Candidates in Astroviruses, Seadornaviruses, Cytorhabdoviruses and Coronaviruses for +1 frame overlapping genes accessed by leaky scanning
© Firth and Atkins; licensee BioMed Central Ltd. 2010
Received: 25 November 2009
Accepted: 25 January 2010
Published: 25 January 2010
Overlapping genes are common in RNA viruses where they serve as a mechanism to optimize the coding potential of compact genomes. However, annotation of overlapping genes can be difficult using conventional gene-finding software. Recently we have been using a number of complementary approaches to systematically identify previously undetected overlapping genes in RNA virus genomes. In this article we gather together a number of promising candidate new overlapping genes that may be of interest to the community.
Overlapping gene predictions are presented for the astroviruses, seadornaviruses, cytorhabdoviruses and coronaviruses (families Astroviridae, Reoviridae, Rhabdoviridae and Coronaviridae, respectively).
Overlapping genes (whereby the same nucleotide sequence codes for two or more proteins in different reading frames) are particularly common in RNA viruses, where they may serve as mechanisms to optimize the coding potential of compact genomes, regulate gene expression, or circumvent the host cell's canonical - though not ubiquitous - rule of 'one functional protein per mRNA'. However, such genes can be difficult to detect using conventional gene-finding software.
MLOGD is a gene-finding program which was designed specifically for identifying overlapping coding sequences (CDSs) through the incorporation of explicit models for sequence evolution in multiply-coding regions [1–3]. One caveat is that de novo overlapping CDSs are often considerably less conserved than the ancestral genes that they overlap (the ancestral gene is usually the known gene as it tends to be the longer of the two, while the de novo gene is often very short). The explicit 'coding signal' of such a CDS may be swamped by the 'coding signal' of the ancestral CDS. Thus there are a number of known overlapping CDSs which MLOGD fails to detect. Another caveat of MLOGD is that, if an overlapping CDS is very short and highly conserved (e.g. due to coding in two different reading frames and perhaps also harbouring an RNA secondary structure for stimulating ribosomal frameshifting into the overlapping CDS; ), then there may be too few base variations to obtain a useful signal (either coding or non-coding). On the other hand, for overlapping CDSs that are subject to a reasonable degree of purifying selection and that are not too short, MLOGD can provide a robust detection with just two input sequences provided that they are sufficiently divergent.
A sometimes more sensitive, but generally less specific, approach involves analysis of conservation at synonymous sites within known CDSs. This method is particularly useful when a large and diverse input sequence alignment is available [5, 6]. Enhanced conservation may be associated with overlapping functional elements. However such elements may be either coding or non-coding, so additional evidence (e.g. conservation of an overlapping open reading frame and a potential translation mechanism over a sufficiently divergent sequence alignment) is required in order to use this method to identify overlapping CDSs. Care also needs to be taken to discriminate dual-coding sequences from regions of enhanced conservation that may arise from recombination.
Over the past few years we have engaged in a systematic survey of viral genomes for previously undetected overlapping genes. While many of these merit detailed analysis and experimental follow-up (either because they are in important and well-studied viruses or because they involve novel non-canonical translation mechanisms), there remain a miscellany of promising candidate new overlapping genes that we are not currently in a position to follow up experimentally but, nonetheless, may be of interest to the community. The purpose of this article is to communicate five of these candidates.
Results and Discussion
Summary of candidate overlapping CDSs
Human, porcine and feline astroviruses etc
5.9 × 10-28
Banna virus, Kadipiro virus etc
[GenBank: NC_004204, GenBank: NC_004209]
2.3 × 10-10
Lettuce necrotic yellows virus etc
[GenBank: NC_007642, GenBank: NC_011532]
Group 3c coronaviruses
[GenBank: NC_011548, GenBank: NC_011549, GenBank: NC_011550]
Bat coronaviruses 1A, 1B, HKU8
[GenBank: NC_010436, GenBank: NC_010437, GenBank: NC_010438]
Mamastrovirus (human, porcine, feline astrovirus clade)
Astroviruses have monopartite positive-sense ssRNA genomes and are associated with gastroenteritis and viral diarrhoea in humans and other vertebrates. The non-structural polyprotein (ORF1a and, via ribosomal frameshifting, an ORF1a-ORF1b fusion) are translated from the genomic RNA (gRNA) while the structural polyprotein (ORF2) is translated from a sub-genomic RNA (sgRNA) [8, 9]. ORFX overlaps the 5' end of ORF2 in the +1 reading frame. A 112-codon AUG-initiated +1 frame ORF is present in nearly all human astrovirus sequences in GenBank with complete or partial coverage of the ORFX region (~780 sequences; ~150 with full coverage of ORFX; 25 Oct 2009). Just a very small number of sequences are ORFX-defective: one partial sequence contains a premature termination codon, one sequence contains a 16-codon 3' extension, and three sequences have a CUG codon (which may, nonetheless, allow a low level of initiation ) instead of an AUG codon at the proposed ORFX initiation site.
