Molecular and functional characterization of bacterial communities in organic farm and pasture soils.

摘要

Culture-dependent denaturing gradient gel electrophoresis (CD DGGE) fingerprinting of the 16S rRNA gene was developed to characterize mixed soil bacterial communities recovered on agar plates. CD DGGE analyses resulted in clear banding patterns of sufficient complexity (16--32 major bands) and reproducibility to investigate differences in bacterial communities in silt loam soils. Replicate CD DGGE profiles from plates inoculated with less-dilute samples (10 -3) had a higher band count and were more similar (72--77%) than profiles from more-dilute samples (51--61%). When CD DGGE fingerprints were compared to profiles constructed from 16S rRNA genes obtained from culture-independent clone libraries (CB DGGE profiles) 34% of the bands were unique to the culture-dependent profiles, 32% were unique to the culture-independent profiles and 34% were found in both communities. CD DGGE fingerprints of the 16S rRNA gene were used to compare culturable microbial communities in soils from an organic farm and pastures that were subjected to differing agronomic treatments. Multiple and simple correlation analyses revealed significant relationships between microbial biomass carbon (MBC), potentially mineralizable nitrogen (PMN), and variables derived from principal component analysis (PCA) of community-level physiological profiling (CLPP), as well as non-metric multidimensional dimensional scaling (NMDS) of CD DGGE data. All measures clearly separated soil types, but varied in their ability to distinguish among treatments within a soil type. Overall, MBC, PMN, and CLPP were most responsive to additions of compost (organic farm) and manure (pasture); while CD DGGE resolved differences in legume cropping (organic farm) and inorganic fertilization (pasture). Bacteria from filtered ( 0.45 mum) and non-filtered slurries of soils from an organic farm were cultivated on R2A agar for up to one month. According to CD DGGE analyses, filtered and unfiltered communities were highly dissimilar and distinct small-celled bacteria were isolated from soils collected one year apart. Filterable bacteria were classified by phylogenetic analyses as alpha-, beta-, gamma-, or delta-Proteobacteria or Bacteroidetes with the genera Polaromonas and Ramlibacter representing 23% of the non-chimeric sequences in clone libraries. Other sequences were associated with known copiotrophic ultramicrobacteria, endophytic nitrogen-fixing bacteria in the genera Herbospirillum and Azospirillum , and bacteria adapted to inter- or intracellular growth. Sequences obtained from slow-growing, pinpoint-sized colonies were highly similar to an amoeba-resistant bacterium affiliated with the genus Rhodoplanes , a putative diazotroph that may play a potentially important role in soil nitrogen dynamics. Finally, the spatial variability and heterogeneity of microbial substrate utilization in tilled and untilled organic farm soils were explored using CLPP coupled with multivariate and geostatistical analyses. Robust semivariograms based on the kinetics of substrate utilization in BIOLOG GN2 plates were constructed. Microbial substrate utilization was spatially structured in some untilled soils and subsurface tilled soils over ranges > 4 m. Rotary tillage completely disrupted the spatial structure of substrate utilization in surface soils. Amines and amides were the only substrate classes that exhibited a tillage-induced loss of spatial structure at lower soil depths. Spatial structure at lower soil depths in untilled soils was most pronounced for utilization of amino acids, amines and amides. Collectively, data presented in this dissertation support arguments that culturable soil microbial communities provide nonredundant and useful information for characterization of selected agronomic perturbations.