Posts Tagged ‘origin of viruses’

A feeling for the Molechism* – revisited

10 July, 2012

This is an update of a post I did on Alan Cann’s MicrobiologyBytes back in 2007, before i started ViroBlogy: I am doing this because (a) it’s mine, (b) I want to update it – and the MB version is archived, so I can’t.  So here we are again:

I think it’s permissible, after working on your favourite virus for over 20 years, to develop some sort of feeling for it: you know, the kind of insight that isn’t directly backed up by experiment, but that may very well be right. Or not – but in either case, it would take a deal of time and a fair bit of cash to prove or disprove, and would have sparked some useful discussion in the meantime. And then, of course, the insights you have into (insert favourite virus name here) – if correct – can usually be extended into the more general case, and if you are sufficiently distinguished, people may actually take them on board, and you will have contributed to Accepted Wisdom.

I can’t pretend – at least, outside of my office – to any such Barbara McClintock-like distinction; however, I have done a fair bit of musing on my little sphere of interest as it relates (or not) to the State of the Viral Universe, and I will share some of these rambles now with whomever is interested.

I have been in the same office now, and teaching the same course, more or less, for 32-odd years. In that time I have worked on the serology and epidemiology of the bromoviruses, cucumovirus detection, potyvirus phylogeny, geminivirus diversity and molecular biology, HIV and papillomavirus genetic diversity, and expressing various bits of viruses and other proteins in plants and in insect cells. However, much of my interest (if not my effort) in that time has been directed towards understanding how grass-infecting mastreviruses in particular interact with their environment and with each other, in the course of their natural transmission cycle.

Maize streak virus

Maxwell’s Demon (left, lower) and Martian Face (right, upper) visible on a MSV virion

Fascinating little things, mastreviruses: unique geminate capsid architecture, and at around a maximum of 2.8 kb of single-strand circular DNA, we thought they were the smallest DNA genomes known until the circoviruses and then the zoo of anello- and anello-like viruses were discovered. Their genomes code for only 4 proteins – two replication-associated, one movement and one capsid – yet we have managed to work on just one subgroup of mastrevirus species for 27 years, without exhausting its interest – at least, to us… (see PubMed list here). We also showed that one could see Martian faces quite distinctly on virions – and possibly even Maxwell’s Demon. But I digress….

Maize streak

Severe symptoms of MSV on sweetcorn

We have concentrated on the “African streak viruses” – related species Maize streak virus, Panicum streak virus, Digitaria streak virus, Sugarcane streak virus and friends – for two very simple reasons:
1. They occur in Africa, near us, and nowhere else;
2. Maize streak virus is the worst viral pathogen affecting maize in Africa.

So we get situational or niche advantage, and we get to work on an economically-important pathogen. One that was described – albeit as “…not of…contagious nature” – as early as 1901, no less.

Maize streak virus

Maize streak virus or MSV, like its relatives, is obligately transmitted by a leafhopper (generally Cicadulina mbila Naudé): this means we have a three-party interaction – of virus-host-vector – to consider when trying to understand the dynamics of its transmission. Actually, it’s more complicated than that: we have also increasingly to consider the human angle, given that the virus disease affects mainly the subsistence farming community in Africa, and that human activity has a large influence on the spread of the disease. So while considering just the virus – as complicated as that is – we have to remember that it is only part of the whole picture.

So how complicated is the virus? At first sight, not very: all isolates made from severe maize infections share around 97% of their genome sequence. However, however…that 3% of sequence variation hides a multitude of biological differences, and there is a range of relatives infecting grasses of all kinds, some of which differ by up to 35% in genome sequence. Moreover, maize is a crop plant first introduced to Africa a maximum of 500 years ago, so it is hardly a “natural” host – yet, all over Africa, it is infected by only a very narrow range of virus genotypes, from a background of very wide sequence diversity available.

So here’s an insight:

the host selects the virus that replicates best in it.

And lo, we found that in the Vaalharts irrigation area in the north of South Africa that the dominant virus genotype in winter wheat was a different strain – >10% sequence difference – to the one in the same field, in summer maize. Different grass species also have quite different strains or even species of streak viruses best adapted to them.

DendrogramNot all that profound a set of observations, perhaps, but they lead on to another insight:

streak viruses travel around as a cloud of variants or virus complex.

