key: cord-337707-xbwilp1w
authors: Kin, Nathalie; Miszczak, Fabien; Lin, Wei; Ar Gouilh, Meriadeg; Vabret, Astrid
title: Genomic Analysis of 15 Human Coronaviruses OC43 (HCoV-OC43s) Circulating in France from 2001 to 2013 Reveals a High Intra-Specific Diversity with New Recombinant Genotypes
date: 2015-05-07
journal: Viruses
DOI: 10.3390/v7052358
sha: 
doc_id: 337707
cord_uid: xbwilp1w

Human coronavirus OC43 (HCoV-OC43) is one of five currently circulating human coronaviruses responsible for respiratory infections. Like all coronaviruses, it is characterized by its genome’s high plasticity. The objectives of the current study were to detect genetically distinct genotypes and eventually recombinant genotypes in samples collected in Lower Normandy between 2001 and 2013. To this end, we sequenced complete nsp12, S, and N genes of 15 molecular isolates of HCoV-OC43 from clinical samples and compared them to available data from the USA, Belgium, and Hong-Kong. A new cluster E was invariably detected from nsp12, S, and N data while the analysis of nsp12 and N genes revealed the existence of new F and G clusters respectively. The association of these different clusters of genes in our specimens led to the description of thirteen genetically distinct genotypes, among which eight recombinant viruses were discovered. Identification of these recombinant viruses, together with temporal analysis and tMRCA estimation, provides important information for understanding the dynamics of the evolution of these epidemic coronaviruses.

AY391777), coding for the glomerular part of the S protein (sub-unit S1) of 7 HCoV-OC43s from clinical samples collected in 2003 at the University Hospital of Caen. They observed four genetically distinct subgroups. One of the subgroups constitutes an outlier group located between HCoV-OC43 and Bovine coronavirus (BCoV), containing a 12-nucleotide deletion in common with BCoV but not with other HCoV-OC43s [32] . More recently, four genetically distinct HCoV-OC43 genotypes have been identified from respiratory specimens sampled at the Public Hospital of China over a 7-year period (2004 to 2011) [33] . In this study, Lau et al. based their genotype definition upon the complete sequence of nsp12, S and N genes [33] . Genotype A contains the prototype VR759 strain isolated in 1967. Genotypes B and C are two naturally circulating HCoV-OC43s. Genotype D is a recombinant virus of genotypes B and C, obtained from a specimen from 2004 [33] . Based on a bootscan analysis of the complete genome of the 3 HCoV-OC43s belonging to the circulating genotypes B, C, and D, it was assumed that a hot spot was likely located between the nsp12 and S genes, more precisely at the nsp12/nsp13 junction. Our objective was to investigate the presence, the temporal distribution and the recombination patterns of HCoV-OC43 in Lower Normandy over 13 years (2001 to 2013), using the methodology and the HCoV-OC43 genotype definition elaborated by Lau et al. in 2011 [33] . This study focuses on the sequences of the nsp12, S, and N genes of 15 HCoV-OC43s detected in respiratory specimens sampled from 2001 to 2013.

Of the 15 HCoV-OC43s and the prototype strain VR759, six, eight, and three overlapping sequences were obtained for the nsp12, S, and N genes, respectively. After assembly, these overlapping fragments encompassed the full nsp12, S, and N genes. In this study, all HCoV-OC43s including the VR759 prototype strain are associated with three accession numbers in GenBank, for nsp12, S, and N genes ( Table 1 ). The sequences of HCoV-OC43s were named according to the following nomenclature: Virus/FRA-EPI/location of sampling/year of sampling/specimen number. EPI is an abbreviation for EPICOREM consortium. For the prototype strain, the specimen number is replaced by VR759. 

