Article pubs.acs.org/est
Detecting Human Bacterial Pathogens in Wastewater Treatment Plants by a High-Throughput Shotgun Sequencing Technique Lin Cai and Tong Zhang* Environmental Biotechnology Laboratory, Department of Civil Engineering, The University of Hong Kong, Hong Kong SAR, China S Supporting Information *
ABSTRACT: Human pathogens are one of the major threats to global public health. Wastewater treatment plants (WWTPs) serve as city guts to receive and digest various human pathogens. Several techniques have been developed to detect human pathogens in WWTPs and to assess potential environmental risks. In this study, we employed 24 metagenomic DNA data sets derived from a high-throughput shotgun sequencing technique to more accurately and efficiently detect human bacterial pathogens in influent, activated sludge, and effluent of two Hong Kong WWTPs. Each data set was quality-filtered and normalized to 12 500 000 DNA sequences with a length of 150−190 bp. Then, a BLASTN search against Greengenes general 16S rRNA gene database and human pathogenic bacteria 16S rRNA gene database, a BLASTX search against human pathogenic bacteria virulence factor database, as well as MetaPhlAn analysis were conducted to survey the distribution, diversity, and abundance of human bacterial pathogens. The results revealed that (i) nine bacterial pathogens were detected; (ii) the overall pathogenic bacteria abundance was estimated as 0.06−3.20% in the total bacteria population using 16S rRNA gene fingerprinting; (iii) pathogenic bacteria detected in activated sludge and effluent shared similar profiles but were different from influent based on both 16S rRNA gene and virulence factor fingerprintings; (iv) Mycobacterium tuberculosis-like species may present potential pathogenic risks because it was detected with high abundance in both activated sludge and effluent. These findings provided a comprehensive profile of commonly concerned human pathogens in two Hong Kong WWTPs and demonstrated that the high-throughput shotgun sequencing technique is a feasible and effectual approach for environmental detection of human bacterial pathogens.
■
INTRODUCTION Human pathogens are disease-causing agents such as viruses, bacteria, fungi, protozoa, and helminths, which impose major threats to public health worldwide.1,2 Pathogens in wastewater are mainly derived from faecal matter, most of which enters the environment through domestic wastewater contaminated by infected people.3,4 Bacterial pathogens are frequently detected in wastewater because they are able to multiply rapidly under suitable growth conditions. For example, many people died of cholera in London during the middle of the 19th century because Vibrio cholerae spread from sewage to humans.5,6 In addition, several years ago three outbreaks of legionnaires disease in Norway were caused by the release of Legionella pneumophila from a local wastewater treatment plant (WWTP).7 In recent decades, a broad range of human bacterial pathogens, such as Clostridium perfringens, Escherichia coli, Legionella pneumophila, Mycobacterium tuberculosis, Pseudomonas aeruginosa, Salmonella enterica, Shigella flexneri, Staphylococcus aureus, Vibrio cholerae, and Yersinia enterocolitica were reported in wastewater.8,9 Most of the commonly found viral pathogens in WWTPs are enteric viruses, e.g. hepatitis A virus (HAV), which is a major threat to WWTP workers.10 A case report described an HAV infection in three sewage workers at a WWTP that spread to a small community involving 16 cases.11 © XXXX American Chemical Society
WWTPs are designed to remove nutrients and pollutants in sewage through the action of various microorganisms. A major concern is that some human pathogens may survive and grow rapidly under favorable conditions, especially for pathogenic bacteria.2 Hence, monitoring pathogens in WWTPs is routine work, and several techniques have been developed,12 such as colony count,13 PCR,14 qPCR,15−17 microarray,15,16,18 and others. However, most of these techniques are specifically designed to target one or several pathogens and cannot give a full picture of the presence of commonly encountered pathogens in WWTPs. The Illumina high-throughput sequencing technique, a newly established molecular platform, employs shotgun sequencing to profile the diversity and abundance of various pathogens at species level accurately and efficiently. 19 Several pioneering efforts have already confirmed the feasibility of this technique in detection of both bacterial and viral pathogens for medical applications.20−23 Medically, human pathogens can be disastrous and detrimental for human health because of their high virulence Received: January 24, 2013 Revised: April 15, 2013 Accepted: April 17, 2013
A
dx.doi.org/10.1021/es400275r | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
Figure 1. Schematic representation of experimental design used in this study. A total of 15 DNA samples were sequenced. A 10× deep sequencing (35 Gb) was conducted on the ST08-7 sample to obtain 10 subdata sets while all other samples were sequenced at a depth of 3−5 Gb. Raw reads were quality-checked and merged to generate iTags with a length of 150−190 bp. Then, 24 metagenomic data sets were obtained to subsample the equal depth of 12 500 000 iTags for downstream bioinformatic analysis using BLASTN, BLASTX, and MetaPhlAn.
