Article pubs.acs.org/est
Relationship of Microbiota and Cyanobacterial Secondary Metabolites in Planktothricoides-Dominated Bloom Shu Harn Te,† Boon Fei Tan,‡ Janelle R. Thompson,‡,§ and Karina Yew-Hoong Gin*,†,∥
Downloaded via UNIV DE BARCELONA on January 2, 2019 at 20:44:06 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
†
NUS Environmental Research Institute, National University of Singapore, 5A Engineering Drive 1, No. 02-01 T-Lab Building, Singapore 117411 ‡ Centre for Environmental Sensing and Modelling, Singapore-MIT Alliance for Research and Technology Centre, 1 CREATE Way, #09-03 CREATE Tower, Singapore 138602 § Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, United States ∥ Department of Civil and Environmental Engineering, National University of Singapore, 1 Engineering Drive 2, E1A 07-03, Singapore 117576 S Supporting Information *
ABSTRACT: The identification of phytoplankton species and microbial biodiversity is necessary to assess water ecosystem health and the quality of water resources. We investigated the short-term (2 days) vertical and diel variations in bacterial community structure and microbially derived secondary metabolites during a cyanobacterial bloom that emerged in a highly urbanized tropical reservoir. The waterbody was largely dominated by the cyanobacteria Planktothricoides spp., together with the Synechococcus, Pseudanabaena, Prochlorothrix, and Limnothrix. Spatial differences (i.e., water depth) rather than temporal differences (i.e., day versus night) better-explained the short-term variability in water quality parameters and bacterial community composition. Difference in bacterial structure suggested a resource-driven distribution pattern for the community. We found that the freshwater bacterial community associated with cyanobacterial blooms is largely conserved at the phylum level, with Proteobacteria (β-proteobateria), Bacteroidetes, and Actinobacteria as the main taxa despite the cyanobacterial species present and geographical (Asia, Europe, Australia, and North America) or climatic distinctions. Through multivariate statistical analyses of the bacterial community, environmental parameters, and secondary metabolite concentrations, we observed positive relationships between the occurrences of cyanobacterial groups and off-flavor compounds (2-methyisoborneol and β-ionone), suggesting a cyanobacterial origin. This study demonstrates the potential of 16S rRNA gene amplicon sequencing as a supporting tool in algal bloom monitoring or water-resource management. temperature, pH, alkalinity, and organic carbon.7,8 Similarly, heterotrophic bacteria can also alter the water environments,9 leading to positive or negative impacts on blooms. For instance, Flavobacterium, Streptomyces, and Oxalobacteraceae can enhance cyanobacterial growth; Pedobacter and Arthrobacter may cause negative impacts,6,10 while Limnobacter, Rhizobium, and Lacibacter develop symbiotic relationships with cyanobacteria.11 Previous studies demonstrated that bloom-associated microbial communities are highly varied at phylum level.8,12 However, the functional potential revealed by genes could be highly conserved despite differences in community assemblage,12 such as nitrogen fixing and assimilation,8 organic matter degradation,6,11 and iron uptake.13
1. INTRODUCTION Proliferation of cyanobacteria in freshwater systems has increased in scale and frequency in past decades as a consequence of eutrophication and climate change, becoming one of the major challenges in protecting aquatic ecosystem health and maintaining safety of water resources.1 The production of secondary metabolites and exudates that are toxic to organisms and humans or having unpleasant smells is one of the key concerns related to cyanobacterial blooms.2 These compounds can be synthesized by different taxa or genera,3,4 which complicates the water-management strategy in controlling the sources of these pollutants. Extensive effort has been carried out to explain bloom formation with regards to nutrients and physiochemical factor.5 Recently, more studies have focused on the bacterial community that thrives alongside the cyanobacterial population.6 Indeed, heterotrophic bacteria can be influenced by biotic and abiotic factors induced by blooms such as nutrients, © 2017 American Chemical Society
Received: Revised: Accepted: Published: 4199
November 26, 2016 March 7, 2017 March 27, 2017 March 27, 2017 DOI: 10.1021/acs.est.6b05767 Environ. Sci. Technol. 2017, 51, 4199−4209
Article