The infectivity of a mutant astrovirus in which ORFX expression was inadvertently abolished was reduced by only 50% relative to wild-type virus  (a reduction which may be due to amino acid changes in the polyprotein frame besides the absence of ORFX product), thus demonstrating that the putative ORFX product is non-essential - at least for replication in cell culture. However, this does not imply that ORFX is not a CDS since the conditions or functions for which the presumed ORFX product is important may not have been directly tested (e.g. for comparison, the infectivity of a mutant alphavirus in which expression of the experimentally verified TF protein was abolished was also reduced by only ~50% ).
Seadornavirus segment 7
There are currently two other seadornavirus species with sequence coverage of the ORFX region - Kadipiro virus (KDV; 1 sequence), and Liao ning virus (LNV; 2 sequences) [20, 25]. When MLOGD was applied to an alignment of the three species BAV, KDV and LNV, the results were ambiguous due to the high divergence between the different sequences. However, there is the potential, at least, for a functional ORFX in KDV, and possibly also LNV. In KDV, the VP7 CDS utilizes the first AUG codon, which has a weak Kozak context, while the second AUG (separated from the first AUG by 16 nt) is in the +1 reading frame and heads a 65-codon potential ORFX (Figure 4). In LNV, however, the VP7 CDS has two closely spaced AUG codons in one sequence, and a medium Kozak context in both sequences ('G' at +4), so is sub-optimal for leaky scanning. Moreover, although the next AUG codon is in the +1 reading frame, it is 64 nt downstream and only heads a 42-codon ORF. Thus, although there is a very strong case for a coding ORFX in BAV, whether or not this ORF is also present in KDV, and especially in LNV, can not be reliably assessed with the currently available sequence data.
Cytorhabdovirus (Lettuce necrotic yellows and Lettuce yellow mottle viruses)
In fact these two sequences are the only two distinct sequences with coverage of ORFX currently available in GenBank (2 Nov 2009). The mean nucleotide identity between the two sequences within the ORFX region is only ~50%, which is below the ideal range for MLOGD (substitution saturation at high divergences makes it more difficult for MLOGD to distinguish between the single- and dual-coding models and causes a high-divergence 'turnover' in the MLOGD score ). Nonetheless, and notwithstanding the very limited sequence data, there is a good coding signal for ORFX (Figure 5). In fact the presence of this ORF has already been noted in both viruses [27, 28] (and designated P' by Ref. ) though, so far as we are aware, this is the first evidence (apart from its conserved presence) that it is likely to be coding.
CDSs in the Rhabdoviridae are translated from a series of mRNA transcripts produced via a transcription termination-reinitiation mechanism, with conserved junction sequences containing the transcription stop and start signals located between consecutive CDSs so that mRNAs are generally monocistronic [26, 28, 29]. In the case of the P mRNA of LNYV and LYMoV, the P CDS utilizes the second AUG codon on the mRNA while the first AUG codon is in the correct frame for ORFX translation (Figure 6). However, the ORFX AUG codon has poor Kozak context ('U' at - 3, 'A' or 'C' at +4; cf. ), which presumably allows a significant proportion of ribosomes to translate the P CDS via leaky scanning.
A similarly positioned overlapping CDS (generally referred to as 'C' and generally initiating downstream rather than upstream of the P initation site) occurs in certain paramyxovirus genera (e.g. Morbillivirus, Respirovirus) besides Vesicular stomatitis virus (Vesiculovirus, family Rhabdoviridae), though the C gene is likely to have arisen independently in the two families . It has been suggested that the highly variable nature of the P protein facilitates the evolution of novel genes overlapping its N-terminal regions and, in the absence of discernable sequence homology, it cannot be assumed that the resulting proteins have similar functions [31, 32].
