widely exists in natural freshwater and artificial water sources, and belongs in the class of opportunistic pathogens [1
]. The species Legionella pneumophila
, accounts for approximately 90% of human infection with the Legionella
genus that has 16 serogroups identified. Serogroup 1 (LP1) is the most common group implicated in Legionnaires’ disease. Other serogroups were also detected in past outbreaks with the second most common being serogroup 6 (LP6) [2
]. Bemander et al. studied an outbreak of L. pneumophila
in a Swedish hospital where they detected and isolated serogroups 4, 6, 7, and 10 [5
]. The main sources of infection causing outbreaks and legionellosis epidemic are water from the air-conditioning cooling system, condensed water system, cold and hot water supply systems, hot springs, and spa baths, amongst others. Possible sources of human infection are faucets, shower heads, fountains and evaporative condensers [6
The European working group for Legionella
infections (EWGLI) reported a total of 243 cases of legionellosis from 2007 to 2008, with 87 of them associated to hot and cold water supply system, thirty-two of them related to cooling tower system contamination, and 6 cases connected to fountain and spa baths [7
]. In Spain, Legionella
from the hotel spa pool and domestic water supply system were responsible for the outbreak of Legionnaires’ disease from 2011 to 2012 [8
]. In the US, engineered water systems such as cooling towers are known to be major sources for legionellosis and outbreaks [9
]. Legionellosis had also been an increasing concern in other countries. In Poland, a study carried out from 2001 to 2008 in Warsaw revealed that the frequency of Legionella
was 78% in the hospitals’ hot water systems, 68% in industrial plants, 93% in residential buildings, and 68% in hotels [6
]. In four provinces of Gabon, Africa, 29 isolates of Legionella
spp. were frequently found in hospitals particularly in hot water systems at a rate of 11.6% in a 2013 investigation [10
]. In China, since the first case of legionellosis in 1982, the disease has been reported sporadically in many cities. However, these were rarely reported as outbreaks, which could be attributed to the lack of a nationwide monitoring network of L. pneumophila
and its pathogenicity as well as incomplete epidemic data on legionellosis [11
]. Previous cases in China mostly arose from water in the cooling tower and condensation water of air conditioners. Unfortunately, L. pneumophila
is also prevalent in the shower, pipelines and hot springs waters.
Strain differentiation is necessary for the identification of sources of contamination and determination of routes of transmission in water distribution systems, which could in turn enable us to more accurately detect outbreaks and limit the spread of L. pneumophila
infections. A variety of subtyping techniques have been used for epidemiological typing, including pulsed field gel electrophoresis (PFGE) and sequence-based typing (SBT). PFGE is a highly discriminative epidemiological method for subtyping L. pneumophila
], and is the most commonly applied approach to investigate Legionnaires’ disease outbreaks and trace the source of infection [15
]. SBT could not only meet the need for distinguishing outbreak isolates but also provide a rapid, highly discriminatory, and reproducible seven-gene molecular typing method that has now become an internationally recognized procedure for genotyping L. pneumophila
In Wenzhou, located in the southeastern part of China, L. pneumophila monitoring of water in cooling towers has been going on for 7 years. However, the other possible sources of infection remained uninvestigated. Thus, the aim of this study was to investigate other water sources that could expose large number of people to L. pneumophila (e.g., hotels, hospitals and hot spring resorts). Pulsed-field gel electrophoresis (PFGE) and sequence-based typing method (SBT) were used to analyze the pathogen’s genetic characteristics. The pathogenicity was measured by intracellular growth ability of the isolates. This study is critical in guiding policies for the management of possible sources harboring L. pneumophilato prevent outbreaks of legionellosis.
2. Materials and Methods
2.1. Sample Collection
Water samples from various sources were collected from June 2015 to March 2016 in Wenzhou. For the collection of water from the cooling towers and showers, we randomly selected 20 hotels and 4 hospitals as collection points. In each hotel, 1 water sample from the cooling tower and 3 water samples (one sample from each room shower totaling 3 rooms) were collected. In each hospital, 2 or 3 water samples (one from each cooling tower) were collected per the number of hospital cooling towers. Ten to fifteen water samples from hospital wards were collected per the number of respiratory wards. For the collection of water from the hot springs, we collected water from four hot springs hotels situated in Wenzhou. Two samples were collected in one year, once in summer and another in winter; each time, 5 pools were randomly chosen for collecting water samples.
