Increased infectivity of Ebola virus glycoprotein from West Africa

filovirionWhen viruses cross species, serial transmission may lead to the selection for mutations that confer improved replication or transmission in the new host. Identifying such mutations in human viruses is extremely difficult: we cannot conduct the appropriate experiments in humans, and often do not have viral isolates spanning the time from spillover through prolonged circulation. The 2013-2016 outbreak of Ebola virus in West Africa is unique because viral genome sequences were obtained early and throughout the epidemic. The results of two new studies (link to paper one, link to paper two) suggest that some of the observed mutations increase infectivity for human cells. The impact of these mutations on infection of humans, and their role in the West African outbreak, remain unknown.

Many mutations have been identified among the many hundreds of genome sequences obtained during the recent Ebola virus epidemic. One stands out: a mutation that leads to a single amino acid change in the viral glycoprotein, from alanine to valine at position 82 (A82V). This change arose early in the outbreak (it was first observed in Guinea in March 2014) and was subsequently found in most of the isolates. It has never been observed in previous Ebolavirus outbreaks.

The effect of the A82V change on viral infectivity was determined by building pseudotyped viral particles – in this case, HIV particles with the Ebola virus glycoprotein. Human cells in culture were infected with pseudotyped viruses with the Ebola virus glycoprotein with either alanine or valine at position 82. Infectivity was measured by quantifying the production of a protein from the HIV genome. The results show that A82V increases infectivity by twofold. The effect is also observed in cells from non-human primates, but not from rodents, dogs, or cats. However, the A82V change decreased infectivity in bat cells.

The A82V change is located at the binding site of the Ebola virus glycoprotein with the cell fusion receptor, NPC1. It appears to increase the fusion activity of the viral glycoprotein.

Other amino acid changes in the Ebola virus glycoprotein were also observed to increase infectivity in human cells, and decrease infectivity in bat cells.

The pattern of increased infectivity in primate cells, and decreased infectivity in bats, is consistent with the hypothesis that the outbreak virus came from bats, and after circulation of the virus in humans, it lost some ability to infect bat cells while becoming better at infecting human cells. However there is still no solid proof that bats are a reservoir of Ebolaviruses.

What does increased infectivity have to do with infection of humans? The idea is that the mutation increases the efficiency of virus entry into cells, and hence increased viral gene expression is observed. Fewer viruses needed to infect a cell, the better chance of initiating an infection. But is the two-fold increase observed in cells enough to impact infection in humans?

The assays used in these papers measure protein production from an HIV genome. The experiments need to be repeated using bona fide Ebola virus, to make sure that the mutations have the same effect. The changes might have impacts on other stages of viral replication. Furthermore, the impact of the changes in the viral glycoprotein should be assessed in animal models, to determine if improved infectivity has any impact on pathogenesis and transmission. Ultimately, we can’t prove that these mutations have any effect in humans – the needed experiments cannot be done.

I’m curious about why the A82V change was not seen in previous Ebola virus outbreaks. Those were in different parts of Africa – could the changes be driven by population genetics, ecology, or other factors? It will be important to determine if the same change is selected in future outbreaks.

The authors are sufficiently cautious in their conclusions. From paper #2:

Despite the experimental data provided here, it is impossible to clearly establish whether the adaptive mutations observed were in part responsible for the extended duration of the 2013–2016 epidemic. Indeed, it seems likely that the prolonged nature of the outbreak in West Africa was primarily due to epide- miological factors, such as an increased circulation in urban areas that in turn led to larger chains of transmission.

From paper #1:

Our findings raise the possibility that this mutation contributed directly to greater transmission and thus to the severity of the outbreak. It is difficult to draw any conclusions from this hypothesis, though…

As I feared, press coverage of these findings has been inaccurate. For example, a BBC headline proclaims “Ebola adapted to easily infect people”. Even the journal Cell, which published both papers, made an incorrect conlcusion: see the screen capture below from the journal website.key mutations ebola virusBoth Cell and the BBC might have taken too literally the unfortunate title of one of the papers,  “Human adaptation of Ebola virus during the West African Outbreak.” The results suggest adaptation to human cells, not to humans. The title of the second paper is sufficiently careful: “Ebola virus glycoprotein with increased infectivity dominated the 2013-2016 epidemic”. But that’s not a BBC headline.

