Viral Protein and Nucleic Acid Synthesis
Up to this point in the replication process, the virus particle has been somewhat passive as no biosynthetic activity directed by the viral genome has occurred. The preliminary steps of the infection process have placed the viral genome in position to take active control of the replication cycle and to remodel the cell to assist in the production of progeny virus particles. The details of the next phases of the replication cycle, which include the expression of viral proteins and replication of the viral genome, differ markedly between viruses, and play a major role in determining the evolutionary relationships between viruses and in the placement of viruses into proper taxonomic groupings. Examples of four different replication strategies will be described in succeeding pages in order to emphasize specific aspects of virus replication and to demonstrate the diversity of replication strategies.
Representative Examples of Virus Replication Strategies
Picornaviruses
The family Picornaviridae includes a number of important pathogens of animals and humans, for example poliovirus, hepatitis A virus, and foot and mouth disease virus (see Chapter 26: Picornaviridae). Picornaviruses are small, relatively simple, nonenveloped viruses. The virus particle has an icosahedral symmetry and consists of a protein capsid and a genome comprised of a single strand of positive sense RNA. The genomic RNA contains a small virus-encoded protein (VPg) covalently bound to its 5’ end and a genetically encoded 3’ poly A tail. The picornavirus entry process differs depending on the specific virus. The capsid of some picornaviruses (eg, poliovirus) undergoes conformational changes at the plasma membrane in response to receptor binding, and these changes are thought to create a transmembrane protein pore through which the virus genome is extruded from the virion into the cytoplasm of the host cell (as described above, Fig. 2.5). Other picornaviruses, such as foot and mouth disease virus, enter cells via receptormediated endocytosis and release their genome into the cytoplasm following conformational changes induced by the acidification of the endosome. Regardless of the mechanism used, the entry process results in release of the genomic RNA into the cytoplasm of the host cell where it will be used as a template for protein synthesis and for the replication of new viral genomes as depicted in Fig. 2.8. Shortly after the genomic RNA is released into the cytoplasm, the VPg protein is removed from the RNA by a cellular enzyme that normally functions in the repair of cellular DNA. Following the removal of VPg, the RNA associates with the cellular translational system and is used as a template for synthesis of the viral proteins. However, unlike most host-cell mRNAs, the picornavirus genomic RNA lacks a standard 5’ cap structure which is normally required to initiate the assembly of a ribosome onto the mRNA template (cap-dependent translation). Therefore, picornaviruses have had to evolve a mechanism for assembling host cell ribosomes onto viral mRNAs in the absence of a 5’ cap (cap-independent translation). This function is provided by RNA sequences located near the 5’ end of the genomic RNA itself. In picornaviruses, the AUG codon that is used to initiate translation is located an unusually long distance from the 5’ end of the RNA (743 nt in the case of poliovirus). The long nontranslated region between the 5’ end and the AUG start site assumes multiple secondary and tertiary structures due to extensive intramolecular base pairing. The majority of this region is referred to as the internal ribosome entry site (IRES) based on its ability to interact with cellular components of the translational machinery and to assemble ribosomes internally on the RNA a short distance upstream of the start codon. As virus replication proceeds, translation of cellular proteins decreases markedly as ribosomes are assembled almost exclusively onto viral mRNAs. The restriction of cellular protein synthesis is due to the cleavage and subsequent inactivation of the translation initiation factor eIF4G by a virus-encoded protease (designated L protease or 2A protease depending on the virus). eIF4G is required for cap-dependent translation and in its absence ribosomes are not assembled on capped mRNA. Translation of viral mRNA is not affected by cleavage of eIF4G as these mRNAs lack a cap and ribosomes are assembled internally on viral mRNAs by the IRES. The selective inhibition of cellular translation reduces competition for ribosomes and reduces the ability of the cells to produce an array of antiviral molecules such as type I interferons that are made in response to the viral infection (see Chapter 4: Antiviral Immunity and Virus Vaccines).