The phylogenetic analysis was conducted by comparing the topology of the three phylogenetic trees obtained from the nsp12, S, and N genes. Figure 1 shows the three trees obtained by the neighbor joining method [35] . On each tree, five to six clusters were observed, including an outlier group compounded with the three BCoVs sequences Mebus, Kakegawa, and Quebec that root HCoV-OC43 sequences [36] [37] [38] . The clustering is supported with high bootstrap values. We used part of the nomenclature of genotypes proposed by Lau et al. in 2011 [33] . These authors described the genotypes A, B, C, and D from nsp12, S and N genes. We used this nomenclature to name each genotype using three letters corresponding to clusters in which the different sequences are located in nsp12, S, and N trees, respectively. Following this nomenclature, genotype D is a recombinant genotype B/C/C. The three previously described clusters-A, B, and C-are observed in our three gene trees in addition to a newly described cluster E. The nsp12 tree depicts a new cluster we named "F" that roots all other HCoV-OC43s. According to the N gene, a sixth cluster we named "G" separates from cluster E and roots clusters A, B, and C. Among the set of sequences obtained from the 16 HCoV-OC43s of our study, three are distributed in two non-recombinant genotypes as follows: the VR759 sequence belongs to the genotype AAA; HCoV-OC43/FRA-EPI/Caen/2003/05 and HCoV-OC43/FRA-EPI/Caen/2006/08 belong to genotype BBB. The 13 other sets of sequences are allocated among six recombinant genotypes as follows: five genotypes BCC, defined as genotype D by Lau et al.; two genotypes CEE, three genotypes CBE, one genotype BCG, one genotype CCE, and one genotype CEB. Among the 14 American sets of sequences, seven are distributed into two non-recombinant genotypes as follows: one CCC and six EEE. The seven other sets of sequences are distributed among four recombinant genotypes as follows: three genotypes CCA, two genotypes FCB, one genotype CCB, and one genotype CEE. Among the two sets of sequences from Hong-Kong, one is a non-recombinant genotype CCC and the other is a recombinant genotype BCC. Among the two HCoV-OC43s from Belgium, one is a non-recombinant genotype BBB and the other is a recombinant genotype BCC. Finally, the two last sequences of HCoV-OC43 VR759 (accession number: AY391777 and NC005147) belong to the non-recombinant genotype AAA [28, 39] . The results of the phylogenetic analysis are summarized in Table 2 .

The two alignments constructed from the 39 dated sequences of S and N genes have been used to set up a molecular clock calibration in order to estimate the date of emergence of mean clusters. Figure 2 shows Bayesian trees inferred from the alignments of S and N genes. The dates of emergence of the clusters and the corresponding 95% Highest Posterior Density (95% HPD) are indicated. Based on the sequence data of the S gene, the tMRCA of BCoV and HCoV-OC43 is estimated in 1885, with a 95% HPD from 1858 to 1907. For the HCoV-OC43, cluster E is predicted to have emerged in 1943 (95% HPD 1933 to 1952), and genotype A is estimated to have emerged in 1951 (95% HPD, 1943 to 1959). Genotypes B and C are predicted to have emerged from their tMRCA in 1982 (95% HPD, 1980 to 1983). The molecular clock conducted from sequence data of N gene allows us to estimate the tRMCA of BCoV and HCoV-OC43 in 1879 (95% HPD, 1822 to 1923). Genotypes E and A are predicted to have emerged in 1932 (95% HPD, 1904 to 1952) and 1957 (95% HPD, 1950 to 1972), respectively. Genotypes B and C are predicted to have emerged from their tMRCA in 1984 (95% HPD, 1976 to 1991).

Phylogenetic analysis of the complete nsp12, S, and N genes of the 36 HCoV-OC43 strains and 3 BCoV strains. The phylogenetic trees were constructed by the neighbor joining method [35] . Bootstrap values were calculated from 1000 replicates. Bootstrap values over 70% are shown [40] . The evolutionary distances were computed using the Kimura 2-parameter method (kimura) and units are the number of base substitutions per site. Evolutionary analyses were conducted in MEGA, version 6.06. [41] . Blue triangle, isolates from Caen; red circle, isolates from USA; green circle, isolates from Hong-Kong; black diamond, isolates from Belgium; purple square, ATCC-VR759 strains, empty black square, BCoV strains.