For all water sources, sodium thiosulfate (0.1 mol/L) was added in empty pre-sterile containers to inactivate chlorine. Five hundred milliliters of water were collected each time and is referred to as one sample. All procedures for collection and pretreatment of samples were based on the recommendations found in ISO 11731:1998. The samples were delivered to the laboratory for testing immediately after collection.
Prior study of L. pneumophila isolation was conducted with the same sample collection procedure. Twenty-one water samples from cooling tower of hotels and hospitals were collected from 2009 to 2014.
2.2. Legionella Isolation
Two hundred milliliters of water from each sample was filtered through a 0.45 μm membrane. The membrane was cut into pieces by sterile scissors followed by the addition of 5 mL of sterile dilution buffer, which was placed on a vortex bath for 2 min thereafter. After re-suspension, acid treatment was applied (ISO 11731:1998). Each sample (100 μL processed solution) was plated on GVPC agar (Oxoid Microbiology Products, Hampshire, UK). The plates were cultured in an incubator at 37 °C with 5% CO2 for 10 days. We examined plates and screened the colonies each day for Legionella. Legionella colonies were identified using Gram staining, the l-cysteine requirement test and serum agglutination using polyclonal antisera (Denka Seiken Co. Ltd., Tokyo, Japan), and were counted.
2.3. Real Time PCR Assay
The primers and TaqMan probes were designed per L. pneumophila
5S rRNA genes and the conservative dot
]. The primers and probe for the 5S rRNA are as shown in Table 1
The pre-screened colonies were reconfirmed upon the detection of these two genes of interest. Pure cultures of L. pneumophila
were obtained as per the reference method [18
] and the genomic DNA extraction was performed according to the manufacturer’s instruction (QIAamp DNA Mini Kit (QIAGEN, Venlo, The Netherlands)). The real time PCR mixture contained primers (0.2 μM for each primer), 0.4 μM probe, 100 ng bacteria DNA as templates and 10 μL PCR mixture (Premix Ex Taq, Takara Co. Ltd., Kusatsu, Japan) in a final volume of 20 μL. Amplification reactions were performed using a ViiA™ 7 Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA). The thermal profile was 2 min at 95 °C (activation of the TaKaRa Ex Taq HS polymerase), followed by 15 s at 95 °C, and 1 min at 60 °C for 40 cycles.
2.4. Pulsed Field Gel Electrophoresis (PFGE)
We followed a standardized PFGE protocol for L. pneumophila
subtyping as previously described [21
]. Cell suspensions were prepared in polystyrene tubes with optical density of 3.8 to 4.0 measured using a Densimat photometer (BioMe’rieux, Marcyl’Etoile, France). The cell suspensions were then embedded into 1% SeaKem Gold molten agar. Thereafter, the agar was soaked in the cell lysis buffer with proteinase K (20 mg/mL) so that the embedded cells were lysed completely. Legionella
slices were digested using AscI 20U per slice (New England Biolabs, Ipswich, MA, USA) for 4 h at 37 °C. The strain serogroup H9812 of Salmonella braenderup
was adopted as a molecular standard. Electrophoresis was run with a switch time of 6.8 s to 54.2 s for 19 h, voltage of 6.0 V/cm, and electric field angle of 120 degrees using a CHEF-DRIII system (CHEF-Mapper®
XA, Bio-Rad, Hercules, CA, USA). Images were captured using a Gel Doc 2000 system (Bio-Rad, Hercules, CA, USA) and converted to TIFF files. The TIFF files were analyzed using the BioNumerics version 7.5 software (Applied Maths, Kortrijk, Belgium). Similarity analysis of the PFGE patterns was evaluated by calculating the Dice coefficients (SD) [22
] and clustering was built by applying the unweighted-pair group method with average linkages (UPGMA). Sixty-two strains in total were selected for PFGE analyzation, 52 strains were isolated from the time period 2015 to 2016 by PFGE, while 10 strains were isolated and identified previously, from the time period 2009 to 2014.