The genomic RNA of picornaviruses includes only a single open reading frame that is translated into a single large polyprotein that is subsequently cleaved by virusencoded proteases (which are embedded within the polyprotein) into the individual structural and nonstructural proteins of the virus. Intermediate cleavage products are designated P1, P2, and P3 (Fig. 2.8). Proteins that are used to assemble the capsid (VP1, VP2, VP3, and VP4) are ultimately derived from P1. Proteins required for genome replication and interference with host cell processes are ultimately derived from P2 and P3. The input genomic RNA will be translated repeatedly to generate virus proteins but eventually it will be used as a template for replication. Replication of the picornavirus RNA is performed in close association with remodeled cellular membranes and requires most of the proteins derived from the P2 and P3 precursor proteins as well as several cellular proteins. The host cell does not provide an RNAdependent RNA polymerase (RdRp) enzyme capable of replicating the viral RNA genome; and therefore, picornaviruses (and nearly all other RNA viruses) have evolved their own RdRp enzyme for this purpose. Picornaviruses encode an RdRp enzyme called 3Dpol, which is derived from the P3 precursor protein. 3Dpol is a primerdependent polymerase and the primer that is used in the replication process is the VPg protein itself. A tyrosine residue within VPg donates the hydroxyl group onto which two uridine nucleosides are added by 3Dpol to form VPgUU2OH. The addition of the uridine nucleosides is templated by two adenosine nucleosides located in the non-base paired region of a RNA stem loop structure located internally on the genomic RNA. This stem loop structure is referred to as the cis-acting replication element (CRE). The actual sequence and internal location of CRE varies among different picornaviruses but all contain two or more adjacent adenosine residues within their loop structure which serve as the template for the uridylylation of VPg. Following its synthesis, the VPgUU2OH primer is translocated to the terminal sequences of the 3’ poly A tract where it is hybridized to the RNA through A:U base pairing. The VPgUU2OH primer is then extended by 3Dpol to form a full-length complementary negative strand. The negative strand terminates in at least two adenosine residues, which facilitates base pairing with another VPgUU2OH primer and the synthesis of positive strand RNAs. Many of the newly synthesized positive strand RNAs will be used as mRNAs following the enzymatic removal of VPg. Other positive strand RNAs will retain VPg and be packaged into progeny virions. The capsid structure of picornaviruses consists of multiple copies of a structural subunit called the protomer. The protomer of most picornaviruses contains single copies of the structural proteins VP1, VP3, VP0 (a precursor to VP2 and VP4), each of which is derived from P1. Sixty protomers associate through noncovalent interactions to form the icosahedral capsid. The exact mechanism by which the RNA genome is incorporated into the developing capsid remains unclear, but two primary models have been proposed. The first model proposes that individual protomers assemble on a genomic RNA and incorporation of the genome occurs coincident with the capsid assembly process. The second model proposes that protomers interact in the absence of RNA to form empty capsid structures into which the genomic RNA is then somehow inserted. In both models, the final step of capsid maturation involves cleavage of VP0 into VP2 and VP4 by what is believed to be an autoproteolytic process. The ratelimiting process for particle maturation appears to be the availability of VPg-containing RNA. All steps of picornavirus virion assembly occur intracellularly, and late in infection crystalline arrays of virus particles form in the cytoplasm of infected cells. Ultimately, these virus particles are released from the cell en mass following dissolution of the cell structure. The replication cycle of picornaviruses illustrates several properties that are common to many positive strand RNA-based viruses. First, the RNA genome is infectious, meaning that the genomic RNA itself is capable of initiating a productive infection when introduced into a host cell in the absence of any viral proteins. Second, the positive sense genomic RNA is able to associate with ribosomes and serve as a template for the production of viral proteins which then perform the processes of replicating the viral RNA and of manipulating critical host-cell metabolic and defense-related processes. Third, viral proteins can be synthesized as larger precursor proteins (polyproteins) that are subsequently resolved into the individual structural and nonstructural proteins by virus-encoded proteases. Finally, these viruses often induce the remodeling of cellular membrane structures that provide sites for viral RNA synthesis.
Rhabdoviruses
Vesicular stomatitis virus is the prototypical member of the Rhabdoviridae family (see Chapter 18: Rhabdoviridae), and the following description of rhabdovirus replication is based on the replication cycle of this virus (Fig. 2.9). Rhabdoviruses are enveloped viruses that have a distinctive bullet-shaped morphology. The rhabdovirus genome consists of a single strand of negative sense RNA. Unlike the genomic RNA of picornaviruses, the genomic RNA of rhabdoviruses is not naked, but instead exists as a nucleocapsid consisting of an RNA complexed throughout its length with repeating copies of the nucleocapsid (N) protein (1 N protein:9 nt of RNA). Infection of a host cell is initiated by attachment of the virus glycoproteins (G) to receptors expressed on the plasma membrane, and cell entry via receptor-mediated endocytosis. Decreasing pH within the endosomal vesicle induces conformational changes in the G proteins, which in turn mediate fusion of the viral envelope with the endosomal membrane. Membrane fusion results in the release of the helical nucleocapsid into the cytoplasm. In contrast to picornaviruses, the genomic RNA of rhabdoviruses cannot serve as a template for protein synthesis. Consequently, the first biosynthetic process initiated following release of the nucleocapsid is transcription of the genomic RNA into translatable mRNAs. Positive strand RNA viruses such as the picornaviruses do not package their RdRp enzyme as a structural component of the virus particle as their genome can be readily translated to produce the enzyme components soon after entry into the cytoplasm. In contrast, negative strand RNA viruses such as the rhabdoviruses do package their RdRp enzyme within the virus particle because the synthesis of viral proteins cannot proceed until the viral genome has been transcribed into mRNAs, and no host cell enzyme capable of performing this function is available in the cytoplasm. The rhabdovirus RdRp enzyme is a multisubunit complex consisting of the large (L) protein, which possesses the catalytic activity of the complex, and the phosphoprotein (P) which functions as an essential, but noncatalytic cofactor. The RdRp complex enters the cytoplasm as a component of the nucleocapsid. The genetic organization of the genomic RNA is highly conserved among the different rhabdoviruses. The 3’-terminal sequences encode for a short nontranslated RNA (“leader”), followed by the coding sequences for 5 genes in the order of N, P, matrix (M), G, and L, and concludes with the 5’-terminal sequences that encode a short, nontranslated “trailer” RNA. Each of these sequences is separated from adjacent sequences by a short, highly conserved intergenic region that plays an important role