Coronaviruses (family Coronaviridae) belong to the order Nidovirales. At 26-32 kb, coronavirus genomes are among the largest of all RNA viruses. As with other members of the order, these viruses have a monopartite positive-sense ssRNA genome encoding a large replicase polyprotein that is expressed from the genomic RNA (ORF1a and, via ribosomal frameshifting, an ORF1a-ORF1b fusion product), and a number of other proteins which are translated from a nested set of 3'-coterminal sgRNAs [33, 34]. The coronaviruses are currently classified into three main groups (recently elevated to genera) which are further divided into subgroups, though there are also a large number of species that await formal classification. Although a core set of sgRNA-encoded structural proteins (S, E, M, N) are conserved throughout all groups, a variable number of auxilliary proteins are also encoded by sgRNAs - including a number of known overlapping CDSs (e.g. the I CDS that overlaps the N CDS of some Group 2 coronaviruses ). We have identified two new candidates - one in Group 3 coronaviruses of the proposed subgroup 3c , and one in certain Group 1b coronaviruses. (Note that, while the evidence for the coding status of the preceding three candidates is very strong, the evidence for these two candidates is less certain but should become clearer as more sequence data become available.)
Coronavirus Group 3c
Bat coronaviruses 1A, 1B, HKU8
Overlapping genes are difficult to identify and are often overlooked. However, it is important to be aware of such genes as early as possible in order to avoid confusion (otherwise functions of the overlapping gene may be wrongly ascribed to the gene they overlap), and also so that the functions of the overlapping gene may be investigated in their own right. Computational analysis of sequence data is a time- and cost-efficient way to find such genes and help direct experimental follow-up.
The list of new candidate overlapping genes presented here is by no means complete as we have omitted several candidates that we are currently following up experimentally, and a number of candidates that are less certain - for example candidates where a conserved potential translation mechanism has not been identified, or candidates where a significant fraction of isolates contain premature termination codons, or candidates with too little phylogenetic support within the currently available sequence databases. To the best of our knowledge (except as noted for cytorhabdovirus) these candiates have not previously been described or annotated elsewhere, and we apologize if we have accidently omitted any previous references to any of these candidates.
Virus sequences were downloaded from GenBank and alignments were generated using standard bioinformatics software (blast , clustal  and EMBOSS ). Candidate overlapping CDSs were identified using either MLOGD or analysis of conservation at synonymous sites of the annotated CDSs as described previously [1, 2, 5]. The following astrovirus sequences with complete coverage of ORF2 were used for the alignment and statistics illustrated in Figure 1: [GenBank: AB000283, AB000284, AB000285, AB000286, AB000287, AB000288, AB000289, AB000290, AB000291, AB000292, AB000293, AB000294, AB000295, AB000296, AB000297, AB000298, AB000299, AB000300, AB000301, AB009984, AB009985, AB013618, AB025801, AB025802, AB025803, AB025804, AB025805, AB025806, AB025807, AB025808, AB025809, AB025810, AB025811, AB025812, AB031030, AB031031, AB037272, AB037273, AB037274, AB290149, AB308374, AB496913, AF056197, AF117209, AF141381, AF248738, AF260508, AY720891, AY720892, DQ028633, DQ070852, DQ344027, DQ630763, EF138823, EF138824, EF138825, EF138826, EF138827, EF138828, EF138829, EF138830, EF138831, EF583300, FJ755402, FJ755403, FJ755404, FJ755405, FJ890352, FJ890355, FM213330, FM213331, FM213332, GQ405855, GQ405856, GQ405857, GQ495608, GQ901902, L06802, L13745, L23513, S68561, U15136, Y08632, Y15938, Z25771, Z33883, Z46658, Z66541]. The seadornavirus, cytorhabdovirus and coronavirus sequences used are listed in Figures 4, 6, 8 and 10. Additional sequences with partial coverage of the annotated CDS which each overlapping gene candidate overlaps were available for the seadornaviruses ([GenBank: EU265679] and [GenBank: EU265722]) and the mamastroviruses (~700 sequences; not listed).
