Figure 1. Structure of adenovirus particles. (a) Surface view of the Ad2 virion obtained by cryoelectron microscopy and image reconstruction. The fibres project some 33 nm, but only the portions proximal to the virion surface are seen, because these structures are bent. This view, which is orientated along an icosahedral axis of 3-fold rotational symmetry, has a nominal resolution of 30 Å. From Stewart PL et al. (1991) Cell 67: 145–154, with permission. Copyright © 1991 Cell Press. (b) Schematic section through the virion, illustrating the locations of the virion proteins and DNA genome. The organization of the capsid proteins shown is based on biochemical and structural studies. However, protein VIII has not been localized, even though it is known to be internal, and is shown associated with hexons because it is believed to stabilize the capsid. In the core, the viral DNA is shown associated with protein VII (dashed lines), the major DNA-binding protein of the virion. There have been reports of organization of the DNA into spherical domains, but in the absence of more detailed structural information, no specific structure is shown for the core. Adapted from Flint SJ et al. (2000) Principles of Virology: Molecular Biology, Pathogenesis and Control, ASM Press, Washington, DC, with permission.

Figure 2. Organization of the human adenovirus 2 (Ad2) genome. The linear double-stranded DNA genome is represented by the pair of solid horizontal lines in the centre of the figure. The terminal protein (TP) that is covalently linked to the 5¢ end of each strand and the adjacent sequence required for initiation of viral DNA synthesis (Ori) are indicated. The origins are included within an inverted terminal repeat sequence of 202 bp present at the ends of the genome, designated A and A¢. The locations of the eight RNA polymerase II transcription units are represented by barbed arrows drawn in the direction of transcription, with the immediate early, early and late transcription (ML) units shown in pink, blue and red, respectively. Viral proteins encoded within each transcription unit are listed above or below the genome, using the unsystematic nomenclature adopted in the field (see Tables 2 and 3). Several virion proteins are synthesized as precursors that are proteolytically processed by the viral protease only following virion assembly. Such proteins are indicated by the prefix p. The five families, L1 to L5, of ML proteins, which are defined by the locations of the 3¢ ends of the mRNAs that specify them (see Figure 5), are indicated. The coding sequences for proteins that perform related functions are often organized together in the viral genome. Thus, all viral replication proteins, including the DNA polymerase (Po1), the protein primer and precursor to the TP (pTP) and the single-stranded DNA-binding protein (DBP), are encoded in the E2 transcription unit, while coding sequence for the core proteins or their precursors, pVII, V and p, lie within the L2 segment of the ML transcription unit. The positions of the virus-associated (VA) RNA genes transcribed by RNA polymerase III are indicated by the solid red arrowheads.

Figure 3. Structure and function of human adenovirus 2 (Ad2) E1A proteins. (a) Organization of E1A proteins. Alternative splicing of E1A pre-mRNA, in which introns are indicated by the caret symbols, in the infected cell nucleus produces the abundant 13S and 12S E1A mRNAs. Because splicing does not change the translational reading frame, the 289R and 243R E1A proteins translated from these mRNAs differ only in the internal sequence of 46 amino acids unique to the 289R protein. This unique sequence contains most of one of the three highly conserved regions (CR1 to 3) identified by comparison of E1A sequences of human adenoviruses. The regions of the E1A proteins necessary for binding to the cellular proteins Rb and p300 are indicated. (b) Model for countermanding Rb protein-mediated regulation of transcription by E1A proteins. The E2f proteins are sequence-specific transcriptional activators first identified by virtue of their binding to the adenoviral E2 early promoter. As indicated, these are heterodimeric proteins, comprising one member of the E2f family, which contains six known members, and a Dp protein, such as Dp1. In actively growing mammalian cells, E2f proteins are bound to hypophosphorylated Rb protein for most of the cell cycle, but free from mid-G₁ through S phase, when Rb becomes heavily phosphorylated. The Rb-E2f complexes retain DNA-binding activity and bind to E2f recognition sites in specific promoters. However, the Rb protein actively represses transcription, as indicated by the red bar (top right). The adenoviral E1A proteins made in infected (or transformed) cells bind to the Rb protein by means of a CR2 sequence and, via CR1 sequences, actively dismantle Rb-E2f complexes. Thus, the E1A protein-Rb interaction releases E2f that can then stimulate transcription (green arrow, bottom right). This mechanism would make E2f available for transcription from the viral E2 promoter and from those of cellular genes normally expressed in the late G₁ and S phases of the cell cycle. Such genes include those whose products ensure progression through the cell cycle (e.g. cyclin-dependent kinase (Cdk) 2 and cyclins E and A), carry out replication of the cellular genome or produce substrates for DNA synthesis, such as thymidine kinase and dihydrofolate reductase. Adenoviruses supply viral replication proteins, but depend on host cells for the latter enzymes. The ability of E1A proteins to abrogate the negative regulatory function of the Rb protein is required for their transforming activity. Adapted from Flint SJ et al. (2000) Principles of Virology: Molecular Biology, Pathogenesis and Control, ASM Press, Washington, DC, with permission.