Environmental Science & Technology
space solid-phase microextraction (HS-SPME) and gas chromatography−tandem mass spectrometry (GC−MS/MS) method using an Agilent GC 7890A coupled with a 7000B Triple Quad series mass spectral detector, as described previously.20 All secondary metabolite measurements were run in duplicate. 2.4. Quantitative Analysis of Bacterial and Cyanobacteria Abundance. Microbial biomass was harvested within 3 h after each sampling by filtering reservoir water through a 0.45 μm cellulose nitrate filter (Sartorius) and frozen at −20 °C. DNA was extracted from the filter using a PowerWater DNA Isolation Kit (MOBIO) following the manufacturer’s instructions. The abundance of bacteria and cyanobacteria in water were inferred from the 16S rRNA gene copy numbers (GNC) in the samples. Total bacterial 16S rRNA gene was quantified using a droplet digital PCR system (ddPCR) (QX200, Bio-Rad), while cyanobacterial 16S rRNA gene was quantified with a StepOnePlus real-time PCR (qPCR) system. The two most commonly reported cyanobacterial bloom genera, Microcystis (MIC) and Cylindrospermopsis (CYL) were also quantified using qPCR assays developed previously.21 Detailed methods are provided in the Supporting Information (A). To investigate the major off-flavor producers, we amplified and sequenced partial 2-MIB biosynthesis genes using primer sets reported previously,22 and searched for sequence homologues against the database of NCBI by using BLAST.23 2.5. 16S rRNA Amplicon Sequencing and Data Analyses. Isolated genomic DNA (gDNA) was amplified using the universal pyrotaq primer set (926wF and 1392R) targeting the small subunit 16 rRNA variable regions 6−8 of bacterial communities.24 Details on Illumina Miseq sequencing and sequence preprocessing are available in the Supporting Information (B). All 16S rRNA sequences have been deposited at the National Centre for Biotechnology Information (NCBI) Sequence Read Archive (SRA) under accession no. SRR3996072-3996095. The standardized biological data (relative abundance of bacterial OTUs and cyanobacterial genera) were square-roottransformed to reduce the dominance of OTUs in high abundances in analysis, while all environmental data sets were normalized using means and standard deviations of the variables. The similarity (resemblance) of each pair of samples in terms of biological and physicochemical characteristics was calculated using the Bray−Curtis similarity and Euclidean distance, respectively. Relationships between OTUs and environmental variables were analyzed using RELATE-BEST, and a simple model indicating variance explained by selected environmental variables was calculated using the distance-based linear model (DistLM) implemented in Primer-7.25 To detect spatial and temporal differences in the microbial community between samples, permutational multivariate analysis of variance (PERMANOVA) and analysis of similarity test (ANOSIM) were performed. Samples were separated into four groups and compared according to the sampling time and depth−day versus night; surface versus bottom. This was followed by similarity percentage calculations (SIMPER) to indicate OTUs that contributed the most to the variation. Diel and vertical differences in water quality variables and biological variables (qPCR and ddPCR) were determined using paired-t test and general linear model (GML) in the statistical software package PASW18 (SPSS Inc.). Non-normally distributed data sets were transformed with a two-step
Riding on the development of high-throughput sequencing (HTS), information regarding bacterial groups and microbial diversity in bloom-affected water bodies is fast expanding, but the existing literatures report only a few genera, such as Microcystis, Anabaena, and Cylindrospermopsis.8,11,14,15 In addition, most of the relevant studies focused on temperate or subtropical regions, with very limited investigations done in the tropical climate zone.16 In the current study, we carried out an intensive short-term sampling in one of the reservoirs in an urbanized area of Singapore, during which a rarely reported filamentous cyanobacterium, Planktothricoides, dominated the waterbody. This study aims to (i) determine the bacterial community associated with Planktothricoides bloom, (ii) investigate diel and vertical variations of physiochemical variables, cyanobacterial metabolites and microbial community, and (iii) identify potential metabolite producers using 16S rRNA amplicon sequencing combined with correlation analysis.