This work was supported by National Institutes of Health Grant R01 GM079523 and an award from Science Foundation Ireland, both to JFA.
- Firth AE, Brown CM: Detecting overlapping coding sequences with pairwise alignments. Bioinformatics 2005, 21: 282-292. 10.1093/bioinformatics/bti007PubMedView ArticleGoogle Scholar
- Firth AE, Brown CM: Detecting overlapping coding sequences in virus genomes. BMC Bioinformatics 2006, 7: 75. 10.1186/1471-2105-7-75PubMedPubMed CentralView ArticleGoogle Scholar
- Chung BYW, Miller WA, Atkins JF, Firth AE: An overlapping essential gene in the Potyviridae. Proc Natl Acad Sci USA 2008, 105: 5897-5902. 10.1073/pnas.0800468105PubMedPubMed CentralView ArticleGoogle Scholar
- Firth AE, Chung BY, Fleeton MN, Atkins JF: Discovery of frameshifting in Alphavirus 6K resolves a 20-year enigma. Virol J 2008, 5: 108. 10.1186/1743-422X-5-108PubMedPubMed CentralView ArticleGoogle Scholar
- Firth AE, Atkins JF: A conserved predicted pseudoknot in the NS2A-encoding sequence of West Nile and Japanese encephalitis flaviviruses suggests NS1' may derive from ribosomal frameshifting. Virol J 2009, 6: 14. 10.1186/1743-422X-6-14PubMedPubMed CentralView ArticleGoogle Scholar
- Balmori Melian E, Hinzman E, Nagasaki T, Firth AE, Wills NM, Nouwens AS, Blitvich BJ, Leung J, Funk A, Atkins JF, Hall R, Khromykh AA: NS1' of flaviviruses in the Japanese encephalitis virus serogroup is a product of ribosomal frameshifting and plays a role in viral neuroinvasiveness. J Virol 2010, 84: 1641-1647. 10.1128/JVI.01979-09View ArticleGoogle Scholar
- Belshaw R, Pybus OG, Rambaut A: The evolution of genome compression and genomic novelty in RNA viruses. Genome Res 2007, 17: 1496-1504. 10.1101/gr.6305707PubMedPubMed CentralView ArticleGoogle Scholar
- Jiang B, Monroe SS, Koonin EV, Stine SE, Glass RI: RNA sequence of astrovirus: distinctive genomic organization and a putative retrovirus-like ribosomal frameshifting signal that directs the viral replicase synthesis. Proc Natl Acad Sci USA 1993, 90: 10539-10543. 10.1073/pnas.90.22.10539PubMedPubMed CentralView ArticleGoogle Scholar
- Willcocks MM, Brown TD, Madeley CR, Carter MJ: The complete sequence of a human astrovirus. J Gen Virol 1994, 75: 1785-1788. 10.1099/0022-1317-75-7-1785PubMedView ArticleGoogle Scholar
- Touriol C, Bornes S, Bonnal S, Audigier S, Prats H, Prats AC, Vagner S: Generation of protein isoform diversity by alternative initiation of translation at non-AUG codons. Biol Cell 2003, 95: 169-178. 10.1016/S0248-4900(03)00033-9PubMedView ArticleGoogle Scholar
- Finkbeiner SR, Li Y, Ruone S, Conrardy C, Gregoricus N, Toney D, Virgin HW, Anderson LJ, Vinjé J, Wang D, Tong S: Identification of a novel astrovirus (astrovirus VA1) associated with an outbreak of acute gastroenteritis. J Virol 2009, 83: 10836-10839. 10.1128/JVI.00998-09PubMedPubMed CentralView ArticleGoogle Scholar
- Zhu HC, Chu DK, Liu W, Dong BQ, Zhang SY, Zhang JX, Li LF, Vijaykrishna D, Smith GJ, Chen HL, Poon LL, Peiris JS, Guan Y: Detection of diverse astroviruses from bats in China. J Gen Virol 2009, 90: 883-887. 10.1099/vir.0.007732-0PubMedView ArticleGoogle Scholar
- Rivera R, Nollens HH, Venn-Watson S, Gulland FM, Wellehan JF: Characterization of phylogenetically diverse astroviruses of marine mammals. J Gen Virol 2010, 91: 166-173. 10.1099/vir.0.015222-0PubMedView ArticleGoogle Scholar