Not intuitively obvious, perhaps…but testable, and when we did, we found we were right: cloning virus genomes back out of maize or from a grass infected via leafhoppers gave a single predominant genotype in each case, with a number of other variants present as well. Looking further, we discovered that even quite different viruses could in fact trans-replicate each other: that is, the Rep/RepA complex of one virus could facilitate the replication of the genome of a virus differing by up to 35% in DNA sequence. We have also – we think – made nonsense of the old fancy that you could observe “host adaptation” of field isolates of MSV: we believe this was due to repeated selection by a single host genotype from the “cloud” of viruses transmitted during the natural infection cycle.

So, insight number three:

there is a survival benefit for the viruses in this strategy.

This is simple to understand, really, and relates to leafhopper biology as well as to host: the insects move around a lot, chasing juicy grasses, and it would be an obvious advantage to the streak virus complex to be able to replicate as a complex in each different host type – given that different virus genotypes have differential replication potential in the various backgrounds. This is quite significantly different, incidentally, to what happens with the very distantly-related (in terms of geological time) begomoviruses, or whitefly-transmitted geminiviruses: these typically do not trans-replicate each other across a gap of more than 10% of sequence difference.

Boring, I hear you say, but wait…. Add another factoid in, and profound insights start to emerge. In recent years, the cloud of protégés or virologist complex around me has accumulated to critical mass, and one of the most important things to emerge – apart from some frighteningly effective software for assessing recombination in viral genomes, which I wish he’d charge for – was Darren Martin’s finding that genome recombination is rife among African streak viruses. This was unexpected, given the expectation that DNA viruses simply don’t do that sort of thing; that promiscuous reassortment of components between genomes is a hallmark of RNA viruses. Makes sense in retrospect (an exact science), however, because of the constraints on DNA genomes: how else to explore sequence space, if the proof-reading is too good? And if you travel in a complex anyway…why not swap bits for biological advantage?

MSV web

Linkage map of the MSV genome, showing what interacts with what

So Darren swapped a whole lot of bits, in a tour-de-force of molecular virology, to create some 54 infectious chimaeric MSV genomes – and determined that

The pathogenicity of chimeras was strongly influenced by the relatedness of their parental viruses and evidence was found of nucleotide sequence-dependent interactions between both coding and intergenic regions“.

In other words –new insight:

the whole genome is a pathogenicity determinant, and bits of it interact with other bits in unexpected ways.

At this point you could say “Hey, all his insights are in fact hypotheses!” – and you would be partially correct, except for

Profound Insight No. 1hypotheses are the refuge of the linear-thinking.

Or its variant, found on my office wall:

“**c* the hypotheses, let’s just discover something”. I also have

“If at first you don’t succeed, destroy all evidence that you tried” and a number of exotic beer bottle labels on my wall – but I digress….

As an aside here, I am quite serious in disliking hypothesis-driven science: I think it is a irredeemably reductionist approach, which does not easily allow for Big Picture overviews, and which closes out many promising avenues of investigation or even of thought. And I teach people how to formulate them so they can get grants and publications in later life, but I still think HDS is a tyranny that should be actively subverted wherever possible.

Be all this as it may, now follows

Profound Insight No. 2genome components may still be individually mobile even when covalently linked.

Now take a moment to think on this: recombination allows genes to swap around inside genetic backgrounds so as to constitute novel entities – and the “evolutionary value of exchanging a genome fragment is constrained by the number of ways in which the fragment interacts with the rest of the genome*“. Whether or not the genome is RNA, DNA, in one piece or divided. All of a sudden, the concept of a “virus genome” as a gene pool rather than a unitary thing becomes obvious – and so does the reductionism inherent in saying “this single DNA/RNA sequence is a virus”.

So try this on for size for a brand-new working definition of a virus – and

Profound Insight No. 3a virus is an infectious acellular entity composed of compatible genomic components derived from a pool of genetic elements.

Sufficiently paradigm-shifting for you? Compare it to more classical definitions – yes, including one by AJ Cann, Esq. – and see how much simpler it is. It also opens up the possibility that ANY virus as currently recognised is simply an operational assembly of components, and not necessarily the final article at all.

