Introduction

Picornaviruses are ubiquitous and globally distributed, and they pose a threat to the health of humans and livestock. Picornavirus infections may induce diseases of the central nervous system, of the respiratory and gastrointestinal tracts, and of the heart, liver, pancreas, skin and eye [26, 92, 95, 98]. For most picornaviruses, the majority of infections are asymptomatic, and only a fraction of infections produce severe clinical symptoms, but due to their ubiquitous prevalence and the great number of types, hundreds of millions of infections occur annually, causing morbidity and mortality as well as substantial costs to national health care systems and economic losses [19, 25].

The family Picornaviridae is one of only two virus families that were listed in the First Report of the International Committee on Nomenclature of Viruses (ICNV) in 1971 [101]. Before that, in 1963, the Subcommittee on Nomenclature of Viruses (SCNV) had adopted the term ‘picornavirus group’ following a proposal of the International Enterovirus Study Group, then chaired by Joseph L. Melnick [35]. André Lwoff, Nobel Prize winner of 1965 and elected member of the ICNV, groused about the name ‘picornavirus’: “Pico is a prefix in the metric system meant to indicate a submultiple of a unit, namely 10 −12 , and rna stands for RNA. Thus, picorna means 10 −12 ribonucleic acid. Moreover, it is stated in the minutes of the subcommittee that the initial letters of picorna may be taken to refer to poliomyelitis, insensitivity to ether, Coxsackie, orphan, and rhinovirus. A disease, a chemical property, a virus, a state, and again a virus. This is ridiculous.” [62]. However, despite such criticism, the term ‘picornavirus’ gained general acceptance in the scientific community, and today the family Picornaviridae presents itself as a highly diverse virus family comprising 80 species divided in 35 genera [110]; https://talk.ictvonline.org/ictv-reports/ictv_online_report/positive-sense-rna-viruses/picornavirales/w/picornaviridae). In 2017, an additional 15 species within five new and three existing genera have been proposed, and more than 50 unassigned picornaviruses are awaiting classification. The fast-growing field of picornavirology may be illustrated by the following fact: as of September 2017, the nucleotide sequence database of NCBI (GenBank) registers c. 100,000 entries with the search term ‘[organism] Picornaviridae’, and this number increases by roughly 1000 entries each month. This review aims to give a very brief overview of picornaviruses and their life cycle and focuses on picornavirus diversity and the state of picornavirus taxonomy in 2017.

Picornavirus genome and capsid structure

Picornaviruses are small, non-enveloped RNA viruses [88]. Their genomes have a positive-strand polarity [32] with lengths ranging from 6.7 kilobases (seal picornavirus) to 10.1 kilobases (goose megrivirus). The 5′ end of the viral RNA is covalently linked to the tyrosine-3 residue of a small virus-encoded peptide, termed 3A or VPg (viral peptide, genome-associated) [1, 23, 54, 86]. The 3′ end has a poly(A) tail [17, 108]. The typical picornavirus genome codes for a single polyprotein [36, 41] that is co- and posttranslationally processed by virus-encoded proteinases to yield the capsid proteins [63] and a number of nonstructural proteins as well as various intermediate proteins that are also necessary for replication of viral RNA [64, 93]; reviewed in [73]. Only the recently discovered dicipiviruses have a dicistronic genome organisation with one capsid-protein-encoding open reading frame (ORF) and one ORF coding for the nonstructural proteins [102].

A first, simple model of a picornavirus, poliovirus, based on x-ray diffraction patterns was presented by [21], but the near-atomic three-dimensional structure of poliovirus at a 2.9 Å resolution did not follow until 1985 [21, 31]. Meanwhile, more than 200 crystal structures and cryoelectron microscopy structures of picornaviruses representing seven genera and 17 species have been deposited in the RCSB Protein Databank. The icosahedral T=1/pseudoT=3 capsid is composed of 60 protomers, each comprising either one copy of the four capsid proteins 1A (VP4), 1B (VP2), 1C (VP3) and 1D (VP1), or—if the 1AB precursor remains uncleaved—of 1AB (VP0), 1C and 1D [85].

A systematic nomenclature of picornavirus proteins was proposed on basis of an idealized genome layout, called the L434 diagram (Fig. 1). The L434 convention was agreed on at the third meeting of the European Study Group on the Molecular Biology of Picornaviruses [87] and refers to an idealized polyprotein and the picornaviral polyprotein processing paradigm, which summarized the knowledge of that time: the idealized polyprotein is cleaved into a leader protein (L) and three polyproteins—P1, P2, P3—in the primary processing steps. The next processing steps generate three capsid proteins from the precursor P1 (i.e., 1AB, 1C, 1D), three nonstructural proteins from the midpiece polyprotein P2 (2A, 2B, 2C), and four nonstructural proteins from P3 (3A, 3B, 3C, 3D). A final cut of 1AB yields the mature capsid proteins 1A and 1B. This occurs in some picornaviruses after RNA incorporation. The maturation cleavage is thought to be an RNA-catalyzed step [5, 29]; reviewed in [96] and induces the formation of stable, infectious virions [28]. It was further agreed that the designations VP4, VP2, VP3, VP1, VP0 and VPg for 1A, 1B, 1C, 1D, 1AB and 3B may be retained [87]. Noteworthily, the systematic nomenclature denotes the picornavirus proteins according to their order in the polyprotein, whereas the former designations of the capsid proteins, VP1 to VP4, refer to their apparent molecular weights in SDS polyacrylamide gel electrophoresis, with VP1 being the largest protein and VP4 the smallest (Fig. 1). For many of the newer picornaviruses, experimental and biological data are lacking, and only sequence data are available. In these cases, putative proteins have been proposed solely on the basis of their similarity to known picornavirus proteins. This has lead to a considerable variation in the sizes of the proposed proteins compared to their reference proteins with available biochemical data. Accordingly, designations like VP1, VP2, etc. may no longer correspond to the sizes as originally intended. Therefore, it is recommended by the Picornaviridae Study Group to prefer the systematic nomenclature. An idealized picornavirus genome layout may be described as VPg+5′UTR[(L/)1A-1B-1C-1D/2A-2B-2C/3A-3B-3C-3D]3′UTR-poly(A), with “[“ and ”]” indicating the open reading frame. The “/” symbols indicate primary cleavage sites leading to the three polypeptides P1, P2, P3, and “-” symbols indicate the final processing sites. Functions of the nonstructural proteins or characteristic sequence motifs may be presented (e.g., Lpro, 2Apro, 2Anpgp, 2AH-box/NC, 2ANPTase). If multiple copies of a protein are present, these copies are numbered (e.g., 2A1, 2A2, 2A3, etc.). If certain proteins are not present in all members of a genus, the respective genome region is presented in round brackets. There are many exceptions to the idealized genome layout, e.g., the absence of a leader protein in many genera, an uncleaved 1AB precusor (VP0), the presence of several distinct 2A proteins, or the presence of additional copies of 3B (VPg).