TWiV 413: Partnerships not parachutes

From the EIDA2Z conference at Boston University, Vincent, Alan and Paul meet up with Ralph Baric, Felix Drexler, Marion Koopmans, and Stacey Schultz-Cherry to talk about discovering, understanding, protecting, and collaborating on emerging infectious diseases.

You can find TWiV #413 at microbe.tv/twiv, or watch or listen here.

Click arrow to play
Download TWiV 413 (44 MB .mp3, 72 min)
Subscribe (free): iTunesRSSemail

Become a patron of TWiV!

Genome recombination across viral families

coronavirus and reovirus A novel coronavirus isolated from bats in China is unusual because the genome contains a gene from a virus in a completely different family, the Reoviridae (link to paper). The finding suggests that recombination occurred between a (+) strand RNA virus and a virus with a segmented, double-stranded RNA genome.

The unusual recombinant virus was identified in rectal swabs from Rousettus leschenaulti bats in Yunnan Province, China. Sequence analysis revealed a typical coronavirus genome with the exception of a small region near the 3′-end of the viral RNA with homology to a bat reovirus. This sequence, called p10, was also detected in viral mRNAs from infected bats, further demonstrating the presence of the reovirus-like gene in the coronavirus genome.

In bat reoviruses, the p10 gene is known to encode a protein that causes cell fusion. When the p10 gene from the bat coronavirus was expressed in cells, formation of syncytia (fused cells) was observed. Furthemore, the p10 protein was detected by western blot analysis of feces from infected bats. These results indicate that the p10 protein is produced from the viral genome and that the protein is functional.

This report is not the first suggesting recombination between viruses of different families – we discussed one example here previously (link to article), and there are a handful of other examples. The important question is how such inter-family recombinants arise. It must begin with co-infection of a host with two different viruses – in this case, likely a bat – but the precise molecular events are unknown. It might be useful to attempt to isolate such recombinants in cell culture to understand the underlying mechanisms.

TWiV 412: WO, open the borders and rig the infection

The TWiVome reveal the first eukaryotic genes found in a bacteriophage of Wolbachia, and how DNA tumor virus oncogenes antagonize sensing of cytoplasmic DNA by the cell.

You can find TWiV #412 at microbe.tv/twiv, or listen below.

Click arrow to play
Download TWiV 412 (73 MB .mp3, 121 min)
Subscribe (free): iTunesRSSemail

Become a patron of TWiV!

A new function for oncoproteins of DNA tumor viruses

oncoproteinsOncogenes of DNA tumor viruses encode proteins that cause cells to divide incessantly, eventually leading to formation of a tumor. These oncoproteins have now been found to antagonize the innate immune response of the cell (link to paper).

Most cells encountered by viruses are not dividing, and hence do not efficiently support viral DNA synthesis. The genomes of adenoviruses, polyomaviruses, and papillomaviruses encode proteins that cause cells to divide. This effect allows for efficient viral replication, because a dividing cell is producing the machinery for DNA synthesis. Under certain conditions, infections by these viruses do not kill cells, yet they continue to divide due to the presence of viral oncoproteins. Such incessant division gives the cells new properties – they are called transformed cells – and they may eventually become a tumor.

These so-called viral oncoproteins include large T antigen (of SV40, a polyomavirus); E6 and E7 (papillomavirus), and E1A (adenovirus). These viral proteins kick cells into mitosis by inactivating cell proteins (such as Rb, pictured) that are normally involved in regulating cell growth. The cells divide, and in the process produce proteins involved in DNA replication, which are then used for viral replication. These oncoproteins accidentally cause tumors: the replication of none of these viruses is dependent on transformation or tumor formation.

Cells transformed with T, E6/E7, or E1A proteins are commonly used in laboratories because they are immortal. An example is the famous HeLa cell line, transformed by human papillomavirus type 18 (which originally infected Henrietta Lacks and caused the cervical tumor that killed her). Another commonly used transformed cell line is 293 (human embryonic kidney cells transformed by adenovirus E1A). It’s been known for some time that when DNA is introduced into normal (that is, not transformed) cells, they respond with an innate response: interferons are produced. In contrast, when DNA is introduced into the cytoplasm of a transformed cell, there is no interferon response.