Figure 4. Replication of adenoviral DNA. Assembly of a complex of the viral DNA polymerase (Po1) and pTP at the viral origins (step 1) at each end of the genome, is facilitated by the cellular, sequence-specific transcriptional activators Nf1 and Oct-1, which bind to the origins and to the viral replication proteins. Pol then catalyses covalent linkage of dCMP to a specific serine residue in pTP to provide the 3¢-OH primer needed by all known DNA polymerases (box, top left). This enzyme synthesizes viral DNA from this primer in the 5¢3¢ direction (step 2), in a reaction that requires the viral DNA-binding protein (DBP) (Table 3), which coats the displaced strand and unwinds the double-stranded template, and a cellular topoisomerase to relieve supercoiling of the DNA ahead of the replication fork and torsional stress. As the genome carries an identical origin at each end, each parental strand can be replicated by this continuous mechanism, with displacement of its complement. Such displaced strands carry the complementary, terminal sequences of the inverted terminal repetition, designated A and A¢. Their reannealing forms a short duplex stem identical to the terminus of the parental genome (step 3). Thus, origins are reformed, allowing protein priming and continuous synthesis of the parental strands initially displaced (steps 4 and 5). From Flint SJ et al. (2000) Principles of Virology: Molecular Biology, Pathogenesis and Control, ASM Press, Washington, DC, with permission.

Figure 5. Alternative processing of major late (ML) pre-mRNA. The Ad2 genome is represented by the solid horizontal lines at the top of the figure, 0-100 map units, with the site of initiation of ML transcription by RNA polymerase II indicated by the jointed arrow drawn in the direction of transcription. During the late phase of infection (a), ML transcription proceeds from this initiation site to close to the right-hand end of the genome. The large (nearly 30 kb) primary transcript shown below the genome, with the 5¢ cap designated (C), is processed into over 15 mRNAs by polyadenylation at one of five possible sites, L1 to L5 (vertical arrows), used at approximately equal frequency. The polyadenylation sites define the L1 to L5 families of 3¢ coterminal mRNAs. The tripartite leader sequence present at the 5¢ ends of all ML mRNAs is formed by splicing of three small exons, l1 to l3, and is then ligated to alternative 3¢ splice sites, as illustrated for the L1 and L3 mRNAs. During the early phase of infection (b), ML transcription terminates at multiple sites with a large region around the middle of the transcription unit, so that no sequences beyond map unit 70 are transcribed. Even though these ML transcripts contain the L1, L2 and L3 polyadenylation sites, the L1 site is preferentially utilized. As indicated, the splicing of L1 pre-mRNAs is also temporally regulated. Only the 52/55-kDa mRNA is made during the early phase, but the 3¢ splice site for the protein IIIa mRNA is also recognized during the late phase. Changes in the activities of specific, cellular polyadenylation or splicing proteins induced by viral proteins appear to effect these alterations in ML pre-mRNA processing. Adapted from Flint SJ et al. (2000) Principles of Virology: Molecular Biology, Pathogenesis and Control, ASM Press, Washington, DC, with permission.

Figure 6. Assembly of adenovirus particles. Virion proteins (Table 2) are synthesized in large quantities in the cytoplasm of infected cells and imported into the nucleus. The major structural units of the virion, hexons and pentons, then assemble from their monomeric components. Pentons are formed by self-assembly, but hexons can be assembled only with the assistance of the viral L4 100-kDa protein, which binds to hexon monomers. Two possible pathways for capsid assembly are depicted. In the sequential assembly pathway (A), hexons, pentons and capsid-stabilizing proteins self-assemble into empty capsids, which contain the L1 52/55-kDa proteins. These L1 proteins may form a scaffold for capsid assembly, and are required for encapsidation of the genome. In this mechanism, newly-synthesized viral DNA is inserted into preformed empty capsids upon recognition of a packaging sequence that lies at the left end of the genome. Core proteins enter the capsid with viral DNA to form noninfectious, immature particles (young virions). Cleavage of the precursors to the six virion proteins listed at the right by the L3 protease, produces infectious virions. The failure of an Ad5 mutant with a deletion within the packaging sequence to direct assembly of any capsid-like structures indicates that assembly of the capsid and encapsidation of the genome may be concerted reactions (pathway B). The incomplete particles shown in pathway A would then represent ‘dead-end’ products that cannot complete assembly, or artefacts of the methods of extraction of particles from infected cells. Adapted from Flint SJ et al. (2000) Principles of Virology: Molecular Biology, Pathogenesis and Control, ASM Press, Washington, DC, with permission.