in transcription as explained below. The ability of the RdRp enzyme to utilize the genomic RNA as a template for transcription or replication is dependent upon two critical parameters. First, the RdRp complex can only access the genomic RNA via the highly conserved 3’-terminal sequences, thus the transcription and replication processes only initiate at this site. Second, the RdRp can only access and utilize viral RNA that is complexed with N protein; naked RNA cannot serve as a functional template for any viral process mediated by the RdRp. Transcription of the viral nucleocapsid results in the synthesis of a series of capped, polyadenylated monocistronic mRNAs, which is achieved as follows. Transcription is initiated at the 3’ end and continues until the RdRp enzyme enters the first intergenic region. Each intergenic region contains a sequence that signals the end of transcription of the upstream gene and a sequence that signals the start of transcription of the downstream gene. After the RdRp encounters the first intergenic region it stops transcribing and releases the short leader RNA. The RdRp then scans to the next transcription start signal and begins transcription of the first gene (N gene). In addition to functioning as an RdRp, the L protein has capping activity and will synthesize a methylated 5’ cap structure on the nacent mRNA. When the RdRp encounters the next intergenic region it will pause over a short poly-uridine tract and through a process involving iterative slippage, will use these residues repetitively to synthesize a long poly-adenosine sequence before releasing the mRNA (now containing both a methylated 5’ cap and a poly A tail). This process will be repeated until individual mRNAs representing each viral gene have been transcribed. Reinitiation of transcription following release of an mRNA is an error prone process and some RdRp complexes detach from the template before successfully reinitiating transcription of the downstream gene. RdRp complexes that do detach from the template are unable to re-access the template at an internal site and must reinitiate transcription at the 3’ end of negative strand. This situation results in transcriptional attenuation, in which the gene nearest the 3’ end of the genome (N gene) is transcribed at the highest level, and transcription of downstream genes decreases progressively with transcription of the L gene (5’ terminal gene) occurring at the lowest level. Replication of the viral genome requires a full-length positive sense RNA that can serve as template for synthesis of the negative sense genomic RNA. Each of the viral mRNAs is of subgenomic length, thus none of these RNAs can serve this purpose. As protein synthesis progresses, a full-length plus strand (antigenome) of viral RNA is produced and this RNA is used in the process of genome replication. The switch from transcription of monocistronic mRNAs to synthesis of genome-length positive strands appears to occur once the cytoplasmic concentration of N protein reaches a critical threshold level. Viral mRNAs are devoid of protein, but the fulllength plus and minus strand RNAs are bound throughout their length by repeating copies of N protein. The N protein is an RNA binding protein and is maintained in a soluble, RNA-free form through an association with dimers of the P protein. Once a sufficient level of N protein is achieved, N protein is transferred from the soluble N/P2 complexes onto the nascent leader RNAs as soon as they are synthesized by the RdRp. Additional N proteins will continue to bind to the RNA as it is synthesized. The presence of N protein on the nascent RNA has a profound effect on RdRp function which under these conditions is unaffected by the regulatory signals of the intergenic regions and continues to synthesize a full-length positive strand. The full-length positive strand RNA (in complex with N protein) in turn serves as the template for synthesis of full-length negative strand RNAs. The newly synthesized negative strand RNAs can serve as templates for more mRNA (secondary transcription), as templates for replication, or as genomes for incorporation into progeny virions. Maturation of rhabdoviruses occurs by budding of newly forming virions through the plasma membrane of the host cell. This process requires specific interactions between components of the three major structural elements of the virus particle; specifically, the G proteincontaining envelope, the matrix, and the nucleocapsid. Budding of virus particles occurs through regions of the plasma membrane that contain a high concentration of G protein. The G proteins are synthesized in association with the rough endoplasmic reticulum and are transported to the plasma membrane through the exocytotic pathway where they become concentrated in so-called membrane microdomains. The M protein is initially synthesized as a soluble monomer but as the infection proceeds many copies of the M protein localize to the cytoplasmic side of the plasma membrane where they too assemble into M-rich membrane microdomains. Nucleocapsids interact with the M proteins in the microdomains through noncovalent interactions between N and M proteins. In the case of vesicular stomatitis virus, the M protein appears to be primarily responsible for driving the actual budding process, but this is thought to be enhanced by interactions between M proteins and the cytoplasmic portion of the G proteins. Though not detailed here, the budding process also requires functions provided by cellular proteins. In simple terms, virion budding involves the association of the internal components of the virus (nucleocapsid and matrix) with the G-rich membrane microdomains, evagination of the plasma membrane at these sites, and eventual membrane scission. Rhabdoviruses produce RNA molecules that are functional ligands for several different cellular pattern recognition receptors [PRRs, eg, retinoic acid-inducible gene 1 (RIG-I), melanoma differentiation-associated protein 5 (MDA5), and toll-like receptor 7 (TLR7)], and their recognition can stimulate a type I interferon response by the host cell (see Chapter 4: Antiviral Immunity and Virus Vaccines). For example, the leader RNA that is produced during the transcription process and the full-length genome and antigenome RNAs possess a 5’ triphosphate, and these uncapped RNAs serve as ligands for RIG-I. In addition, viral RNA of positive or negative sense can be bound by TLR7 following delivery of viral products to the endosome as occurs during autophagy. Rhabdoviruses are sensitive to the antiviral effects of type I interferons; however, like the picornaviruses, rhabdoviruses have evolved strategies for inhibiting this innate antiviral defense system of the host cell. Inhibition of the interferon system by vesicular stomatitis virus is mediated by the M protein which limits the synthesis of type I interferon and the products of interferon stimulated genes (ISGs) by globally suppressing the transcription of host cell genes and by inhibiting the export of cellular mRNAs out of the nucleus. Rabies virus has evolved an alternative strategy for interfering with the interferon response that is mediated by the P protein. The P protein interferes with the RIG-I signaling pathway which prevents activation of the type I interferon genes. In addition, rabies virus P protein inhibits the nuclear localization of phosphorylated STAT 1 and STAT 2 proteins, which limits activation of ISGs and subsequent establishment of the antiviral state [by proteins such as protein kinase R (PKR) and 2’5’ oligoadenylate synthetase (OAS) as described in detail in Chapter 4, Antiviral Immunity and Virus Vaccines].