2. MATERIALS AND METHODS 2.1. Site Description and Sampling. This study was conducted in a freshwater reservoir located in a fully urbanized catchment covered mainly by commercial, residential, and industrial areas. Having a total catchment area of 100 km2, a water body surface area of 2.4 km2, and an average depth of nearly 4 m, the water body is the largest multifunctional reservoir in Singapore, providing recreation, flood control, and water storage to the country. Sampling was carried out over 2 days from January 7th to 9th, 2014 at the confluence of three main tributaries feeding water to the reservoir. To study possible diel vertical variations caused by thermal or nutrient stratifications,17,18 samples were collected at different times of the day (early morning, late morning, afternoon, evening, and night, i.e., 6 a.m., 10 a.m., 2 p.m., 6 p.m., 10 p.m., and 2 a.m.) from two depths (surface, 0.1 m; bottom, 3.2 m) using a subsurface grab sampler. A 4 h sampling interval was used to improve detection of fast-degrading off-flavor compounds, which may have half-lives on the order of hours to days.19 All samples were kept in sterile polypropylene bottles and transported to the lab immediately for processing or storage. 2.2. Physical and Chemical Water Quality Variables. Temperature, pH, conductivity, total dissolved solid (TDS), salinity, and dissolved oxygen (DO) were measured with a YSI 6600 V2 multiparameter water quality sonde (YSI Inc.). Other water quality metrics, including chlorophyll-a (chl-a), total phosphorus (TP), orthophosphate (PO4), total nitrogen (TN), ammonium (NH4), nitrate (NO3), nitrite (NO2), total organic carbon (TOC), dissolved organic carbon (DOC), calcium (Ca), and magnesium (Mg), were determined with APHA standard methods (10200H, 4500-P (B,E), 5310B, 4110B, and 3120B) and the JIS method (K 0102). Light intensity was measured with a LI-250A light meter and LI-192 sensor (LICOR). 2.3. Cyanobacterial Secondary Metabolites. Water samples for cyanotoxin analysis were stored at −20 °C upon arrival to the lab. Concentrations of two bloom-associated cyanotoxins, i.e., microcystin and cylindrospermopsin, were measured using the microcystin and nodularin (ADDA) and cylindrospermopsin ELISA kits (Abraxis LLC), respectively. Water samples were sonicated without heat for 15 min to release cell-bound toxin, and final development of color was measured with a microtiter plate reader. A total of four offflavor compounds, i.e., 2-methylisoborneol (2-MIB), geosmin, β-cyclocitral, and β-ionone, were determined with the head4200
DOI: 10.1021/acs.est.6b05767 Environ. Sci. Technol. 2017, 51, 4199−4209
Article
Environmental Science & Technology
Figure 1. Box plot of water physicochemical characteristics measured at the upper and lower water layers (surface and bottom). Unit for parameters: μg/L (Cyn, Mcyst, and Chl-a); mg/L (TN, TP, DOC, TOC, NO2, NO3, NH4, Salinity, Ca, Mg, and DO); m·S/cm (Cond); g/L (TDS); ng/L (2MIB, β-cyclocitral, Geosmin, and β-ionone); μmol/m2.s (light); °C (Temp); pH unit (pH); and NTU (Turbidity). Asterisks (*) indicate parameters with significant differences determined using a paired t test and GLM; P < 0.05.
transformation approach26 prior to the tests. Day and night samplings were determined from the ambient light intensity, of which a q quantum measurement of >8 umol/m2·s was considered as a day sample. To investigate changes of bacterial community composition, nonmetric multidimensional scaling (nMDS) was conducted for sample comparisons based on OTUs, followed by Spearman correlation to determine environmental variables that correlated most highly with the first and second ordination axes. Results were visualized using nMDS and vector plots.
December of 2013 to February of 2014. Compared to nonbloom periods (Chl-a < 20 μg/L; phytoplankton count 0.8), with the results obtained from amplicon sequencing. Relative abundances of cyanobacteria derived from HTS were consistently lower by 35% (Figure 5) compared to those of qPCR and ddPCR. While amplicon sequencing is still powerful in the overall prediction of population dynamics, especially for the major taxa,47 large discrepancies could occur while estimating the relative abundance of bacteria especially for minor taxa (i.e., Microcystis and Cylindrospermopsis), as shown in our study when qPCR and sequencing results were compared. Although 16S rRNA gene is by the far the most-popular approach to study the diversity and abundance of bacterial community,48 several intrinsic limitations, such as bias in DNA extraction method,49 selectivity of primer, choice of target hypervariable regions,48,50 and availability of reference sequences,51 could lead to significant difference to both 16S rRNA gene sequencing and qPCR results. Limitations on molecular-based analyses should be taken into consideration in data interpretation. 4.3. Planktothricoides Bloom. The genus of Planktothricoides is a relatively new cyanobacterial group that used to be classified as Planktothrix. It was separated from the genus of Planktothrix based on their distinctive features found in morphology, pigment and fatty acid compositions, and 16S rRNA sequences.52 Worldwide, the number of Planktothricoides blooms reported to date is limited, presumably due to inaccurate taxonomy assignment. Planktothrix, the closest genus related to Planktothricoides (
