The entry of PRRSV and EAV requires a low pH, suggesting that it occurs via the standard endocytic route¹⁰¹,¹³⁷,¹⁴¹ (Fig. 29.4). Clathrin heavy-chain knockdown suppressed EAV infection¹⁴¹ and electron microscopy revealed arterivirus particles contained in relatively small vesicles that appeared to be clathrin coated.¹⁰⁰,¹⁰¹

The host factors required for arterivirus entry have been studied in detail only for PRRSV (e-Fig. 29.3). Several viral and cellular players have been implicated in binding, entry, and uncoating, although their exact roles remain to be defined in more detail (for recent reviews see 203,224). Sialoadhesin (or sialic acid-binding immunoglobulin [Ig]-like lectin 1 [CD169]; 216), a macrophage-restricted membrane protein, mediates the internalization of the virus by porcine alveolar macrophages (PAMs), the primary target cells of PRRSV.⁶⁴,⁶⁶ In addition, glycosaminoglycans (heparan sulfate) on the cell surface⁵¹ and sialic acids on the virion surface⁵⁰ were implicated in the initial binding step. It is believed that the virus initially binds heparin-like molecules on the cell surface and that subsequently internalization via clathrin-mediated endocytosis is triggered by the interaction of CD169 with sialic acids on the ectodomains of the GP5/M dimer.²⁰⁴ Expression of porcine CD169 in nonsusceptible cell lines can mediate PRRSV internalization, but not disassembly and productive infection,²¹⁶ indicating that additional factors must be required for successful infection. This notion is further supported by the fact that MARC-145 cells, which are commonly used to grow PRRSV, do not express CD169 on their surface.⁶⁴ In particular CD163, a member of the scavenger receptor cysteine-rich (SRCR) family, was implicated in the early stages of PRRSV infection.²⁶ Although normally a macrophage-specific antigen, CD163 is aberrantly expressed on MARC-145 cells, possibly explaining their unique susceptibility to PRRSV infection among nonengineered cell lines. Expression of CD163 from various species rendered a variety of nonpermissive cell lines susceptible to PRRSV infection, in the absence
of detectable CD169 expression. It was postulated that in PAM, CD169 and CD163 work together, with the former serving as receptor for internalization and the latter playing a key role in virus uncoating and genome release, which are thought to occur in association with the early endosome, following its acidification²¹¹ (e-Fig. 29.3).

Which virion proteins direct the fusion between viral envelope and endosomal membrane remains one of the key questions to be addressed. A role for the minor glycoproteins in arterivirus receptor recognition and tropism has not been rigorously excluded.⁴⁴ In fact, recent data show that a chimeric PRRSV carrying E, GP2, GP3, and GP4 of EAV acquired the broad tropism for cultured cells that is typical of the latter virus.¹⁹³ These findings are in line with the previously observed phenotypes of recombinant viruses in which the GP5 or M ectodomain was replaced, which did not result in an altered tropism in cell culture.⁶⁰,²¹⁹ These studies, together with the identification of several other host factors as potential “PRRSV entry mediators”,²⁰³,²²⁴ illustrate that several questions and controversies regarding arterivirus entry remain to be addressed.

The arterivirus replication cycle (Fig. 29.4) is presumed to be entirely cytoplasmic, despite the fact that at least two viral proteins are (in part) targeted to the nucleus (see below). The
incoming genome is translated into the two large replicase polyproteins pp1a (1,727 to 2,502 amino acids) and pp1ab (3,175 to 3,959 amino acids), which comprise all functions required for viral RNA synthesis.¹³¹ Despite the relatively large 5′ nontranslated region (NTR), translation presumably initiates following “conventional” ribosomal scanning of the genomic 5′ NTR.²⁰⁶ ORF1b translation requires a −1 ribosomal frame shift (estimated efficiency of 15% to 20%) just before ORF1a translation is terminated⁵⁴ (Fig. 29.3). The ORF1a/1b overlap region contains two signals that are assumed to promote this event: a so-called “slippery” sequence, which is the actual ribosomal frame shift site, and a downstream RNA pseudoknot structure.

