Filming fluorescent Marburg virus

For some Marburg is a city in Germany. It's also the name of a virus closely related to the much more widely known Ebola virus (a name which people tend to associate with a virus as opposed to the small river it's named after). What they both have in common, beyond both being members of the Filoviridae family,  is a propensity to induce highly unpleasant, and often lethal, haemorrhgagic fevers. Marburg virus (MARV) first surfaced in 1967 in laboratory workers in Marburg and Yugoslavia and, just like Ebola, has caused sporadic cases and outbreaks since then, the most horrifying of which was in 2004-2005 in Angola, where 227 of 252 (90%) of those known to be infected died.

The worm-like form of Marburg virus particles. 

As per many viruses, much about the MARV lifecycle within the cell remains a mystery. A recent paper in PNAS used live cell imaging to dissect some of the events involved in making new viruses and how they shuttle to a point of release. 

Live cell imaging is often based upon fluorescence, so one of the first things was to make the tools. Essentially, they made versions of structural viral proteins, VP30 and VP40, that are tagged with a fluorescent molecule. To VP30 they added green fluorescent protein (GFP) and showed that when expressed from a plasmid it behaved like the untagged VP30. Similarly, they inserted a red (RFP) version of VP40 into the genome of the virus, such that wild type (wt, = unmodified) and tagged VP40 were produced. The new virus behaved similarly to the unmodified virus, at least early during infection. In infected cells, RFP-VP40 colocalised with wt VP40, implying that this modification didn't alter its localisation. Tools made.

The first step of virus production/release they looked at was the exit of nucleocapsids from the inclusions where nascent viruses are thought to assemble. When they filmed inclusions, VP30-GFP was seen to be leaving, confirming this is where nucleocapsids are assembled, but not with RFP-VP40 (despite VP40 being present at the inclusion body), leading to the conclusion that VP40 is added elsewhere.

Nucelocapsids leaving the inclusion. Individual nucleocapsids could be seen leaving (top) and of those leaving, VP30 but not VP40 was present (bottom) 

If the VP40 component of particles is being added somewhere other than the inclusion, then the obvious question to ask is 'where does VP40 get added?'. Again, they turned to microscopy. When they counted the number of nucleocapsids containing both VP40 and VP30, the number increased towards the plasma membrane, implying that it is here that VP40 associates with the nucleocapsids.

VP40 gets added near the plasma membrane: the closer a nucleocapsid is to the plasma membrane, the greater the likelihood that it also contains VP40, suggesting that it is at the cell periphery that VP40 becomes associated with nucleocapsids.

A  bonus of filming an infection is that there is an additional parameter, i.e. time. This means that the speed at which things happen can be worked out. In this case the authors were able to work out that the nucleocapsids moved at up to 500 nm/sec. On top of that, they were able to figure out that the movement was quicker towards the centre of the cell, as opposed to more remote regions, possibly because nucelocapsids are using different motor proteins in different regions to surf the cytoskeleton. But which component of the cytoskeleton? Two approaches were used. First, they filmed VP30-GFP labelled nucleocapsids in cells with either red tubulin or red actin: only in the case of actin was the movement consistent with riding a particular filament .

In the second approach, treating cells with nocodazole, which disrupts microtubules, had no effect whereas cytochalasin D disruption of the actin filaments brought MARV movement to a halt. In both cases actin appears to be the answer.

Lastly, they looked at the presence of the nucleocapsids in the filopodia extruded from the infected cells. From their observations, they concluded both that VP40 must be associated with the nucleocapsids, and that the motor protein Myosin 10 (Myo10) is involved in the transport of nucleocapsids in the filopodia.

This work is impressive for many reasons. most immediately obvious is the reliance upon live cell imaging. Very often there are the (reasonable) requirements for observations in the microscope to be backed up by biochemical data. It's a great example of what can be achieved via deductions made from observations in careful experiments. The difficulty in doing this work is also easy to overlook. I get the impression that doing this project in BSL-4 would be tricky. Conveniently, they had a remote controlled microscope that they could operate from a more comfortable location, something that's rather handy if you're going to film fluorescent Marburg.

Gordian Schudt, Larissa Kolesnikova, Olga Dolnik, Beate Sodeik, and Stephan Becker (2013). Live-cell imaging of Marburg virus-infected cells uncovers actin-dependent transport of nucleocapsids over long distances Proceedings of the National Academy of Science DOI: 10.1073/pnas.1307681110