Fig. 1
figure 1

Schematic diagram of an idealized picornavirus genome (A) and of the dicistronic Dicipivirus genome (B). Genome lengths vary from 6.7 kb to 10.1 kb. Genomic RNA is polyadenylated and linked to a virus-encoded VPg peptide at its 5′ end. It consists of a 5′ noncoding region (NCR) up to 1450 nt (sometimes with a poly(C) tract), one (A) or two (B) open reading frames, and a 3′-NCR of variable length (25-795 nt). The 5′-NCR contains one of five known IRES types; the intergenic region (IGR) of dicipiviruses (B) is thought to contain another, yet undefined IRES type. All picornaviruses except dicipiviruses encode a single polyprotein ranging from 2050 to 2886 amino acids; dicipiviruses code for two polyproteins. Primary polyprotein processing yields the leader protein L, if present, and three polypeptides, P1, P2, P3. Further processing steps are carried out by the viral 3C proteinase (3Cpro) and generate the mature proteins 1AB, 1C, 1D, 2A, 2B, 2C, 3A, 3B, 3C, and 3D. A final processing step, the maturation cleavage of 1AB, occurs in some picornaviruses, either after assembly of empty capsids or after RNA packaging, presumably by autocatalytic mechanisms

The capsid proteins 1B, 1C, 1D as well as 2C (a helicase), 3C (a chymotrypsin-like cysteine proteinase) and 3D (a RNA-dependent RNA polymerase) are orthologous and thus conserved in all picornaviruses. Due to their marked divergence, the proteins 1A, 2B, 3A, 3B often lack detectable similarity, although there may be a functional analogy and possibly a common evolutionary origin with their cognates of other picornaviruses. At least the proteins L and 2A may have different evolutionary origins in the various picornavirus genera. The functions of the processed picornavirus proteins and some intermediates are summarized in Table 1.

Table 1 Picornavirus proteins

Eighteen of 35 picornavirus genera exhibit highly divergent L proteins, but only for members of the genera Aphthovirus, Cardiovirus and Erbovirus has their function been elucidated. The aphthovirus and erbovirus L protein is a papain-like cysteine proteinase [27] that cleaves at its C-terminus to release itself from the nascent polyprotein. In addition, the Lpro of FMDV is involved in cleavage of eukaryotic elongation factor 4G, leading to a host cell shut-off of translation [14, 43]. The cardiovirus L proteins inhibits the assembly of stress granules and nucleocytoplasmic transport [4, 7, 79]. The function of all other L protein is obscure. Various versions of 2A exist, two of which release the P1 polypeptide either by proteolytic cleavage (2Apro of Enterovirus and possibly Sapelovirus and Rabovirus) or by a translation elongation arrest followed by a re-initiation of translation at the next in-frame codon (2Anpgp of Aphthovirus, Aquamavirus, Avihepatovirus, Avisivirus, Cardiovirus, Cosavirus, Erbovirus, Hunnivirus, Kunsagivirus, Limnipivirus, Mischivirus, Mosavirus, Parechovirus B-D, Pasivirus, Potamipivirus, Rosavirus, Senecavirus, Teschovirus and Torchivirus plus several unclassified picornaviruses) [16]. A third type of 2A, 2AH-box/NC, has similarity to the H-rev107 family of proteins and the N-terminus of the viral polyprotein (pfam08405) of caliciviruses [34], but its function in the virus life cycle is unclear. The 2AH-box/NC protein is found in members of the genera Avihepatovirus, Avisivirus, Gallivirus, Kobuvirus, Megrivirus, Parechovirus, Passerivirus, Potamipivirus, Sakobuvirus, Salivirus, Sicinivirus and Tremovirus. Members of the genera Avihepatovirus and Avisivirus have a fourth type of 2A. This protein, 2ANTPase, shares similarity with the AIG1 (avrRpt2-induced gene 1)-type guanine binding domain found in P-loop NTPases; one characteristic feature is its NTP-binding motif GxxGxGKS [15]. In addition, putative 2A proteins without similarity to any known protein in the GenBank database and of unknown function have been described in members of the genera Ampivirus, Aquamavirus, Dicipivirus, Harkavirus, Hepatovirus, Kunsagivirus, Megrivirus, Oscivirus and Pasivirus.

The coding region of picornaviruses is flanked by 5′ and 3′ nontranslated regions, which give rise to RNA structures. Some of these RNA structures promote cap-independent initiation of translation [60] and—at least for enteroviruses—initiation of negative- and positive-strand RNA synthesis [2]. Five distinct types of internal ribosome entry sites (IRESs), which recruit ribosomes and other host factors to direct initiation of translation, have been shown in picornaviruses [37]. The type I IRES is found in members of the genera Enterovirus and Harkavirus, type II in members of the genera Aphthovirus, Avisivirus, Cardiovirus, Cosavirus, Erbovirus, Gallivirus, Hunnivirus, Mischivirus, Mosavirus, Parechovirus, Rabovirus, Rosavirus, and Sicinivirus, and type III in members of the genus Hepatovirus. The type IV IRES is similar to the hepatitis C virus IRES and is present in members of the genera Aquamavirus, Avihepatovirus, Kunsagivirus, Limnipivirus, Megrivirus, Pasivirus, Sakobuvirus, Sapelovirus, Senecavirus, Teschovirus and Tremovirus. Type V has been described in members of the genera Kobuvirus, Oscivirus and Salivirus. Due to its dicistronic genome, the only known member of the genus Dicipivirus presumably has two IRESs. The first one is a type II IRES, whereas the IRES type of the intergenic region is unknown. Also unknown is the IRES type of the ampiviruses, passeriviruses, potamipiviruses and torchiviruses.