Following proteolytic processing of the replicase polyproteins, a complex for viral RNA synthesis is formed that generates a genome-length minus strand (or “anti-genome”), the template for genome replication. In addition, a complex transcription mechanism operates to produce complementary nested sets of sg-length minus-strand RNAs and sg mRNAs⁵⁵ (see below and Fig. 29.5). The RNA signals involved in arterivirus genome replication remain to be studied in detail. The coding regions of the genomes are flanked by 5′ and 3′ NTRs of 156 to 221 and 59 to 117 nucleotides, respectively. However, natural and synthetic defective interfering RNAs of EAV invariably require at least 300 nucleotides from both genome termini for efficient replication, indicating that replication signals extend into the coding sequences.¹³⁰,¹⁹⁸ Likewise, in the case of PRRSV, a so-called “kissing interaction” between the loop sequences of RNA hairpin structures in the 3′ NTR and the N protein gene was found to be crucial for viral RNA synthesis.²¹⁸

Using a combination of approaches, detailed RNA secondary structure models were developed for the EAV 5′ and 3′ NTRs. In the 5′ NTR, a region involved in translation, replication, and transcription,²⁰⁵ (e-Fig. 29.4B), one domain in particular was found to be crucial for sg RNA production (see below). This so-called “leader TRS [transcription-regulating sequence] hairpin” (LTH) is potentially conserved in the 5′ NTR of all arteriviruses (e-Fig 29.4C). The importance of the other structural features of the EAV 5′ NTR and, for example, their involvement in RNA–protein interactions, remains to be investigated because few of these elements are conserved in other arteriviruses.²⁰⁵ A possible exception is EAV hairpin C (termed SL2 in PRRSV; 115) that was reported to be crucial for PRRSV replication and subgenomic RNA synthesis in particular (e-Fig. 29.4C).

The 3′ NTR of the arterivirus genome does not contain obviously conserved primary sequences. For EAV, the 3′-terminal CC motif immediately upstream of the poly(A) tail plays a critical role in viral RNA synthesis.¹⁵ Furthermore, a stem-loop structure near the 3′-terminus of the EAV genome is also required for RNA synthesis¹⁴ (e-Fig. 29.5A) and its loop was implicated in an essential pseudoknot interaction with an upstream stem-loop structure residing in the N protein gene.¹⁵ This conformation was predicted to be conserved in all arteriviruses and proposed to constitute a molecular switch that could regulate the specificity or timing of viral (minus strand) RNA synthesis (e-Fig. 29.5B).

Various proteins from MA-104 cells bind to arterivirus-derived RNA sequences.⁹⁰,¹²² and the in vitro RNA-synthesizing activity of semi-purified EAV replication and transcription complexes depends on the presence of a soluble host protein of 59 to 70 kD, which remains to be identified.²¹²

One of the hallmarks of the replication cycle of arteriviruses (and other nidoviruses) is the synthesis of a 3′-co-terminal nested set of sg mRNAs (Fig. 29.5) from which the genes in the 3′ end of the genome are expressed. In the case of arteriviruses, all these genes encode structural proteins. Arterivirus sg mRNAs also have a common 5′ end, the so-called “leader sequence,” which is derived from the 5′ end of the genome.⁵⁵ This property is shared with coronaviruses, but—remarkably—not with some other nidoviruses (toroviruses and roniviruses; for reviews, see 83,146,168). Supported by the presumed common ancestry of the arterivirus and coronavirus replicase genes, leader-to-body fusion during arterivirus sg RNA synthesis was proposed to rely on a mechanism of discontinuous RNA synthesis similar to that previously proposed for coronaviruses. In both virus groups, short conserved TRSs are present at the 3′ end of the leader sequence (“leader TRS”) and at the 5′ end of each of the transcription units specifying a sg mRNA “body” (“body TRS”; reviewed in 146,168,175). The observation that arterivirus-infected cells contain a nested set of sg-length minus-strand RNAs, complementary to the sg mRNAs, is another important parallel with coronaviruses.³⁰,⁵³

