Molecular and biological characterization of poliovirus 3 strains isolated in adriatic seawater samples

In a previous study [Muscillo, M., Carducci, A., La Rosa, G., Cantiani, L., Marianelli, C. (1997a) Enteric virus detection in adriatic seawater by cell culture, polymerase chain reaction and polyacrylamide gel electrophoresis. Water Res. 31, 1980–1984] enterovirus strains were isolated from Adriatic seawater and estuarine water from the Foglia River, by infecting susceptible cells with ultrafiltrated water samples. In the present work we have studied three of those samples, in which routine reverse transcriptase–polymerase chain reaction (RT–PCR) and sequencing analysis had identified the presence of poliovirus type 3 (P3). In order to better estimate the risk to human health of such occurrence in bathing water (having bacteriological standards in line with the EEC directive 76/160), we set up a protocol to distinguish wild from Sabin P3 strains. Three sets of RT–PCR primers were engineered and their predicted products were: 593 nucleotides (nt) in the 5′ noncoding (5′NC) region (11–603), 350 nt at the Vp3–Vp1 junction (2438–2787) of the capside protein genes, and 420 nt in the 2C (4209–4628) region, which is regarded as the hotspot of recombinant polioviruses. Eight reference ATCC strains, whose sequences were known, were also tested under the same experimental conditions in order to verify the accuracy of the RT–PCR reactions. The amplicons were directly sequenced by Big-dye™ terminator sequencing using a capillary automatic sequencer. The latter two regions found the same viral species Polio 3 in all the sample strains, with no meaningful distinction between P3/Leon/37 and P3/Leon/12a1b, the vaccine strain. The analyses in the 5′NC region were more useful, where genetic relationships and the predicted secondary structure suggested that the viruses were of vaccinal sources. Molecular data were confirmed by in vitro phenotypic marker tests rct/40, where all the examined samples displayed a temperature sensitive phenotype rct/40(−). Our results suggest that the 472U→C transition alone, is not a predictive marker of reversion to neurovirulence. Finally, we conclude that the 220U constantly found in the consensus sequences of the samples can serve as a good predictor of rct/40(−) phenotype.