High throughput imaging of influenza using super-resolution microscopy A) Schematic of the labelling protocol. Virus samples were dried directly onto glass coverslips pre-coated with poly-L-lysine before being fixed, permeabilised and stained with antibodies using a standard immunofluorescence protocol. B) A representative field of view (FOV) of a widefield image of labelled A/Udorn/72 influenza, imaged in the green channel. Scale bar 10 μm. C) A representative FOV of a widefield image of a virus negative sample, imaged in the green channel. Scale bar 10 μm. D) The corresponding dSTORM image of the FOV in B), where HA is labelled in green and NA is labelled in red. Scale bar 10 μm. E-G) Zoomed in images from D) showing individual filaments and spherical particles. Scale bar 5 μm.

Super-resolution imaging and size analysis of SARS-CoV-2 virus particles


Unfortunately, we have become familiar with the burden of viral diseases over the past two years, with coronavirus disease 2019 (COVID-19) claiming the lives of millions of people worldwide.

Even though influenza is not usually in the headlines, it is not uncommon for outbreaks of the virus to result in over 650,000 deaths per year – or even more when irregular non-annual epidemics are taken into account.

It is common for viruses of the same species to show significant differences in shape and size – pleomorphism. For example, in influenza, viral particles can be seen as anything from spheres 100nm in diameter to long, outstretched filaments that can reach several microns across.

Other viruses known to exhibit pleomorphism include Ebola, measles, and Newcastle Disease Virus. Unfortunately, this pleomorphism is poorly understood and presents a barrier to further understanding the structure and morphology of the viruses.

Now, researchers from the University of Oxford and the University of Warwick have collaborated to develop a method to better understand and evaluate these pleiomorphisms. The researchers’ work is currently available on the bioRxiv* preprint server while awaiting peer review.

Influenza particles tend to form into one of three shapes – spheres of around 120nm diameter, ellipsoidal, capsular, or kidney-bean shaped bacilliform around 120-250nm in diameter, or filaments over 250nm in length. These are surrounded by lipid bilayers that contain embedded surface protein spikes, around 375 in total. These are mostly hemagglutinin (HA), but approximately one-seventh are neuraminidase (NA). The NA spikes tend to form clusters, often at the opposite end of the pole to the viral genome. Beneath this surface layer lies a matrix of structural proteins, and beneath this lies the genome. Most of this information was gathered through electron microscopy – fluorescence microscopy does not usually have the resolution to detect most viruses, making detailed viewing of individual particles near impossible.

High throughput imaging of influenza using super-resolution microscopy A) Schematic of the labelling protocol. Virus samples were dried directly onto glass coverslips pre-coated with poly-L-lysine before being fixed, permeabilised and stained with antibodies using a standard immunofluorescence protocol. B) A representative field of view (FOV) of a widefield image of labelled A/Udorn/72 influenza, imaged in the green channel. Scale bar 10 μm. C) A representative FOV of a widefield image of a virus negative sample, imaged in the green channel. Scale bar 10 μm. D) The corresponding dSTORM image of the FOV in B), where HA is labelled in green and NA is labelled in red. Scale bar 10 μm. E-G) Zoomed in images from D) showing individual filaments and spherical particles. Scale bar 5 μm.

High throughput imaging of influenza using super-resolution microscopy A) Schematic of the labeling protocol. Virus samples were dried directly onto glass coverslips pre-coated with poly-L-lysine before being fixed, permeabilized and stained with antibodies using a standard immunofluorescence protocol. B) A representative field of view (FOV) of a widefield image of labeled A/Udorn/72 influenza, imaged in the green channel. Scale bar 10 μm. C) A representative FOV of a widefield image of a virus negative sample, imaged in the green channel. Scale bar 10 μm. D) The corresponding dSTORM image of the FOV in B), where HA is labeled in green and NA is labeled in red. Scale bar 10 μm. E-G) Zoomed in images from D) showing individual filaments and spherical particles. Scale bar 5 μm.