- Kapoor A, Li L, Victoria J, Oderinde B, Mason C, Pandey P, Zaidi SZ, Delwart E: Multiple novel astrovirus species in human stool. J Gen Virol 2009, 90: 2965-2972. 10.1099/vir.0.014449-0PubMedPubMed CentralView ArticleGoogle Scholar
- Monroe SS, Jiang B, Stine SE, Koopmans M, Glass RI: Subgenomic RNA sequence of human astrovirus supports classification of Astroviridae as a new family of RNA viruses. J Virol 1993, 67: 3611-3614.PubMedPubMed CentralGoogle Scholar
- Sedman SA, Gelembiuk GW, Mertz JE: Translation initiation at a downstream AUG occurs with increased efficiency when the upstream AUG is located very close to the 5' cap. J Virol 1990, 64: 453-457.PubMedPubMed CentralGoogle Scholar
- Kozak M: A short leader sequence impairs the fidelity of initiation by eukaryotic ribosomes. Gene Expr 1991, 1: 111-115.PubMedGoogle Scholar
- Lewis TL, Greenberg HB, Herrmann JE, Smith LS, Matsui SM: Analysis of astrovirus serotype 1 RNA, identification of the viral RNA-dependent RNA polymerase motif, and expression of a viral structural protein. J Virol 1994, 68: 77-83.PubMedPubMed CentralGoogle Scholar
- Geigenmüller U, Ginzton NH, Matsui SM: Studies on intracellular processing of the capsid protein of human astrovirus serotype 1 in infected cells. J Gen Virol 2002, 83: 1691-1695.PubMedView ArticleGoogle Scholar
- Attoui H, Billoir F, Biagini P, de Micco P, de Lamballerie X: Complete sequence determination and genetic analysis of Banna virus and Kadipiro virus: proposal for assignment to a new genus (Seadornavirus) within the family Reoviridae. J Gen Virol 2000, 81: 1507-1515.PubMedView ArticleGoogle Scholar
- Attoui H, Mohd Jaafar F, de Micco P, de Lamballerie X: Coltiviruses and seadornaviruses in North America, Europe, and Asia. Emerg Infect Dis 2005, 11: 1673-1679.PubMedPubMed CentralView ArticleGoogle Scholar
- Attoui H, Charrel RN, Billoir F, Cantaloube JF, de Micco P, de Lamballerie X: Comparative sequence analysis of American, European and Asian isolates of viruses in the genus Coltivirus. J Gen Virol 1998, 79: 2481-2489.PubMedView ArticleGoogle Scholar
- Mohd Jaafar F, Attoui H, Mertens PP, de Micco P, de Lamballerie X: Structural organization of an encephalitic human isolate of Banna virus (genus Seadornavirus, family Reoviridae). J Gen Virol 2005, 86: 1147-1157. 10.1099/vir.0.80578-0PubMedView ArticleGoogle Scholar
- Matsuda D, Dreher TW: Close spacing of AUG initiation codons confers dicistronic character on a eukaryotic mRNA. RNA 2006, 12: 1338-1349. 10.1261/rna.67906PubMedPubMed CentralView ArticleGoogle Scholar
- Attoui H, Mohd Jaafar F, Belhouchet M, Tao S, Chen B, Liang G, Tesh RB, de Micco P, de Lamballerie X: Liao ning virus, a new Chinese seadornavirus that replicates in transformed and embryonic mammalian cells. J Gen Virol 2006, 87: 199-208. 10.1099/vir.0.81294-0PubMedView ArticleGoogle Scholar
- Wetzel T, Dietzgen RG, Dale JL: Genomic organization of lettuce necrotic yellows rhabdovirus. Virology 1994, 200: 401-412. 10.1006/viro.1994.1204PubMedView ArticleGoogle Scholar
- Dietzgen RG, Callaghan B, Wetzel T, Dale JL: Completion of the genome sequence of Lettuce necrotic yellows virus, type species of the genus Cytorhabdovirus. Virus Res 2006, 118: 16-22. 10.1016/j.virusres.2005.10.024PubMedView ArticleGoogle Scholar