Retroviruses
The Retroviridae family includes pathogens of both humans and animals (see Chapter 14: Retroviridae). The retrovirus particle is enveloped and contains envelopeassociated glycoprotein spikes. The spike is a multiprotein structure that consists of transmembrane (TM) subunits and surface (SU) subunits. The TM and SU subunits associate with one another to form heterodimers. Three identical TM/SU heterodimers then assemble to form the functional trimetric spike. The interior structures of the virion include a matrix that underlies the envelope and is constructed from repeating copies of the matrix protein (MA), a capsid that is constructed from repeating copies of the capsid protein (CA), and two RNA-based nucleocapsids. The RNA components of the nucleocapsids are identical and consist of single stranded, positive sense RNA (retroviruses are diploid for every virus gene). Each RNA is capped at its 5’ end, contains a poly A tail and is complexed throughout its length by multiple copies of the nucleocapsid protein (NC). Although retroviruses are technically positive strand RNA viruses, their replication cycle is markedly different from that of other positive strand RNA viruses such as the picornaviruses that were discussed earlier. The following description of the retrovirus replication cycle, and depicted in Fig. 2.10, is based on that of a simple retrovirus, and some details will not apply to all members of the Retroviridae family. Infection of a cell by a retrovirus begins with virus binding to receptors (and to coreceptors in some instances) on the host cell. In general, receptor binding is mediated by the SU component of the spike. Most retroviruses appear to enter the host cell at the plasma membrane and no change in pH is required to initiate or complete this process. However, some retroviruses appear to enter host cells via receptor-mediated endocytosis in a manner similar to that described for the rhabdoviruses. For retroviruses that enter the cell at the plasma membrane, receptor binding stimulates conformational changes in the SU subunit, which in turn induce conformational changes in the TM subunit that then mediates fusion between the virus envelope and the plasma membrane. Membrane fusion causes the loss of the envelope, disassembly of the matrix, and release of the virus core. The core consists of the capsid, the nucleocapsids and two viral enzymes [reverse transcriptase (RT) and integrase (IN)] that are required early in the infection process. Although the details of the next step in the infection process are not entirely understood, evidence suggests that the core undergoes structural changes, probably mediated by cellular proteins, and these structural changes are required to initiate the process of reverse transcription by the core-associated RT enzyme. RT is a multifunctional enzyme that possesses RNA-dependent DNA polymerase (RdDp) activity. RT is a primer-dependent polymerase and uses the viral single stranded RNA as a template to synthesize a linear, complementary double stranded DNA (cDNA) product. The primer used to initiate the synthesis of DNA is a cell-derived tRNA that is base paired to a primer binding site on the RNA. This tRNA was acquired from the cell that generated the virus particle and it enters the newly infected cell already bound to the viral RNA. The molecular details of the reverse transcription process are not presented here, but three important outcomes of the process should be noted. First, the viral RNA is degraded in the process, thus the viral gene segments, which are fully capable of functioning as mRNA, are never translated into proteins. Second, the process causes the duplication and transposition of specific viral sequences which together result in the formation of repeated sequences at the termini of the cDNA. These direct repeats are referred to as the long terminal repeats (LTRs) and they perform important replicationrelated functions as described below. Third, upon completion of the reverse transcription process the cDNA product exists as a component of a nucleoprotein complex called the preintegration complex (PIC). The complex also contains the virus protein (eg, the IN enzyme) and cellular proteins that are now poised to mediate the integration of the cDNA into a chromosome of the host cell if access to cellular DNA can be achieved. Most retroviruses are not able to transport the PIC into the nucleus, and therefore, these viruses can only integrate their cDNA into a chromosome of an actively dividing cell as cell division involves the temporary dissolution of the nuclear membrane. Retroviruses belonging to the genera Lentivirus and Spumavirus have evolved mechanisms for transporting their PIC into the nucleus that make it possible for these viruses to integrate their cDNA into chromosomes of nondividing cells. The integration reaction is initiated by IN which cleaves host cell DNA at the site selected for integration, and the process is dependent on interactions between IN and the LTR sequences of the cDNA. The final steps of the integration process are performed by cell-derived DNA repair enzymes. Integration of the cDNA into a host cell chromosome is essentially random and does not require specific host DNA sequences, but integration generally occurs within regions of a chromosome that are transcriptionally active, and consequently, more readily accessible. The integrated viral cDNA is referred to as the provirus, and establishment of the provirus must be achieved before the expression of viral genes can occur. The DNA sequences that control transcription of the viral genes are located within the LTRs. These sequences are similar to those that regulate the expression of cellular genes (eg, TATA box, binding sites for cellular transcription factors, etc.). Therefore, the LTR sequences are accessible to the transcription machinery of the