2.5. Sequence-Based Typing
Seventy-three strains were analyzed by SBT: Fifty-two strains were isolated during 2015 to 2016 from various water sources, and 21 strains were sporadically isolated previously from 2009 to 2014 from the cooling tower in hotels and hospitals. The European Working Group for Legionella
Infections (EWGLI) suggests seven genes (flaA
) for genotyping using the standard sequenced-based typing (SBT) method; the 5.0 version can be found here: http://www.hpa-bioinformatics.org.uk/legionella/legionella_sbt/php/sbt_homepage.php
. The genomic DNA extraction method and the PCR procedure were according to the Sequence-Based Typing protocol for epidemiological typing of Legionella pneumophila
version 5.0. The amplicons were sent for sequencing by Sangon Bio-technique Company (Beijing, China). The SBT database that was available on the EWGLI website (http://www.ewgli.org/
) was used for nucleotide analysis, and the sequences were compared with those in the SBT database, which were also available on the website. BioNumerics software version 7.5 (Applied Maths, Brussel, Belgium) was applied to cluster analysis of related sequence types (STs), representing their genetic relationships.
2.6. Intracellular Growth Assay
Ten isolated Legionella
strains from different sources and different SBT types were selected and subjected to an intracellular growth assay repeated three independent times with L. pneumophila
serogroup 1 strain (ATCC33152) as a positive control. All the strains were added to J774 cells, and the bacteria were stained using the Gimenez stain method after 0, 1, 2 and 3 days of infection and counted [23
]. J774 cells were incubated in RPMI 1640 tissue culture medium including 10% calf serum at 37 °C with 5% CO2
. Phosphate buffer saline (PBS) was used to dilute the bacteria to 1 × 108
CFU/mL to obtain the primary dilution, which was then 10-fold diluted with culture medium containing J774 cells (2 × 105
CFU/mL) to reach a multiplicity of infection (MOI) of approximately 10. Twenty-four-well dishes containing culture medium and bacteria were placed in the incubator at 37 °C, 5% CO2
, and saturated humidity conditions. For the continuous culture, the culture medium was replaced by fresh medium each day. At days 0, 1, 2, and 3, we counted intracellular bacteria CFUs to determine growth. To measure internalization, PBS was used to wash away the extracellular bacteria and other potentially interfering substances. One milliliter of sterile, distilled water was added to the wells to release internal bacteria from the host cells, and the CFUs were determined by plating dilutions on BCYE agar plates (Oxoid Microbiological Products, Hampshire, UK).
2.7. Statistical Analysis
SPSS 19.0 for Windows (SPSS Inc., Chicago, IL, USA) was used to analyze the data. The chi square test or Fisher exact test were used for analyzing qualitative data, while the MANOVA of repeated measures or Mann-Whitney test were used to analyze quantitative data. The bacterial concentrations of L. pneumophila in the intracellular growth assay were analyzed by MANOVA of repeated measures. The differences were considered statistically significant when p < 0.05.
This study suggests a high prevalence of Legionella
in Wenzhou’s hot springs. Legionella
was detected in more than 60% of collected hot springs water samples both in summer and winter, and the serogroups of LP1 and LP3 mostly accounted for the group of detected Legionella
spp., which in fact are the two main serogroups causing Legionnaires’ disease. The viable count assay revealed the high abundance of Legionella
in some hot springs. The concentration of Legionella
was observed to be over 1000 CFU/100 mL in 4 samples. Among them, 1 sample had up to 10,720 CFU/100 mL. When above normal levels or an increase in Legionella
spp. number greater than 100 CFU/100 mL of water is calculated, policies should be enforced to eradicate the bacteria from the contaminated water source [24
]. To date, the infective dose has not yet been precisely determined. Estimated data indicates that in the case of water contamination by Legionella
CFU/100 mL may lead to the disease occurring sporadically, but when the counts exceed 106
CFU/100 mL, a Legionnaire’s disease outbreak can be expected [25
]. In this study, although only one hot spring sample had enough L. pneumophila
(10,000 CFU/100 mL) to potentially cause sporadic disease, three other hot spring samples had elevated amounts of the bacteria. Although the 2011 tourism industry standard of China, Classification and accreditation for service-rated hot springs enterprise, claims that L. pneumophila
has not been detected in the hot springs, the detection range was not being defined and the standard detective method was not established, indicating that the industry standard for detection of L. pneumophila
in China should be improved.
