Because dsRNA genomes cannot be replicated by the host replication machinery dsRNA, dsRNA viruses must carry their own polymerase. Therefore, most dsRNA viruses employ similar architecture, a polymerase complex (also called a core), for sequestering the genome from the cellular environment and for effective replication of the genome [1].

The polymerase complex constitutes the innermost component of the mature virion and it is formed from multiple copies of a small number of virally encoded structural protein and enzymes. Typical lifecycle of a dsRNA virus is exemplified in figure 1 using bacteriophage φ6, a model organism for studying assembly and replication of other dsRNA viruses.

φ6 is a triple-layered icosahedral phage infecting Pseudomonas syringae [2]. The polymerase complex is composed of four protein species (copy number in parentheses), P1 (120), P2 (12), P4 (72) and P7 (60) together with three dsRNA segments. P1 forms a dodecahedral framework of the complex (a T = 1 lattice, triangulated with dimers). Each five-fold vertex contains a P4 hexamer, which is the packaging ATPase [3].

Figure 1: Life cycle of bacteriophage φ6. The virion, containing 3 dsRNA segments S (small), M (medium), L (large) attaches to the host cell (a) and the polymerase complex penetrates the host membranes and loses both the envelope and the nucleocapsid coat (b). The polymerase complex is activated and produces transcript mRNA (c) which are subsequently translated into viral proteins (d1-d2). An empty polymerase complex (procapsid, PC) is self-assembled from proteins P1, P2, P4 and P7 (e). The icosahedral procapsid specifically packages the three single stranded RNA molecules s+, m+, and l+ (f). The packaged ssRNA (g) is replicated by viral RNA polymerase P2 into the double stranded form inside the polymerase complex (h). The polymerase complex is coated by a shell of protein P8 to form a nucleocapsid (i), which is subsequently enveloped and mature virions (j) leave the cell by lysis.

It is assumed that one P2 monomer (the RNA-dependent RNA polymerase) is present at each five-fold vertex of PC. Similarly, the stoichiometry of P7 (30 dimers) indicates that one dimer is associated with each of the two-fold symmetry positions. In neither case the localisation has been confirmed experimentally. The intrinsically low contrast of proteins within assemblies and the breakdown of strict icosahedral symmetry within the procapsid made it difficult to localise P2 and P7 by electron microscopy (EM).

In order to circumvent the shortcomings of electron microscopy (EM) in localising minor capsid components, we have resorted to small-angle neutron scattering (SANS) and contrast variation to determine the position of P2 and P7. Selective deuterium labelling has been facilitated by the φ6 in vitro assembly system [4]. The neutron scattering experiments were performed at the high flux beamline D22 of ILL. Three sets of procapsid samples were measured, unlabelled (PC), PC containing deuterated P2 (designated PC-P2-d) or P7 (designated PC-P7-d), respectively (figure 2).

Figure 2: Model of P2 and P7 location. Left: Surface representation of the procapsid electron densities [3]. Right: Vertices shown in a cross-section through the procapsid EM density with radial positions of P2 and P7 indicated by the white and yellow circles, respectively. Using the SANS constraints five P2 atomic models were placed within the EM electron densities under the five-fold vertices in the cross-section.

The deuterated sub-units were modelled as a collection of spheres. Based on the crystallographic structure of P2 [5] and stoichiometric considerations one P2 (modelled as a sphere of 30 Å radius) was positioned on each of the twelve icosahedral 5-fold axes. The parameter of the model was the distance of the sub-units from the center of the procapsid. The 60 P7 sub-units were modelled as 30 dimers, each dimer consisting of a linear string of 10 beads (string length 200 Å, bead diameter 10 Å) with the string axes perpendicular to the 2-fold symmetry axes; i.e. P7 dimers were positioned along the dodecahedral edges of the P1 framework. The fitted parameter was the distance of the string center from the procapsid center. The scattering from unlabelled procapsid gives a smeared shell model whose parameters are in very good agreement with the procapsid radius determined by cryo-EM (diameter 460 Å) [3]. The fit of the P2 model to the data collected from the procapsids with deuterated P2 also showed good correspondence. The distance of P2 from the procapsid center was 110 Å. A sphere of radius 110 Å intersects the EM densities on the inner surface of the procapsid at positions just under the five-fold vertex (figure 2).

Thus, the most likely position of P2 is under the five-fold vertex in contact with the P1 framework and in close proximity to the packaging machinery (e.g. P4 hexamer). A similar arrangement has been seen within the rotavirus core [6] and inferred from the low-resolution X-ray models of BTV [7]. This indicates that the position of the polymerase machinery is conserved across a wide range of dsRNA virus species and is dictated by similar genomic organisation and replication mode.
The P7 model yielded an average radial distance of 160 Å, but rather noisy data makes this estimate less reliable. Nevertheless, the present estimate of radial position indicates that P7 might be located inside the particle at a site that is distal to the replication and packaging machinery. P7 may act as an internal stabilising clamp of the dodecahedral framework as recently suggested [4]. These studies demonstrate the unique ability of neutron small-angle scattering combined with deuterium labelling to locate in situ the presence of specific sub-units within a multi-protein complex.