host cell and viral transcripts are synthesized by cellular RNA pol II, capped by cellular capping enzymes, and polyadenylated by the cellular poly A polymerase enzyme. Transcription is initiated within the LTR that is positioned upstream of the viral genes and continues through the entire provirus sequence. A polyadenylation signal encoded by the downstream LTR is utilized for the addition of a poly A tail. Due to their sequence similarity, both LTRs are capable of initiating transcription; however, transcription activity initiated by the upstream LTR typically interferes with the initiation of transcription from the downstream LTR. This phenomenon, which is referred to as promoter occlusion, normally prevents the downstream LTR from initiating transcription of downstream cellular sequences. The single viral mRNA contains the sequence of all viral genes in the order of gag (encoding MA, CA, NC proteins, and the protease (PR) enzyme), pol (encoding the RT and IN enzymes), and env (encoding the glycoprotein precursor of SU and TM). In some retroviruses, the PR enzyme in encoded in the pol region. Depending on the particular retrovirus, the open reading frames that encode Gag, Pol and Env can be in frame with one another or out of frame, and the open reading frames that encode Gag and Pol can be continuous or overlapping. The capped and polyadenylated transcripts produced by the simple retroviruses experience one of two alternative fates. If the transcript is exported from the nucleus without being spliced it will serve as a transcript for the synthesis of the Gag polyprotein (encoding only MANCCAPR) and/or a larger GagPol polyprotein (encoding MANCCAPRRTIN). The Gag open reading frame terminates with a stop codon; and therefore, the majority of ribosomes that translate the unspliced transcript will only synthesize the Gag polyprotein. However, a small percentage of translating ribosomes will synthesize the larger GagPol polyprotein using one of two mechanisms. If the Gag and Pol open reading frames are continuous and in frame then the GagPol polyprotein can be produced if the translating ribosome reads through the Gag stop codon. This process, in which the ribosome treats the stop codon as a sense codon, is referred to as stop codon suppression. If the Gag and Pol open reading frames are overlapping and out of frame, then the GagPol polyprotein can be produced if the translating ribosome shifts from its original reading frame (the Gag reading frame) into the Pol reading frame. The ribosomal frame shift is facilitated by sequences within gag and occurs just upstream of the Pol open reading frame. Typically, the Gag and GagPol polyproteins are not resolved into their individual protein components within the cell, but instead only undergo proteolytic processing after they have been incorporated into progeny virions during the virus assembly process. This process will be described in more detail below. Alternatively, the transcript can be spliced prior to being exported to the cytoplasm. Splicing removes an intron that includes the sequences encoding Gag and Pol; and therefore, spliced transcripts only retain the Env open reading frame and can only be translated into the Env glycoprotein that serves as the precursor to SU and TM. Translation of this transcript occurs in association with the endoplasmic reticulum and the Env glycoproteins are processed and routed to the cell surface using the endoplasmic reticulum/ Golgi apparatus protein export system. The Env glycoprotein is cleaved into its SU and TM components by the host cell enzymes called furin (or by a furin-like enzyme) during its transits through the trans-Golgi or after its arrival at the cell surface. Splicing of viral transcripts is not unique to the retroviruses. For example, splicing of viral transcripts occurs during the replication of influenza A virus (Family Orthomyxoviridae), Borna disease virus (Family Bornaviridae), and of most DNA viruses. In addition to functioning as mRNA for the production of the Gag and GagPol polyproteins, the unspliced transcript can also be incorporated into progeny virions as genomic RNA. These full-length RNAs possess a packaging signal located within the Gag sequence. Packaging signals of two RNAs interact with one another, facilitating the formation of RNA:RNA dimers. The spliced viral mRNAs lack this sequence and are unable to participate in dimer formation. Similar to the rhabdoviruses, formation of the retrovirus particle generally requires specific and coordinated interactions between components of the three major structural elements of the virus particle. With respect to retroviruses these structural elements include the spike-modified membrane microdomains, the Gag and GagPol polyproteins, and the dimeric RNAs. Key interactions responsible for virion assembly and budding include those that take place between the MA-component of the polyproteins and the cytoplasmic tails of TM (and with the membrane), and those that take place between the NC-component of the polyproteins and the dimeric RNA. These interactions help to drive the budding process by which immature virions are formed. The newly budded immature virions contain the Gag and GagPol polyproteins and lack defined internal structures such as a matrix, capsid, or nucleocapsids. To this point in the process, the PR enzyme has been inactive; however, soon after formation of the immature virion, the PR enzyme is activated and proceeds to process the Gag and GagPol polyproteins into their individual constituent proteins. Once released, these proteins then assemble into the matrix, capsid, and nucleocapsid structures that are characteristic of the mature, infectious virus particle. Proteolytic processing of the GagPol polyproteins also releases RT and IN which are now available to perform the early replication events required to infect the next cell.
