Why RNA viruses?

RNA viruses are among the most simple organisms on Earth. For virologists, RNA viruses are a great model to test many of biology's unknowns, their molecular precision and simplicity fascinate us! But we must not forget that many viruses are the cause of severe diseases in humans, animals and plants. So RNA viruses are both models and targets, as they cause the majority of existing and emerging infectious diseases: polio, influenza, Zika, MERS, dengue, chikungunya, Ebola, Foot and Mouth Disease, West Nile, and many more.

RNA viruses mutate, A lot

Higher organisms, including humans, store their genetic information as DNA, a very stable nucleic acid that is proofread and corrected every time it is copied. This means that large volumes of information can be saved without serious mistakes. RNA viruses, on the other hand, store information as RNA, a less reliable, error-prone version. This RNA is used as a template to make thousands of copies without proofreading. In result, the message quickly changes through mutations, additions or deletions. Mutations often result in non-sense instructions that can kill a virus, but can sometimes make the virus better at infecting, transmitting, or escaping our immune responses.

The mutant swarm

While other organisms, like bacteria or even the cells in our bodies, replicate by dividing into two, and so on and so on; RNA viruses make hundreds or thousands of copies from a single parent. Each copy carries mutations. This creates a swarm or cloud of variants or mutants, that are similar but distinct from one another. When the environment changes, some of these variants might be better at surviving the new conditions and evolution will select them, they will grow out of the population to become the new majority. Because there are thousands of these variants in each viral population, it is difficult to identify them all; but our belief is that if we can characterize and monitor the behavior of this mutant cloud, we may eventually be able to predict a virus' evolution, at least in the short term.

RNA virus replication fidelity: mutators and antimutators

The swarm of mutants created by RNA viruses is mostly the result of the error-riddled copying of virus genomes by the viral RNA polymerase - a copying enzyme that cannot proofread or correct its mistakes. Our lab studies how and why this viral protein makes mistakes, by comparing normal polymerases to mutant ones that make more mistakes (low-fidelity mutators of chikungunya or Coxsackie virus, for example) or less mistakes (like the high-fidelity antimutators of chikungunya virus). We discovered that if we change a virus' ability to make mutations, the virus' fitness (ability to survive) is changed. These fidelity variants are excellent tools to study evolution, and can also help design better vaccines and antivirals.

RNA virus evolution in vitro and in vivo

A lot of what we learn about viruses and how they affect us can be studied in vitro - in test tubes or tissue culture, using cell lines that represent different organs or organisms (comparing insect versus human cells, for example). While these models are not perfect, they allow us to have better control of an experiment. We study how viruses grow and evolve in these cells. Then, we use new sequencing technologies and mathematical modeling to understand how adaptation is occurring. This helps us know why viruses prefer certain cells or tissues over others, and what kind of mutations allow the virus to replicate better or worse -  revealing strengths or weaknesses that we can hope to target in future drug or vaccine designs. In other studies, we work in vivo, looking at how a virus adapts and evolves during natural infection. For example, we can examine the kinds of mutations present in nasal swabs from humans infected by influenza virus, or monitor how a virus like chikungunya evolves over time in infected mosquitoes.

Population dynamics of viral infection

For a virus to successfully infect and transmit to a new host, it must jump over a lot of hurdles. The kinds of obstacles a virus encounters depends on the route of infection and mode of transmission. For example, arboviruses (like dengue or chikungunya virus) are viruses that are carried by mosquitoes and are transmitted to humans. These viruses have complicated life cycles and population dynamics. We study how viruses move from one tissue to another, and whether the mutant swarm changes as this happens. These population dynamics can considerably affect how a virus evolves and one of the goals of our lab is to identify at which point during infection, the viral army is at its most vulnerable stage, and whether targeting these steps can disrupt the transmission cycle. Current work focuses on emerging arboviruses such as Mayaro and Usutu.

Deep sequencing and the mutant swarm

Until ten years ago, we could only sequence one virus at a time - impossible then to describe the thousands of genomes in a sample. With the next generation sequencing (NGS) technologies, we can now identify all of the mutants in a virus population. We routinely use NGS to genetically map the RNA viruses in our lab. But with added power comes the challenge of processing large datasets. In collaboration with partners at the University of Tel Aviv (Noam Shomron lab), we developed a bioinformatic pipeline, ViVan (virus variance analysis), that enables us to fully characterize the mutant swarm and monitor changes in mutant compositions.

Modeling the kinetics and mechanisms of adaptation

While we are an experimental research group (wet lab), we collaborate with computational teams (dry labs) to help us understand the biology that we observe. As an example, we collaborate with computational labs studying protein structure and dynamics (Ben Tal and Haliloğlu labs) to understand how changes in viral proteins improve a virus' ability to infect and survive in the environment. With the help of an applied mathematics group (Fontes lab), we predicted which set of mutations must have emerged together during the adaptation of a virus to a new cell type. These works show that, while virology and mathematics may be quite different disciplines, they can come together to solve questions that would otherwise go unanswered.