Picornavirus replication cycle

Picornaviruses enter their host cells by receptor-mediated endocytosis. Various mechanisms have been described that lead to endosome formation followed by alteration of capsid structures and penetration of the membrane by the viral RNA (reviewed in [96]). In the cytoplasm, viral RNA is translated to yield one major polyprotein (all genera except Dicipivirus), which is co- and posttranslationally processed into the L protein (if present), capsid proteins (1AB, 1C, 1D), mature nonstructural proteins, and some stable intermediates, namely 2BC, 3AB and 3CD (Fig. 1). Not only the RNA-dependent RNA polymerase [3] but most if not all nonstructural proteins are necessary for replication of viral RNA. They modify the cellular environment and promote synthesis of negative-strand RNA molecules, which in turn serve as template for positive-strand RNA production. VPg molecules are uridylated at cis-replicative elements (cre) and VPg-pU-pU serves as a primer for both positive- and negative-strand RNA synthesis (reviewed in [77]). It is assumed that viral RNA synthesis occurs at a multimeric replication complex comprising several of the viral nonstructural proteins and their precursors as well as host factors. The components of the replication complex are concentrated and assembled at the surface of the viral replication organelle (RO). Presumably, the RO protects viral RNA from host nucleases and cytosolic defense mechanisms of the innate immunity. Several membrane-associated viral proteins (2B, 2C and 3A) are involved in generation of an RO of the protrusion type, which is achieved by the massive virus-induced rearrangement of intracellular membranes. This membrane rearrangement presumably involves Golgi components and autophagic mechanisms. ROs are highly dynamic structures and consist initially of a network of single-membrane vesicles, which develop into double-membrane and, at late stages, multilamellar agglomerations. Picornaviruses have developed various mechanisms to accumulate sterols, glycerophospholipids and sphingolipids at the ROs. For sterols, some mechanisms involve the viral 3A protein and phosphatidylinositol 4-kinases (PI4Ks), phosphatidylinositol 4-phosphate (PI4P), and oxysterol-binding protein (OSBP), as shown for enteroviruses, Aichi virus, encephalomyocarditis virus, and Saffold virus. Other viruses are independent of OSBP (foot-and-mouth disease virus, hepatitis A virus, human parechovirus). Virus-induced membrane contact sites (vMCSs) are formed to direct lipids from the endoplasmic reticulum to viral membranes and modulate the lipid composition of the RO membranes, which is necessary for their function (reviewed in [69, 93, 97]).

Progeny picornavirus particles are generally released by lysis; however, for hepatitis A virus, poliovirus and coxsackievirus B3, alternative, non-lytic mechanisms of viral spread have been described [6, 13, 20, 84].

Picornavirus taxonomy

A virus species is defined as “a monophyletic group of viruses whose properties can be distinguished from those of other species by multiple criteria” (International Code of Virus Classification and Nomenclature, ICVCN; https://talk.ictvonline.org/information/w/ictv-information/383/ictv-code). Accordingly, a picornavirus species contains viruses that are phylogenetically related and share a significant degree of homology of all proteins. They have essentially identical genome maps and thus are thought to have a significant degree of compatibility in proteolytic processing, replication, encapsidation and genetic recombination. Picornavirus genera—the next taxonomic level—may subsume species that cluster on the same branch in phylogenetic trees and share a significant degree of homology of the orthologous proteins P1 (divergence <0.66), 2Chel, 3Cpro and 3Dpol (divergence each < 0.64). Thus they conform to the ICVCN rule that a virus genus comprises “a group of species sharing certain common characters”. The many picornavirus genera differ by distinctive features of their genome maps and exhibit significant divergence (number of differences per site between sequences) of the orthologous proteins. The remaining proteins (L, 2A, 2B, 3A, 3B) usually lack detectable similarity. For sequence comparisons, the Picornaviridae Study Group recommends to align only the orthologous genome regions of picornavirus sequences that share significant similarity to each other. Species demarcation rules vary depending on the genus under investigation and consider the observed intra- and interspecies variability; moreover, they are based on current sequence data and may be amended when additional data are available in the future. Usually, the amino acid sequence divergence within a genus may vary from 0.4 to c. 0.65 depending on the protein. Intergenus comparisons, however, yield divergences greater 0.65. Within a species, sequence variation of the VP1 gene region may be used to define genetic types. For Enterovirus, it has been shown that genetic types correspond to the antigenic serotypes [66, 71]. However, typing is species-specific; for Enterovirus A, B, C, the border was estimated with a nucleotide sequence divergence of 0.25 (corresponding to an amino acid sequence divergence of 0.12), whereas the nucleotide sequence divergence thresholds of Rhinovirus A, B, C were 0.12-0.13 (0.095-0.1 amino acid sequence divergence). Since the available picornavirus sequence data meanwhile include isolates collected in a period of eight decades, a considerable accumulation of substitutions over time can be observed. Whether this genetic drift also affects serological properties is unknown.

In recent years, the binomial nomenclature of virus names has been adopted for all but one of the picornavirus genera. A typical picornavirus species name is composed of the genus name and the adjunct capital letters A, B or C etc. (e.g., genus Ampivirus, species Ampivirus A). Exceptions are the genera Dicipivirus, with the species Cadicivirus A, genus Enterovirus with species Enterovirus A to J and Rhinovirus A to C, genus Kobuvirus with the six species Aichivirus A to F, and genus Megrivirus with the species Melegrivirus A. Genus Aphthovirus is the only picornavirus genus that does not conform the International Committee on Taxonomy of Viruses (ICTV) nomenclature rules, as the names of the four species were derived from the host species and/or the disease they induce (Foot-and-mouth disease virus, Equine rhinitis A virus, Bovine rhinitis A virus, Bovine rhinitis B virus).

Since the invention of next-generation sequencing techniques, the majority of novel picornaviruses have been identified in metagenomics studies rather than by classical virus isolation methods and subsequent experimental characterization. Based on typical hallmarks of picornaviruses, i.e., the presence of an IRES, three capsid protein domains with drug-binding sequence motifs (so-called rhv domains), a helicase with an NTP-binding site (GxxGxGKS/T), a VPg with a tyrosine-3 residue, a chymotrypsin-like proteinase with a GxCGx10-18GxH motif, and an RNA-dependent RNA polymerase with KDE, PSG, YGDD, FLKR motifs, picornaviruses can reliably be identified solely on basis of their sequences, even if sometimes the one or other feature may be absent. Moreover, bioinformational techniques may help to identify the possible hosts of new picornaviruses. Metagenomics studies investigating the intestinal viromes of humans and various animals have revealed numerous viruses that are likely of dietary origin (e.g., [55, 57, 112]). Nucleotide composition analysis (NCA), or demonstration of characteristic sequence patterns may provide evidence of the identity of the picornavirus host. Whereas NCA can clearly distinguish between mammal, insect, and plant hosts, discrimination of mammal, avian and fish viruses is less conclusive [49, 107]. An example of a host-specific sequence pattern is the apical “8” structure of the type IV IRES of many avian picornaviruses, which is either a part of domain III (megriviruses, unclassified sapelo-like viruses) or part of a separate domain I (avihepatoviruses, aaliviruses) (reviewed in [9]). Because picornavirus sequences can be reliably identified, the Picornaviridae Study Group supports proposals that are premised on sequence data, phylogeny, divergence metrics and similarity detection. This practice conforms to the requirements of the ICTV. Correspondingly, in 2016, the Executive Committee of the ICTV endorsed a consensus statement of an expert panel that proposed a classification pipeline based on metagenomic sequence data [90].