To understand why HeLa and HEK 293 cell lines did not respond to cytoplasmic DNA, the authors silenced the viral oncogenes by disrupting them with CRISPR/Cas9. The altered cells produced interferon in response to cytoplasmic DNA. Furthermore, they produced new transformed lines by introducing genes encoding E6, E7, E1A, or T into normal mouse embryonic fibroblasts. These new transformed cells failed to respond to cytoplasmic DNA.

Cytoplasmic DNA is detected in cells by an enzyme called cGAS (cyclic guanosine monophosphate-adenosine monophosphate synthase) together with an adaptor protein known as STING (stimulator of interferon genes). When cytoplasmic DNA is detected by this system, the antiviral interferons are produced. The viral oncoproteins were found to directly bind STING, but not cGAS. A five amino acid sequence within E1A and E7 proteins was identified that is responsible for overcoming the interferon response to cytoplasmic DNA. When this sequence was altered, interaction of the oncoprotein with cGAS was reduced, and antagonism of interferon production in response to cytoplasmic DNA was blocked.

These findings provide a new function for the oncoproteins from three DNA tumor viruses: antagonism of the interferon response to cytoplasmic DNA. Normally DNA is present in the cell nucleus, and when it is detected in the cytoplasm, this is a signal that a virus infection is underway. The cytoplasmic DNA is sensed by the cGAS-STING system, leading to interferon production and elimination of infection. A herpesvirus protein has been identified that binds to STING and blocks interferon responses to cytoplasmic DNA. Clearly antagonism of the cGAS-STING DNA sensing system is of benefit to DNA viruses.

An interesting question is what selection pressure drove the evolution of viral oncogenes. One hypothesis, described above, is that they are needed to induce a cellular environment that supports viral DNA synthesis. The other idea, favored by the authors of this new work, is that oncogenes arose as antagonists of innate immune signaling. But I can’t imagine these DNA viruses without oncogenes, because they would not be able to replicate very efficiently. Could both functions have been simultaneously selected for? Why not – the same five amino acid sequence that binds cGAS also binds cellular proteins (such as Rb), disrupting their function and leading to uncontrolled cell growth!

TWiV 411: Chicken runs

The TWiVeroos examine a reverse spillover of Newcastle disease virus vaccines into wild birds, and identification of a protein cell receptor for murine noroviruses.

You can find TWiV #411 at microbe.tv/twiv, or listen below.

Click arrow to play
Download TWiV 411 (59 MB .mp3, 95 min)
Subscribe (free): iTunesRSSemail

Become a patron of TWiV!

Eukaryotic genes in a bacteriophage

Wolbachia

Wolbachia in an insect cell. Image credit: PLoS/Scott O’Neill.

Viruses are tidily categorized into three groups according to the hosts they infect – bacteriophages, eukaryotic viruses, and archaeal viruses. Viruses do not infect hosts in another domain of life, and therefore lateral gene transfer is limited (giant DNA viruses might be exceptions). Now there is evidence for lateral gene transfer between eukaryotes and bacteriophages.

Proof of this unusual movement of DNA comes from studies of the obligate intracellular bacteria Wolbachia, which infects 40% of arthropods (pictured). Wolbachia are in turn infected with a bacteriophage called WO; nearly all sequenced Wolbachia genomes contain integrated WO DNA. Analysis of complete WO genome sequences revealed the presence of mutiple eukaryotic genes (link to paper) that comprise about half of the phage genome!

Ten different protein domains were identified in the eukaryotic genes of WO phage with four functions: toxins, host-microbe interactions, host cell suicide, and protein secretion through membranes.

One eukaryotic gene in phage WO is a black widow spider toxin called latrotoxin-CTD. Sequence analysis suggests that the spider toxin gene was transferred to phage WO within a Wolbachia genome (these bacteria are known to infect widow spiders).

It is not surprising that a virus of a bacterium that infects a eukaryotic cell might acquire eukaryotic genes, but the exact mechanism of gene transfer is unknown. Eukaryotic DNA might enter the WO genome while the particles are in the insect cell cytoplasm, or during packaging of viral DNA in the presence of animal DNA. Another possibility is transfer of eukaryotic DNA to the Wolbachia genome, and then to phage WO.

The fact that eukaryotic-like DNA sequences make up half of the phage WO genome suggests that they serve important functions for the virus. The functions ascribed to these eukaryotic genes suggest roles in cell lysis, modification of host proteins, and toxicity.

