Towards in cellulo virus crystallography

Abstract

Viruses are a significant threat to both human health and the economy, and there is an urgent need for novel anti-viral drugs and vaccines. High-resolution viral structures inform our understanding of the virosphere, and inspire novel therapies. Here we present a method of obtaining such structural information that avoids potentially disruptive handling, by collecting diffraction data from intact infected cells. We identify a suitable combination of cell type and virus to accumulate particles in the cells, establish a suitable time point where most cells contain virus condensates and use electron microscopy to demonstrate that these are ordered crystalline arrays of empty capsids. We then use an X-ray free electron laser to provide extremely bright illumination of sub-micron intracellular condensates of bacteriophage phiX174 inside living Escherichia coli at room temperature. We have been able to collect low resolution diffraction data. Despite the limited resolution and completeness of these initial data, due to a far from optimal experimental setup, we have used novel methodology to determine a putative space group, unit cell dimensions, particle packing and likely maturation state of the particles.

Figure 1

Schematic of the stages in the assembly and maturation of phiX174, a T = 1 bacteriophage. The procapsid (pdb 1cd3) contains proteins B, D, F, G & H and develops into the provirion by losing at least some of the B proteins and gaining protein J and DNA. Subsequent formation of the mature virion (pdb 2bpa) occurs through the loss of D proteins and the remaining B proteins.

Figure 2

Electron microscopy. (a) Ultrathin section of a E. coli C990 slyD1 cell infected with wt phiX174 at 4.5 h p.i. Scale bar shows 200 nm. (b) Two-dimensional projection image of similarly infected E. coli C990 slyD1 cell. Scale bar shows 500 nm. (c) Fourier transform of the area shown in (b). (d) Section through averaged tomogram density. Scale bar shows 100 nm.

Figure 3

Diffraction data and analysis. (a) Example of a strong diffraction pattern from phiX174 wild-type with a clear lattice. Panel shadowing was a consequence of a silicon support used to protect the detector. The black streak was observed on many patterns, although at different angles and is presumably derived from the beam reflecting from the jet edge. (b) Histogram of vector distances illustrating selection process of correct cubic space group. Blue fill peaks correspond to combined data from wild type and mutant data sets. Patterned peaks show predicted frequency of vector distances for face- (F), body- (I) and primitive- (P) centred structures. The closest match to our observed data (pale blue fill, with blue outline) is the face-centred cubic space group (green stars).