Shower water may carry pathogens that come into contact with people, thus is directly related to a person’s health. The Legionella-positive rate of hotel and hospital showers was 15.5%. Currently, the cleaning regulations and monitoring standards for L. pneumophila in central air conditioning of public areas are made available, however, disinfection and detection standards for shower systems have not been set up. The concentration of the bacteria in Legionella-positive shower samples ranged from 20 to 100 CFU/100 mL. Shower water originates from hot water shower tanks and are piped to each room’s showers, therefore, once the bacteria colonizes in tanks or pipes, it would spread to every part of the building. All the isolated strains from the hospital showers were serogroup LP1; the situation was very different from that of hotels where the most common serogroups were LP1, LP5, and LP6.Cancer patients, postoperative patients, and patients under long-term immunosuppressive treatment that are more susceptible to infection can be at risk once L. pneumophila invades the hot water system.
The PFGE results revealed that the patterns of isolates from different rooms with tin the same collection point were not different, suggesting that they share the same pipe system. It was observed that some L. pneumophila strain patterns were detected in certain hospitals’ and hotels’ showers in different years, indicating that the same strain patterns may be persisting for a long time in those particular water sources. However, the patterns of isolates from different hot springs resorts varied widely, with some samples from the same collection point exhibiting several different strain patterns.
SBT results showed that the population of STs was highly diverse in the Wenzhou area. Fifty-two isolates from the year 2015 to 2016 were divided into 18 STs, including ST7, ST9, ST87, ST114, ST1226, ST1230, and ST1469. Isolates from 2009 to 2016 were divided into 25 STs, 4 clonal groups and 4 singletons. ST1226 was the most common type found in local hot springs and also existed in the cooling tower over the years. ST1 was the most dominant ST-type of L. pneumophila
in the cooling tower from 2009 to 2014 and this result was consistent with ST-type distribution of L. pneumophila
in the cooling tower in Shijiazhuang (China) [28
]. However, from 2015 to 2016, the ST1 type was not detected, but ST7, which differs only in one allele from ST1, was observed instead.
According to the SBT database, the strains of ST87, ST7, ST114, ST9, and ST961 isolated in our study previously caused Legionnaires’ diseases in some countries. ST87 was detected in all three hot spring resorts and their serogroups were all LP3, which agrees with the strain noted in the SBT database and was also reported to be present in a hot springs resort in Beijing [23
]. 188 strains of ST9 were available in the SBT database (last accessed: 18 May 2016), with 125 of them isolated from clinic samples, which corresponds with one ST9 strain we isolated from the cooling tower in this study. ST7 and ST114 were detected in shower water samples with the serogroups LP1 and LP6, respectively, of which the results are consistent with the available strains in the SBT database. Nine strains of ST type (from ST2196 to ST2205) specific to the Wenzhou area were first detected in this study; of them, 7 strains including ST2196 and ST87 belong to one clonal cluster. However, the newly isolated ST type ST2197, is in the same clonal cluster belonging to ST2199 and ST114.
It was found that the serogroup and ST type of L. pneumophila from the hotels’ and hospitals’ showers were almost unchanged with that of the same sampling points of recent years. Although the PFGE types of strains from two hotels varied within the recent years, the similarities within these strains were up to 91.2%. The results may be due to the higher resolution of PFGE for typing bacteria groups vs. the capacity of multilocus sequence typing (MLST). Some mutations may have occurred during the bacteria’s generational propagation and caused the PFGE variations that were observed, but may not be detected by MLST. Thus, PFGE and SBT are complementary approaches for group differentiation of L. pneumophila.
The pathogenic mechanism of L. pneumophila
is that the bacteria has the ability to invade into macrophages of the host’s lung and reproduce until the cells collapse and release cellular contents to ultimately initiate cascades of immune reactions leading to lung damage. Thus, the intracellular growth ability assay could be an effective way for measuring the pathogenicity of L. pneumophila
]. We studied the intracellular growth ability of isolates from artificial water sources for investigating their pathogenicity. All the isolates from all the various water sources: hot springs, cooling tower, and showers revealed that they could infect mouse J774 macrophages, indicating a possible high risk for people living in the Wenzhou area to contract Legionnaires’ disease.