Molecular evaluation of spatial and temporal variation in bacterial communities and clinically relevant antibiotic resistance genes in an aquatic system
Freshwater ecosystems such as rivers and lakes contain up to 10% of all species and are an important repository of genetic information. They are important sources of drinking water and food. They are also extensively used for different agricultural and industrial processes. One major health concern on the use of water resources from rivers is the presence of antibiotics resistance genes (ARGs). ARGs are found in bacterial genomes and can spread between environmental bacteria and ultimately into the human microflora. ARGs can render antibiotics ineffective, increasing the morbidity and mortality rates of human diseases. Unfortunately, this health concern is poorly assessed in rivers systems including those used for agricultural processes, such as the Werribee River, Victoria which is the focus of the research in this thesis. Therefore, this study aims to study the bacterial community structure, diversity and composition and the influence of physico-chemical factors on this community in the Werribee River in Melbourne using 16S rRNA-based Next-Generation Sequencing (NGS) tool. The prevalence of ARGs using PCR, qPCR and ARG-based targeted Next-Generation Sequencing approaches was also investigated to explore variation in the antibiotic resistance gene pool along the river and over time.
Variation in bacterial community structure, composition and diversity along the river was investigated at four selected sites; Ballan and Bacchus Marsh (upstream sites) and Cobbledicks Ford and Riversdale (downstream sites). At these sites, both spatial and/or temporal variation was observed in the numbers of Operational Taxonomic Units (OTUs), Shannon diversity and taxonomic composition (phylum, class and genus levels) in most samples. The highest Chao1 OTU richness value of 2332 was recorded in Cobbledicks Ford while the lowest value of 1645 was observed in Bacchus Marsh. Similarly, the lowest Shannon diversity value of 2.5 was reported at Bacchus Marsh while the highest diversity value of 4 was recorded in Ballan. Bacterial Shannon diversity (December-February; 4.2 and 3.9) and Chao1 OTU richness (December; 2408, February; 2976) were significantly higher in the summer months when compared to most other months. Three key phyla (Proteobacteria, Actinobacteria and Bacteroidetes) were observed at all sites with the Proteobacteria being the most dominant phylum across all sites and at all sampling months. At the class level, four different classes, Actinobacteria, Alphaproteobacteria, Betaproteobacteria and Cytophagia, were the dominant classes based on SIMPER analysis. The Betaproteobacteria was the most abundant class at all sites except at Riversdale where it was replaced by the Actinobacteria. Different bacterial genera were important at different sites. At upstream sites, the key genera were Polynucleobacter, Arcicella, Limnohabitans, Acidovorax and Methylotenera, whilst in downstream sites, the key genera were Acidovorax, Candidatus Pelagibacter, Limnohabitans and Demequina. Temperature and dissolved oxygen concentrations were commonly associated with changes in the bacterial community at upstream and downstream sites. Bacterial diversity and OTU numbers were strongly and positively correlated to water temperature but negatively correlated to dissolved oxygen concentrations. Other factors such as turbidity, suspended solids, electrical conductivity, total bound nitrogen and ammonia concentrations were also important at the (downstream) Cobbledicks Ford site.
The prevalence of 12 ARGs (blaNDM-1, mecA, tet(M), ampC, VanA, mcr-1, tet(B), erm(B), aac (6)-Ie-aph (2)-Ia, SulII, catII and dfrA1) was assessed using PCR in upstream and downstream samples from the Werribee River. Out of these ARGs, only three ARGs, blaNDM-1, tet(B) and catII were detected. Substantial spatial and temporal variation in the frequency of detection of these ARGs was observed. These ARGs were detected more frequently at the downstream sites of Cobbledicks Ford and Riversdale (detection frequency of 12.5%) than at upstream sites Ballan (4.2 %) and Bacchus Marsh (1.4 %). Only the catII gene was detected at Bacchus Marsh. There was, therefore, a greater risk of exposure to ARGs from water samples in downstream sites than in upstream sites.
The relative abundance of detected ARGs as evaluated using quantitative PCR showed limited spatial variation. This was because apart from the abundance of the ARGs in Riversdale samples being significantly higher, than at Bacchus Marsh, there was no significant difference in ARGs relative abundance at the remaining sites. However, there was substantial temporal variation in ARG abundance at most sampling times. This variation was significant in some instances, for example, when the relative abundance was higher in summer (December and February) compared to autumn (April). The detection of ARGs, especially at downstream sites highlights a potential health risk of using water from Werribee River for recreational and agricultural purposes. However, this study did not determine whether the greater frequency of detection and increasing ARG relative abundance at the downstream sites was a natural occurrence or due to anthropogenic sources.
The application of targeted ARG-based NGS to selected samples was partly successful. The ARGs evaluated were blaNDM (sample from Riversdale) and catII (samples from Cobbledicks Ford, Riversdale, Ballan and Bacchus Marsh). Data analysis showed a significant similarity of blaNDM sequence from Riversdale to a blaNDM gene from Escherichia coli indicating that the target gene was successfully amplified in the Riversdale sample. However, the target gene/enzyme (chloramphenicol O-acetyltransferase) was identified at only one site (from Cobbledicks Ford). At the remaining sites, chloramphenicol resistance genes were not identified and instead, other acetyltransferase genes were detected, namely; acetyl-CoA C-transferase (Bacchus Marsh and Riversdale) and 3-oxoadipyl-CoA thiolase (Ballan) which are not associated with chloramphenicol resistance.
This study has therefore shown substantial spatial and temporal variation in the bacterial community structure and diversity in the Werribee River with temperature and DO levels being some of the significant drivers of changes in the bacterial community structure. This spatial and temporal variation was reflected in the frequency of detection of the important ARGs (blaNDM, tet(B) and catII) in the Werribee River. Finally, the targeted ARG-based NGS approach used in this study is a promising approach for determining the diversity, identity and relative abundance of ARG sequences in environmental samples but requires more optimizations with more samples from different sources in future investigation.
History
Degree Type
Doctorate by ResearchImprint Date
2019-01-01School name
School of Science, RMIT UniversityFormer Identifier
9921970711201341Open access
- Yes