Adenoviruses
Adenoviruses belong to the family Adenoviridae (see Chapter 10: Adenoviridae). Unlike picornaviruses, rhabdoviruses and retroviruses, the adenovirus genome consists of DNA. The adenovirus particle is nonenveloped and consists of an icosahedral capsid that is constructed from hexon (trimers of protein II) and penton (pentamers of protein III) subunits. Prominent structures called fibers (trimers of protein IV) are associated with the penton subunits and project outward from each of the icosahedron’s 12 vertices. The adenovirus particle also has numerous proteins located internally; some of which are in contact with the penton and hexon subunits, and others that are associated with the DNA genome. The genomic DNA consists of 3036 kbp of linear dsDNA, contains terminal inverted repeat sequences that play an important role it the DNA replication process, and is covalently bound to a virus-encoded protein (terminal protein) at each 5’ end. The following description of the adenovirus replication cycle, and the representation of the process that is depicted in Fig. 2.11, is based on that of human adenovirus 2. The initial interaction of adenovirus with a host cell is mediated by the fibers that bind to a host cell protein called the coxsackievirus and adenovirus receptor (CAR). High-affinity binding between the fiber and this receptor allows the penton base proteins to make contact with cellular integrins, whose normal function is to bind the host cell to components of the extracellular matrix. This binding initiates the process of clathrin-mediated entry with subsequent internalization of the virion into clathrincoated pits, and initiates the first steps of virion uncoating. The interactions that occur between the fibers and CARs and between pentons and integrins, and perhaps other factors that are not yet adequately characterized, induce substantial changes in the capsid structure. These changes include the shedding of the fibers and externalization of a lytic factor (protein VI) from the virion interior into the endosome lumen. Protein VI mediates disruption/fragmentation of the endosomal membrane which allows the modified capsid to enter the cytoplasm. After release from the endosome the virions associate with the molecular motor dynein which then transports them along microtubules to a nuclear pore. At the nuclear pore the capsid establishes interactions with a number of host cell proteins, including the nuclear pore filament protein Nup214, kinesin-1, and histones, which further destabilize the virion structure and result in the release of the viral DNA into the nucleus. Gene expression programs of most DNA viruses are temporally regulated with specific genes being expressed at different times. The expression of adenovirus genes occurs in three phases, which are referred to as immediate early, early, and late. The adenovirus genes are arranged in sets called transcription units. Each transcription unit is controlled by a single promoter that is used by the transcriptional machinery of the host cell, and polyadenylation signals that define the 3’ ends of the viral transcripts. Each transcription unit directs the synthesis of a single primary RNA; however, alternative splicing yields a population of mRNAs that encode multiple different proteins. The first transcription unit to become transcriptionally active is E1A which encodes the immediate early proteins. The E1A primary transcripts are alternatively spliced to form transcripts that encode a family of E1A proteins. The major E1A proteins (289R and 243R) perform several critical functions that are required during the initial stages of the infection. First the E1A proteins interfere with the type I interferon response as will be discussed below. Second, they induce the host cell to enter the S phase of the cell cycle by directly interacting with the retinoblastoma (Rb) tumor suppressor protein. Adenoviruses typically infect terminally differentiated cells that are not actively dividing. By inducing the host cell to enter the S phase the virus creates a cellular environment that is more conducive to replication of the viral DNA. The major E1A proteins also activate the transcription units for the early genes, as well as activate some cellular promoters. Collectively, the proteins expressed from the early genes perform three major functions; including the inhibition of apoptosis, replication of viral DNA, and inhibition of host immune defenses. Two proteins expressed from the E1B transcription unit (E1B-19K and E1B-55K) inhibit apoptosis, which is a normal cellular response to unscheduled entry into the S phase (as induced by E1A) and to cellular stress induced by virus infection. The E1B-19K protein is a homolog of the antiapoptotic cellular protein Bcl-2. Like Bcl-2, E1B-19K binds to the pro-apoptotic protein Bax and inhibits its ability to mediate mitochondrial release of cytochrome C, which is a potent inducer of the intrinsic apoptosis pathway (see Chapter 3: Pathogenesis of Viral Infections and Diseases). The E1B-55K protein induces the rapid turnover of the tumor suppressor protein called p53 which becomes stabilized in the infected cell as a consequence of E1Amediated inactivation of Rb. Under normal conditions, stabilized p53 activates transcription of cellular genes that cause cell cycle arrest (eg, p21) and of genes such as Bax, which promote apoptosis. Three proteins expressed from the E2 transcription unit cooperate to replicate the viral DNA. The precise details of the genome replication process will not be addressed here, but the general functions of these three proteins will be described. One of these proteins is the DNA polymerase that catalyzes the replication of the DNA. The second protein is the preterminal protein (Pre-TP) which serves as a primer for DNA replication in much the same way as VPg served as a protein primer for replication of the picornavirus RNA genome, except that no cis-acting replication-like element is required. At the conclusion of the DNA replication process a Pre-TP remains covalently attached to each 5’ end of the DNA. Later in the replication process as DNA genomes are incorporated into newly assembling virions, Pre-TP is cleaved by a virus-encoded protease into a smaller form called the terminal protein (TP). The third E2 protein is the DNA binding protein (DBP) which binds to the single stranded DNA that is displaced from the dsDNA template during the replication process. After being displaced and bound throughout its length by DBP, the ssDNA serves as a template for the synthesis of a genome-length dsDNA. The final transcription unit to become active is that which controls expression of the late genes. The late genes encode the major structural proteins of the virus and nonstructural proteins that function in the virus assembly process. Late gene expression does not begin until after the onset of DNA replication and it is enhanced by a virus-encoded protein called IVa2. Transcription of the late genes is controlled