Progress in sequencing techniques not only led to a remarkable increase in the number of known picornaviruses but has also revealed an unexpected degree of diversity. As of 2017, 35 genera and 80 species have been approved by the ICTV. More than 500 types can be distinguished genetically or by serological means. Moreover, the Picornaviridae Study Group of the ICTV has recently proposed five novel genera with six species plus nine new species within existing genera; Table 2 summarizes these picornaviruses. An additional 56 picornaviruses have been described but are yet unassigned (Table 3). In order to improve the clarity of this fast-growing family, there is a need to define subfamily criteria.

Table 2 New picornavirus genera and species proposed in 2017
Table 3 List of unassigned picornaviruses

Genus Ampivirus

Number of species: 1 (Ampivirus A, 1 type; GenBank acc. no. of reference strain or exemplar: KP770140).

Genome layout: VPg+5′UTRIRES[1AB-1C-1D/2A-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: 9246 nt (5′-UTR: 970 nt; ORF: 8142 nt; 3′-UTR: 134 nt), length of polyprotein: 2713 aa.

Genus Aphthovirus

Number of species: 4 (Bovine rhinitis A virus, 2 types; Bovine rhinitis B virus, 5 types; Equine rhinitis A virus, 1 type; Foot-and-mouth disease virus, 7 types; GenBank acc. nos. of reference strains or exemplars: KP236128, EU236594, X96870, AY593829).

Genome layout: VPg+5′UTRIRES-II[Lpro/1A-1B-1C-1D-2Anpgp/2B-2C/3A-3B(-3B2-3B3)-3C-3D]3′UTR-poly(A).

Genome length: 7250-8203 nt (5′-UTR: up to 1112 nt with poly(C) tract; ORF: 6657-7020 nt; 3′-UTR: 32-110 nt); length of polyprotein: 2218-2339 aa.

Genus Aquamavirus

Number of species: 1 (Aquamavirus A, 1 type; GenBank acc. no. of reference strain or exemplar: EU142040).

Genome layout: VPg+5′UTRIRES-IV[1AB-1C-1D-2A1npgp/2A2-2B-2C/3A-3B1-3B2-3C-3D]3′UTR-poly(A).

Genome length: c. 6700 nt (5′-UTR: 506 nt, ORF: 6154 nt, 3′-UTR: 34 nt); length of polyprotein: 2050 aa. Aquamaviruses have the shortest known picornavirus genomes and polyproteins.

Genus Avihepatovirus

Number of species: 1 (Avihepatovirus A, 3 types; GenBank acc. no. of reference strain or exemplar: DQ226541).

Genome layout: VPg+5′UTRIRES-IV[1AB-1C-1D-2A1npgp/2A2NTPase-2A3H-box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: up to 7792 nt (5′-UTR: 625-655 nt, ORF: 6750-6756 nt, 3′-UTR: c. 318 nt); length of polyprotein: 2249-2251 aa.

Genus Avisivirus

Number of species: 3 (Avisivirus A, 1 type; Avisivirus B, 1 type; Avisivirus C, 1 type; GenBank acc. nos. of reference strains or exemplars: KC465954, KF979333, KF979334).

Genome layout: VPg+5′UTRIRES-II[1AB-1C-1D-(2A1npgp)/2A2npgp/2A3NTPase-2A4H-box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: > 7167-7532 nt (5′-UTR: > 457-609; ORF: 6441-6720; 3′-UTR: 194-269 nt); length of polyprotein: 2146-2239 aa.

Genus Cardiovirus

Number of species: 3 (Cardiovirus A, 2 types; Cardiovirus B, 15 types; Cardiovirus C, 2 types; GenBank acc. nos. of reference strains or exemplars: M81861, X56019, JQ864242).

Genome layout: VPg+5′UTRIRES-II[L/1A-1B-1C-1D-2Anpgp/2B-2C/3A-3B-3C-3D]3′UTR-poly(A) Despite a significant similarity, the 2A protein of Cardiovirus C lacks the NPGP motif.

Genome length: c. 7730-8530 nt (5′-UTR: up to 1451 nt with poly(C) tract, ORF: 6879-6924 nt, 3′-UTR: 121-144 nt); length of polyprotein: 2293-2308 aa. Cardioviruses C have the longest known poly(C) tracts of picornaviruses.

Genus Cosavirus

Number of species: 5 (Cosavirus A, 24 types; Cosavirus B, 1 type; Cosavirus D, 5 types; Cosavirus E, 2 types; Cosavirus F, 1 type; GenBank acc. nos. of reference strains or exemplars: FJ438902, FJ438907, FJ438908, FJ555055, JN867758). There is one tentative cosavirus species with one type (compare Table 3).

Genome layout: VPg+5′UTRIRES-II[1A-1B-1C-1D-2Anpgp/2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 7632-7802 nt (5′-UTR: up to 1361 nt; ORF: 6366-6402 nt; 3′-UTR: 75-93 nt); length of polyprotein: 2113-2133 aa.

Genus Dicipivirus

Dicipiviruses are the only known picornaviruses with a dicistronic genome layout. Two open reading frames encoding a capsid protein precursor and a nonstructural polyprotein are separated by an intergenic region thought to have IRES function.

Number of species: 1 (Cadicivirus A, 1 type; GenBank acc. no. of reference strain or exemplar: JN819202). There is one tentative dicipivirus (compare Table 3).

Genome layout: VPg+5′UTRIRES-II[1A-1B-1C-1D]-IGRIRES[2A-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 8785 nt (5′-UTR: at least 982 nt; ORF1: 2535 nt; IGR: 588 nt; ORF2: 4221 nt; 3′-UTR: 549 nt); lengths of polyproteins: 845 aa (P1), 1407 aa (P2-P3).

