Eric Delwart's lab at the Blood Systems Research Institute is part of a growing community of labs hunting new human viruses using the most advanced technologies for biodiscovery. But in applying metagenomics to novel biologies, Delwart has found that contaminating nucleic acids can be a costly distraction, and he believes the development of DNA-free reagents could help alleviate this problem.
In an interview with PCR Insider, Delwart discussed the recent diatom virus that vexed researchers by masquerading as a human pathogen, as well as his ongoing work to characterize the human plasma virome.
He pointed out that "contamination issues are a problem for deep sequencing, [but] they're not really a problem for PCR because … you don't really care if there's this strange exotic virus in there because you're not going to amplify it anyway. It's not going to interfere with your PCR studies," he said. "But when you make metagenomic libraries where you have to find everything, that's going to come along and be quite a bit misleading."
Delwart, like many of his colleagues, is not as interested in discovering viruses that infect microorganisms. "We're more in the business of looking for new human viruses, not diatom viruses," he said.
With collaborator Charles Chiu at the University of California, San Francisco, with which BSRI is affiliated, as well as other labs, Delwart recently deduced that Qiagen's QiaAmp kits' spin columns may be made with silica derived from virus-infected diatoms. As he related it, "We'd been seeing [the viral sequence] in all kinds of unrelated libraries ─ sewage water, animal samples ─ we thought, this is suspicious because it's everywhere," he said. "We kind of put it aside, [but] when the sequence came out [in] the PNAS paper, the first thing we did was look for homology, and it was the same virus that we'd seen before. We basically quickly published that to correct the literature, so that people don't spend their time chasing a virus which turns out to be from the extraction columns."
Delwart's lab collaborated on the publication of this discovery, described recently in PCR Insider. The finding has since been replicated by another group.
His work has thus highlighted the need for nucleic acid-free reagents for the metagenomics community. "There's a lack of what I would call really clean reagents that have an extremely low amount of DNA contamination," said Delwart. "I think it would be interesting for people to try to develop DNA-free or nearly DNA-free reagents." For example, he said, "Whenever we use reverse transcriptase for metagenomic sequencing, we tend to find reverse transcriptase sequences, and this clearly comes from the protein purification, which also carries along some of the DNA." He added, "In these days of very sensitive deep sequencing, you want to remove the sources of DNA [contamination] in your input sample," he said.
The diatom case also called to mind XMRV, the murine leukemia virus which caused false alarms as a possible causative agent of chronic fatigue syndrome. That case took much longer to resolve, and required extensive studies before the research community could self-correct. According to Delwart, that situation was "a little bit like the diatom virus, in terms of contamination issues; in that case it was more like a straight-up DNA contamination happening in many labs," he said. Fortunately, "the issue was resolved much more quickly; within a week of seeing the genome sequence we had confirmed this was a Qiagen contaminant," he said.
In terms of reagents, Delwart suggested there is a real need for suppliers to eliminate contaminants. "I'm not blaming Qiagen," he said. "Other vendors probably have other contaminants, but they should try to improve their process as much as possible." Delwart affirmed what Qiagen has told PCR Insider – that the company is offering researchers low-DNA replacement kits – but, he added, "they call it low DNA, but 'low' is not 'no,'" so his lab has chosen to go in another direction.
Delwart and colleagues now use a viral RNA isolation kit from Life Tech. He said that although they have not done extensive product comparisons, and he'd be surprised if this kit is completely free of nucleic acids, it does seem to "produce less of these extra sequences." He noted that the product materials do not detail the nucleic acid binding surface on the magnetic beads, but he suspects it is also silica, perhaps treated more harshly to release nucleic acids during manufacturing.
The human plasma virome
Delwart's lab at the BSRI is in the fourth year of a nearly $2 million grant from the National Heart, Lung, and Blood Institute at the National Institutes of Health aimed at detecting emerging viruses in samples from blood banks. Their study uses 1,200 plasma pools, each made up of thousands of individual plasma donor specimens.
The protocol for prospecting viruses in plasma pools involves first filtering out cell-sized particles, then using DNase to eliminate cell-free DNA, leaving behind the precious few encapsulated viruses.
Viral RNAs are converted to cDNA, then these sequences are subjected to what Delwart called "random RT-PCR." This technique is "a bit of an oxymoron," he said, but essentially involves randomly amplifying 3' ends, "just so we have enough material so we can throw it all into a sequencer," he said. The lab does massively parallel sequencing using Illumina MiSeq. This is followed by BlastX, where "we take the DNA sequences, we translate them every six ways possible, from both strands, starting at position one, two, and three, and get all these sort of virtual protein translation products, then we take those and we Blast those, so we're looking for homology at the protein level," he said.
Delwart remarked that the "disruptive technologies" of the past 10 years have made his lab spoiled for data. "Deep sequencing has changed the way we do business," he said, adding, "I started this by subcloning these random PCR products into plasmids and sequencing 50 plasmids; now I can take these random PCR products and generate 10 million sequences for a tenth of the price. The data flow is humongous. Part of the problem now is actually analyzing the data, just moving the data from one computer to the next," he said.
