TWiV 419: The selfless gene

The TWiVrific gang reveal how integration of a virophage into the nuclear genome of a marine protozoan enhances host survival after infection with a giant virus.

You can find TWiV #419 at, or listen below.

Click arrow to play
Download TWiV 419 (64 MB .mp3, 105 min)
Subscribe (free): iTunesRSSemail

Become a patron of TWiV!

Altruistic viruses

Cafeteria roenbergensisVirophages (the name means virus eater) were first discovered to replicate only in amoeba infected with the giant mimiviruses or mamaviruses.  They reduce yields of the giant viruses, and also decrease killing of the host cell. Another virophage called mavirus has been found to integrate into the genome of its host and behaves like an inducible antiviral defense system (link to paper).

The host cell of the virophage mavirus is Cafeteria roenbergensis, Cro (pictured), a marine phagotropic flagellate, that is infected with the giant virus CroV (Cafeteria roenbergensis virus). When Cro cells are infected with a mixture of mavirus and CroV, the virophage integrates into the host cell genome. There it remains silent; the cells survive, and no virophage particles are produced. Such cells can be called lysogens, a name applied to bacteria containing integrated bacteriophage genomes, or prophages.

How does the mavirus genome integrate into the Cro cell? The viral genome encodes an integrase, an enzyme that cuts host DNA and inserts a copy of the viral genome. Retroviruses achieve the same feat via an integrase.

When Cro-mavirus lysogens are infected with CroV, the integrated mavirus genome is transcribed to RNA, the viral DNA replicates, and new virus particles are formed. These virophages inhibit the replication of CroV by 100-1000 fold. As a consequence, the host cell population survives.

These findings suggest that the virophage mavirus is altruistic: induction of the integrated genome leads to killing of the host cell, but other members of the cell population are protected. Altruism is not unknown in Nature, but how it evolved is an intriguing question.

All this work was done in a laboratory. It will be necessary to determine if integration of mavirus into Cro cells in the wild has any influence on the ecology of these organisms.

Do giant viruses have a CRISPR-like immune system or a protein restriction factor?

Zamilon virophageA battle is brewing between two research groups in Marseille, France that are involved in the discovery and study of giant viruses. Didier Raoul and colleagues believe that they have discovered a CRISPR-like, DNA based defense system in mimivirus that confers resistance to virophage (paper link). Claverie and Abergel disagree: they think that the defense system involves proteins, not nucleic acids (paper link).

Virophages are DNA viruses that can only replicate in cells infected by giant viruses like mimivirus. Their name, which means ‘virus eater’, comes from the observation that they inhibit mimivirus replication. A specific virophage called Zamilon was discovered that can inhibit the replication of lineage B and C mimivirus but not lineage A.

Examination of the DNA sequences of 60 different mimivirus strains revealed that the genomes of lineage A contained a 28 nucleotide sequence identical to Zamilon virophage. This sequence was not found in any lineage B mimivirus and in only one out of 19 lineage C mimiviruses. In addition, a 15 nucleotide subset of this sequence is repeated four times in the lineage B and C mimivirus genomes.

Near the 15 nucleotide Zamilon-derived repeated sequences in lineage B and C mimivirus genomes are genes encoding several proteins related to components of the bacterial CRISPR-Cas system. These include a nuclease, an RNAse, and an ATP-dependent DNA helicase.

The CRISPR system provides defense against invading DNA. When a foreign DNA, such as a bacteriophage genome, enters a bacterial cell, some is fragmented and integrated into the CRISPR locus as a ’spacer’ (sequences in the foreign DNA are called ‘protospacers’). Following transcription, CRISPR RNAs (crRNA) are processed by a multiprotein complex to produce ~60 nucleotide RNAs. When the spacer of a crRNA base pairs with a complementary sequence in an invading DNA molecule, CRISPR-associated endonucleases cleave the DNA. The integration of the sequences of the invading DNA into the host cell genome, from which they can be mobilized in the form of crRNAs, provides a form of “memory” and acquired immunity. It should be noted that there are six known types of CRISPR systems that differ in their components and mechanisms.