- Heim F, Lot H, Delecolle B, Bassler A, Krczal G, Wetzel T: Complete nucleotide sequence of a putative new cytorhabdovirus infecting lettuce. Arch Virol 2008, 153: 81-92. 10.1007/s00705-007-1071-5PubMedView ArticleGoogle Scholar
- Barr JN, Whelan SP, Wertz GW: Transcriptional control of the RNA-dependent RNA polymerase of vesicular stomatitis virus. Biochim Biophys Acta 2002, 1577: 337-353.PubMedView ArticleGoogle Scholar
- Joshi CP, Zhou H, Huang X, Chiang VL: Context sequences of translation initiation codon in plants. Plant Mol Biol 1997, 35: 993-1001. 10.1023/A:1005816823636PubMedView ArticleGoogle Scholar
- Jordan IK, Sutter BA, McClure MA: Molecular evolution of the Paramyxoviridae and Rhabdoviridae multiple-protein-encoding P gene. Mol Biol Evol 2000, 17: 75-86.PubMedView ArticleGoogle Scholar
- Rancurel C, Khosravi M, Dunker AK, Romero PR, Karlin D: Overlapping genes produce proteins with unusual sequence properties and offer insight into de novo protein creation. J Virol 2009, 83: 10719-10736. 10.1128/JVI.00595-09PubMedPubMed CentralView ArticleGoogle Scholar
- Gorbalenya AE, Enjuanes L, Ziebuhr J, Snijder EJ: Nidovirales: evolving the largest RNA virus genome. Virus Res 2006, 117: 17-37. 10.1016/j.virusres.2006.01.017PubMedView ArticleGoogle Scholar
- Pasternak AO, Spaan WJ, Snijder EJ: Nidovirus transcription: how to make sense...? J Gen Virol 2006, 87: 1403-1421. 10.1099/vir.0.81611-0PubMedView ArticleGoogle Scholar
- Senanayake SD, Hofmann MA, Maki JL, Brian DA: The nucleocapsid protein gene of bovine coronavirus is bicistronic. J Virol 1992, 66: 5277-5283.PubMedPubMed CentralGoogle Scholar
- Woo PC, Lau SK, Lam CS, Lai KK, Huang Y, Lee P, Luk GS, Dyrting KC, Chan KH, Yuen KY: Comparative analysis of complete genome sequences of three avian coronaviruses reveals a novel group 3c coronavirus. J Virol 2009, 83: 908-917. 10.1128/JVI.01977-08PubMedPubMed CentralView ArticleGoogle Scholar
- Dong BQ, Liu W, Fan XH, Vijaykrishna D, Tang XC, Gao F, Li LF, Li GJ, Zhang JX, Yang LQ, Poon LL, Zhang SY, Peiris JS, Smith GJ, Chen H, Guan Y: Detection of a novel and highly divergent coronavirus from asian leopard cats and Chinese ferret badgers in Southern China. J Virol 2007, 81: 6920-6926. 10.1128/JVI.00299-07PubMedPubMed CentralView ArticleGoogle Scholar
- Powell ML, Brown TD, Brierley I: Translational termination-re-initiation in viral systems. Biochem Soc Trans 2008, 36: 717-722. 10.1042/BST0360717PubMedView ArticleGoogle Scholar
- Chu DK, Peiris JS, Chen H, Guan Y, Poon LL: Genomic characterizations of bat coronaviruses (1A, 1B and HKU8) and evidence for co-infections in Miniopterus bats. J Gen Virol 2008, 89: 1282-1287. 10.1099/vir.0.83605-0PubMedView ArticleGoogle Scholar
- Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ: Basic local alignment search tool. J Mol Biol 1990, 215: 403-410.PubMedView ArticleGoogle Scholar
- Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, Thompson JD, Gibson TJ, Higgins DG: Clustal W and Clustal X version 2.0. Bioinformatics 2007, 23: 2947-2948. 10.1093/bioinformatics/btm404PubMedView ArticleGoogle Scholar
- Rice P, Longden I, Bleasby A: EMBOSS: The European Molecular Biology Open Software Suite (2000). Trends in Genetics 2000, 16: 276-277. 10.1016/S0168-9525(00)02024-2PubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.