Results of Bayesian analysis based on S and N gene sequence data. Inferences were calculated with the one parametric coalescent model with a constant size, under the TN93+G substitution parameter, using the BEAST package (version 1.8.2) [42] . The length of MCMC was fixed at 10 8 states for S and N genes. The dates of emergence of mean clusters are indicated, associated with the 95% HPD. 

The 4 HCoVs-OC43, 229E, NL63, and HKU1-belong to the viral molecular panel tested during the virological routine diagnostics of acute respiratory infections in humans. Among these four circulating HCoVs, OC43 and NL63 seem to show the greatest prevalence and incidence. These viruses also proved to be the viruses encountered at the earliest age of childhood [43] . These HCoVs circulate widely in epidemic form in the general population, causing infections that are most often benign. Infants, immunosuppressed patients, and very elderly patients may however develop very severe pathologies. These four circulating HCoVs must be distinguished from the emerging HCoVs, SARS-CoV, and MERS-CoV, causes of much graver and potentially fatal respiratory pathologies. Control of the latter HCoVs requires the implementation of international health measures [44] [45] [46] .

The evolutionary potential of coronaviruses brings into play point mutations and recombination events. Such genetic diversity generated in this way may have an impact on the performance of molecular detection techniques used in the virological diagnostic process. The study of intra-specific diversity is thus useful in the monitoring of means of detection. Specifically, it allows for the detection of new circulating variants, and provides information about the evolutionary dynamics of the family of respiratory viruses being monitored, with special attention given to coronaviruses. Our study focuses solely on the HCoV-OC43.

The first analyses of the HCoV-OC43 S gene were conducted in 2005 and revealed a potential spatial and temporal distribution of genetic clusters [34] . In 2011, S. Lau and his colleagues were the first to propose a genotypic classification of HCoV-OC43 from the complete sequencing of three genes: Nsp12 (RdRp), S (spike), and N (nucleocapsid). Four genotypes-A, B, C, and D-were identified: genotype A corresponds to the sequences of the HCoV-OC43 prototype strain VR759, adapted to culture cell line HRT-18 (human adenocarsinoma rectal) or RD (rhabdomyosarcoma); genotypes B and C correspond to contemporary circulating strains, and genotype D is described as a recombinant B/C virus. This study was performed on 29 HCoV-OC43s detected in respiratory samples collected in Hong Kong over a period of 7 years (2004 to 2011), in a non-homogenous temporal distribution. More precisely, the majority of these samples (18 of 29) were collected in 2004, and the others (11 of 29) fall between 2005 and 2011 [33] . Our study defines the genotype of 15 HCoV-OC43s detected from upper respiratory samples frozen at −80 °C, collected over a period of 13 years from 2001 to 2013 at the rate of one to two samples per year. The 15 HCoV-OC43s were detected in patients hospitalized in our university hospital. Of these 15 patients, 10 were male and five were female, at ages from 4 months to 67 years. The symptomatology proved variable ( Table 3 ). The infections were essentially located in upper and lower respiratory airways. The presence of associated gastrointestinal symptoms in some patients should be noted. The prototype strain HCoV-OC43 ATCC VR759, grown on cell line HRT-18, was used as a control, allowing for the validation of the methodology used by Lau et al. [33] . This methodology had to be adapted since some of the primers published by the authors did not allow for amplification of all the HCoV-OC43s. Of these 41 sequences, we selected 14 that reflect the genetic diversity and temporal distribution of the whole, and we introduced them into the alignment. In the end, three respective alignments of complete genes-nsp12, S, and N-of 36 HCoV-OC43s identified in three different regions of the world (Hong Kong, Europe, and the USA) over a period of 30 years (from 1983 to 2013) were generated and analyzed. They served as the basis for the construction of phylogenetic trees. Within the three phylogenetic trees-nsp12, S, and N-we identified six different clusters named A, B, C, E, F, and G. A comparative study of the topology of phylogenetic trees corresponding to several genes of the same virus is often used in phylogeny to define recombinant viruses [33, 47, 48] . We thus compared the topology of nsp12, S, and N trees by observing the contextual position of the corresponding sequences of the same HCoV-OC43 molecular isolate. The complexity of the results of our analysis led us to establish a nomenclature allowing for the simplest expression of the genotypes: each genotype is expressed in the form of a three-letter code, linking each one to the member cluster of the different nsp12, S, and N sequences. From here, we defined 13 different genotypes, of which four were non-recombinant (AAA, BBB, CCC, and EEE). The nine remaining genotypes correspond to different recombinant genotypes (Table 2) . Among the 36 HCoV-OC43 sequences examined in this study, 14 HCoV-OC43s of so-called non-recombinant genotypes were found: three HCoV-OC43s of genotype AAA corresponding to prototype strains VR759; three HCoV-OC43s of genotype BBB (origin, France 2003 and 2006, and Belgium 2003); two HCoV-OC43s of genotype CCC (origin, USA 1997 and Hong Kong 2004); and lastly, six HCoV-OC43s of genotype EEE, all originating in the USA, and having circulated from 1983 to 1995. The foremost important point to underline is that these results suggest a high level of recombination among circulating HCoV-OC43 epidemic strains. The limited number of sequences studied in time and space does not allow for conclusive evidence of a possible temporal and spatial distribution of different genotypes HCoV-OC43. However, it should be noted that the E cluster has not yet been described. In the present study, it is identified only in HCoV-OC43 sequences from the USA and France and data suggest that it has been circulating since the 1940s after being the first known group to differentiate from BCoV.

