Article pubs.acs.org/est
Your Garden Hose: A Potential Health Risk Due to Legionella spp. Growth Facilitated by Free-Living Amoebae Jacqueline M. Thomas,*,†,⊥ Torsten Thomas,‡ Richard M. Stuetz,† and Nicholas J. Ashbolt†,§,∥ †
School of Civil and Environmental Engineering, The University of New South Wales, Sydney, NSW 2052, Australia Centre for Marine Bio-Innovation and School of Biotechnology and Biomolecular Sciences, The University of New South Wales, Sydney, NSW 2052, Australia § National Exposure Research Laboratory, Office of Research and Development, United States Environmental Protection Agency, 26 West Martin Luther King Drive, Cincinnati, Ohio 45268, United States ‡
S Supporting Information *
ABSTRACT: Common garden hoses may generate aerosols of inhalable size (≤10 μm) during use. If humans inhale aerosols containing Legionella bacteria, Legionnaires’ disease or Pontiac fever may result. Clinical cases of these illnesses have been linked to garden hose use. The hose environment is ideal for the growth and interaction of Legionella and free-living amoebae (FLA) due to biofilm formation, elevated temperatures, and stagnation of water. However, the microbial densities and hose conditions necessary to quantify the human health risks have not been reported. Here we present data on FLA and Legionella spp. detected in water and biofilm from two types of garden hoses over 18 months. By culturing and qPCR, two genera of FLA were introduced via the drinking water supply and reached mean densities of 2.5 log10 amoebae·mL−1 in garden hose water. Legionella spp. densities (likely including pathogenic L. pneumophila) were significantly higher in one type of hose (3.8 log10 cells·mL−1, p < 0.0001). A positive correlation existed between Vermamoebae vermiformis densities and Legionella spp. densities (r = 0.83, p < 0.028). The densities of Legionella spp. identified in the hoses were similar to those reported during legionellosis outbreaks in other situations. Therefore, we conclude that there is a health risk to susceptible users from the inhalation of garden hose aerosols.
■
FLA hosts15,16 and replicate, and at least some emerge more virulent to humans.17 FLA also enhance intracellular pathogens’ environmental survival, particularly when the host’s cyst form protects associated bacteria from chlorine disinfection,18 desiccation,19,20 and exposure to ultraviolet radiation.21,22 ARB isolated directly from FLA in drinking water include pathogenic Legionella spp.,6,23 Mycobacterium spp.,6 and Pseudomonas spp.24,25 However, there are also likely to be many non-human-pathogenic ARB, including unculturable Legionella spp.26 FLA and opportunistic pathogens present in drinking water are aerosolized during events such as showering, spa baths, and drinking from water faucets, as well as by cooling towers and humidifiers.27,28 Exposure models have established the health risks of inhalation of pathogenic Legionella in aerosols ≤10 μm from water systems generally,29 following an outbreak at a public spa,30 and while showering.31 A garden hose is a very common household item that generates aerosols during use that can be inhaled.32 Garden
INTRODUCTION Certain pneumonia and respiratory tract infections are caused by the inhalation of water droplets (aerosols) containing waterbased bacterial pathogens from the genera Legionella,1,2 Mycobacterium,3,4 other amoebae-resisting bacteria,5,6 and possibly amoebal viruses.7 Legionellosis (Legionella infection) alone accounts for 29% of all reported disease outbreaks from drinking water in the United States since it became a reportable disease in 2001. The elderly are most at risk, and, together with nontuberculosis mycobacterial infections that require hospitalizations, legionellosis cases account for the majority of hospital costs associated with drinking water disease.8 Legionella and Mycobacterium can survive and parasitize eukaryotic microorganisms, particularly free-living amoebae (FLA).9 FLA are thought to be one key biological factor that allows water-based pathogens to grow, break through drinking water treatment processes,10 and avoid residual disinfectants.11,12 Furthermore, FLA are ubiquitous in engineered drinking water systems across the world,13 and a subset may themselves cause human infections, including nonfatal amoebic keratitis and potentially fatal amoebic meningoencephalitis.14 The focus of concern in the current study is FLA’s ability to host a large range of intracellular bacterial pathogens.9 Amoebae-resisting bacteria (ARB) avoid digestion by their © 2014 American Chemical Society