Again, my favourite organisms supply good object examples: the begomoviruses – whitefly-transmitted geminiviruses –

  • may have one- or two-component genomes;
  • some of the singleton A-type components may pick up a B-type in certain circumstances;
  • some doubletons may lose their B without apparent effect in model hosts;
  • some A components may apparently share B components in natural infections;
  • the A and B components recombine like rabbits with cognate molecules (or Bs can pick up the intergenic region from As);
  • in many cases have one or more satellite ssDNAs (β DNA, or nanovirus-related components) associated with disease causation;

…and so on, and on…. An important thing to note here is the lab-rat viruses – those isolated early on, and kept in model plant species in greenhouses – often don’t exhibit any of these strangenesses, whereas field-isolated viruses often do.

Which tells you quite a lot about model systems, doesn’t it?

But this is not only true of plant viruses: the zoo of ssDNA anello-like viruses found in humans and in animals – with several very distantly-related viruses to be found in any individual, and up to 80% of humans infected – just keeps on getting bigger and weirder. Added to the original TT virus – named originally for the initials of the Japanese patient from whom it was isolated, and in a post hoc exercise of convoluted logic, named Torque teno virus (TTV) [why don’t people who work with human or animal viruses obey ICTV rules??] – are now Torque teno minivirus (TTMV) and “small anellovirus” SAV) – all of which have generic status. And all of which may be the same thing – as in, TTVs at a genome size of 3.6–3.8 kb may give rise to TTMVs (2.8-29 kb) and SAVs (2.4-2.6 kb) as deletion mutants as part of a population cloud, where the smaller variants are trans-replicated by the larger. Thus, a whole lot of what are being described as viruses – without fulfilling Koch’s Postulates, I might point out – are probably only “hopeful monsters” existing only as part of a population. Funnily enough, this sort of thing is much better explored in the ssDNA plant virus community, given that working with plant hosts is so much easier than with human or animal.

And now we can go really wide, and attempt to be profound on a global scale: it should not have escaped your notice that the greatest degree of diversity among organisms on this planet is that of viruses, and viruses that are found in seawater in particular. There is a truly mind-boggling number of different viruses in just one ml of seawater taken from anywhere on Earth, which leads respectable authors such as Curtis Suttle to speculate that viruses almost certainly have a significant influence on not only populations of all other marine organisms, but even on the carbon balance of the world’s oceans – and therefore of the planet itself.

Which leads to the final, and most obvious,

Profound Insight (No. 4)in order to understand viruses, we should all be working on seawater…. 

That is where the diversity is, after all; that is where the gene pool that gave rise to all viruses came from originally – and who knows what else is being

Hypolith – cyanobacteria-derived, probably – under a piece of Namib quartzite from near Gobabeb Research Station

cooked up down there?

And this is the major update: not only have I managed to get funded for a project on “Marine Viromics” from our local National Research Foundation – a process akin to winning the lottery, and about as likely to succeed – I am also collaborating with friends and colleagues from the Institute for Microbial Biotechnology and Metagenomics at the University of the Western Cape on viruses in desert soils, and associated with hypoliths– or algal growths found under quartzite rocks in extreme environments.

Thus, I shall soon be frantically learning how to deal with colossal amounts of sequence data, and worse, learning how to make sense of it.  We should have fun!


* And as a final curiosity, I find that while I – in common with the World Book Encyclop[a]edia and Learning Resources – take:mol|e|chism or mol|e|cism «MOL uh KIHZ uhm», noun. to mean any virus, viewed as an infective agent possessing the characteristics of both a living microorganism and a nonliving molecule; organule.
[molechism < mole(cule) + ch(emical) + (organ)ism; molecism < molec(ule) + (organ)ism] –
There is another meaning… something to do with sacrifice of children and burning in hellfire eternally. This is just to reassure you that this is not that.

The largest marine virus yet

13 November, 2010

This is another welcome guest post from Gillian de Villiers, a Scientific Officer in our Vaccine Group.  This was presented as a Journal Club article recently, and fit so well into my continuing theme of “viral diversity from water” that I asked her to write it up.  Thanks Gillian!

Giant virus with a remarkable complement of genes infects marine zooplankton

Matthias G. Fischer, Michael J. Allen, William H. Wilson, and Curtis A. Suttle

PNAS published ahead of print October 25 2010

This publication covers the sequencing of the genome of Cafeteria roenbergensis virus(CroV).  This nucleocytoplasmic large DNA virus (NCLDV) is the largest marine virus described to date, and its closest relative is Acanthamoeba polyphaga Mimivirus.

