Principles of Virology. Jane Flint

      DISCUSSION

       Remarkable architectural relationships among viruses with double-stranded DNA genomes

      Viruses with double-stranded DNA genomes are currently classified by the International Committee on the Taxonomy of Viruses into 31 families on the basis of the criteria described in Chapter 1. As might be expected, these viruses exhibit different morphologies and infect diverse organisms representing all three domains of life. They span a large size range, with genomes from a few kilobase pairs (members of the Polyomaviridae) to >2,500 kbp (Pandoravirus). Nevertheless, consideration of structural properties indicates that these very disparate virus families in fact represent a limited number of architectural types.

      Structural information is now available for the major capsid proteins of representatives of some two-thirds of families of known double-stranded DNA viruses. Based on the fold of the proteins, most of these families can be assigned to one of just five structural classes. It is noteworthy that the two most common major capsid protein folds, the double β-barrel jelly roll and the HK97-like, are found in viruses that infect Bacteria, Archaea, and Eukarya (including mammals), as summarized in the figure.

      The small number of building blocks seen in the major capsid proteins of these viruses might indicate convergent evolution, the compatibility of only a tiny fraction of the >1,400 distinct protein folds described to date with assembly of an infectious virus particle. However, viruses that infect hosts as divergent as bacteria and humans share more than the architectural elements of their major capsid proteins. This property is exemplified by the bacteriophage PRD1 and human adenoviruses, in which the major structural unit comprises a trimer of monomers each with two jelly roll domains and hence exhibiting pseudohexagonal symmetry. These icosahedral capsids also share a structural unit built from different proteins at the positions of fivefold symmetry, from which project proteins that attach to the host cell receptors; features of their linear double-stranded DNA genomes, such as the presence of inverted terminal repetitions; and mechanisms of viral DNA synthesis. Extensive similarities in morphology and the mechanisms of particle assembly and active genome packaging are also shared by tailed, double-stranded DNA viruses that infect bacteria, e.g., phage T4, and herpesviruses. It is therefore difficult to escape the conclusion that these modern viruses evolved from ancient common ancestors (see also Volume II, Chapter 10).

       Abrescia NG, Bamford DH, Grimes JM, Stuart DI. 2012. Structure unifies the viral universe. Annu Rev Biochem 81:795–822.

       Benson SD, Bamford JK, Bamford DH, Burnett RM. 1999. Viral evolution revealed by bacteriophage PRD1 and human adenovirus coat protein structures. Cell 98:825–833.

       Koonin EV, Krupovic M, Yutin N. 2015. Evolution of double-stranded DNA viruses of eukaryotes: from bacteriophages to transposons to giant viruses. Ann N Y Acad Sci 1341:10–24.

      The simplified evolutionary tree shows just some of the branches within each domain of life, with archaeal, bacterial, and eukaryote hosts of viruses described in this chapter indicated. Viruses with major capsid proteins with the double jelly roll and HK97-like folds are listed in red and blue, respectively. STIV, sulfolobus turreted icosahedral virus.

Simian virus 40: an alternative icosahedral design. The capsids of the small DNA polyomaviruses simian virus 40 and mouse polyomavirus, ~50 nm in diameter, are organized according to a different design that is not based on quasiequivalent interactions. The structural unit is a pentamer of the major structural protein, VP1. The capsid is built from 72 such pentamers engaged in one of two kinds of interaction. Twelve pentamers occupy positions of fivefold rotational symmetry, in which each is surrounded by five neighbors. Each of the remaining 60 pentamers is surrounded by six neighbors at positions of sixfold rotational symmetry in the capsid (Fig. 4.14A). Consequently, the 72 pentamers of simian virus 40 occupy a number of different local environments in the capsid, because of variations in packing around the five- and sixfold axes.

      Like the three poliovirus proteins that form the capsid shell, simian virus 40 VP1 contains a large central β-barrel jelly roll domain, in this case with an N-terminal arm and a long C-terminal extension (Fig. 4.14B and C). However, the arrangement and packing of VP1 molecules bear little resemblance to the organization of poliovirus capsid proteins. The VP1 β-barrels in each pentamer project outward from the surface of the capsid to a distance of about 50 Å, in sharp contrast to those of the poliovirus capsid proteins, which tilt along the surface of the capsid shell. As a result, the surface of simian virus 40 is much more “bristly” than that of poliovirus (compare Fig. 4.13A and 4.14A). Furthermore, the VP1 molecules present in adjacent pentamers in the simian virus 40 capsid do not make extensive contacts via the surfaces of their β-barrel domains. Rather, stable interactions among pentamers are mediated by their N- and C-terminal arms. The packing of VP1 pentamers in both pentameric and hexameric arrays requires different contacts among these structural units. In fact, there are just three kinds of interpentamer contact, which are the result of alternative conformations and noncovalent interactions of the long C-terminal arms of VP1 molecules. The same capsid design is also exhibited by human papillomaviruses.

Figure 4.13 Interactions among the proteins of the poliovirus capsid. (A) Ribbon representation of the particle, with four pentamers removed from the capsid shell and VP1 in blue, VP2 in green, VP3 in red, and VP4 in yellow, as in Fig. 4.12A, and myristate chains in white. Note the large central cavity in which the RNA genome resides; the dense protein shell formed by packing of the VP1, VP2, and VP3 β-barrel domains; and the interior location of VP4, which decorates the inner surface of the capsid shell. (B) Space-filling representation of the exterior surface showing the packing of the β-barrel domains of VP1, VP2, and VP3. Interactions among the loops connecting the upper surface of the β-barrel domains of these proteins create the surface features of the virion, such as the plateaus at the fivefold axes, which are encircled by a deep cleft or canyon. The particle is also stabilized by numerous interactions among the proteins on the inner surface of the capsid. (C) These internal contacts are most extensive around the fivefold axes, where the N termini of five VP3 molecules are arranged in a tube-like, parallel β-sheet. The N termini of VP4 molecules carry chains of the fatty acid myristate, which are added to the protein posttranslationally. The lipids mediate interaction of the β-sheet formed by VP3 N termini with a second β-sheet structure, containing strands contributed by both VP4 and VP1 molecules. This internal structure is not completed until the final stages of, or after, assembly of virus particles, when proteolytic processing liberates VP2 and VP4 from their precursor, VP0. This reaction therefore stabilizes the capsid. Panels A and B were created by Jason Roberts, Doherty Institute, Melbourne, Australia.

Figure 4.14 Structural features of simian virus 40. (A) View of the simian virus 40 particle showing the organization of VP1 pentamers. One of the 12 5-coordinated pentamers is shown in purple and 10 of the 60 pentamers present in hexameric arrays are in light gray. The individual VP1 molecules in the pentamers surrounding a