Recovery and concentration of microbial cells may be an important first step if nucleic acid analysis is to be performed in environments with relatively low cell densities, such as aquatic or phyllosphere habitats. Moreover, even for environments with high population densities, it can be advantageous to prepare a cell fraction prior to nucleic acid extraction, in particular if intracellular targets are to be assessed. Bacterial diversity assessments based on DNA reassociation kinetics21 even depend on the DNA being primarily of bacterial origin. Paul recently reviewed the available methods for cell concentration from aquatic environments;22 a variety of approaches has been used, the most effective being those including the filtration of large volumes through membrane filters.23-25 Tangential flow filtration26 or Vortex flow filtration27 have the advantage that cells experience less meiianical stress and that large volumes (up to 100 L) can be processed within short times. Prefiltration through 3 ^m filters can be used to remove large particles which would clog filters.26 Alternatively, high-speed centrifugation is applicable for concentrating bacteria in smaller water volumes. Filters with microbial biomass or bacterial pellets obtained after highspeed centrifugation can be stored frozen until further processing.
As most bacterial cells in nature, including aquatic environments, occur attached to surfaces (forming microcolonies or bofilms28), cell dislodgement methods are often required. Cells can actually adhere quite strongly to surfaces by bonding mehanisms such as via bacterial polymers, pili or flagellae, electrostatic forces or water bridging.29-31 Bacteria can also be entrapped in soil aggregates which are formed through gluing by bacterially-produced polysaccharides and physico-chemical interactions between silica/day surfaces and decomposed organic matter.32, 33
The representative extraction of surface-attached bacteria from environmental matrices requires that the cells bound by the various modes and with different strengths are efficiently dislodged. Dissociation of cells from surfaces is generally achieved by repeated homogenization steps. For soil, homogenization can be achieved by shaking soil suspensions with gravel or beads, or by blending in Stomacher or Waring blenders.34 Mild ultrasonication in a low energy bath has also been used.29 However, the amount of energy needed to completely disperse soil aggregates can result in considerable cell death.29,35 During extraction, bacterial cells experience not only mechanical stress but also changes in their physico-chemical conditions. Thus, the physiological state of cells following dislodgment from surfaces, reflected in the diversity of intracellular RNA molecules, might not represent their in situ physiological state. Lindahl et al36 found that the activity of extracted bacterial fractions is often lower than that in soil slurries. Furthermore, Prieme et al31 reported that treatment of soil samples in a Waring blender decreased methane oxidation activity, with a profound difference between different soils. Soil slurries in sodium pyrophosphate showed less than 10% cf the methane oxidation observed with soils resuspended in water. Furthermore, the sodium pyrophosphate used might be problematic in microbiological incubations due to the enhanced supply of phosphorus.33
Following dislodgment, separation of bacterial cells from the particulate matter is usually achieved by low-speed centrifugation, with their subsequent recovery in a pellet by high-speed centrifugation (differential centrifugation37). Recent studies revealed that the centrifugal force applied is highly critical;38 centrifugal forces over 500 xg reduced cell recoveries dramatically. The use of forces < 100 xg was therefore recommended, but at these reduced speeds, recovery rates did not exceed 45% of the total discernable cells.38
An alternative method to disperse soil particles and dissociate microbial cells from soil, sediment or root parts involves the use of cation exchange resins. Shaking cf soil with, for instance, Dowex39 or Chdex-10029,40,41 is used to remove the bivalent and polyvalent cations responsible for electrostatic bonding between like-charged bacterial cells and soil particles. Detergents are also used to overcome adhesive interactions. A comparison of five different treatments (Fig. 3.2) showed that cell extraction could be improved by using Stomacher blending instead of shaking. Consistent results and a fairly rapid treatment were possible with the automated paddle action of the Stomacher. On the other hand, Lindahl and Bakken42 did not find any positive effect of the Chelex-100 treatment on cell extraction efficiency and, thus, recommended simple, threefold repeated, blending with water. Shaking in low-electrolyte concentrations, e.g. in distilled water, increases the interactive free energy between like-charged soil particles. In a recent comparison of both methods with agricultural soil, the total cfu numbers obtained were found to be comparable between the methods (unpublished). However, the water-based protocol yielded a higher number of different colony types, suggesting that the extraction protocol strongly affects the diversity of bacterial types obtained.
In particular for soils with high clay contents, separation of bacterial cells and soil particles is necessary. This can be achieved by flotation in the density media Percoll or
Nycodenz,31,43 but not with simple centrifugation steps, because day particles and bacterial cells show similar sedimentation behavior. The efficiency of different cell extraction protocols thus depends on the oil type; hence, different protocols may prove suitable for different soils.43 However, even with the most optimized protocol, complete extraction of all bacterial cells, in particular of cells bound to soil particles, is probably impossible. Thus, for studies on the diversity of microbial communities in soil, it seems reasonable that the cells recovered represent, in their relative abundance, the surface-attached microbial community. As microbial cells may occur both at the outside of soil aggregates and in the inferior regions, the extent to which these soil aggregates are disrupted will determine the sites from which the inhabiting bacterial cells are dislodged. To monitor the population dynamics of bacterial inoculants, it is, thus, primordial to understand their localization, as the protocol to be used and the limit cf detection are affected by it. Moreover, inoculant cells may preferentially occur at different locations than indigenous microbes, and hence the requirements with respect to extraction are different.
The bacterial cell fractions obtained can be used for the extraction of genomic DNA4,5 or for the parallel extraction of both DNA and RNA (Fig. 3.1).44,45 Furthermore, the cell fractions can serve for assessments of cell numbers using fluorescent dyes or immunofluorescence enumeration. In general, nucleic acids recovered from soil/rhizosphere bacterial cell fractions are less contaminated with (»extracted humic acids, fulvic acids, polyphenols, polysaccharides, or other plant-derived substances which can hinder molecular analyses, than these recovered directly.7 For polluted soils, separation of bacterial cell fractions may even be an absolute requirement. For instance, DNA extracted directly from a zinc-contaminated soil was not PCR amplifiable even after several purification steps, whereas DNA extracted from the recovered bacterial fraction was amplifiable without any problems (Brim et al, unpublished).
Was this article helpful?