findings provided a comprehensive understanding of human bacterial pathogens commonly found in WWTPs and demonstrated the feasibility of this technique in pathogen detection and its potential application to evaluate environmental health.
and mortality rates, and their rapid airborne or contact transmission. Emerging and re-emerging human infectious disease always raises new concerns about disease prevention and treatment. Environmentally, monitoring human pathogens in influent could be applied as a health indicator of urban residents. Also, the fate of pathogens in different sections of WWTPs could be tracked to assess potential environmental risks. Hence, detecting human pathogens is not only a medical issue, but also an environmental issue. In this study, we mainly directed our efforts toward the detection of human bacterial pathogens due to their prevalence in WWTPs and high risks to human health. The aim of this study is to detect and track human bacterial pathogens in two Hong Kong WWTPs, including influent, activated sludge, and effluent, using a high-throughput shotgun sequencing technique. The distribution, diversity, and abundance of human bacterial pathogens were examined by three different analytical methods using BLAST searches against several established databases and the MetaPhlAn tool. The
■
MATERIALS AND METHODS Sampling. A total of 23 field samples were collected from July 2007 to March 2012 in Hong Kong Shatin and Stanley WWTPs. The characteristics of the two WWTPs were previously described.24 The samples used in this study and the relevant information are listed in Supporting Information Table S1 and shown in Figure 1. There were nine and two activated sludge samples were taken from Santin WWTP and Stanley WWTP, respectively. In addition, six influent and six effluent samples were collected from the Statin WWTP and were eventually pooled in the laboratory to form four composite samples as described below. At sampling, suspended activated sludge was collected in 50 mL sterilized tubes and B
dx.doi.org/10.1021/es400275r | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
Figure 2. Detecting human pathogenic bacteria using 16S rRNA gene fingerprinting. All figures shared the same x axis labels at the bottom. (A) 16S hit number and abundance based on a BLASTN search against the Greengenes general 16S rRNA gene database. (B) 16S hit number based on a BLASTN search against the human pathogenic bacteria 16S rRNA gene database. (C) Human pathogenic bacteria abundance calculated using part B data divided by part A data accordingly. (D) Human pathogenic bacteria relative abundance calculated using part B data only. Rare species included S. boydii, S. aureus, B. anthracis, S. dysenteriae, Y. enterocolitica, C. tetani, E. coli, L. pneumophila, and C. novyi.