Among the questions raised in this paper are:

  • what is the evolutionary origin of big viruses?
  • Did they get their genes from horizontal gene transfer (including from eukaryotes), or
  • are the “eukaryotic” genes viral in origin?

Spoiler alert: the authors do not answer this question.

Please note: this is a virus from a seawater host.  It is the largest marine virus yet found, but how hard has anyone been looking?  This ties in with Ed’s theme that we should be looking for viral diversity and interesting things in the water, because interesting things have been found there.

Some background…

This lytic virus strain was isolated off the coast of Texas in the 1990s.  The host, Cafeteria roenbergensis was originally misidentified as a Bodo species.  It is a major micro flagellate grazer (microzooplanton = major ocean predator) a 2-6um “bicoecid heterokant phagotrophic flagellate” and has been found in multiple marine environments including surface waters, deep sea sediment and hydrothermal vents.

In other words, the host is an extremely significant part of the ocean ecosystem, and has been found in most places.  The authors note that protists host the largest viruses known and that other giant viruses probably are widespread in the oceans, but so far only the Acanthamoeba-infecting giant viruses have been characterised (Acanthamoeba does not live in the ocean). Viral infections of cyanobacteria play a significant role in global oxygen production; in a similar way the viral infections by CroV may have implications for carbon and other nutrient cycling and the “food chain” in the oceans, although this is beyond the scope of the article.]


The genome is the second-largest viral genome described and at 730kb is very AT rich.  Approximately 618kb is thought to be coding with 544 predicted protein-coding genes.  At least 274 genes are expressed during infection.  22 percent of CroV CDSs (coding sequences) were probably best related to eukaryotic genes.  Most CroV CDSs had unknown function, but 32% of CDSs could be assigned a putative function.

For enzymatic functions that have not previously been reported in any other viruses you can refer to Table S1 of the Supplemental materials.

This is similar to CroV’s closest known relative, Mimivirus, where of 911 predicted genes only 300 were assigned a predicted function (see table).  Only 1/3 of their genes are common to these two viruses!  This suggests tremendous diversity within the nucleocytoplasmic large DNA viruses, as they may have common evolutionary origins for some genes, but not for others.  As viruses are not monophyletic (although the NCLDVs may be) and can be considered to be bags of protein that contain genetic material and share a strategy (rather than an origin) this may not be particularly surprising.  But I find it amazing that so many potential genes, and so many unique potential genes, have been found in these organisms.

Included in the genes assigned function are genes involved in translation.  CroV encodes an isoleucyl-tRNA synthetase and putative homologs of eukaryotic translation initiation factors.  22 tRNA genes and two putative tRNA-modifying enzymes: tRNA pseudouridine 5S synthase and tRNAIle lysidine synthetase were found.  Mimivirus also has four tRNA synthetases and several putative translation factors.

Cafeteria roenbergensis virus Acanthamoeba polyphaga Mimivirus
~730kb dsDNA genome ~1200kb dsDNA genome
300nm capsid 500-750nm capsid (publications differ)
Largest marine virus yet described Largest virus yet described
Second-largest virus yet described
544 predicted genes 911 predicted genes
174 genes with predicted function 300 genes with predicted function
Host: Cafeteria roenbergensis Host: Acanthamoeba castellani (amoeba)
Habitat: marine environment Habitat: soil (?freshwater)
Genes shared with Mimivirus ~ 1/3 Genes shared with CroV ~ 1/5

Similarly to other large DNA viruses a number of DNA repair genes were found.  This includes a base excision repair pathway that appears complete.  In addition crov115’s gene product is predicted to be a CPD class 1 photolyase, the first viral homologue in its class.  Crov149 appears to be part of a recently described photolyase/cytochrome group found in several bacterial phyla and euryarchaeotes, but not among established types of photolyase.  The authors suggest that the only eukaryote with this gene, Paramecium tetraurelia may have acquired it by horizontal gene transfer from a giant virus

CroV also has transcription-related genes including eight DNA-dependent RNA polymerase II subunits, six transcription factors involved in transcription initiation, elongation, and termination, a tri-functional mRNA capping enzyme, a poly (A) polymerase, as well as helicases.  Mimivirus provides considerably more genes for protein transcription and translation than most viruses, and sets up its own ‘virus factory’ in the cytoplasm of the cell.  It is possible that CroV has a similar strategy, with viral gene transcription independent of the host and occurring in the cytoplasm.

