S-Space College of Natural Sciences (자연과학대학) Dept. of Biological Sciences (생명과학부) Theses (Master's Degree_생명과학부)
Microbial community structure and assembly processes in Mediterranean ecosystems: a case study of Fynbos (South Africa)
- Itumeleng Petros Moroenyane
- Jonathan Adam
- 자연과학대학 생명과학부
- Issue Date
- 서울대학교 대학원
- Microbial Ecology
- 학위논문 (석사)-- 서울대학교 대학원 : 생명과학부, 2015. 8. Jonathan Miles Adams.
- The Mediterranean heathland (Fynbos) is world renowned for its levels of plant diversity, endemism, and structure, but also for its mild Pleistocene and poor soils. Majority of microbial studies from the Fynbos have only examined soil sampled from one or two Fynbos vegetation types. In this study, soils were sampled from five Fynbos vegetation types to examine the overall microbial community structure, diversity, and assembly processes. Using 454 Pyrosequencing platform (18S rRNA gene marker for soil nematodes) and. the Illumina HiSeq platform (16S rRNA gene marker for bacteria and ITS 1 region for fungi). A Non-metric dimensional scaling (NMDS) drawn from the environmental Euclidian distance showed that soils from the different Fynbos vegetation types were significantly different. Firstly, this study looked at the nematode community structure, diversity, and which assembly processes were important. The detected phylogenetic signal showed that nematodes in the Fynbos were ecologically coherent (tend to co-occur more than by chance when sampled at random), and the community based on the Bray-Curtis matrix revealed that both SO4- and K+ were delimiting the community structure in the Fynbos. The most abundant feeding group of nematode was the bacteria feeding (BF), however, only the abundance of plant feeding group (PF) was influenced by NH4 .Nearest taxon index (NTI) revealed that the community was phylogenetically clustered and that ses.MNTD showed that at the local scale deterministic processes are important in assembling the nematode community. Both UniFrac and MNTD matrix were highly influenced by geographical distance and NH4 respectively. We conclude that in the Fynbos there is niche overlap of nematodes, the phylogenetic community structure also corroborates this, and deterministic processes are important in delimiting the community assembly. A regression analysis showed that OTU richness was significantly influenced by K concentration in Fynbos soil. A look at the fungal community revealed that unlike the nematode community, the NMDS was clustered significantly by Fynbos vegetation type. A Mantel and partial Mantel tests performed on the Bray-Curtis matrix showed that the fungal community in the fynbos was significantly influenced and delimited by both environment and elevation. Furthermore
elevation, N, pH, and NH4+ were shown to be significantly delimiting the community structure. To further capture the community composition, the most dominant phyla were Ascomycota, Basidiomycota, and Zygomycota, with only Ascomycote and Zygomycota showing significant difference across the sites. Conversely, the EcM community in the Fynbos also to an extend clustered by Fynbos vegetation type, however pairwise comparison showed that some Fynbos vegetation were indistinguishable overall there was a significant difference. In the Fynbos the environment is important in delimiting fungal community structure and abundance. This corroborates previous studies on fungi and our study further sheds more light on importance of fungi in ecological process that occur in the soil.
The eukaryote community in Fynbos seems to show some predictability in their community structure and diversity. However, the prokaryote community revealed unique patterns. Firstly, this study looked at what were the diversity patterns before unrevealing the assembly processes. A NMDS of the bacterial community drawn from Bray Curtis distance did not cluster by Fynbos soils. Although, there was no clustering of the bacterial community, Ca was the only edaphic variable that influenced community structure, this is in accordance with previous studies that showed that divalent cations dictate bacterial community structure in Fynbos soils. Mantel and partial Mantel test revealed that both environment and geographic distance influenced bacterial community structure. However, NMDS of the bacterial community drawn from Bray Curtis distance did not cluster by Fynbos soils type. Ca was the only edaphic variable that influenced community structure at the OTU level .Regression analysis revealed that both silt and clay content together with Ca were highly correlated with OTU richness and 16S rRNA gene copy numbers. Of the three most dominant bacterial phyla, only the relative abundance of Acidobacteria differed significantly across all Fynbos soil types, whilst, Actinobacteria and Proteobacteria did not. Multiple regression on matrices showed that relative abundance of Bacteroidetes, Chloroflexi, Gemmatimonadetes, and Protebacteria were significantly correlated with total organic Carbon (TOC), silt and clay, SO4 and Ca respectively. Interestingly, at class level members of Acidobacteria and Gemmatimonadetes relative abundance was influenced by Ca and silt and clay. Overall, this study identifies the edaphic variables that influence and shape the bacterial community at different taxonomic levels in the Fynbos. Next, this study looked at community assembly processes that were delimiting the bacterial community in the Fynbos. It was shown by the UniFrac analysis that the community clustered strongly by vegetation types, suggesting a history of evolutionary specialisation into certain vegetation types. The standardised beta mean nearest taxon distance (ses. βMNTD) index, showed no association with vegetation type. However, the overall phylogenetic signal indicates distantly related OTUs tend to co-occur and only a small proportion of closely related OTUs were ecologically coherent. Both NTI (nearest taxon index) and ses.βMNTD significantly deviated from the null models, indicating that deterministic processes governed phylogenetic community structure assembly. Furthermore, ses.βMNTD was significantly higher than null expectations, indicating that over-dispersion in phylogenetic beta diversity is explicable by the differences in environmental conditions across the sites. The community was over-dispersed, attributable to the differences in environmental conditions across sites, even though only weak correlation soil texture could be proved.
Overall, our study indicates that microbial ecology of the Fynbos system is similar to other ecological systems, where both deterministic and stochastic factors govern community structure and assembly.