Because the CRISPR-Cas system is an adaptive immune system that protects bacteria and Archaea from virus infections and invasion of foreign DNA, the authors propose that they have discovered a new adaptive immune system that protects mimiviruses from virophage infection. They call this system mimivirus virophage resistance element, or MIMIVIRE.

The authors provide experimental support for their hypothesis by showing that silencing the genes encoding the endonuclease, the helicase, and the repeated insert using siRNA allows Zamilon replication in mimivirus-infected cells.

Claverie and Abergel think that Raoult and colleagues are wrong (paper link). They provide three reasons to dispute their findings, and ‘propose a simpler protein-based interaction model that explains the observed phenomena without having to extend the realm of adaptive immunity to the world of eukaryotic viruses, a revolutionary step that would require stronger experimental evidences.’

The first problem is that mimivirus and Zamilon virophage replicate in the same location in the infected cell, making a CRISPR-like defense system difficult to conceptualize. In contrast, CRISPR sequences reside in the bacterial genome, from which RNAs are produced that target the destruction of invading DNAs elsewhere in the cell.

The second problem is that the Zamilon sequences in the mimivirus genome are not regularly spaced or flanked by recognizable repeats, a hallmark of the CRISPR system (the name stands for ‘clustered regularly interspersed short palindromic repeats). However it should be noted that type VI CRISPR systems have no CRISPR locus and likely function via mechanisms that are different from other CRISPR systems.

Finally, Claverie and Abergel argue that there is no way for the proposed nucleic acid defense system to distinguish between the virophage and the virophage sequences in the mimivirus genome. In some CRISPR systems this discrimination is achieved by protospacer adjacent motifs (PAMs), short (2-5 nt) sequences next to the invader protospacer sequences that are recognized by the endonuclease complex guided by the crRNA. PAMs are not present in the bacterial genome, sparing it from endonucleolytic cleavage. Nevertheless, non PAM-based mechanisms of discriminating invader from host are known, for example, in the type III CRISPR system.

If Raoult and colleagues have not discovered a CRISPR-like mimivirus defense system, then why would silencing the genes encoding CRISPR-like proteins allow Zamilon replication? Claverie and Abergel think that it is not the 15 nucleotide Zamilon repeats that are important to mimivirus, but the encoded amino acids: Asp-Asn-Glu-Ser (DNES in one letter code). They believe that DNES is a motif present in proteins that block Zamilon replication by as yet unidentified mechanisms.

Many cellular proteins have been identified that interfere with virus replication, such as those encoded by interferon induced genes (ISGs). DNES containing proteins that inhibit Zamilon replication would be conceptually analogous, except that they are encoded by a virus, not the host.

Claverie and Abergele appear to have a strong case that mimivirus defense against Zamilon virophage is mediated by protein, not nucleic acid, but further experimentation is certainly needed to support their position. Nevertheless they recognize that the discovery by Raoult and colleagues “remains fascinating even if it falls short of demonstrating the existence of a CRISPR-Cas-like adaptive immune system”.

TWiV 380: Viruses visible in le microscope photonique

TWiVOn episode #380 of the science show This Week in Virology, the TWiVeroos deliver the weekly Zika Report, then talk about a cryoEM structure of a plant virus that reveals how the RNA genome is packaged in the capsid, and MIMIVIRE, a CRISPR-like defense system in giant eukaryotic viruses.

You can find TWiV at, or you can listen below.

Click arrow to play
Download TWiV 380 (80 MB .mp3, 110 min)
Subscribe (free): iTunesRSSemail

Satellites – the viral kind

Hepatitis delta satellite genomeSatellites are subviral agents that differ from viroids because they depend on the presence of a helper virus for their propagation. Satellite viruses are particles that contain nucleic acid genomes encoding a structural protein that encapsidates the satellite genome. Satellite RNAs do not encode capsid protein, but are packaged by a protein encoded in the helper virus genome. Satellite genomes may be single-stranded RNA or DNA or circular RNA, and are replicated by enzymes provided by the helper virus. The origin of satellites remains obscure, but they are not derived from the helper virus.

