Legionella spp. isolation and quantification from greywater

Graphical abstract


Method details
There are several ISO methods for Legionella isolation from water [1,2]. However, none of them is suitable for Legionella isolation from GW samples due to the fact that GW is contaminated with a very high bacterial load. Thus, almost no data can be found regarding Legionella presence in GW.
Here we describe a modification of the ISO 11731:1998 protocol for Legionella isolation from GW. Our modified protocol allows the isolation of Legionella on GVPC selective Legionella medium (glycinevancomycin-polymyxin-cycloheximide medium, Beckton Dickinson GmbH, Heidelberg, Germany) without the massive bacterial contamination that develops on the media when ISO 11731:1998 is applied.

Pre-filtration of the greywater sample
Filter a 100 ml GW sample (pre-filtration) to remove coarse matter, using a 100 mm pore size cell strainer (Becton Dickinson, USA) placed in one 50 ml tube (two 50 ml tubes are needed) (Fig. 1).

Filtration of the GW sample
The 100 ml pre-filtered GW sample is filtered again through a 0.2 mm cellulose nitrate filter using a vacuum filtration system attached to a 2511 Dry Vacuum Pump (WELCH, Germany) (Fig. 2).

Combined acid-thermal treatment
Each sample is then subjected to a combined acid-thermal treatment as follows: 1 ml of the sample is centrifuged at 6000 Â g for 10 min. For the acid treatment, 0.5 ml of the supernatant is replaced with 0.5 ml of acid buffer (HCl 27 mM; KCl 173 mM, pH 2.2). The sample is then vortexed and immediately subjected to thermal treatment for 30 min at 50 8C.

GVPC Legionella media inoculation
Following the ISO 11731:1998 recommendations, two 0.5 ml sub-samples are plated on a GVPC Legionella selective media immediately after the thermal treatment. The plates are incubated at 37 8C. Presumptive Legionella colonies are counted after 7 and 15 days of incubation.

Legionella identification
The presumptive Legionella colonies are then identified using a Legionella latex test (Oxoid, Basingstoke, UK). This test allows separate identification of Legionella pneumophila serogroup 1 and  Method validation: efficiency of the Legionella isolation protocol and limit of detection (LOD) We used the method described above to successfully isolate and quantify Legionella along a one year GW monitoring campaign. The results of this study have been already published [3]. Briefly, a total of 16 greywater samples were analyzed. Legionella was isolated from 81% of the samples, with a mean of 1.2 Â 10 5 [ 1 _ T D $ D I F F ] cfu/l. Details about the efficiency and the limit of detection of this method can also be found in the mentioned publication. This method is highly aggressive, so the recovery rates of Legionella were very low (2.5%, SD = 1.5%) and the LOD established from this average recovery rate was 4.0 Â 10 3 cfu/l. Nevertheless, the results were consistent. It should be noted that this modified methods is the only way to isolate Legionella from GW, as using the current ISO protocols does not allow the isolation of this bacteria. [ ( F i g . _ 2 ) T D $ F I G ]

Recommendations
This method is highly aggressive for the sampled bacteria, including Legionella. For that reason, the LOD of the method is high and the efficiency of Legionella isolation is low. We recommend using this method only with problematic samples in which Legionella can't be isolated using the methods described in the ISO protocols 11731:1998 and 11731-2:2004 [1,2] due to massive contamination with other bacterial species.