by the major late promoter which defines the 5’ end of all late mRNAs. The 3’ end of late transcripts is determined by any one of 5 different polyadenylation signals that are present within the primary transcript. The use of alternative polyadenylation signals leads to the production of a nested set of five different transcripts, some of which retain one or more internal polyadenylation sites. The polyadenylated transcripts are then alternatively spliced into multiple unique transcripts, each of which is then translated into a different late protein. The late proteins are then transported to the nucleus where they participate in the virion assembly process. Unlike the picornaviruses which possess a simple icosahedral structure, the adenovirus virion is a large and complex structure. Simple self-assembly models cannot account for this degree of complexity. Accordingly, viral proteins have been identified that act as chaperones for moving structural proteins to maturation sites and others that act as scaffolds for assembling the virion subunits. A virus-encoded protease that requires DNA as a cofactor to prevent premature proteolysis participates in the maturation process by degrading scaffold proteins and cleaving precursor proteins. Late in infection, inclusion bodies composed of large crystalline arrays of newly assembled virions appear in the nucleus of the host cell. Release of progeny virions occurs following lysis of the host cell. As with other viruses, adenovirus infections are detected by microbial pattern recognition receptors (PRRs) of the host cell and infection initiates a type I interferon response (see Chapter 4: Antiviral Immunity and Virus Vaccines). However, adenoviruses actively limit the effectiveness of the response in several ways. First, the E1A proteins inhibit the activity of a cellular protein complex (hBRe1) that preferentially activates transcription of interferon stimulated genes (ISGs). Consequently, this activity of E1A greatly reduces the intracellular levels of the effector proteins that normally contribute to the interferon-induced antiviral state. Adenoviruses also express high levels of noncoding virusassociated RNAs that inhibit the activity of PKR, an ISG product that inhibits virus replication by globally suppressing protein synthesis within the cell. The majority of the virus-associated RNA exists in dsRNA form due to extensive intramolecular sequence complementarity. These virus-associated RNAs bind to the dsRNA binding site of inactive PKR but this interaction does not activate the enzyme, and PKR that has been bound by virusassociated RNA is unable to bind other dsRNAs and, therefore, remains in an inactive form. Late in infection, adenoviruses selectively inhibit the synthesis of host-cell proteins by preventing the export of cellular mRNAs from the nucleus and by blocking translation of host-cell mRNAs through modifications of key translation initiation factors. The late mRNAs encoded by the virus are exempt from these effects due to a unique sequence called the tri-partite leader that is present at the 5’ end of all adenovirus late mRNAs.
Assembly and Release
Near the end of the replication cycle the newly synthesized structural proteins and genomic molecules are assembled in a step-wise manner into new virus particles. Depending on the virus, the release of newly assembled virus particles from the host cell occurs as a separate step following the assembly process, or occurs concurrently with the assembly process. Similar to other aspects of the replication cycle, details of the virus assembly and release processes differ significantly between viruses. Based on the obvious structural differences between viruses that possess an envelope and those that lack an envelope, and on the profound effect that envelope acquisition has upon the assembly and release processes, this section of the chapter will discuss the assembly and release of nonenveloped viruses and enveloped viruses separately. The external capsid structure of virtually all nonenveloped animal viruses consists of an icosahedron, and these exist in variable levels of complexity. For the structurally simple icosashedral viruses such as those belonging to the Parvoviridae, Polyomaviridae, Papillomaviridae, and Picornaviridae families, the structural proteins spontaneously associate into the repeating structural subunit of the capsid called the protomer. A defined number of protomers are then used to assemble the mature icosahedral capsid. The protomers of some nonenveloped viruses are able to assemble into capsids in the absence of the genomic molecule (eg, canine parvovirus and human papilloma virus); for other viruses the genome appears to serve as the nucleation site for protomer assembly and capsids only form in the presence of a genomic molecule (eg, SV40, Polyomaviridae). The structure of some icosahedral viruses is more complex and does not consist of just a single type of repeating subunit. For example, the external structure of the adenovirus particle is assembled from individual hexamer subunits, pentamer subunits and trimeric fibers. Assembly of the adenovirus particle requires chaperone proteins that facilitate the proper assembly and folding of other structural components of the virus particle, and scaffold proteins that serve as temporary components of intermediate structures that are generated during the assembly process. The proteins that serve as chaperones and/or scaffolds can be displaced during the assembly process and may not remain as a structural component of the mature virion. In addition, cleavage of some virion-associated proteins is required to convert the intermediate structures into the mature, infectious form of the virus particle. Nonenveloped viruses that contain an RNA genome replicate and complete their assembly process in the cytoplasm of the host cell. Nonenveloped viruses that contain a DNA genome typically replicate and complete their assembly process in the nucleus. Most nonenveloped viruses are released only when the cell lyses, thus for these viruses the processes of virus assembly and virus release are separate and sequential events. Enveloped viruses acquire their envelope as the internal structures of the virus (eg, nucleocapsid(s) and matrix components) bud through a cellular membrane. Depending on the virus, budding can occur at the plasma membrane, through the membranes of the endoplasmic reticulum or Golgi apparatus, or through the inner membrane of the nucleus. For most enveloped viruses budding occurs at regions of the membrane in which cellular glycoproteins have for the most part been displaced by the glycoproteins of the virus (Fig. 2.12). This ensures that the viral glycoproteins are incorporated into the virion during the budding process. However, displacement of host cell glycoproteins is not an absolute requirement and some viruses (eg, human immunodeficiency and vesicular stomatitis viruses) readily incorporate host cell glycoproteins into virus particles during