Satellite viruses may infect plants, animals, or bacteria. An example of a satellite virus is satellite tobacco necrosis virus, which encodes a capsid protein that forms an icosahedral capsid that packages only the 1,260 nucleotide satellite RNA. The helper virus, tobacco necrosis virus, encodes an RNA polymerase that replicates its genome and that of the satellite.

Satellite RNAs do not encode a capsid protein and therefore require helper virus proteins for both genome encapsidation and replication. Satellite RNA genomes range in length from 220-1500 nucleotides, and have been placed into one of three classes. Class 1 satellite RNAs are 800-1500 nucleotide linear molecules with a single open reading frame encoding at least one non-structural protein. Class 2 satellite RNAs are linear, less than 700 nucleotides long and do not encode protein. Class 3 satellite RNAs are 350-400 nucleotide long circles without an open reading frame.

In plants, satellites and satellite viruses may attenuate or exacerbate disease caused by the helper virus. Examples of disease include necrosis and systemic chlorosis, or reduced chlorophyll production leading to leaves that are pale, yellow, or yellow-white. The symptoms induced by satellite RNAs are thought to be a consequence of silencing of host genes. For example, the Y-satellite RNA of cucumber mosaic virus causes systemic chlorosis in tobacco. This syndrome is caused by production of a small RNA from the Y-satellite RNA that has homology to a gene needed for chlorophyll biosynthesis. Production of this small RNA leads to degradation of the corresponding mRNA, causing the bright yellow leaves.

The giant DNA viruses including Acanthamoeba polyophaga mimivirus, Cafeteria roenbergensis virus, and others are associated with much smaller viruses (sputnik and mavirus, respectively) that depend upon the larger viruses for reproduction. For example, sputnik virus can only replicate in cells infected with mimivirus, and does so within viral factories. Whether these are satellite viruses or something new (they have been called virophages) has been a matter of controversy.

Like satellite viruses, sputnik and others have similar relationships with their helper viruses: they require their helper for their propagation, but their genomes are not derived from the helper, and they negatively impact helper reproduction. Others argue that the definition of satellite viruses as sub-viral agents cannot apply to these very large viruses. For example, sputnik virophage contains a circular dsDNA genome of 18,343 bp encoding 21 proteins encased in a 75 nm t=27 icosahedral capsid. Sputnik is dependent upon mimivirus not for DNA polymerase – it encodes its own – but probably for the transcriptional machinery of the helper virus. Those who favor the name virophage argue that dependence upon the cellular transcriptional machinery is a property of many autonomous viruses – the only difference is that Sputnik depends upon the machinery provided by another virus. It seems likely that a redefinition of what constitutes a satellite virus will be required to solve this disagreement.

Most known satellites are associated with plant viruses, but hepatitis delta satellite virus is associated with a human helper virus, hepatitis B virus. The genome (illustrated) is 1.7 kb – the smallest of any known animal virus – of circular single-stranded RNA that is 70% base paired and folds upon itself in a tight rod-like structure. The RNA molecule is replicated by cellular RNA polymerase II. These properties resemble those of viroid genomes. On the other hand, the genome encodes a protein (delta) that encapsidates the RNA, a property shared with satellite nucleic acids. The hepatitis delta satellite virus particle comprises the satellite nucleocapsid packaged within an envelope that contains the surface protein of the helper, hepatitis B virus.

Infection with hepatitis delta satellite virus only occurs in individuals infected with hepatitis B virus: it is globally distributed, present in about 5% of the 350 million carriers of hepatitis B virus. Acute co-infections of the two viruses can be more severe than infection with hepatitis B virus alone, leading to more cases of liver failure. In chronic hepatitis B virus infections, hepatitis delta satellite virus aggravates pre-existing liver disease, and may lead to more rapid progression to cirrhosis and death than monoinfections. Why co-infection with both viruses leads to more serious outcomes is not known.