Sequence space and fitness landscapes

Sequence space is a multidimensional framework that represents all of the possible variants in a population, and how each of these variants relate to each other. Our goal is to apply mathematics to our experimental data and develop ways to reduce the dimensionality of viral sequence space, so that we can visualize the information in 2 or 3 dimensions. In this way, each variant could be located as a single coordinate in sequence space, much like a map. Fitness landscapes are visual representations, first described by Sewall Wright in the 1930s and conceptually refined over the decades, where 2-dimensional sequence space is given a fitness topology - like adding mountains and valleys to a 2-D map. Mountains represent areas of high fitness that evolving populations want to reach; while valleys are areas of low fitness that populations want to avoid. While helpful, it has been nearly impossible to determine the real fitness landscape of a population, using experimental data. The expectation is that if we are able to reconstruct empirical fitness landscapes, we could monitor a population's evolutionary movement and potentially predict where it will go. In our lab, we use molecular engineering to explore how a virus' sequence space determines its ability to move along these fitness landscapes. These models have recently been developed in collaboration with an applied mathematics lab from Lund University, published as the DISSEQT pipeline.

We find new mutations, and then what?

One of the great things about working in virus evolution is that a new mutation can happen anywhere in the genome and in the protein sequences, which leads us to increase our knowledge and expertise on specific viral proteins accordingly. Indeed, we consider the identification of a new mutation as the starting point, rather than end point, in our studies. Once a new mutation is found, we develop a panel of experiments that will dissect the molecular basis or mechanism of how it works, what it does better than the original protein and why it appears. In doing this, we gain more knowledge on how the normal version of the protein works, and find potential new targets in fighting the viral disease.

Structural proteins and virus entry

The structural proteins of a virus come together to form the shell that will contain the genetic information (RNA genome). The purpose of the shell is to protect the genome while the virus is in transit from one host cell to another, and to find a new target cell to penetrate. Since structural proteins are exposed, it is no wonder that these proteins undergo many mutations. Studying these proteins allow us to understand how a virus physically interacts with the host cell.

Non-structural proteins and virus replication

The non-structural proteins of a virus are the work horses once a virus has entered the cell - they are responsible for hijacking the cell's machinery to make more viruses, and to replicate the virus' RNA genome into thousands of copies. One of the most important non structural proteins we study in our lab is the RNA dependent RNA polymerase, the copying enzyme that makes a single virus into a cloud of mutants. But other non structural proteins also play important roles in replication and infection, that we study through mutation. Finally, non-coding sequences can also play an important role in evolution and the biology of a virus. For example, we found that a duplication in the 3' non-coding region of chikungunya virus increases the ability of the American strain to replicate in mosquito cells.

Viruses are running an obstacle course as they disseminate

When a virus infects a host, it faces an obstacle course. A human, for example, is an incredibly complex environment with many different tissues and anatomical barriers to cross. Each time a virus crosses a physical boundary, its population size is reduced (population bottleneck), and presumably it must regenerate the population that it lost along the way. All along this course, the immune system is chasing after it. Not an easy task for such a simple organism trying to disseminate and transmit. While we know that population bottlenecks exist, we do not know whether the viruses that cross the barrier are specialized, or just make it through by chance. We study how the genetic structure of a virus population may change and may regenerate after it encounters a bottleneck.

What can we learn from transmission studies?

A virus disseminates to ultimately reach a tissue or site where it can be transmitted to a new host and continue its spread. While bottlenecks in a single host are poorly defined for most viruses, even less is known about what occurs when a virus jumps from one host to another. While these experiments are very difficult to perform and control, transmission studies give us the chance to test in the lab, what is happening in nature. Our recent work on transmission of chikungunya virus shows how such studies can help predict the emergence and epidemic potential of new strains. These experimental data could then be coupled with computational modeling to determine how this risk translates to the human, real-world context.

Fidelity variants and lethal mutagenesis

Lethal mutagenesis is an antiviral approach, based on increasing RNA virus mutation rates beyond a threshold of viability, leading to population extinction through deleterious mutations and defective interference. In most studies, mutagenic compounds, such as ribavirin and 5-fluoruracil, are used; but other compounds are being examined as potential RNA mutagen antivirals. Because high and low fidelity variants would be more resistant or more sensitive to such drugs, they constitute excellent tools with which to screen for mutagenic activity. Recently, our work in characterizing high and low fidelity variants of Coxsackie virus B3, uncovered a previously unknown, moderate mutagenic activity for the amiloride compounds. In collaboration with Mark Denison's group, we helped confirm by deep sequencing that a mutator strain of coronavirus was hyper-sensitive to RNA mutagens. We are also examining how the mutational robustness of a given virus strain can render the virus more or less sensitive to increases in mutation rates, as demonstrated in a recent collaboration with Craig Cameron's group.

Other work in the lab has uncovered the role of polyamines in the replication and translation of RNA viruses, and shown that polyamine inhibitors can be used as broadly active antiviral compounds.

Altering virus evolution to attenuate infection

We have used our computational and experimental approaches to identify the ‘good’ and ‘bad’ neighbourhoods in viral sequence space. Using these approaches, we showed that RNA viruses, such as Coxsackie virus and influenza virus, can be forced into attenuation by changing their evolution towards regions of low fitness. We’ve since extended this approach to arboviruses, showing that transmission of chikungunya virus can be impeded.

Poisoning the viral population from within

While our lab has focused for many years on the adaptive mutations that are to the virus’ benefit, we have become more and more interested in the bad mutations too. Most recently, we’ve been looking at other genetic changes, including deletions, and the resulting defective genomes. Many members of our team are now studying how defective genomes are generated and whether they can be used as an antiviral approach. By increasing the number of defective genomes, we hope to ‘poison’ the virus population from within, using these low fitness versions of the virus against infection.