The analysis of alignments shows that cluster E is characterized by a deletion of 12 nucleotides in the S gene, more precisely in the S1 gene that codes the glomerular part of the S protein (position 24,091-24,102, GenBank accession number NC_005147). This deletion of 12 nucleotides in the coding phase results in a deletion of 4 amino acids, TQDG, in the corresponding protein sequence. This deletion was also present in the bovine coronavirus strains used in the alignment, as shown in Figure 3 . The analysis of 105 S1 BCoV sequences available in GenBank shows that this deletion is expressed in all BCoVs, and described in bovine coronaviruses associated with both enteric and respiratory tropisms. This deletion is therefore not a specific marker of viral tropism. It is located in the lectin domain of the S1 protein of BCoV, and is involved in the attachment of the virus to the cellular receptor, identified as a derivative of neuraminic acid (N-acetyl-9-O-acetylneuraminic acid or Neu5,9Ac2). No protein receptor has yet been described for the BCoV or, more broadly, for Betacoronavirus bovine-like clade A. We analyzed all the sequences of this S1 HCoV-OC43 region available in GenBank, for a total of 108 sequences, and the deletion of 12 nucleotides is found in only 27 of them. Of these 27 sequences, 19 correspond to HCoV-OC43s detected in respiratory samples from the USA and sequenced by Town et al. Four correspond to HCoV-OC43s detected in France (three in the current study, and one published by our team in 2006); two correspond to HCoV-OC43 sequences deposited by Browlin et al. under patent in the United Kingdom (EP2182066 and WO2004011651), and finally, the remaining two correspond to HCoV-OC43s adapted to Vero cells and MDCK I, published in 1996 by F. Künkel and Herrler G [49] . The HCoV-OC43 E cluster is phylogenetically close to the BCoV cluster, and the presence of the genetic marker common to these two clusters provides an argument supporting the hypothesis put forth by Vijgen in 2006 that would identify a crossing of the species barrier occurring from cattle to human [39] . A molecular clock analysis was inferred using the BEAST software version 1.8.1 (http://beast.bio.ed.ac.uk) based on sequence alignments of S and N genes. Sequences of the nsp12 gene were not used because they were uninformative due to the low variability of this gene. Results were compared with those previously obtained by Vijgen from 20 Betacoronavirus clade A sequences (nine BCoVs, three PHEVs, and eight HCoV-OC43s). It should be noted that eight sets of HCoV-OC43 sequences used in the analysis of In 2013, Wertheim et al. demonstrated that a strong purifying selection could lead to the underestimation of the evolutionary rate of coronaviruses. In this study, the authors proposed an alternative theory of origin for all coronaviruses. They estimate the tMRCA of coronaviruses infecting mammalian species (Alpha-and Betacoronavirus genus) and avian species (Gamma-and Deltacoronavirus genus) at about 293 million years, as opposed to 10,000 years as previously estimated [5] . These authors found a general agreement in the inferred branch lengths between evolutionary models for the shorter branches (inferior to 0.05 substitution per site). As expected in the phylogeny of two species of the same coronavirus genus involved in a relatively recent zoonotic transmission event, all the branches of neighbor joining and Bayesian trees were shorter than 0.03 substitutions per site, as shown in Figures 1 and 2 . It can be speculated that the phylogenetic analyses presented in this study did not underestimate the date of emergence of the different clusters nor the estimation of tMRCA for HCoV-OC43 and BCoV. The molecular clock analysis indicated that HCoV-OC43 and BCoV strains have emerged from their tMRCA around 1890 (1859-1952), which was consistent with the data published by Vijgen et al. in 2006 [39] . The emergence of cluster E, characterized by a TQGD amino acid deletion, is estimated at around 1944 , and would have occurred at the same time as that of cluster A, in the 1942-1958 time range. The cluster A includes the first HCoV-OC43 strain identified in 1967 from a respiratory sample [13] . The B and C clusters would have circulated later, since the 1980s. The molecular clock results therefore suggest that, since their emergence in the late nineteenth century, the evolution of circulating HCoV-OC43 epidemic strains would have been marked by the occurrence of an insertion of four amino acids, TQDG, in the N-terminal region of the S1 attachment protein, as opposed to the occurrence of the deletion of these four amino acids. The pursuit of this study should lead to a means of determining the functional impact of this insertion, especially with regard to tropism and the clinical expression of infection by HCoV-OC43. Note that the frequent presence of gastrointestinal symptoms in the clinical landscape of HCoV-OC43 respiratory infections has yet to be accounted for.

