During 2007–2014, we collected 27 fecal and 24 intestinal samples from pigs with suspected PEDV or TGEV infections; the pigs came from swine farms in northern Italy (Po Valley), which contains the regions of Piemonte, Lombardia, Emilia Romagna, and Veneto (Technical Appendix Figure 1). The Po Valley contains 70% of Italy’s swine. Clinical signs included watery diarrhea in sows and a death rate in piglets of 5%–10%, lower than is typical with PEDV or TGEV infections. Samples were submitted for testing by electron microscopy, PEDV ELISA, viral isolation, pan-CoV reverse transcription PCR (RT-PCR), and RT-PCR for PRCV and TGEV; selected positive pan-CoV samples were sequenced (12–14) (Technical Appendix).

Results of electron microscopy showed that 25 (49%) of the 51 samples contained CoV-like particles, but all samples were negative for viral isolation. Although only 38 samples (74%) were positive by pan-CoV RT-PCR, 47 (92%) were positive by the PEDV ELISA (Table 1) (12,13). Of the 38 pan-CoV–positive samples, 18 were selected for partial RNA-dependent RNA polymerase (RdRp), spike (S1) (14), and membrane (M) sequencing (Table 1). All samples were negative for PRCV and TGEV by RT-PCR, ruling out co-infection with PEDV and TGEV or PRCV (15).

Figure 1

Figure 1. Phylogenetic analyses of swine enteric coronaviruses in Italy. A) Analysis performed on the basis of the nucleotide sequence of the complete spike (S1) gene of 4 representative strains from the 3...

On the basis of the partial sequences from RdRp and the S1 and M genes, the strains from Italy clustered into 3 temporally divided groups, suggesting 3 independent virus entries. Cluster I represents strains circulating from 2007 through mid-2009; cluster II represents strains circulating from mid-2009 through 2012; and cluster III represents strains circulating since 2014 (Technical Appendix Figure 2, panels A–C). Cluster I was identified in Emilia Romagna (n = 1), Lombardia (n = 5), and Veneto (n = 1). Cluster II was identified in Emilia Romagna (n = 1) and Lombardia (n = 8). Cluster III was identified in Emilia Romagna at 2 swine farms. To help explain the temporal clustering, a single S1 gene segment was sequenced from clusters I and II (PEDV/Italy/7239/2009 and SeCoV/Italy/213306/2009, respectively). Because of the recent outbreak of PEDV in Europe, the 2 positive samples from cluster III (PEDV/Italy/178509/2014 and PEDV/Italy/200885/2014) were sequenced (Figure 1, panel A).

Figure 2

Figure 2. Potential recombination points in the SeCoV strains in study of swine enteric coronaviruses in Italy. The potential parent strains H16 (TGEV) and CV777 (PEDV) are shown in teal and purple, respectively....

One strain from each cluster was selected for whole genome sequencing (Technical Appendix). Unfortunately, the whole genome was obtained from only clusters I and II (PEDV/Italy/7239/2009 and SeCoV/Italy/213306/2009, respectively; Figure 1, panel B). Recombination analysis was conducted on the 2 whole genomes and was not detected in PEDV/Italy/7239/2009. Recombination was detected in SeCoV/Italy/213306/2009 at position 20636 and 24867 of PEDV CV777 and at position 20366 and 24996 of TGEV H16 (Figure 2), suggesting the occurrence of a recombination event between a PEDV and a TGEV. The complete S gene of SeCoV/Italy/213306/2009 shared 92% and 90% nt identity with the prototype European strain PEDV CV777 and the original highly virulent North American strain Colorado 2013, respectively, and the remaining genome shared a 97% nt identity with the virulent strains TGEV H16 and TGEV Miller M6. Whole-genome analysis of PEDV/Italy/7239/2009 showed that it grouped with the global PEDV strains (6) and shared ≈97% nt identity with PEDV strains CV777, DR13 virulent, and North American S-INDEL strain OH851 (Table 2).

During 2007–2014, most (92%) samples collected from the Po Valley in Italy were positive for PEDV by ELISA; only 72% were positive by pan-CoV PCR. However, because we were investigating the presence of PEDV or TGEV in samples with clinical signs of diarrhea, the high occurrence of PEDV may not reflect the actual prevalence of PEDV in Italy. The increased percentage of PEDV found in samples tested by ELISA, compared with the proportion found by PCR, may be explained by the number of ambiguous bases in the pan-CoV primers; the ambiguous bases severely reduce the efficiency of the reaction. The swine enteric CoV strains from Italy in our study, including the recombinant strain, were reported in pigs with mild clinical signs, indicating that PEDV and SeCoV have been circulating in Italy and likely throughout Europe for multiple years but were underestimated as a mild form of diarrhea.

To understand the evolution of PEDV in Italy, the partial RdRp, S, and M genes were sequenced from 18 samples and grouped in 3 different temporal clusters. Cluster I (2007–mid 2009) resembles the oldest PEDV strains; cluster II resembles a new TGEV and PEDV recombinant variant; and cluster III, identified from 2 pig farms in northern Italy in 2014, resembles the PEDV S-INDEL strains identified in Germany, France, Belgium, and the United States. The >99.3% nt identity of the S1 gene within cluster III and in previously identified strains could suggest the spread of the S-INDEL strain into Europe. However, directionality of spread cannot be determined because of a lack of global and temporal PEDV sequences.

Although our findings could indicate 3 introductions of PEDV in Italy, the results more likely suggest the high ability of natural recombination among CoVs and the continued emergence of novel CoVs with distinct pathogenic properties. Further investigation is needed to determine the ancestor of the SeCoV strain or to verify whether the recombinant virus was introduced in Italy. Recombinant SeCoV was probably generated in a country in which both PEDV and TGEV are endemic, but because the presence of these viruses in Europe is unclear and SeCoV has not been previously described, it is difficult to determine the parental strains and geographic spread of SeCoV. Future studies are required to describe the pathogenesis of SeCoV and its prevalence in other countries.