Genus Enterovirus

Number of species: 13 (Enterovirus A, 25 types; Enterovirus B, 63 types; Enterovirus C, 23 types; Enterovirus D, 5 types; Enterovirus E, 5 types; Enterovirus F, 7 types; Enterovirus G, 20 types; Enterovirus H, 1 type; Enterovirus I, 1 type; Enterovirus J, 6 types; Rhinovirus A, 80 types; Rhinovirus B, 32 types; Rhinovirus C, 55 types; GenBank acc. nos. of reference strains or exemplars: AY421760, M16560, V01149, AY426531, D00214, DQ092770, AF363453, AF326759, KP345887, FJ007373, FJ445111, DQ473485, EF077279). There are two tentative species with 3 types (compare Table 2).

Genome layout: VPg+5′UTRIRES-I[1A-1B-1C-1D/2Apro-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 7100-7450 (5′-UTR: 610-822 nt; ORF: 6417-6645 nt; 3′-UTR: 37-99 nt); length of polyprotein: 2138-2214 aa. Recently, porcine enteroviruses (Enterovirus-G) have been identified in the USA and Germany that express a torovirus-like proteinase inserted between 2C and 3A [12, 46, 89], in press). The function and implication of this insertion for the viral life cycle are still unclear, ditto its prevalence.

Genus Erbovirus

Number of species: 1 (Erbovirus A, 3 types; GenBank acc. no. of reference strain or exemplar: X96871).

Genome layout: VPg+5′UTRIRES-II[Lpro/1A-1B-1C-1D-2Anpgp/2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 8828 nt (5′-UTR: up to 905 nt; ORF: 7752-7770 nt; 3′-UTR: 164 nt); length of polyprotein: 2583-2589 aa.

Genus Gallivirus

Number of species: 1 (Gallivirus A, 1 type; GenBank acc. no. of reference strain or exemplar: JQ691613). Additionally, there is one tentative type (compare Table 3).

Genome layout: VPg+5′UTRIRES-II[L/1AB-1C-1D/2AH-box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 8500 nt (5′-UTR: 761 nt; ORF: 7425 nt; 3′-UTR: 310 nt); length of polyprotein: 2474 aa.

Genus Harkavirus

Number of species: 1 (Harkavirus A, 1 type; GenBank acc. no. of reference strain or exemplar: KP230449).

Genome layout: VPg+5′UTRIRES-I[1A-1B-1C-1D/2A-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: 8003 nt (5′-UTR: 673 nt; ORF: 7263 nt; 3′-UTR: 67 nt); length of polyprotein: 2420 aa.

Genus Hepatovirus

Number of species: 9 (Hepatovirus A, 1 type; Hepatovirus B, 1 type; Hepatovirus C, 2 types; Hepatovirus D, 2 types; Hepatovirus E, 1 type; Hepatovirus F, 2 types; Hepatovirus G, 2 types; Hepatovirus H, 3 types; Hepatovirus I, 1 type; GenBank acc. nos. of reference strains or exemplars: M14707, KR703607, KT452742, KT452637, KT452735, KT452685, KT452730, KT452691, KT452658).

Genome layout: VPg+5′UTRIRES-III[1A-1B-1C-1D-2A/2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: up to 7810 nt (5′-UTR: c. >225-773 nt; ORF: 6597-6909, nt; 3′-UTR: 35-257 nt); length of polyprotein: 2199-2303.

Genus Hunnivirus

Number of species: 1 (Hunnivirus A, 6 types; GenBank acc. no. of reference strain or exemplar: JQ941880). There is one tentative species (Table 3).

Genome layout: VPg+5′UTRIRES-II[L/1A-1B-1C-1D-2Anpgp/2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: 7469-7588 nt (5′-UTR: > 600-732 nt; ORF: 7432-6822 nt; 3′-UTR: > 14-122 nt); length of polyprotein: 2244-2274 aa.

Genus Kobuvirus

Number of species: 6 (Aichivirus A, 6 types; Aichivirus B, 3 types; Aichivirus C, 2 types; Aichivirus D, 2 types; Aichivirus E, 1 type; Aichivirus F, 2 types; GenBank acc. nos. of reference strains or exemplars: AB010145, AB084788, EU787450, LC055961, KT325853, KJ641686). There are three tentative kobuviruses (Table 3).

Genome layout: VPg+5′UTRIRES-V[L/1AB-1C-1D/2AH-box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: up to 8280 nt (5′-UTR: up to 744 nt; ORF: 7299-7467 nt; 3′-UTR: 136-240 nt); length of polyprotein: 2432-2488 aa. Porcine kobuviruses (Aichivirus C) have no type V IRES but have a type IV IRES.

Genus Kunsagivirus

Number of species: 1 (Kunsagivirus A, 1 type; GenBank acc. no. of reference strain or exemplar: KC935379). There are two tentative species with 2 types (Table 2).

Genome layout: VPg+5′UTRIRES-IV[1AB-1C-1D-2A1npgp/2A2-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 7092-7429 nt (5′-UTR: 393-532 nt; ORF: 6639-6729 nt; 3′-UTR: 25-177 nt); length of polyprotein: 2213-2248 aa. Kunsagiviruses A have the shortest known 3′-UTR of picornaviruses.

Genus Limnipivirus

Number of species: 3 (Limnipivirus A, 1 type; Limnipivirus B, 1 type; Limnipivirus C, 1 type; GenBank acc. nos. of reference strains or exemplars: JX134222, KF306267, KF183915).

Genome layout: VPg+5′UTRIRES-IV[1AB-1C-1D-2A1npgp/2A2npgp/-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 7700-8050 nt (5′-UTR: > 501-712 nt; ORF: 6810-7011 nt; 3′-UTR: 312-363 nt); length of polyprotein: 2270-2337 aa.

Genus Megrivirus

Number of species: 1 (Melegrivirus A, 3 types; GenBank acc. no. of reference strain or exemplar: KF961188). A revision of megrivirus taxonomy was proposed in 2017 as sequence data indicate that all known members of Melegrivirus A1 are interspecies recombinants. There are five tentative species with 11 types (compare Table 2).

Genome layout: VPg+5′UTRIRES-IV[(L/)1AB-1C-1D/(2A0-2A1-2A2-)2A3H-box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A). The tentative megriviruses exhibit a considerable diversity of the 2A1 and 2A2 proteins. The tentative ‘Megrivirus D’ has a type II IRES.