from two separate aliquots of each fixed sample was pooled to minimize the potential DNA extraction bias. A Thermo NanoDrop 1000 Spectrophotometer was used to determine the purity and yield of each DNA extract. The 12 DNAs extracted from the influent and effluent field samples were combined to form the following 4 samples. ST11-7-Inf and ST11-7-Eff were the composite samples of DNA extracted from influent and effluent field samples collected in July, August, and September 2011, respectively. Similarly, ST12-1-Inf and ST121-Eff represented the composite influent and effluent field samples collected in November 2011, December 2011, and January 2012, accordingly. Finally, a total of 15 prepared DNA samples (11 activated sludge DNA samples, 2 pooled influent DNA samples, and 2 pooled effluent DNA samples, as depicted in Figure 1 and Supporting Information Table S1) were sent
mixed with 100% ethanol at a 1:1 volume ratio for fixation. Samples of influent and effluent were collected in sterilized containers cooled in an ice bath and transported to the laboratory. Influent samples were centrifuged at 4000 rpm for 20 min to collect the cell pellets which were suspended and fixed using 50% ethanol. Effluent samples were filtered with 0.2 μm membrane to collect cell pellets and fixed using 50% ethanol. All the fixed samples were kept at −20 °C before DNA extraction. DNA Preparation and High-Throughput Sequencing. Comparing five commercially available kits, FastDNA SPIN Kit for Soil (MP Biomedicals, France) was proven to be the most suitable kit for DNA extraction of these samples.25 Therefore, FastDNA SPIN Kit was used to isolate and purify total DNA according to the manufacturer’s instructions. DNA extracted C
dx.doi.org/10.1021/es400275r | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
Figure 3. Detecting human pathogenic bacteria using virulence factor fingerprinting. Both figures shared the same x axis labels at the bottom. (A) VFs hit number based on BLASTX against human pathogenic bacteria virulence factor database. (B) Human pathogenic bacteria relative abundance calculated using part A data. Rare species included S. dysenteriae, Y. enterocolitica, E. faecalis, S. aureus, B. melitensis, S. boydii, S. agalactiae, B. pertussis, S. pneumoniae, H. influenzae, H. pylori, B. henselae, L. monocytogenes, and S. pyogenes.
out to Beijing Genomics Institute (BGI) for shotgun library construction and Illumina high-throughput sequencing using HiSeq 2000 platform. Databases and Tools. A total of four databases and two tools were used for metagenomic analysis in this study, as summarized in Figure 1. All of the following databases only covered commonly concerned and medically significant pathogens. (1) The human pathogenic bacteria virulence factor database was obtained from the VFDB Web site (http://www. mgc.ac.cn/VFs/).1,26,27 Only those proteins which were experimentally demonstrated and published in the literature were used in this study.1 A total of 2295 VF proteins were downloaded, which were then grouped at the species level for taxonomic statistical analysis. (2) Three 16S rRNA gene databases were used for BLASTN searches. Both Greengenes general and gold strains 16S rRNA gene databases were downloaded directly from the Greengenes Web site (http://greengenes.lbl.gov/).28 Human pathogenic bacteria 16S rRNA gene database was constructed based on the taxonomic list derived from human pathogenic bacteria virulence factor database, i.e. bacteria were selected if they possessed one or
more virulence factors (VFs) in the database. In this study, all the selected human pathogenic bacteria 16S rRNA genes are available in GenBank because their genomes have already been released. Finally, a total of 259 16S rRNA gene sequences were retrieved and assigned to 35 species (Supporting Information Table S2). The self-constructed human pathogenic bacteria 16S rRNA gene database was uploaded onto our homepage for public use (http://web.hku.hk/∼zhangt/ZhangT. htm). (3) The MetaPhlAn tool (Metagenomic Phylogenetic Analysis, http://huttenhower.sph.harvard.edu/ metaphlan/) is a computational tool for profiling microbial populations using high-throughput metagenomic data sets, which relies on BLAST searches against unique clade-specific marker genes identified from over 3000 available whole genomes in GenBank.19 (4) The MEGAN tool (MEtaGenome Analyzer, http://ab. inf.uni-tuebingen.de/software/megan/) was employed to annotate the 16S rRNA gene BLAST hits generated by Best Hit and LCA (lowest common ancestor), respectively.29 D
dx.doi.org/10.1021/es400275r | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
Figure 4. Detecting human pathogenic bacteria using MetaPhlAn tool. Relative abundance of bacteria at each taxonomic unit was calculated by MetaPhlAn. Pathogenic bacteria diversity and abundance at both genus and species levels were extracted and log10-transformed. Scales of ND, −4, −3, −2, and −1 represented the abundance of 0, 0.01%, 0.1%, 1%, and 10%, respectively. The detected pathogenic bacteria in the left column were labeled with dark red.
Bioinformatic Analysis. As shown in Figure 1, low quality reads (ambiguous nucleotides and quality value