With the exception of the smallest species, the arterivirus sg mRNAs are structurally polycistronic, but most of them are assumed to be functionally monocistronic. Notable exceptions are mRNAs 2 and 5 (in EAV, LDV, and PRRSV; Fig. 29.3), which are functionally bicistronic transcripts from which the partially overlapping gene sets E/GP2 and ORF5a/GP5 are expressed.⁷⁵,⁹⁴,¹⁷⁶ The mRNAs tentatively numbered 4 and 6 are thought to be used to translate the corresponding SHFV gene sets, and also mRNA2 of this virus was proposed to be functionally bicistronic.⁸¹

A substantial number of models for coronavirus and arterivirus sg mRNA synthesis have been proposed and reviewed extensively (see also Chapter 28; 146,168,175, and references therein). The detection of sg-length minus strands indicated that the discontinuous step in sg RNA synthesis likely occurs during minus-strand RNA synthesis. This concept was subsequently supported by data from biochemical and genetic studies with coronaviruses and arteriviruses and resulted in a model (Fig. 29.5; e-Fig. 29.4) in which discontinuous extension of minus-strand RNA synthesis yields sg-length minus-strand templates for sg mRNA synthesis.¹⁶⁶,¹⁶⁸

Direct proof for base-pairing between leader TRS and antibody TRS was obtained from reverse genetics studies using an EAV infectious complementary DNA (cDNA) clone.¹⁴⁷,²¹⁴ The mechanism by which the transcriptase is translocated between the body and leader TRS in the genomic template, a step that may resemble copy-choice RNA recombination,¹⁹,¹⁴⁷,²¹⁴ remains to be elucidated. Arterivirus sg RNAs are produced in nonequimolar, but relatively constant amounts, thus providing a mechanism to regulate the expression of the various structural protein genes. EAV reverse genetics studies have rigorously demonstrated that transcription depends on duplex formation between leader TRS and anti-body TRS and that—in general—the relative amount of sg mRNA correlates with the calculated stability of this duplex.¹⁴⁷,¹⁴⁸,²¹⁴ Sequences flanking the body TRS, the relative order and/or location of body TRSs in the genome, and possibly also higher order RNA structure were also shown or postulated to influence transcription.¹⁴⁵ Structural studies on the 5′-proximal part of the EAV genome²⁰⁵ placed the leader TRS in a single-stranded loop of the structure referred to as the “leader TRS hairpin” (or LTH; e-Fig. 29.4B,C), which was characterized as a critical player in transcription.²⁰⁶

At the protein level, transcription-specific functions have been attributed to several replicase subunits, in particular nonstructural protein 1 (nsp1) and nsp10, for which mutations were described that resulted in the (near) complete inactivation of sg mRNA synthesis.¹³⁸,¹⁹⁶,¹⁹⁸,²⁰⁸ EAV nsp1 controls the accumulation levels of viral genome and individual sg mRNAs in the infected cell by determining the levels at which the minus-strand templates for each of these molecules are produced.¹³⁸ An N-terminal zinc finger (ZF) domain was implicated in this function, but also other nsp1 domains appear to be important. Mutagenesis of nsp1 triggered the evolution of numerous nsp1 pseudorevertants with compensatory mutations that invariably rescued both balanced EAV mRNA accumulation and efficient virus production.¹³⁸ In the case of PRRSV, where nsp1 is internally cleaved into nsp1α and nsp1β, the ZF-containing nsp1α subunit is presumed to fulfill a similar role in transcription regulation.¹⁰²

