Efficacité des méthodes d'adsorption-élution utilisant la poudre ou la laine de verre pour la concentration des virus dans les effluents de stations d'épuration Comparative efficiency of glass powder or glass wool adsorption-elution methods for virus concentration in effluents of sewage treatment plant

Résumé

Pour rechercher les virus dans les eaux usées traitées, une nouvelle méthode d'adsorption-élution sur laine de verre a été appliquée comparativement à la méthode d'adsorption-élution sur poudre de verre. Lorsque la technique de concentration sur laine de verre est utilisée, c'est dans les 25 premiers ml de l'éluat que la majorité des virus poliomyélitique est retrouvé (89 à 94 %). La comparaison des méthodes de concentration des virus indigènes à partir d'échantillons d'effluents provenant de deux stations d'épuration biologique de la Côte d'Azur (Cagnes-sur-Mer et Nice), a mis en évidence la supériorité de cette nouvelle méthode : les taux de positivité ont été respectivement de 85 % vs 38 % pour l'effluent de Cagnes-sur-Mer et 100 % vs 44 % pour l'effluent de Nice. De même, les titres en virus indigènes après concentration ont varié de 0 à 250 NPPUC/l pour la méthode sur laine de verre contre 0 à 25,5 NPPUC/l pour la méthode sur poudre de verre. La différence constatée entre les méthodes est statistiquement significative après analyse de variance (p = 0,0119 pour l'effluent de Cagnes-sur-mer et p < 0,0001 pour l'effluent de Nice). De plus, la technique sur laine de verre ne nécessite ni l'abaissement du pH, ni le changement de la composition ionique de l'échantillon d'eau à analyser.

Mots-clés:

• Méthodes,
• concentration,
• virus,
• laine de verre,
• poudre de verre,
• effluent

Abstract

Biological treatment of sewage in waste water plants does not allow elimination of the whole of the microbial load. Discharge of the treated sewage results in viral pollution of river, lakes and seas, a potential hazard for the health that has to be monitored. The amont of virus in waste water beeing low, concentration from the samples brought to the laboratory is rendered necessary. The aim of this study was to evaluate the efficacy of a new adsorption-elution method on glass wool to recover indigenous viruses from effluents of the cities of Cagnes and Nice (Alpes-Maritimes, France). In order to evaluate its efficiency we compared it to the regular adsorption-elution method on glass powder. As a preliminary we determined upon artificially contaminated 5 liter waste waters samples what detection of virus could be performed only in the first 25 ml of the 100 ml eluate, as in the glass powder concentration method. Results show chat virus titers found in that first fraction of eluate were close to those in the total sample. Thus from 3 samples containing 1.60 10⁸ MPNCU/5 l, 1.96 10⁷ MPNCU/5 l and 4.32 10⁴ MPNCU/5 l we found in that first fraction respectively 1.50 10⁴ MPNCU/5 l (94 %), 1.80 10⁷ MPNCU/5 l (92 %) and 3.85 10⁴ MPNCU/5 l (89 %); these recovery rates are not significantly different by comparison of confidence limits. The glass powder method, necessitates preliminary treatment of the sample : acidification to pH 3.5 and adjunction of AICI3 at a final concentration of 5.10^-4 M. After flowing the acidified sample through 100 g of borosilicated glass powder at a rate of 10 l/10 min inside a decantation ampulla. Then adsorbed virus may be eluted from the sedimentated glass powder with 100 ml of borate buffer containing 3 % beet extract, pH9 : the first 25 ml were collected into a flask containing 2.5 ml of a mixture of antibacterial and antifungal antibiotics. For the glass wool adsorption method, a 19 cm³ cartridge was packed with 5 g sodocalcic glass wool at a 0.4 g/cm³ density and rinced sequentially with : 10 ml 1N HCl, 10 ml deionised water, 10 ml 1N NaOH and lastly 40 ml deionised water. It was balanced with 200 ml deionised water. The sample, was pumped at a flow rate ca 10 l/h. Enumeration of viruses was performed by inoculating 40 microplate wells containing KB cells, and performing 3 passages 5 days each, after which the number of wells presenting with CPE was determined. This characteristic number allowed calculation of the most probable number of cytopathic units (MPNCU) with the 95 % confidence limit. The Box and Cox analysis of transformation was applied to the data. Since the calculated value of λ approximated zero (λ = - 0.29 for the Cagnes effluent and λ = - 0,062 for the Nice effluent), transformation of the gross data into logarithm was justified. To allow this transformation, the zero had to be substituted for by a value equal to half the limit sensitivity of the method (I well out of 40), i.e. 0.5. Distribution of the data being roughly log-normal, it was then possible to compare the results of the two methods by two-way analysis of variance, cross classification, without replica. The test for factor method was calculated according to the interaction since this factor is fixed. Overall it appeared that all 31 10-liter samples analysed contained viruses when results from bath methods were combined. Still no single method allowed virus recovery in a 100 % of cases, however the glass wool adsorption method found viruses in 29/31 vs 13/31 with the glass powder method. The new method detected virus in 11/13 (85 %) samples from Cagnes waste waters as well as in 18/18 (100 %) from Nice. Quantitative analysis of the viral titers indicates that, titers were higher following the glass wool adsorption method than following the glass powder adsorption method in 11/13 samples from Cagnes treatment plant and in 17/18 from Nice. Thus virus concentrations varied between 0 and 250 MPNCU/l (MGT= 4.6 MPNCU/l) for the Cagnes effluent and between 2 and 60 MPNCU/l (MGT= 7.5 MPNCU/I) for the Nice effluent. For the same samples virus concentrations obtained following glass powder adsorption method varied between 0 and 8.5 MPNCU/l (MGT 0.9 MPNCU/l) for the Cagnes effluent and between 0 and 25.5 MPNCU/l (MOT= 1.3 MPNCU/l1) for the Nice effluent. This difference is statistically significant (p = 0.0119 for the Cagnes effluent and p < 0.0001 for the Nice effluent). Furthermore, when taking into account the origin of the waters analysed, comparison between observed F_0.95 (7.94 for the Cagnes waters and 45.78 for the Nice waters) and theoretical F_0.95 (4.75 for the Cagnes waters and 4.45 for the Nice waters) leads to the rejection of the hypothesis of identity of the two methods. The discordances observed are an illustration of the fact that concluding to the absence of viruses in a given sample is a matter of method and should be interpreted with prudence. A few drawbacks inherent to the glass powder adsorption method may explain its poorer efficiency : the necessary acidification of the sample to pH 3.5 may be fatal to a proportion of virions; also the flow rate necessary to maintain the fluid layer of glass powder in suspension during the adsorption step is 6 fold higher than that required in the glass wool method (60 l/h vs 10 l/h). Finally the nature of the adsorbing material, sodocalcic vs borosilicated, may be determinent. We can conclude from the present comparative study, to the statistically significant superiority of the glass wool method for virus concentration from treated waste waters.

Keywords:

• Methods,
• concentration,
• virus,
• glass wool,
• glass powder,
• effluent