Genome length: 9040-9840 nt (5′-UTR: > 461-663 nt; ORF: 8235-8520 nt; 3′-UTR: > 175-669 nt); length of polyprotein: 2745-2886 aa. The goose megriviruses have the longest known genomes and polyproteins of picornaviruses.

Genus Mischivirus

Number of species: 3 (Mischivirus A, 1 type; Mischivirus B, 1 type; Mischivirus C, 1 type; GenBank acc. no. of reference strain or exemplar: JQ814851, KP054273, KY512802). There is one tentative new species (Table 3).

Genome layout: VPg+5′UTRIRES-II[L/1AB-1C-1D-2Anpgp/2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 8457 nt (5′-UTR: 1407 nt; ORF: 6846 nt; 3′-UTR: 204 nt); length of polyprotein: 2282 aa.

Genus Mosavirus

Number of species: 1 (Mosavirus A, 2 types; GenBank acc. no. of reference strain or exemplar: JF973687).

Genome layout: VPg+5′UTRIRES-II[L/1A-1B-1C-1D-2Anpgp/2B-2C/3A-3B1-3B2-3C-3D]3′UTR-poly(A).

Genome length: c. 8385 nt (5′-UTR: > 646 nt; ORF: 7653 nt; 3′ UTR: 86 nt); length of polyprotein: 2551 aa.

Genus Oscivirus

Number of species: 1 (Oscivirus A, 2 types; GenBank acc. no. of reference strain or exemplar: GU182408).

Genome layout: VPg+5′UTRIRES-V[L/1AB-1C-1D/2A-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 7640-7678 nt (5′-UTR: 636-646 nt; ORF: 6765-6774 nt; 3′ UTR: 238-256 nt); length of polyprotein: 2255-2258 aa.

Genus Parechovirus

Number of species: 4 (Parechovirus A, 19 types; Parechovirus B, 5 types; Parechovirus C, 1 type; Parechovirus D, 1 type; GenBank acc. nos. of reference strain or exemplar: L02971,AF327920, HF677705, KF006989). There are two tentative parechoviruses (Table 3).

Genome layout: VPg+5′UTRIRES-II[1AB-1C-1D-(2Anpgp/)2AH-box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: 7339-7608 (5′-UTR: 710-730 nt; ORF: 6543-6753; 3′-UTR: 87-111 nt); length of polyprotein: 2180-2250 aa.

Genus Pasivirus

Number of species: 1 (Pasivirus A, 3 types; GenBank acc. no. of reference strain or exemplar: JQ316470).

Genome layout: VPg+5′UTRIRES-IV[1AB-1C-1D-2A1npgp/-2A2-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 6938 nt (5′-UTR: c. 420 nt; ORF: c. 6402 nt; 3′ UTR: 136 nt); length of polyprotein: 2133 aa.

Genus Passerivirus

Number of species: 1 (Passerivirus A, 1 type; GenBank acc. no. of reference strain or exemplar: GU182406).

Genome layout: VPg+5′UTRIRES[L/1AB-1C-1D/2AH-box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 8035 nt (partial) (5′-UTR: > 415 nt; ORF: 7287 nt; 3′ UTR: 334 nt); length of polyprotein: 2428 aa.

Genus Potamipivirus

Number of species: 1 (Potamipivirus A, 1 type; GenBank acc. no. of reference strain or exemplar: KC843627).

Genome layout: VPg+5′UTRIRES[1AB-1C-1D-2A1npgp/2A2H-box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 7496 nt (partial) (5′-UTR: > 488 nt; ORF: 6777 nt; 3′ UTR: 231 nt); length of polyprotein: 2259 aa.

Genus Rabovirus

Number of species: 1 (Rabovirus A, 2 types; GenBank acc. no. of reference strain or exemplar: KJ950883).

Genome layout: VPg+5′UTRIRES-II[L/1A-1B-1C-1D/2Apro-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: 7834 nt (5′-UTR: 751 nt; ORF: 7005 nt; 3′-UTR: 78 nt); length of polyprotein: 2335 aa.

Genus Rosavirus

Number of species: 1 (Rosavirus A, 3 types; GenBank acc. no. of reference strain or exemplar: JF973686). There are two tentative new species with 4 types (Table 3).

Genome layout: VPg+5′UTRIRES-II[1AB-1C-1D-2Anpgp/2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 8931 nt (5′-UTR: up to 828 nt; ORF: 7413 nt; 3′ UTR: up to 795 nt); length of polyprotein: 2468 aa. Rosaviruses have the longest known 3′-UTR of picornaviruses.

Genus Sakobuvirus

Number of species. 1 (Sakobuvirus A, 1 type; GenBank acc. no. of reference strain or exemplar: KF387721). There is one tentative sakobuvirus (Table 3).

Genome layout: VPg+5′UTRIRES-IV[L/1AB-1C-1D/2AH-Box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 7800 nt (5′ UTR: 591 nt; ORF: 7059 nt; 3′-UTR: 157 nt); length of polyprotein: 2352 aa.

Genus Salivirus

Number of species: 1 (Salivirus A, 2 types; GenBank acc. no. of reference strain or exemplar: GQ179640).

Genome layout: VPg+5′UTRIRES-V[L/1AB-1C-1D/2AH-Box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: 7982-8021 nt (5′-UTR: 709-763 nt; ORF: 7110-7125 nt; 3′ UTR: 148 nt); length of polyprotein: 2374 aa.

Genus Sapelovirus

Number of species: 3 (Avian sapelovirus, 1 type; Sapelovirus A, 1 type; Sapelovirus B, 3 types; GenBank acc. nos. of reference strains or exemplars: AF406813, AY064708, AY563023).

Genome layout: VPg+5′UTRIRES-IV[L/1A-1B-1C-1D/2Apro-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 7491-8226 nt (5′-UTR: up to 741 nt; ORF: 6969-7566 nt; 3′ UTR: 82-235 nt); length of polyprotein: 2322-2521 aa. A great diversity of the three species plus the availability of c. 30 sapelo-like viruses urge a major revision of this genus. Creation of several new genera is indicated.

Genus Senecavirus

Number of species: 1 (Senecavirus A, 1 type; GenBank acc. no. of reference strain or exemplar: DQ641257).

Genome layout: VPg+5′UTRIRES-IV[L/1A-1B-1C-1D-2Anpgp/2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 7310 nt (5′-UTR: 666 nt; ORF: 6546 nt; 3′-UTR: 71 nt); length of polyprotein: 2181 aa.