The proteolytic maturation of the arterivirus pp1a and pp1ab replicase polyproteins involves the rapid autoproteolytic release of three or two N-terminal nsps and the subsequent cleavage of the remaining part of both polyproteins by the viral nsp4 “main protease.” The posttranslational processing of the
replicase polyproteins has been studied most extensively for EAV (see 202,248, and references therein; e-Fig. 29.6A,B) for which pp1a and pp1ab are cleaved 8 and 11 times, respectively, by three ORF1a-encoded proteinases (see below). In combination with the ORF1a/1b ribosomal frame shift, this yields 13 processing end products (named nonstructural protein [nsp] 1 to 12, including nsp7α and nsp7β; Fig. 29.3; e-Fig. 29.6E). Of these, nsp1-8 are generated from ORF1a, whereas nsp10-12 are entirely ORF1b-encoded and nsp9, due to the ribosomal frame shift consisting of a small, ORF1a-encoded N-terminal domain (identical to nsp8) and a large C-terminal part that is encoded by ORF1b and includes the viral RNA-dependent RNA polymerase (RdRp) domain. EAV reverse genetics studies with cleavage site mutants underscored the critical importance of replicase polyprotein processing for virus replication.²⁰²,²⁰⁹ The nsp3-8 region of pp1a (and likely also pp1ab) is subject to two alternative processing cascades, with the “major pathway” requiring an interaction with nsp2 as a cofactor, to mediate cleavage of the nsp4/5 site²²² (e-Fig. 29.6D).

The three EAV proteinase domains in nsp1, nsp2, and nsp4⁷⁴,¹⁷⁵,²⁴⁸ (e-Fig. 29.6) and their corresponding cleavage sites are well conserved in the other arteriviruses (Fig. 29.3). EAV nsp1 and nsp2 both contain a papain-like proteinase domain (PLP; formerly referred to as PCP or CP for [papain-like] cysteine protease) that mediates their rapid release from the polyprotein,¹⁷⁸ whereas nsp4 includes a chymotrypsin-like serine proteinase (SP), the arterivirus main proteinase.¹⁸⁰ PRRSV and LDV, in addition to having homologs of these three EAV proteinases, possess a fourth nonstructural proteinase,⁵² which mediates the rapid release of an additional N-terminal cleavage product. This PLPα possibly is a duplication of the proteinase (PLPβ) present in the C-terminal domain of EAV nsp1 and appears to have become inactivated in EAV.⁵² The sequence analysis of the SHFV nsp1 region revealed an even more complex situation, with an array of three potential PLP domains present in the 480-residue region upstream of the (predicted) nsp1/nsp2 junction. The nsp4 SP combines the His-Asp-Ser catalytic triad of classical chymotrypsin-like proteinases with the substrate specificity of the so-called 3C-like cysteine proteinases, a subgroup of chymotrypsin-like enzymes named after the picornavirus 3C proteinases. Specific residues in the substrate-binding region of the SP are assumed to determine its specificity for cleavage sites containing Glu (or sometimes Gln) as the P1 residue and mainly Gly, Ala, or Ser at the P1′ position. Nine such sites were identified in EAV pp1a/pp1ab, and they were all found to be conserved in the other family members.¹⁷⁵,²⁰²,²⁴⁸

Nsp4 structures have been obtained by x-ray crystallography for both EAV and PRRSV¹¹,¹⁹⁵ (e-Fig. 29.7). The protein consists of three domains, with domains I and II forming the typical chymotrypsin-like two-β-barrel fold of the SP. The C-terminal domain III is dispensable for proteolytic activity and may be involved in fine-tuning replicase polyprotein cleavage.²⁰⁰,²⁰¹ Recent structural studies also elucidated the structures of PRRSV nsp1α¹⁸⁶ and nsp1β,²³⁸ including their respective PLP domains (e-Fig. 29.7), which—in line with previous studies—were both confirmed to employ a Cys-His tandem as active site residues. Both PLPα and PLPβ appear to act exclusively in cis and the two structures indeed revealed the presence of the C-terminal region of the proteins in the PLP substrate-binding pocket, suggesting an intramolecular cleavage mechanism that would preclude further proteolytic reactions. Both nsp1α and nsp1β of PRRSV have also been implicated in evasion of the host’s immune response (see below), but this was most directly demonstrated for the PLP that resides in the N-terminal domain of the highly variable nsp2 subunit. This PLP2, which possesses both cis and trans cleavage activities,⁸⁶,¹⁷⁹ not only directs the critical cleavage of the nsp2/3 site in pp1a and pp1ab, but is also able to remove ubiquitin (Ub) and Ub-like modifiers like ISG15 from yet-to-be-identified substrates in the infected cell.²¹³ The protein is distantly related to the ovarian tumor domain (OTU) family of deubiquitinating enzymes.⁷⁶,¹²³,¹⁸⁷