There are other examples of phage-infected obligate intracellular bacteria of Chlamydia, aphids, and tsetse flies. A study of these viral genomes should reveal whether lateral gene transfer between metazoans and bacteriophages is a common mechansim for augmenting functions of the viral genome.

Virus Watch: Cancer Killing Viruses

Guest host Lynda Coughlan reviews how oncolytic viruses, which specifically kill tumor cells, are designed and how they work.

TWiV 410: Hurricane Zika

Sharon and Scott join the TWiV team to talk about their work on dengue antibody-dependent enhancement of Zika virus infection, and identifying the virus in mosquitoes from Miami.

You can find TWiV #410 at microbe.tv/twiv, or listen below.

Click arrow to play
Download TWiV 410 (73 MB .mp3, 121 min)
Subscribe (free): iTunesRSSemail

Become a patron of TWiV!

A Nobel for autophagy, and the importance of fundamental research

autophagyYoshinori Ohsumi has been awarded the 2016 Nobel Prize for Physiology or Medicine for his work on autophagy, a process of eukaryotic cells for degrading and recycling cellular components. Because of his research, we now understand the importance of autophagy in health and human disease. It is another example of the serendipity of science and yes, it is relevant to virology!

The word autophagy was coined by Christian de Duve in 1963 to describe a process that he and others had previously described: when stressed, cells would sequester portions of the cytoplasm in double-membraned vesicles called autophagosomes. These would then fuse with lysosomes (which de Duve had discovered) and the contents were degraded (illustrated; image credit).

In subsequent years it was suggested that autophagy might have roles in human disease, but little progress was made on understanding how the process worked: how it was triggered, what proteins were involved, and its function in health and pathogenesis.

As a young Assistant Professor at Tokyo University in the early 1990s, Ohsumi found that autophagy occurred in the yeast Saccharomyces cerevisiae. He decided to produce yeast strains lacking the proteases that would be involved in digesting the contents of the autophagic vesicles, with the idea that these vesicles would accumulate under stress (under normal conditions the autophagosome exists for a short period of time, making its study difficult).

When Ohsumi stressed the yeasts lacking the vacuolar (equivalent of mammalian lysosome) proteases, autophagosomes accumulated in the cytoplasm which were readily visible by light microscopy. He used this clear phenotype to isolate yeast mutants that could no longer accumulate such vesicles, and identified 15 genes that are essential for induction of autophagy in eukaryotic cells.

Later Ohsumi elucidated the functions of these genes in autophagy induction, and found homologues in mammalian cells. His work stimulated great interest in autophagy, and it became a highly studied pathway. We now understand that autophagy is not only important for embryogenesis and cell differentiation, but plays roles in neurodegenerative diseases, cancer, and in defenses against bacterial and viral infections.

As might be expected, virus infection is a stress that triggers autophagy, which may impact the outcome of infections: it can have both antiviral functions, and it can also stimulate virus replication. As might be expected for a cellular process that degrades cytosolic contents, autophagy can lead to clearance of viruses. Consequently a number of viral genomes encode proteins that inhibit autophagy, including herpes simplex virus. Degradation of viral proteins by autophagy can also provide peptides for presentation to the cellular adaptive immune system, further enhancing clearance.

Autophagy may also benefit virus replication. For example, non-lytic cell to cell spread of poliovirus depends on release of virus particles from autophagosomes, and autophagy of lipids provides metabolic energy for dengue virus replication.

When autophagy was first described in the 1950s, no one knew its significance. Nevertheless, a number of scientists, including Ohsumi, continued to study autophagy because they were curious. The unexpected result was the elucidation of a pathway that has substantial roles in a variety of human diseases.

The lesson is clear: let scientists pursue their curiosity. It’s fine to target specific research problems, like curing cancer or diabetes, but don’t ignore fundamental research on problems that don’t seem to be directly relevant to human diseases. I’m concerned (as are many others) that the US science establishment is moving away from fundamental research (whose benefits may not be apparent for a long time) to translational research.

How do we explain the trend from fundamental to translational research? I don’t have all the answers, but I think part of the problem is that the US Congress likes to spend taxpayer money on targeted problems, like Alzheimer’s disease. But tell them that you want to study a process in cells that looks interesting, but you are not sure what it means, and you can see their reluctance to support this type of work.