Delwart's viral discovery work might eventually impact treatment of diseases, and he suggested that virus hunting also will benefit agricultural and veterinary markets in the future. "Biodiscovery metagenomics is going to have a huge impact on animal health because we don't know nearly as much for animals as we do for humans, and there are a lot of unexplained outbreaks, common diseases in farm animals and pets, that we're finding viruses in. Again, finding a virus doesn't show that it’s a pathogen, but it’s the first step toward that goal."
His roadmap toward the comprehensive description of the human virome, published last year in PLOS One, emphasized the value of this type of exploratory work. He wrote that, in combination with studies of viral prevalence and disease association, this work "will provide a better understanding of which viruses account for the large diagnostic gap seen for many diseases of possible infectious origins." He suggested that "a significant fraction of mild to extremely severe symptoms of likely infectious origin remains unexplained in both developed and developing countries, including respiratory problems, diarrhea, and encephalitis. Auto-immune diseases, such as diabetes, may also be triggered by unknown viral infections, and carcinogenic human viruses may still remain uncharacterized."
The good news so far from Delwart's virus hunt in plasma is "the blood supply is extremely safe – we're not finding common, new viruses, at least ones we can recognize," he said.
However, he cautioned, "That doesn't mean there's none, because there may be viruses that we can't even recognize as viruses, because they're so divergent from the ones we already know," he said. He explained, "The way we recognize viruses is by sequence homology, so if there is very little or no sequence homology to what we already know, it's going to pass through our bioinformatics and be one of those mystery sequences of unknown taxonomic origin."
An unknown virus could also be filtered out in Delwart's procedures because it's a large DNA virus. Another possible reason for not finding new viruses in plasma is, they "just haven't looked at enough samples, and maybe there are new human viruses circulating in the human blood supply but we just haven't screened enough to see," Delwart said. He added that their assay does not pick up "viruses that are not replicating, but are just sort of expressing one or two genes from within a cell," he said. Finally, plasma is among the most studied of human fluids and banked plasma comes from ostensibly healthy donors, so Delwart's lab also hunts for new viruses in highly exposed or immunocompromised people. "They're nice people to look at to try to see emerging viruses that may one day go global," he said. His lab is also virus hunting in other biological specimens, such as ones containing enteric or respiratory viruses that could possibly leak into the blood and affect susceptible people.
While Delwart and collaborators have yet to find any new plasma viruses, their study has uncovered a number of previously identified viruses in the blood supply, a few of which could potentially become emergent. "What we do find are the classic human viruses, [such as] B19 parvovirus, [and] another parvovirus called parv4," he said. "We're finding a lot of a flavivirus called GBV-C, which is a very common flavivirus that is in the same family of hepatitis C but is, as far as we know, non-pathogenic."
They also see anelloviruses, said Delwart. This family of ring-shaped viruses is "probably the most common virus we find, and a very interesting one, but there's very little study done on them," he said, adding that "they're extremely common. Everyone is pretty much infected with those. You get it at birth, you get it from your mom, probably, or your siblings. It’s a chronic viremia, it’s an extremely diverse group of viruses, it's transfusion-transmitted, but because it's thought to be commensual, that is non-pathogenic, then people just tolerate it" in the blood supply, he said. "If you actually excluded donations that were anellovirus positive, you basically would no longer have a blood supply [because] everybody's positive. But it's never been shown to be pathogenic, so people kind of live with it," he said.
Among the assorted non-pathogenic chronic viremias metagenomics studies have uncovered, GBVC is interesting because it may even be advantageous in some situations. Coincidental co-infection of this common virus with HIV has been suggested to increase survival, with the quest for a mechanism ongoing.
Pathogenic viruses on Delwart's radar as potentially emergent include dengue, chikungunya, and Heartland virus. He said, for example, "Heartland virus is not, as far as we know, a concern to the blood supply now, but it might become one." Heartland is a phlebovirus which has so far only infected two people in Missouri, in 2009, causing flu-like symptoms requiring hospitalization, according to the US Centers for Disease Control and Prevention website. However, the virus isolated from these patients was since detected by RT-PCR in pooled samples of ticks gathered at collection sites near the patients' farms. In all, 56,428 ticks were processed, separated into 2,113 pools, and tested for Heartland virus, according to an American Journal of Tropical Medicine and Hygiene study. Ten pools were PCR-positive for Heartland virus, with eight of these yielding enough virus for sequencing.
Viruses transmitted by blood-sucking insects seem to be on the rise and could potentially impact the blood supply in the future. "It's not clear whether there's more of them, or it's got to do with changing ecosystems, or world trade, but there seems to be a lot of these, whether its chikungunya, dengue, or West Nile virus," Delwart said. Blood banks have adapted well thus far, and Delwart asserted that "West Nile was a big concern to the blood supply, and it is [now] actually tested for West Nile viruses during the hot months of [its] spreading, which is pretty much in the summer when mosquitoes are out."