TWiV 275: Virocentricity with Eugene Koonin

On episode #275 of the science show This Week in Virology, Vincent and Rich meet up with Eugene Koonin to talk about the central role of viruses in the evolution of all life.

You can find TWiV #275 at

Pithovirus: Bigger than Pandoravirus with a smaller genome

PithovirusA new virus called Pithovirus sibericum has been isolated from 30,000 year old Siberian permafrost. It is the oldest DNA virus of eukaryotes ever isolated, showing that viruses can retain infectivity in nature for very long periods of time.

Pithovirus was isolated by inoculating cultures of the amoeba Acanthamoeba castellani with samples taken in the year 2000 from 30 meters below the surface of a late Pleistocene sediment in the Kolyma lowland region. This amoeba had been previously used to propagate other giant viruses, such as Mimivirus and Pandoravirus. Light microscopy of the cultures revealed the presence of ovoid particles which were subsequently shown by electron microscopy to resemble those of Pandoravirus. Pithovirus particles are flask-shaped and slightly larger than Pandoravirus – 1.5 microns long, 500 nm in diameter, encased by a 60 nm thick membrane. One end of the virus particle appears to be sealed with what the authors call a cork (photo). This feature, along with the shape of the virus particle,  inspired the authors to name the new isolate Pithovirus, from the Greek word pithos which refers to the amphora given to Pandora. The name therefore refers both to the morphology of the virus particle and its similarity to Pandoravirus.

Although the Pithovirus particle is larger than Pandoravirus, the viral genome – which is a double-stranded molecule of DNA – is smaller, a ‘mere 610,033 base pairs’, to use the authors’ words (the Pandoravirus genome is 2.8 million base pairs in length). There are other viruses with genomes of this size packed into much smaller particles – so why is the Pithovirus particle so large? Might it have recently lost a good deal of its genome and the particle size has not yet caught up? One theory of the origin of viruses is that they originated from cells and then lost genes on their way to becoming parasitic.

We now know of viruses from two different families that have similar morphology: an amphora-like shape, an apex, and a thick electron-dense tegument covered by a lipid membrane enclosing an internal compartment. This finding should not be surprising: similar viral architectures are known to span families. The icosahedral architecture for building a particle, for example, can be found in highly diverse viral families. The question is how many viruses are built with the pithovirus/pandoravirus structure. My guess would be many, and they could contain either DNA genomes. We just need to look for them, a process, as the authors say that ‘will remain a challenging and serendipitous process’.

Despite the physical similarity with Pandoravirus, the Pithovirus genome sequence reveals that it is barely related to that virus, but more closely resembles members of the Marseillviridae, Megaviridae, and Iridoviridae. These families all contain large icosahedral viruses with DNA genomes.  Only 32% of the 467 predicted Pithovirus proteins have homologs in protein databases (this number was 61% for Mimivirus and 16% for Pandoravirus). In contrast to other giant DNA viruses, the genome of Pithovirus does not encode any component of the protein synthesis machinery. However the viral genome does encode the complete machinery needed to produce mRNAs. These proteins are present in the purified Pithovirus particle. Pithovirus therefore undergoes its entire replication cycle in the cytoplasm, much like other large DNA viruses such as poxviruses.

Pithovirus is an amazing virus that hints about the yet undiscovered viral diversity that awaits discovery. Its preservation in a permafrost layer suggests that these regions might harbor a vast array of infectious organisms that could be released as these regions thaw or are subjected to exploration for mineral and oil recovery. A detailed analysis of the microbes present in these regions is clearly needed, both by the culture technique used in this paper and by metagenomic analysis, to assess whether any constitute a threat to animals.

TWiV 261: Giants among viruses

On episode #261 of the science show This Week in Virology, Vincent meets up with Chantal and Jean-Michel at the first International Symposium on Giant Virus Biology in Tegernsee, Germany, to discuss their work on Mimivirus, Megavirus, and Pandoravirus.