Received: Revised: Accepted: Published: 10456
June 3, 2014 July 29, 2014 July 30, 2014 July 30, 2014 dx.doi.org/10.1021/es502652n | Environ. Sci. Technol. 2014, 48, 10456−10464
Environmental Science & Technology
Article
hoses have already been linked to cases of legionellosis.33,34 The garden hose of 60-year-old male diagnosed with L. pneumophila pneumonia tested positive for L. pneumophila serogroup I antigen and was the likely source of infection, as the man had drunk from it and hosed his head.33 In an outbreak of legionellosis in a single family (four adults), identical L. pneumophila serogroup I antigens from patients were detected in two garden hose samples, the garden spa, and the garden shower of the private dwelling.34 A survey of 12 garden hoses not related to clinical cases detected L. pneumophila serogroup I antigen at a rate of 50%.33 A garden hose presents an environment for Legionella growth due to biofilm formation35,36 on its substantial inner surface (a hose of 20 m length and 12.5 mm internal diameter possesses 7.9 × 104 cm2 of internal surface area). Between uses, garden hoses may be filled with warm, stagnant water that is likely to further encourage the growth of L. pneumophila.37 While limited literature could be identified for the presence of waterbased pathogens in garden hoses, other types of hoses in medical settings have shown increased densities of FLA38 and Legionella spp.39,40 There is also an absence of literature describing the microbial ecology and environmental parameters that facilitate water-based pathogen growth in garden hoses. Here we present results on the density and diversity of FLA and Legionella in the water and biofilm of two common garden hose types supplied with treated municipal drinking water. Isolated FLA were then screened for natural infections with Legionella and their ability to facilitate the growth of L. pneumophila (ATCC 33155 and ATCC 33215).
taken for each hose system: supplying drinking water, hose sections (for biofilm), and water from the hose end (20 m final total length). Water samples (50 mL) were collected at flow rates of 11−15 L·min−1 from the tap and hose ends in sterile tubes (Sarstedt, Numbrecht, Germany). Hose sections were collected by cutting a 30 cm length of hose from the inlet end. The outside of each hose piece was disinfected (0.1% sodium hypochlorite, Univar, Downers Grove, IL, USA) and drained of any residual free water, and the ends were sealed with film (Parafilm M, Pechiney Plastic Packaging, Chicago, IL, USA) prior to transportation within a cooler box. Samples were transported to the laboratory within 2 h for processing. Hoses were sectioned into 10 cm lengths and cut in half. Biofilm was removed by scraping with sterile forceps and rinsing with 10 mL of sterile filtered water (Milli-Q, Millipore, Billerica, MA, USA) into individual 50 mL tubes (Sarstedt) as previously described.41 Physical scraping of the biofilm was done because it is reported to be effective in the removal of FLA from environmental samples.42 Heterotrophic plate counts (HPC) were conducted using 0.5 mL aliquots according to the standard (AS/NZS 4276.3.1:2007).43 FLA Detection by Culture. For detection of FLA, 0.5 mL aliquots of samples were plated (in duplicate) onto nonnutrient agar (1% w/v, Oxoid, Cambridge, UK) plates with overlays of Escherichia coli as a food source as previously described.44 Plates were sealed (Parafilm M) and incubated at room temperature (22 °C) out of direct sunlight. Plates were examined daily for the first 7 days for plaque formation and then once every 7 days for 30 days total. Any plaques were confirmed microscopically to be FLA and then subcultured onto fresh plates. Control Cultures. For details of FLA and Legionella control cultures, see the Supporting Information (SI), Table S1. FLA supplied by Australian Water Quality Centre (Adelaide, Australia) were xenic cultures, while those from the American Type Culture Collection (ATCC, Manassas, VA, USA) were supplied as axenic cultures. DNA Extraction from Cultures. DNA was extracted from cultures of FLA for PCR identification and/or screening for L. pneumophila. FLA were picked from plates using a sterile loop and then lysis and DNA extraction using InstaGene Matrix (Bio-Rad, Hercules, CA, USA) as per the manufacturer’s instructions. PCR of FLA. FLA isolated by culture were identified using published PCR primers targeting the 18S rRNA gene45 (see SI, Table S2). All FLA PCR reactions