In 2014, Hu et al. studied the genetic diversity of the HCoV-OC43 circulating in Beijing, China [50] . The authors performed partial sequencing of the S1 and N genes (24,228-24,785 and 30,022-30,580 for S1 and N respectively, using the HCoV-OC43 sequence associated with the accession number NC_005147 as reference) of 50 HCoV-OC43s detected in the respiratory samples, all taken in 2012. They included the sequences published by Lau et al. in their analysis in 2011, and used the nomenclature established by these same authors to identify various clusters obtained [33, 50] . In addition to the A and B clusters, they identified 4 other groups, designated UNT1, UNT(B), UNT(C/D), and (C/D), respectively. Several clusters were found to correspond to those previously described by Lau et al. in 2011 [33] . The UNT(B) cluster or "Untyped B" is a group close to the B cluster. The C/D cluster corresponds to the C cluster, named "C/D", due to the presence of the HK04_02 strain defined by Lau et al. as a genotype D, referring to a recombination virus belonging to the B cluster on the nsp12 sequence and to the C cluster on the S and N sequences [33] . The UNT(C/D) cluster or "Untyped (C/D)" corresponds to a group close to the preceding cluster. Finally, the UNT1 cluster or "Untyped 1", which is close to the BCoV sequences included in this analysis, might correspond to the E cluster describe in the current study. In the report by Hu et al. (2014) , the UNT1 cluster includes only five HCoV-OC43s (accession numbers Z32769, HC734573, CQ772300, DD266155, and Z32768) [50] . We have verified that they all have the characteristic nucleotide deletion of 12 nucleotides. This very recent report supports the results determined in the current study in that they both demonstrate a high genetic diversity of circulating HCoV-OC43s.