Genus Sicinivirus

Number of species: 1 (Sicinivirus A, 5 types; GenBank acc. no. of reference strain or exemplar: KF741227).

Genome layout: VPg+5′UTRIRES-II[L/1AB-1C-1D/2AH-box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: up to 9809 nt (5′-UTR: up to 939 nt; ORF: 8595; 3′-UTR: 308 nt); length of polyprotein: 2864 aa.

Genus Teschovirus

Number of species: 1 (Teschovirus A, 13 types; GenBank acc. no. of reference strain or exemplar: AF231769).

Genome layout: VPg+5′UTRIRES-IV[L/1A-1B-1C-1D-2Anpgp/2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: > 7110 nt (5′-end missing) (5′-UTR: >335 nt with poly(C) tract; ORF: 6600-6711; 3′-UTR: 67 nt); length of polyprotein: 2199-2236 aa.

Genus Torchivirus

Number of species: 1 (Torchivirus A, 1 type; GenBank acc. no. of reference strain or exemplar: KM873611).

Genome layout: VPg+5′UTRIRES[L/1A-1B-1C-1D-2Anpgp/2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: 7077 nt (5′-UTR: 192 nt; ORF: 6654 nt; 3′-UTR: 231 nt); length of polyprotein: 2218 aa.

Genus Tremovirus

Number of species: 1 (Tremovirus A, 1 type; GenBank acc. no. of reference strain or exemplar: AJ225173).

Genome layout: VPg+5′UTRIRES-IV[1A-1B-1C-1D/2AH-box/NC-2B-2C/3A-3B-3C-3D]3′UTR-poly(A).

Genome length: c. 7032 nt (5′-UTR: 494 nt; ORF: 6405; 3′-UTR: 135 nt); length of polyprotein: 2134 aa.

Recombination and the consequences for picornavirus taxonomy

Evidence of RNA recombination in poliovirus and foot-and-mouth disease virus was first presented in the 1960s when cells were infected with mixtures of virus variants and progeny virus showed the wild-type phenotypes [30, 53, 80]. RNA recombination occurs during negative-strand RNA synthesis. The most likely mechanism is described by the “copy choice” hypothesis, which assumes that the RNA polymerase initiates RNA synthesis on one viral RNA molecule, then switches the template to continue replication on another RNA molecule [44]. Intertypic recombination was first demonstrated by King et al. [42], but the vast extent of natural recombination was not figured out until two decades later (reviewed by Lukashev [61]).

RNA recombination occurs in all investigated picornavirus genera and seems to be a very efficient mechanism to create chimeric RNA molecules with new combinations of features, of which the advantageous ones will be selected. Analysis of the speed of adaptation in poliovirus variants with and without the ability to support recombination revealed that the accumulation of beneficial alleles was reduced while detrimental mutations were increased in recombination-defective viruses [106]. It is tempting to assume that new virulent enterovirus variants emerge more efficiently through the interplay of recombination and substitution than by the mere accumulation of exchanges over time, and it is compelling that the high frequency of picornavirus coinfections is one driving force of this mechanism of picornavirus evolution. Several reports describing salient recombinant virus strains support this view (e.g., [24, 40, 48].

Whereas intraspecies/intertypic recombination between picornaviruses is quite common, recombination between members of different species or genera seems to be less frequent. There are a few sequence features that may be interpreted as remnants of previous interspecies recombination events: (i) Whereas the coding region of EV-C109 supports grouping this virus with members of the species Enterovirus C, its 5′-noncoding region has similarity to those of members of the species Enterovirus A [109]. Clustering of 5′ noncoding sequences of pairs of enteroviruses of different species has been observed for Enterovirus A and B, Enterovirus C and D, Enterovirus E and F and Rhinovirus A and C [33, 65, 67, 111]. Accordingly, recombination in the 5′ noncoding region has been shown for Rhinovirus A and C, Enterovirus A and C and Enterovirus A and D [33], McIntyre et al. [67, 91]. It has been suggested that extensive recombination occurred during earlier stages of type diversification (McIntyre et al. [67]. (ii) EV-G strains isolated from sheep and goats exhibit the typical 5′ noncoding region of bovine enteroviruses (Enterovirus E, F) [10, 99]. (iii) Interspecies recombinants with recombination breakpoints in the P2 region have been described by Huang et al. [33] and McIntyre et al. [65] for Rhinovirus A/C recombinants and by Kapusinszky et al. [39] for a Cosavirus D/E recombinant. (iv) Genome sequencing of several megriviruses has revealed evidence of interspecies recombination and prompted the Picornaviridae Study Group to propose a revision of the megrivirus taxonomy in 2017. Available sequence data suggest that Melegrivirus A (MelV-A) comprises three types. Additional megriviruses have been detected in pigeons, chickens, geese, ducks, harriers and penguins [9, 11, 58, 78, 100]; Zell et al. unpublished). Their sequences can be grouped into a total of five species and 11 types. All strains of MelV-A1 were revealed to be recombinant, as well as one of two known goose megriviruses. In order to avoid a recombinant exemplar strain of the genus Megrivirus, the proposed revision intends to abolish the previous species Melegrivirus A. Instead, five proposed species, ‘Megrivirus A’ (MeV-A) to ‘E’, will be created, each with a nonrecombinant strain as an exemplar. The recombinant turkey hepatitis viruses are moved to the proposed species ‘Megrivirus C’, consistent with the phylogenetic relationship of their P3 region. To qualify either origin of the presumed picornavirus recombinants, a novel terminology was also proposed (A1CP-Cpol), where the relationship to both parental viruses is indicated: the capsid proteins (CP) cluster with MeV-A and the polymerase (pol) with MeV-C. Likewise, goose megrivirus 2 can be described by B3CP-Apol. (Suppl. Figs. 1 and 2).

There are also hints suggesting intergenus recombination events: (i) the presence of a type IV IRES in porcine kobuviruses (Aichivirus C) rather than a type V IRES [81] and (ii) the similarity of the hunnivirus IRES to the human parechovirus IRES while the hunnivirus polyprotein is more similar to the teschovirus polyprotein [83]. The presence of the HCV-like IRES, the 2AH-box/NC, and torovirus-like proteinase in picornaviruses [12, 46, 89], in press) suggests additional recombination events involving non-picornavirus donors.