You can find TWiV #261 at

TWiV 246: Pandora, pandemics, and privacy

On episode #246 of the science show This Week in Virology, Vincent, Alan, Rich, and Kathy discuss the huge Pandoravirus, virologists planning H7N9 gain of function experiments, and limited access to the HeLa cell genome sequence.

You can find TWiV #246 at

We recorded this episode of TWiV as a Google hangout on air. Consequently the audio is not the same quality as you might be used to. But the tradeoff is that you can see each of us on video.


Pandoravirus, bigger and unlike anything seen before

pandoravirusThe discovery of the giant Mimivirus and Megavirus amazed virologists (and also many others). Their virions (750 nanometers) and DNA genomes (1,259,000 base pairs) were the biggest ever discovered, shattering the notions that viruses could not be seen with a light microscope, and that viral genomes were smaller than bacterial genomes. Now two even bigger viruses have been discovered, which are physically and genetically unlike any previously known viruses. They have been called Pandoraviruses.

Both new viruses were isolated by culturing environmental samples in the amoeba Acanthamoeba castellaniPandoravirus salinus was isolated from shallow marine sediment in a river at the coast of central Chile, and Pandoravirus dulcis was obtained from mud at the bottom of a freshwater pond near Melbourne, Australia. The P. salinus genome is at least 2.77 megabases in length (there is some uncertainty in the actual length due to the presence of repeated sequences at the ends of the DNA), while the P. dulcis genome is 2.47 megabases in length. The smaller P. dulcis genome is a subset of the P. salinus genome.

These new genomes are twice as large as those of previously described viruses, and bigger than the genomes of intracellular bacteria such as Tremblaya (138,927 base pairs) and Rickettsia (1,111,523 bp), some free living bacteria, and many free living Archaea.

While the huge sizes of the Pandoravirus virion and genomes are amazing, I find three other features of these viruses even more remarkable. The first is their atypical replication cycle. The virions are taken into amoebae by phagocytic vacuoles, and upon fusing with the vacuole membrane, the virion contents are released into the cytoplasm via a pore on the virion apex. Within 2-4 hours the cell nucleus is reorganized, and by 8-10 hours new particles appear where the nucleus once was. Pandoravirus DNA and virions are synthesized and assembled simultaneously, in contrast to eukaryotic DNA viruses and phages which fill pre-formed capsids with DNA. Virions are released by 10-15 hours as the cells lyse.

A second amazing feature is that most of the P. salinus open reading frames encode brand-new proteins. Of the 2,556 putative protein coding sequences in the P. salinus genome, 93% have no recognizable counterparts among known proteins. Some of the genes found in large DNA viruses are present, such as those encoding DNA polymerase and DNA-dependent RNA polymerase, and several amino acyl-tRNA synthetases, like members of the Megaviridae. Curiously, many of the Pandoravirus coding regions contain intervening sequences, which must be removed by RNA splicing. This process is known to occur only in the cell nucleus, suggesting that some Pandoravirus transcription occurs in that organelle. The lack of gene homology leads to authors to conclude that ‘no microorganism closely related to P. salinus has ever been sequenced’.

I am also impressed by what the authors describe as the ‘alien morphological features’ of the virions. The oval-shaped particles are 1 micron in length and 0.5 microns in diameter, easily visible by light microscopy. They are wrapped in a three-layered envelope with a pore at one end of the particle, and resemble nothing that has ever been seen before (see photograph).

How much bigger can viruses get? I don’t know the answer but I would guess even bigger than Pandoraviruses. The membranous Pandoravirus particle could easily accommodate even larger genomes. How big can a virus get and still be a virus? The answer to that question is easy: it is a virus as long as it requires a cell for replication.

These remarkable findings further emphasize the need for scientists to pursue their curiosity, and not only work on problems of obvious medical relevance. As the authors write,

This work is a reminder that our census of the microbial diversity is far from comprehensive and that some important clues about the fundamental nature of the relationship between the viral and the cellular world might still lie within unexplored environments.

Continuing their playful naming of giant viruses, the authors note that the name Pandoravirus reflects their ‘lack of similarity with previously described microorganisms and the surprises expected from their future study’.