generating conjugate (mouse anti-human IgG conjugated to acridinium) added to the reaction vessel.After conjugate binding to the immobilized human immunoglobulins and a washing step to remove unreacted material, the microparticles were incubated with a chemiluminescent substrate.The amount of luminescence was then measured in relative light units (RLU) using a bioluminescence imager.Negative controls were prepared either by pooling a series of samples from individuals at low risk for viral infection and testing negative for several common bloodborne viruses (HIV, HBV, and HCV), or by reactions lacking the addition of peptides listed in S4 Fig Seroreactivity was defined using a signal to cutoff (S/CO) ratio of 15, which corresponded to approximately 6 standard deviations from the population mean of volunteer donors testing negative with the peptide pool.Testing on HCV, HIV, HBV and volunteer donor cohorts was performed using a pool of peptides 4, 9, and 16, the three peptides that showed the best correlation with RNA positivity.

Whole-genome sequencing of multiple strains of HPgV-2
Sequence confirmation of the HPgV-2 genome corresponding to the index case (UC0125.US) and whole-genome recovery of additional strains were performed using NGS.Plasma was clarified of particulate matter by centrifugation at 2650g for 5 min at room temperature, and samples were pre-treated with benzonase (500 U/ml; (Sigma)) for 3 hrs at 37°C in 1X benzonase buffer (20 mM Tris-Cl pH 7.5, and 10 mM NaCl, 2 mM MgCl 2 ) to degrade free DNA and RNA.For extractions, the QiaAMP Viral Mini (Qiagen), MagMax Viral RNA (Ambion), and Abbott RNA or total nucleic acid sample preps on the m2000 instrument were performed according to the manufacturer's instructions, all of which successfully recovered HPgV-2.RNA was concentrated using RNA Clean & Concentrator-5 columns (Zymo Research) and eluted in 7 μl of DNase/RNase-Free water.
Randomly primed cDNA libraries were constructed from 5 μl of RNA using the Ovation RNA-SeqV2 kit (NuGen) according to manufacturer's recommendations.cDNA concentrations were measured on a Qubit Fluorometer using dsDNA BR reagents (Molecular Probes/ Life Technologies) and diluted to 0.2 ng/μl.One nanogram (5 μl) was used as input for Nextera XT tagmentation (Illumina) according to manufacturer's instructions, and barcoded libraries were amplified with 16 cycles of PCR.Libraries were quantified on a BioAnalyzer 2200 TapeStation, pooled to a final concentration of 6-12 pM in HT buffer with 1% PhiX loading control, and sequenced on an Illumina MiSeq as 150 nt paired-ends reads using v2 chemistry.To exclude reagent contamination, additional experiments were performed with the SMARTer PCR cDNA synthesis (Clontech), Ovation Single Cell RNA-seq, and Ovation Human Blood RNA-seq library systems, each of which also generated HPgV-2 sequence.
Where required, gaps in NGS sequence were filled using One-Step RT-PCR (Qiagen) and Sanger sequenced on an ABI3130 instrument.Sequences of primers used to characterize the index case and fill in additional genomes are available upon request.Contigs from multiple runs of NGS or Sanger data were assembled in Sequencher 5.2.3 software to generate strain consensus sequences, and the polyprotein open reading frame was confirmed in DNAStar Seq-Builder software.Finishing the 5' and 3' ends of the HPgV-2 was performed using the First Choice RLM RACE kit (Ambion) according to the manufacturer's instruction.Predicted polyprotein cleavage sites were obtained by multiple sequence alignments of the HPgV-2 genome with other fully sequenced pegivirus reference genomes using Geneious v8.0 [41] and by identification of predicted signal peptidase cleavage use using the SignalP 4.0 server [42].Predicted N-linked and O-linked glycosylation sites in the structural proteins of HPgV-2 were obtained using the online NetNGlyc 1.0 [43] and NetOGlyc 4.0 [44] servers, respectively.
Phylogenetic trees of NS3, NS5A, or the polyprotein sequences for HPgV-2 strains, other pegiviruses, and representative hepaciviruses were constructed in Geneious 8.0 using the Jukes-Cantor model and neighbor-joining algorithm with 10,000 bootstrap replicates used to calculate branch supports.These tree topologies were then refined using a maximum likelihood Bayesian approach with MrBayes V3.2 software (1,000,000 sample trees, 10% of trees discarded as burn-in, convergence defined at an average standard deviation of <0.01).Each tree was rooted with dengue yellow fever virus (YFV) as the outgroup.infected with HPgV-2.This blood-borne virus was approximately 96% identical to the HPgV-2 virus from the index case.

Fig 1 .
Fig 1. Discovery and whole-genome characterization of HPgV-2.(A) Three NGS reads with remote amino acid homology (<60%) to SPgV-A / GBV-A, of which two were overlapping, were detected in a plasma sample from a patient with HCV infection who died of abdominal sepsis (the index case).(B) The initial set of contigs generated from de novo assembly of HPgV-2 reads that were identified using BLASTx alignment to other pegivirus genomes.(C) Gap closure using PCR followed by Sanger sequencing.(D) Coverage plot showing mapping of the initial NGS data to the nearly complete (>98%) assembled draft genome.(E) Coverage plot showing mapping of the NGS data from a subsequent sequencing run to the complete HPgV-2 genome, after the 5' and 3' ends were recovered using RACE [20].(F) Genomic arrangement of HPgV-2.Putative cleavage sites within the polyprotein are indicated with black triangles (structural proteins) or hollow triangles (non-structural proteins).Arrows denote predicted N-linked (red) or O-linked (blue) glycosylation sites.doi:10.1371/journal.ppat.1005325.g001

Table 1 .
Serologic and Molecular Detection of HPgV-2 a Peptides 4, 9, and 16 were utilized as a pool to screen samples.b Two samples were co-infected with HIV and HCV; these two samples were grouped with the HCV Ab+/PCR+ group c Index case (HCV Ab+ /NAT+) has been excluded from totals screened.

Table 2 .
Prevalence of HPgV-2 RNA in HCV and non-HCV infected groups.

Table 3 .
Molecular and Serologic Characterization of HPgV-2 Strains Abbreviations: NA, not applicable; NT, not tested.*These two samples are from the same individual with bleed dates 26 days apart (sample ABT0030P.US is the earlier sample).**These two samples are from the same individual with bleed dates 60 days apart (sample ABT0035P.US is the earlier sample).† Italicized numbers show elevated signals, but below the cutoff value.doi:10.1371/journal.ppat.1005325.t003