The present study highlights the difficulty of investigating intraspecific genetic variability of HCoV-OC43s, though the analysis of partial viral genomes (complete genes) allows the description of clear recombination patterns. Nevertheless, due to the high evolutionary potential of HCoV-OC43, analyzing complete genomes would reveal a more detailed and perhaps more complex recombination pattern. In addition, the data reported and analyzed here illustrate the dynamic diversification of HCoV-OC43 following its emergence in human. This remarkable diversification contrasts with the relative stability of BCoV in cattle (unpublished data). The present study therefore draws attention to the importance of increasing our knowledge of the evolution dynamism of coronavirus in combating the permanent emergence risk that they pose in the human population, evidenced by the SARS pandemic of 2002-2003 and the recently identified MERS-coronavirus in the Middle-East.

Fifteen human respiratory specimens (upper and lower) positive for HCoV-OC43 using molecular detection, sampled from 2001 to 2013 (one to two for each year) were chosen among the specimens available at the virology department of the University Hospital of Caen. These specimens were sampled from patients suffering from acute upper or lower respiratory diseases, and were sent to our laboratory for a viral diagnosis. The detection of HCoV-OC43 was performed using an in-house real-time quantitative RT-PCR, for which the original primers may be found in Table 4 . Clinical signs, age, and sex of patients were investigated and compiled in Table 3 . We also used a culture supernatant of the 

Total nucleic acids of 300 μL of clinical specimens and VR759 cultures supernatant were extracted using the QiAsymphony automate and the QIAsynphony virus/bacteria kit (Qiagen, Holden, Germany), following instructions of manufacturer.

To determine the genomic sequence of nsp12, S, and N genes, a set of 17 overlapping RT-PCR products, from 600 to 900 nucleotides, were generated. These RT-PCR products encompassed the entire three genes. Twelve primers were used to amplify and sequence the nsp12 gene, composed of 2783 nucleotides and located from nucleotides 13,317 to 16,099 (reference to VR759, NC005147). Eighteen primers were used to amplify and sequence the S gene composed of 4062 nucleotides. This gene is located in the last third of the HCoV-OC43 genome, between bases 23,644 and 27,729 (reference to HCoV-OC43 VR759, NC005147). Six primers were used to amplify and sequence the N gene composed of 1347 nucleotides and located at the 3' end of the HCoV-OC43 genome, from nucleotide 29,104 to 30,450 (reference to HCoV-OC43 VR759, NC005147). For both RT-PCR and sequencing reactions, primers were selected in conserved domains from an alignment of the previously published HCoV-OC43. Some of the selected primers were published by Lau in 2011, but others were designed using Primer-Blast software (http://www.ncbi.nlm.nih.gov/tools/primer-blast/) [33] . All of them were tested in silico and in vitro with the VR759 prototype strain to test their efficacy. In vitro experimentations have also been conducted using the VR759 strain to test the primer abilities to provide us with quality sequencing products. Table 4 , associated with the corresponding targeted gene and with the fragment size. RT-PCR reactions were performed using OneStep RT-PCR System (Qiagen, Holden, Germany), from 2.5 μL of nucleic acid in a final volume of 25 μL. The same primers were used to perform the complete bidirectional sequencing of the RT-PCR products, with a Sanger derived method. 2.5 μL of each RT-PCR product were first purified by adding 1 μL of ExoSAP-IT (USB corporation, Cleveland, OH, USA), and by heating it at 37 °C for 15 min, followed by 15 min at 80 °C to disable enzymatic activities. Two sequencing reactions were performed for each RT-PCR product with either a forward or a reverse primer. 1 to 2 μL of purified RT-PCR product were used in a final volume of 10 μL of reaction mixture, made up with the BigDye terminator cycle sequencing kit, version 1.1 (Applied Biosystems, Foster City, CA, USA), in addition to 1 μL of one of the corresponding primers. The cycling parameters for the sequencing RT-PCR were 96 °C for 1 min, followed by 30 cycles of 96 °C for 10 s, 60 °C for 5 s, and 60 °C for 4 min. The analysis of labelled products was performed by the capillary electrophoresis ABI Prism 3500 genetic analyzer (Applied Biosystems, Foster City, CA, USA), by the molecular genetics laboratory of the University Hospital of Caen. In this study, the term of molecular isolate was use to designate the sequence of HCoV-OC43 detected directly from clinical sample.