were performed using a realtime PCR machine (Chromo, Bio-Rad). Reactions contained 5 μL of template DNA, 0.1 μM each of forward and reverse primers, 2.5 U of Taq polymerase, 0.2 mM each of dNTPs, 3 mM MgCl2 in final PCR buffer concentration of 100 mM KCl, and 40 mM Tris-HCl at pH 8.4 (Invitrogen, Carlsbad, CA, USA), which was made up to 50 μL with nuclease-free water (Gibco, Invitrogen).45 DNA extracts, negative controls (DNA replaced with nuclease-free water), and positive controls (from AWQC control cultures, see SI, Table S1) were run in duplicate for every sample assayed. DNA Extraction from Samples. All samples from 6, 9, and 18 month time points were taken for molecular analysis. DNA was extracted from 10 mL water samples and resuspended scraped hose biofilms. Samples were concentrated by centrifugation at 3270g for 15 min at 22 °C (Allegra-X-12R, Beckman Coulter, Brea, CA, USA), and the cell pellet was resuspended in 1 mL of supernatant. DNA was extracted using
■
MATERIALS AND METHODS Experimental Setup. Two brand-new plastic garden hoses, 20 m in length, were attached for 18 months to two outdoor drinking water taps of a commercial building in New South Wales, Australia. Two different hose types were used: garden hose A was a standard 12.5 mm internal diameter (ID) green and black nylon braided poly(vinyl chloride) (PVC) plastic (Pope, Beverly, Australia), while hose B was a 12.5 mm ID purple PVC plastic garden hose recommend for use with recycled water (Dual Irrigation, Dural, Australia). The hoses were connected with standard brass fixtures, and hand valves sealed the end (see the Abstract/Table of Contents graphic). The hoses were operated under conditions typical of garden hoses: used intermittently, left filled with water, and exposed to direct sunlight as in normal use for a period of 18 months. Water Quality Analysis. Water quality samples (200 mL) were taken from the taps supplying the garden hoses and from the end of the hose setups at least once every 2 months. During the summer months, sampling occurred every 2 weeks. Samples were collected post meridiem but before 5 p.m. Water quality parameters analyzed and the instruments used are as follow: temperature, hand-held LCD multi-thermometer (China); pH, Aqua-pH pH meter (TPS, Springfield, Australia); turbidity, 2100N turbidimeter (Hach, Loveland, CO, USA); and total organic carbon, combustion oxidation with a TOC-5000A analyzer (Shimadzu, Kyoto, Japan). Total phosphate, total nitrogen, and free and combined chlorine were measured using Spectroquant cell test methods with a Nova 60 photometer as recommended by the manufacturer (Merck, Frankfurt, Germany). Water and Biofilm Sampling. The tap water supply and garden hoses were sampled five times during the experiment: after 6, 7, 8, 9, and 18 months). Three different samples were 10457
dx.doi.org/10.1021/es502652n | Environ. Sci. Technol. 2014, 48, 10456−10464
Environmental Science & Technology
Article
FLA PCR products were sequenced directly after purification without cloning. DNA sequence-labeling PCR was performed using a PCR instrument from Bio-Rad as per the supplied protocol (see SI, Table S2).53 Sequencing was performed on a capillary electrophoresis DNA sequencer (3730, Applied Biosystems) at the Ramaciotti Centre for Gene Function Analysis, University of New South Wales. Unique sequences of sufficient length (>200 bp) were submitted to GenBank at the National Center for Biotechnology Information (Bethesda, MD, USA). DNA Analysis and Phylogenetic Trees. DNA sequences were trimmed to remove primer sequences (Vector NTI, Invitrogen) and searched against the NCBI database using the basic local alignment search tool (BLASTN). The new sequences and selected sequences with >95% similarity were aligned against the SILVA reference database,54 and phylogenetic trees using the neighbor joining algorithm were constructed using ARB software. 