Although accelerated by research efforts following the emergence of severe acute respiratory syndrome (SARS)-coronavirus, the functional dissection of the complex array of nidovirus nonstructural protein functions is still in its infancy. Even the arterivirus replicase polyproteins are of extraordinary size and complexity, despite their twofold smaller size compared to nidoviruses with larger genomes like coronaviruses and roniviruses. Therefore, future studies will undoubtedly reveal both novel similarities and differences between these two groups. In arteriviruses, with the notable exception of the role of nsp1 in sg mRNA synthesis (see above), the ORF1a-encoded functions mainly appear important for the regulation of replicase gene expression (by proteolytic processing; see above) and formation of the membrane-anchored “scaffold” for the replication/transcription complex. The ORF1b-encoded proteins, on the other hand, appear to be more directly involved in viral RNA synthesis.

Except for the proteins from the nsp1 region,³¹,¹⁹⁷ all replicase subunits localize to the perinuclear region of the infected cell (e-Fig. 29.8A–D; 207), where they are associated with intracellular membranes that are derived from the endoplasmic reticulum (ER). Upon arterivirus infection, these host cell membranes are modified into vesicular double-membrane structures that presumably carry the viral replication complex (e-Fig. 29.9A–F).¹⁴⁹,¹⁷⁷,²¹² The formation of closely paired membranes and double membrane vesicles (DMVs) is a typical feature of arterivirus-infected cells described many years ago.¹⁸,¹⁸⁴,²³⁴ Recent electron tomography studies of EAV-infected cells revealed that these structures are in fact interconnected and form a network of modified ER⁹⁹ (e-Fig. 29.9E). Biochemical and electron microscopy studies have implicated ORF1a-encoded subunits that contain hydrophobic, probable trans-membrane domains (in particular nsp2, nsp3, and nsp5; e-Fig. 29.6C) in the formation of these membrane structures.⁷¹,¹⁴⁹,¹⁵⁸,¹⁷⁷,²⁰⁷

Replicase ORF1b is the most conserved part of the arterivirus genome and encodes the core enzymes for viral RNA synthesis—RdRp (nsp9) and helicase (nsp10).⁵⁴,⁸³,¹⁴⁰ Recombinant EAV nsp9 is able to initiate RNA synthesis de novo in the absence of other viral or cellular proteins, but could not utilize sequences derived from the 3′ end of the viral genome as a template,¹³ suggesting additional requirements for its activity in vivo. The predicted NTP binding and superfamily 1 helicase activities of arterivirus nsp10 were corroborated by in vitro assays with recombinant nsp10. These also revealed the 5′-to-3′ polarity of the unwinding reaction, a property shared with coronaviruses,¹²,¹⁶⁹ although it has not been reconciled with the protein’s presumed role in unwinding local double-stranded
RNA structures that might hinder the RdRp during viral RNA synthesis, which proceeds in the opposite direction. As in all nidoviruses, the helicase is linked to an N-terminal zinc-binding domain that might assist the proper folding of nsp10 and/or mediate interactions of the protein with its substrate RNAs. This domain was also implicated in a remarkable transcription-specific defect.²⁰⁸,²¹⁰