Sequences were assembled in contigs corresponding to the entire nsp12, S or N genes with CodonCode Aligner Software, version 5.0.1 (CodonCode corporation, Centerville, MA, USA). Multiple sequence alignment and phylogenetic analysis were performed using MEGA software, version 6.06 (http://www.megasoftware.net). Twenty-one sequences available in GenBank were added to the alignment, including HCoV-OC43s HKU4_01 and HKU4_02 from Hong-Kong [33] ; HCoV-OC43s BE03 and BE04 from Belgium [36, 38, 51] . The accession numbers corresponding to these sequences are given in Table 3 . Phylogenetic trees were constructed using the method of neighbor joining, with the substitution model of Kimura-2, implemented in MEGA6 [41, 52] . These trees were used to perform the comparative analysis of our results with those of Lau et al in 2011 [33] . In order to support neighbour joining trees with a probabilistic method and to estimate divergence time between groups and subgroups, trees of S and N genes were also inferred using the same sequences and a Bayesian Markov Chain Monte Carlo method, implemented in the BEAST package, version 1.8.1 [42, 53] . Inferences were calculated with the one parametric coalescent model with a constant size, under the TN93+G substitution model according to Bayesian Information Criterion (BIC) and Akaike Information Criterion (AIC) values of a model test carried out on MEGA6 software. The tMRCA was estimated using a relaxed molecular clock with an uncorrelated lognormal distribution. The length of MCMC was fixed at 3 × 10 −8 states for the both genes. The Tracer software, version 1.5, implemented in the BEAST package, was used to assess the fitness of the model used, focusing the Effective Sampling Size (ESS) data, after a burning of 10%. The target trees were obtained with the maximum credibility clades, with a posterior probability limit of 0.05. The tMRCA age for each genotype was compared for S and N genes.

Evolving the largest RNA virus genome

Mesoniviridae: A proposed new family in the order Nidovirales formed by a single species of mosquito-borne viruses

Ratification vote on taxonomic proposals to the International Committee on Taxonomy of Viruses

Virus Taxonomy: Classification and Nomenclature of Viruses: Ninth Report of the International Committee on Taxonomy of Viruses

Discovery of seven novel mammalian and avian coronaviruses in the genus deltacoronavirus supports bat coronaviruses as the gene source of alphacoronavirus and betacoronavirus and avian coronaviruses as the gene source of gammacoronavirus and deltacoronavirus

Bovine-like coronaviruses isolated from four species of captive wild ruminants are homologous to bovine coronaviruses, based on complete genomic sequences

Coronavirus diversity, phylogeny and interspecies jumping

Two amino acid changes at the N-terminus of transmissible gastroenteritis coronavirus spike protein result in the loss of enteric tropism

Comparison of genomic and predicted amino acid sequences of respiratory and enteric bovine coronaviruses isolated from the same animal with fatal shipping pneumonia

Genomic characterization of equine coronavirus

A previously undescribed coronavirus associated with respiratory disease in humans

Identification of a new human coronavirus

Growth in suckling-mouse brain of "IBV-like" viruses from patients with upper respiratory tract disease