budding. The viral glycoproteins typically associate into oligomers (usually homotrimers or homotetramers) to form the spike (peplomer) structures. Viral glycoproteins typically consist of a hydrophilic domain projecting outward from the membrane, a hydrophobic transmembrane domain, and a short hydrophilic domain projecting into the cytoplasm (or virion interior). In general, the icosahedral nucleocapsids (eg, togviruses and flaviviruses) and helical nucleocapsids (eg, orthomyxoviruses and rhabdoviruses) of an enveloped virus assemble prior to budding. These preformed nucleocapsids then localize to the appropriate cellular membrane and participate in the budding process. In the relatively rare case of an enveloped virus with icosahedral symmetry (eg, togaviruses), each nucleocapsid protein (C protein) interacts directly with the cytoplasmic domain of a single membrane glycoprotein (E2), and these interactions help drive the budding process. For viruses that possess helical nucleocapsids, the budding process tends to be driven by interactions between the matrix proteins and the cell membrane and/or the surface glycoproteins, and between matrix proteins and the proteins of the nucleocapsid. For some viruses, the energy and forces that are required to induce curvature of the membrane and eventual membrane scission is provided by protein:protein and protein:lipid interactions mediated by virus constituents alone. However, many enveloped viruses utilize proteins of the cellular endosomal sorting complexes required for transport (ESCRT) system to assist in the budding process. One of the normal functions of the ESCRT proteins is to catalyze the budding of membrane-bound vesicles into the endosome to form multivesicular bodies. This process is physiologically similar to the budding of a virus with respect to the process being initiated on the cytoplasmic side of a membrane and the product of membrane scission (a vesicle or a virion) being formed on the extracytoplasmic side of the membrane. Depending on which proteins of the ESCRT system are involved, these proteins assist in the virus budding process itself, and/or in the final step of membrane scission which is required to release and fully envelop the virus. Viruses that acquire their envelope from an internal membrane enter the secretory pathway of the cell and are transported in exocytotic vesicles to the cell surface where they are released upon fusion of the vesicle with the plasma membrane (exocytosis) (Fig. 2.12). For these viruses the processes of virus assembly and virus release are separable and occur in sequence. Viruses that bud through the plasma membrane are released directly into the extracellular environment, thus, for these viruses the processes of virus assembly and release occur simultaneously and are essentially inseparable. Many glycoproteins encoded by enveloped viruses are synthesized as precursor proteins than are subsequently processed by site-specific proteolysis before or after being incorporated into the mature virion. This is particularly common for glycoproteins that mediate the process of membrane fusion during cell entry. Cleavage of the precursor is most commonly performed by the cellular enzyme called furin (or a furin-like protease), which is an ubiquitously expressed endoprotease that resides in the trans-Golgi compartment and at the cell surface. Furin cleaves its substrates on the carboxyl side of a BXBB sequence motif (B represents Arg or Lys and X represents a nonspecified residue). The precursor of the hemagglutinin glycoprotein of influenza A virus (HA0) is not cleaved by furin, but instead is cleaved into the functional subunits of the HA spike (HA1 and HA2) after the newly budded virion has been released from the host cell. HA0 is cleaved by trypsin-like enzymes present in secretions of the respiratory tract of humans and the gastrointestinal tract of the avian host. This finding was a major discovery in the early 1970s that allowed the routine propagation of influenza virus viruses in cell culture (in which trypsin is added to the growth medium). In general, cleavage of the precursor converts a protein that is not functional for membrane fusion, into its fusion-competent form. Cleavage of the precursor primes the virus to perform the entry and uncoating processes in response to proper stimuli (as discussed earlier), and is generally required to produce an infectious virus particle. Influenza virus also depends on a second enzymatic activity in order to be efficiently released from the host cell and to prevent aggregation of virus particles. The HA spike of influenza A virus binds sialic acid and uses this carbohydrate moiety as a receptor for attaching to host cells. However, proteins that comprise the HA and neuraminidase (NA) spikes also possess sialic acid, thus newly released virions are inclined to bind to the host cell from which they budded, and to neighboring virions. The enzymatic activity of the NA spike inhibits these nonproductive binding events by cleaving sialic acid from the cell surface and from the virion itself. The drug called oseltamivir (Tamiflu) inhibits the enzymatic activity of NA, which causes virion aggregation and restricts cell to cell spread. It should be noted that for most viruses (eg, togaviruses, herpesviruses, and retroviruses) incorporation of genomic molecules into the capsid or nucleocapsid structure is highly selective and is dependent of the presence of a highly conserved sequence called a packaging signal that is only present in the appropriate genomic RNA or DNA molecules. In contrast, other viruses (eg, rhabdoviruses and parvoviruses) are not highly selective with respect to the nucleic acid that is packaged and viral nucleic acids of both polarity (genomic and antigenomic) are packaged into virions. Unlike most other cell types in the body, epithelial cells display polarity, which means that they possess an apical surface that interfaces with the external environment (eg, lumen of respiratory tract or gastrointestinal tract) and a basolateral surface that interfaces with underlying cells. These surfaces are chemically and physiologically distinct. Viruses that are shed to the exterior (eg, influenza A virus) tend to bud from the apical plasma membrane, whereas other viruses (eg, C-type retroviruses) bud through the basolateral membrane, which may enable the virus to enter the bloodstream or lymphatic system as a prelude to establishing systemic infection (Fig. 2.13).
FIGURE 2.13 Sites of budding of various enveloped viruses. Viruses that bud from apical surfaces are in position to be shed in respiratory or genital secretions or intestinal contents. Viruses that bud from basal surfaces are in position for systemic spread via the bloodstream (ie, viremia) or the lymphatics. Some viruses, such as flaviviruses, bunyaviruses, and coronaviruses, take a more circuitous route in exiting the cell (see specific chapters in Part II). Viruses that do not bud usually are released only via cell lysis.