Phylogenetic analysis as the main tool of picornavirus taxonomy

Phylogenetic analysis requires alignments of homologous nucleotide or amino acid sequences. For intergenus comparisons of picornaviruses, a reasonable degree of similarity is only detected in the orthologous proteins P1, 2Chel, 3Cpro and 3Dpol or their corresponding gene regions. Concatenated 3C-3D or 2C-3C-3D sequences yield more-robust results due to the lengths of the investigated regions and the compensatory effects of variable and conserved sequence stretches. Problems arise if picornavirus sequences are investigated without the knowledge of precise processing sites. The 3C proteinases of cardioviruses and entero-/rhinoviruses preferably use glutamine-glycine (Q/G) cleavage sites [73]. This lead to the assumption that Q/G sequences are also the cleavage sites of the 3C proteinases of other picornaviruses. However, aphthoviruses and hepatitis A virus use a number of other dipeptide sequences [73]. Due to the lack of biochemical data for many newer picornaviruses, putative processing sites have been proposed on the basis of sequence comparisons of highly divergent viruses. Such predictions have to be treated with caution until reliable experimental results are available.

Phylogenetic trees of the P1 region (Fig. 2) and the 3CD region (Fig. 3) reveal consistent arrangements of many picornavirus genera, suggesting the existence of subfamilies. For the time being, such clusters are designated as ‘supergroups’ in order to avoid an allusion to a taxonomic relevance of this observation. Five supergroups were proposed originally (Nick J. Knowles; http://www.picornaviridae.com/unassigned/unassigned.htm) and modified with the acknowledgement of new genera. Supergroup (SG) 1 comprises members of the genera Aphthovirus, Cardiovirus, Cosavirus, Erbovirus, Hunnivirus, Mischivirus, Mosavirus, Senecavirus, and Teschovirus. SG2 includes the genera Gallivirus, Kobuvirus, Oscivirus, Passerivirus, Sakobuvirus, Salivirus, and Sicinivirus. SG3 contains the genera Enterovirus, Rabovirus, and Sapelovirus. SG4 includes the genera Aquamavirus, Avihepatovirus, Avisivirus, Kunsagivirus, Limnipivirus, Pasivirus, Parechovirus, and Potamipivirus. SG5 includes the genera Hepatovirus and Tremovirus. Members of the genera Ampivirus, Dicipivirus, Harkavirus, Megrivirus, Rosavirus and a few unclassified viruses do not fit this scheme, suggesting the existence of further supergroups. As sequences of the five supergroups cluster in robust P1 and 3CD trees (Figs. 2 and 3), the supergroups are possible candidates for the creation of picornavirus subfamilies. Noteworthily, supergroups are supported by phylogenetic analysis only; other features, such as common genome layouts or the consistent presence/absence of certain IRES types, L proteins, or 2A variants, have not been identified. Whether the supergroup hypothesis is a sustainable concept has to be seen. The available sequences of several novel picornaviruses, however, fit well into this scheme, e.g., bopivirus, the unassigned lesaviruses, and porcine picornavirus Japan (SG1), the unassigned rafivirus and livupivirus (SG2), numerous unassigned sapelo-like viruses (SG3), aalivirus, crohiviruses, oriviruses, shanbaviruses (SG4) and the unassigned rodent picornaviruses (SG5) (Figs. 2 and 3).

Fig. 2
figure 2

Phylogenetic analysis of the picornavirus P1 genome region. One hundred ninety-nine picornavirus P1 nucleotide sequences representing all approved and proposed picornavirus species plus unassigned picornaviruses were aligned with MEGA5 and adjusted manually. Bayesian MCMC tree inference was conducted with MrBayes3.2 using an optimal substitution method (GTR+G+I). Convergence was reached after 4,000,000 generations. Presented are 35 acknowledged genera (printed in italics) plus five proposed genera (prop.) plus unassigned picornaviruses (unass.). Five picornavirus supergroups (SGs) are indicated in different colours. Ampivirus, Dicipivirus, Harkavirus, Megrivirus and Rosavirus (printed in black) do not match the 5-supergroup scheme, suggesting the existence of further supergroups. Supplementary Figure 1 presents details of the same phylogenetic tree

Fig. 3
figure 3

Phylogenetic analysis of the picornavirus 3CD genome region. Two hundred six picornavirus 3CD nucleotide sequences representing all approved and proposed picornavirus species plus unassigned picornaviruses were aligned with MEGA5 and adjusted manually. Bayesian MCMC tree inference was conducted with MrBayes3.2 using an optimal substitution method (GTR+G+I). Convergence was reached after 6,500,000 generations. Presented are 35 acknowledged genera (printed in italics) plus five proposed genera (prop.) plus unassigned picornaviruses (unass.). Five picornavirus supergroups (SGs) are indicated in different colours. Ampivirus, Dicipivirus, Harkavirus, Megrivirus, Rosavirus, an unassigned bat picornavirus and poecivirus (printed in black) do not match the 5-supergroup scheme, suggesting the existence of further supergroups. Supplementary Figure 2 presents details of the same phylogenetic tree

Outlook

By now, picornaviruses have been detected in five of seven vertebrate classes, but not in invertebrates, fungi, plants, protozoans or procaryotes. If, in future, this observation proves to be true, important conclusions regarding the virus life cycle, dependence on host factors and co-evolution may be drawn. Methodological and computational advances allow virologists to determine a virus structure by cryo-electron microscopy in very short time. It is likely that the number of picornavirus structures will substantially increase in the near future and improve our knowledge of virus capsids, interaction of viruses with certain antivirals or their receptors, or RNA penetration. Tools have to be established to isolate and propagate in cell culture those viruses which are known only from their sequences. Presently, there is a frustrating disparity in our capabilities of virus isolation and genome sequencing. Without viruses at hand, biochemical and other molecular analyses are hampered. Most annoying is the lack of knowledge of the precise polyprotein processing sites. Likewise, it is crucial to characterize the nonstructural proteins (e.g., by proteomics approaches), many of which have been identified by their sequence motifs only, while their function in the viral life cycle is obscure. Another important task will be the improvement of picornavirus diagnostics. Many picornaviruses have been identified in stools of pets (dogs, cats) and bats, birds and various lower vertebrates in individual cases, raising questions regarding their prevalence in host populations and environmental samples as well as their zoonotic potential, spill-over infections and transmission routes. An improved knowledge of the diversity and prevalence of pathogens will be important for risk assessment and sanitation of animal flocks. Finally, the expansion and further development of picornavirus taxonomy is essential for keeping pace with this ever-growing virus family.