A New virus isolated from the human respiratory tract

Growth and intracellular development of a new respiratory virus

Characterization and complete genome sequence of a novel coronavirus, coronavirus HKU1, from patients with pneumonia

Croup is associated with the novel coronavirus NL63

Human coronavirus OC43 causes influenza-like illness in residents and staff of aged-care facilities in

Rhinovirus and coronavirus infection-associated hospitalizations among older adults

Prevalence and molecular epidemiology of human coronavirus HKU1 in patients with acute respiratory illness

An outbreak of coronavirus OC43 respiratory infection in Normandy, France

Clinical and molecular epidemiological features of coronavirus HKU1-associated community-acquired pneumonia

Epidemiological findings from a retrospective investigation

Severe acute respiratory syndrome

Isolation of a novel coronavirus from a man with pneumonia in Saudi Arabia

Human respiratory coronavirus OC43: Genetic stability and neuroinvasion

Complete genomic sequence of human coronavirus OC43: Molecular clock analysis suggests a relatively recent zoonotic coronavirus transmission event

Implications of altered replication fidelity on the evolution and pathogenesis of coronaviruses

Lack of evidence for proofreading mechanisms associated with an RNA virus polymerase

High-frequency RNA recombination of murine coronaviruses

Inter-and intra-variant genetic heterogeneity of human coronavirus OC43 strains in France

Molecular epidemiology of human coronavirus OC43 reveals evolution of different genotypes over time and recent emergence of a novel genotype due to natural recombination

Circulation of genetically distinct contemporary human coronavirus OC43 strains

The neighbor-joining method: A new method for reconstructing phylogenetic trees

Molecular analysis of the S glycoprotein gene of bovine coronaviruses isolated in Japan from 1999 to

Pathology of neonatal calf diarrhea induced by a coronavirus-like agent

Full-length genomic sequence of bovine coronavirus (31 kb). Completion of the open reading frame 1a/1b sequences

Evolutionary history of the closely related group 2 coronaviruses: Porcine hemagglutinating encephalomyelitis virus, bovine coronavirus, and human coronavirus OC43

Confidence limits on phylogenies: An approach using the bootstrap

Molecular evolutionary genetics analysis version 6.0

Bayesian phylogenetics with BEAUti and the BEAST 1.7

The dominance of human coronavirus OC43 and NL63 infections in infants

Severe acute respiratory syndrome coronavirus and Middle East respiratory syndrome coronavirus in travelers

Progress in global surveillance and response capacity 10 years after severe acute respiratory syndrome

The first pandemic of the 21st century

Genomic analysis of 16 Colorado human NL63 coronaviruses identifies a new genotype, high sequence diversity in the N-terminal domain of the spike gene and evidence of recombination

Comparative analysis of 22 coronavirus HKU1 genomes reveals a novel genotype and evidence of natural recombination in coronavirus HKU1

Structural and functional analysis of the S proteins of two human coronavirus OC43 strains adapted to growth in different cells

Prevalence and genetic diversity analysis of human coronavirus OC43 among adult patients with acute respiratory infections in Beijing

A cis-acting function for the coronavirus leader in defective interfering RNA replication

A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences

Prospects for inferring very large phylogenies by using the neighbor-joining method

This work was supported by the French Agence Nationale de la Recherche (ANR) on the behalf of EPICOREM project (Eco-Epidemiology of Coronaviruses, from Wildlife to Human: Emergence

prototype strains VR759 as reference and for validation of our method. This strain was obtained from the ATCC and maintained in HRT-18 cells. Threat Assessment), grant ANR-13-BSV3-0013. We thank Paul Brown for the English proofreading of the manuscript.

A.V. and N.K. design the study. N.K. and W.L. performed the experiments. NK. and M.A.G. performed the analyses. N.K. conceived the manuscript with contributions from F.M., M.A.G. and A.V. All authors contributed toward the preparation, editing and proofreading of the manuscript.

The authors declare no conflict of interest.