Of the three DNA topoisomerases, two are very similar to the counterparts in Mimivirus.  CroV TopoIB is the first viral homolog of the eukaryotic subfamily, but the Mimivirus TopoIB appears to be from the bacterial group.  Although the evolutionary origin appears to differ, the topoisomerases are presumably important in transcription, translation or packaging of giant virus genomes, as they appear in both CroV and Mimivirus genomes.

CroV has four inteins: self-splicing proteins.  They are found in DNA-dependent DNA polymerase B (PolB), TopoIIA, DNA-dependent RNA polymerase II subunit 2 (RPB2) and the large subunit of ribonucleotide reductase (RNR).  Inteins have previously been found in viruses infecting eukaryotes, including Mimivirus PolB.  CroV TopoIIA intein is the first case of an intein in a DNA topoisomerase gene.

Microarray analysis on the 12-18 hr infection cycle showed around half the predicted genes, and 63% of the tested genes were expressed during infection.  Work on Mimivirus and PBSC-1 showed transcription of nearly all predicted genes, so this work may underestimate the true transcriptional activity of CroV.  CroV gene expression has an early phase 0-3 hrs after infection affecting 150 genes, and a late phase affecting 124 genes 6 hrs or later post-infection including all the structural components predicted.  A conserved early promoter motif “AAAAATTGA” was identified in 35% of CDSs and is nearly identical to the Mimivirus early promoter motif “AAAATTGA”.  A promoter element for genes transcribed during the late phase of CroV infection was found that is unrelated to the putative late promoter motif in Mimivirus.

A genomic fragment involved in carbohydrate metabolism was also found.  This 38kb fragment includes enzymes for biosynthesis of 3-deoxy-D-manno-octulosonate (KDO).  This is part of the lipopolysaccharide layer in gram-negative bacteria and is found in the green alga Chlorella and the cell wall of higher plants. Ten of the enzymes involved in carbohydrate metabolism were expressed, suggesting a role in viral glycoprotein biosynthesis, suggesting the virion surface may be coated with KDO- or sialic acid-like glycoconjugates. 

There are no homologs in Mimivirus suggesting this region must have been acquired after the CroV and Mimivirus lineages split (or that the Mimivirus lineage lost it subsequently?).  This may have been acquired from bacteria, however GC content is even lower than for the rest of the CroV genome, and a number of the proteins are phylogenetically between bacterial and eukaryotic homologs.

Phylogenetics and Speculations

Phylogenetic reconstruction of NCLDV members. Redrawn and simplified from Fig. 4. The unrooted Bayesian Inference tree was generated from a 263-aa alignment of conserved regions of DNA polymerase B

CroV is an addition to the group of NCLDVs including Ascoviridae, Asfarviridae, Iridoviridae, Mimiviridae, Phycodnaviridae, Poxviridae and Marseillevirus, which are presumed to be monophyletic. CroV seems to be the closest known relative to Mimivirus although it is substantially smaller.  The topology of the NCLDV tree strongly suggests the five largest viral genomes (all mimiviruses) are more closely related to each other than to other NCLDV families.  They may have originated from an ancestral virus that was already an NCLDV that encoded more than 150 proteins.

Mimivirus is the most studied NCLDV, and is the largest.  Most Mimivirus genes have no cellular homologs and may be very ancient, with 1/3 of genes having originated through gene and genome duplication and less than 15% of the genes having potentially been acquired by horizontal gene transfer from eukaryotes and bacteria.  The CroV genome analysis is consistent with this view of giant virus evolution, with gene duplication and lineage-specific expansion contributing to the size of the CroV genome.  The 38kb carbohydrate metabolism fragment may be a potential case of large-scale horizontal gene transfer from a bacterium.  The PolB gene of CroV has high similarity with those of other marine isolates so it may represent a major group of marine viruses, that despite being virtually unknown have ecological significance.

CroV again shows overlap between large viruses and cellular life forms, adding to questions about the evolutionary history of giant viruses as well as what life itself is.