Unfortunately, electron microscopy is expensive and slow. The researchers have created a technique to promote high-throughput imaging of filamentous virions by combining direct stochastic optical reconstruction microscopy (dSTORM) – a method with a resolution of under 20nm – and rapid automated analysis software to allow thousands of virions to be identified and analyzed at the same time.

The scientists focused on an already well-characterized strain of influenza-  influenza A/Udorn/72, which displays both spherical and filamentous phenotypes.

By coating the glass with positively charged linear polymers poly-L-lysine or chitosan, they were able to immobilize virus particles on the slides. Virus particles were dried directly onto coverslips coated with poly-L-lysine and imaged them using a widefield total internal reflection fluorescence microscope. Ten thousand frames of a single field of view (FOV) were taken to generate a high-resolution image of a large number of both spherical and filament particles of different lengths. These images showed a seemingly random population of filament particles with remarkable degrees of variability, ranging between 250nm to several microns long. Some viral filaments were larger than the diffraction limit used and required wide-field images with diffraction-limited signals.

The virus was labeled with an anti-HA antibody that allowed a rapid automated analysis pipeline to measure the length of these filaments.

After adjusting the images and excluding any particles smaller than 234nm (as these would be below the diffraction limit), the researchers used software to fill in the gaps of virus particles with areas not labeled, and took the simplest shape possible.

The lengths of these shapes were assumed to be the lengths of the virus particles. Nearly 500 individual FOVs were imaged, and over 46,000 filamentous particles were measured.

Analysis revealed the vast majority of filaments were under 1000nm long. This is supported by previous studies showing an increased risk of filaments breaking at longer lengths. dSTORM and a DBSCAN clustering algorithm were used to analyze smaller molecules, and surface protein organization was investigated through intensity analysis and Fourier transforms.

Super-resolution imaging and size analysis of SARS-CoV-2 virus particles. A) A representative super-resolution image of SARS-CoV-2 virions dual-labelled with anti-spike and anti-nucleoprotein (N) primary antibodies and secondary antibodies labelled with Alexa647 (red) and Alexa546 (green) respectively. Scale bar 10 µm. B&C) Zoomed images of individual SARS-CoV-2 particles (highlighted in the white boxes in A). Scale bar 100 nm. D) Super-resolution localisations in the green channel (labelling the N protein) were clustered and each cluster fitted with an ellipse to extract particle dimensions. A histogram of the major axis lengths fitted with a Gaussian function shows that virions fall into a single population, centred at 143.8nm. E) Histogram of the major/minor axis ratio shows a single distribution.

Super-resolution imaging and size analysis of SARS-CoV-2 virus particles. A) A representative super-resolution image of SARS-CoV-2 virions dual-labeled with anti-spike and anti-nucleoprotein (N) primary antibodies and secondary antibodies labeled with Alexa647 (red) and Alexa546 (green) respectively. Scale bar 10 µm. B&C) Zoomed images of individual SARS-CoV-2 particles (highlighted in the white boxes in A). Scale bar 100 nm. D) Super-resolution localisations in the green channel (labeling the N protein) were clustered and each cluster was fitted with an ellipse to extract particle dimensions. A histogram of the major axis lengths fitted with a Gaussian function shows that virions fall into a single population, centred at 143.8nm. E) Histogram of the major/minor axis ratio shows a single distribution.

The scientists highlight the use that the aforementioned techniques have in analyzing viral particles quickly and en masse while maintaining a high resolution. To further show the utility of their method, they rapidly analyzed the already-well-characterized severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus to significant effect. This method could be invaluable in assisting with urgent work such as virus purification by identifying the number of filamentous virions or quickly identifying the size and morphology in order to speed up vaccine production.

*Important notice

bioRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behaviour, or treated as established information



https://www.news-medical.net/news/20210928/Super-resolution-imaging-and-size-analysis-of-SARS-CoV-2-virus-particles.aspx