55 For the Legionella phylogenetic trees, 50 DNA sequences in total were selected (Legionella spp. and Legionella-like amoebal pathogens), using the bacterium Chlamydophila pneumoniae (GenBank accession number Z49873) as an outgroup. FLA Uptake of Legionella pneumophila. Fresh cultures of L. pneumophila (ATCC 33152 and ATCC 33155) cells were concentrated (2000g for 15 min) and then resuspended in phosphate-buffered saline (PBS, Univar) and stained with 10 μM fluorescent cell tracing dye (Vybrant CFDA SE Cell Tracer Kit, Invitrogen) for 15 min at 37 °C. Stained L. pneumophila cultures were individually incubated with hose B biofilm isolate (V. vermiformis SDB7) and control FLA cultures. Each well of a 24-well tissue culture plate (Sarstedt) contained 500 mL of sterile water (Milli-Q) and 1 × 106 stained L. pneumophila cells mixed with 2 × 105 fresh FLA (ratio of 5:1). Duplicate treatments were run with controls of L. pneumophila, FLA, or water only. Plates were incubated at room temperature (22 °C) in the dark for 7 days. Six samples (10 μL) were taken every 1− 2 days from each well. Cells were enumerated using an improved Neubauer hemocytometer (Cole-Parmer, Vernon Hills, IL, USA) at 400× magnification with a fluorescent microscope (DM400B, Lieca, Wetzlar, Germany, or Standard 2S, Zeiss, Oberkochen, Germany) using a green-filter cube (I3 or FT580). In total, the whole experiment was repeated three times. FLA containing stained L. pneumophila were imaged using a FITC filter and differential interference contrast (DIC) on a confocal scanning laser microscope (CSLM, Fluoview FV1000, Olympus, Tokyo, Japan) at the Biomedical Medical Imaging Facility, University of New South Wales. Images were taken in three dimensions at 1000× magnification using an oil immersion lens. Statistical Analysis. All statistical analyses were done and graphs produced using the statistical software program Prism (GraphPad Software, La Jolla, CA, USA). Tests for normalcy were conducted,56 and non-Gaussian-distributed data were transformed to log10 and then tested again. Water quality parameters, culture FLA, and Legionella qPCR were analyzed using Student t test (two data sets) or one-way and two-way ANOVA tests (three or more data sets). Non-Gaussiandistributed data (qPCR FLA quantification) were analyzed using Mann−Whitney test (two data sets) or Kruskal−Wallis test (three or more data sets). Correlation was performed using the Pearson or Spearman’s rank correlation for Gaussian and non-Gaussian data, respectively. All results were considered statistically significant at the level of p ≤ 0.05.
the FastDNA SPIN Kit for Soil as per the manufacturer’s instructions using a FastPrep FP120 machine (Bio101, Qbiogene, MP Biomedicals, Solon, OH, USA). DNA was quantified on the basis of its specific absorbance at 260 nm (A260) on a spectrophotometer (NanoDrop 2000, Thermo Scientific, Wilmington, DE, USA).46 DNA concentrations were used as a measure of biofilm quantities.47 Quantitative Polymerase Chain Reaction (qPCR). Previously published qPCR methods were used to identity Acanthamoeba spp.,48 Vermamoeba vermiformis49 (previously named Hartmanella vermiformis50), Naegleria spp.,51 and Legionella spp.52 All FLA reactions were performed using the real-time PCR machine Chromo (Bio-Rad), while all Legionella reactions were performed using the real-time PCR machine ICycler iQ Multi-Color (Bio-Rad). Reactions contained 5 μL of template DNA, 0.2 μM each of forward and reverse primers, and 12.5 μL of qPCR master mix (iQ SYBR Green Supermix or SYBR green PCR master mix, Applied Biosystems, Life Technologies, Carlsbad, CA, USA), and volume was made up to 25 μL with nuclease-free water (Gibco or Fisher Scientific, Pittsburgh, PA, USA). DNA extracts, negative controls (DNA replaced with nuclease-free water), and positive controls (from standards) were run in duplicate for every sample assayed. A melt curve analysis of the products was performed from 75 to 95 °C with 20 s holds and a plate read at every 0.5 °C increment. All qPCR products were visualized by gel electrophoresis and a selection sequenced (see below) to confirm specificity. qPCR Standards and Efficiencies. Cell-based calibration curves of FLA control cultures (Acanthamoeba sp. AC362, V. vermiformis HM061, and N. lovaniensis NG1020) were generated by 8-fold serial dilutions of known concentrations of whole cell (trophozoite) cultures (see SI, Figures S1 and S2). For all curves, the linear regression and the goodness-of-fit (r2) were ≥0.99. The limit of quantification (LOQ) was 0.1 amoeba per reaction. For L. pneumophila, cell-based calibration curves were generated from an 8-fold serial dilution from fresh growth cultures of two different L. pneumophila (ATCC 33155 and ATCC 33215) (see SI, Figure S3). Using linear regression, the goodness-of-fit was calculated (r2 = 0.91). The LOQ was 5 Legionella cells per reaction, which occurred at the maximum cycle threshold (Ct = 40). Recovery efficiencies during upstream processing (sample processing and DNA extraction) were calculated at >80% using spiked drinking water samples. Any qPCR inhibition with undiluted DNA extracts was low (
