TEXT

Top Abstract Text References

Typically, soil and sediment bacteria are stained in solution with acridine orange or 4',6-diamidino-2-phenylindole (DAPI), collected onto polycarbonate membrane filters, and enumerated by using epifluorescence microscopy. However, DAPI is replacing acridine orange as the bacterial stain of choice (6). High background fluorescence or unspecific staining occurs frequently in soils and sediments and can hamper the enumeration of bacteria (6). Moreover, not all cells are stained with DAPI and DAPI is less specific for DNA than previously thought (6, 15, 19). Recently, novel high-affinity nucleic acid dyes such as Yo-Pro-1 and SYBR Green I (subsequently called SYBR I) and Anodisc aluminum oxide membrane filters were applied to enumerate bacteria or even viruses in the water column of freshwater and marine systems (4, 12, 18). Moreover, the stains SYBR I, SYBR II, and PicoGreen were used to count bacteria in aquatic systems by flow cytometry (8-10). However, to our best knowledge these and other dyes such as SYBR Gold and RiboGreen have not been used so far for epifluorescence counts of soil and sediment bacteria. Thus, we tested whether these new green fluorescent dyes (GFD) would improve the visualization of bacteria in soils and sediments, and we also evaluated Anodisc filters.

Staining of reference bacteria. To determine the fluorescence yield of formaldehyde-fixed cells upon staining with nucleic acid dyes, Escherichia coli K-12 strain DSM429 was grown on liquid nutrient broth medium (8 g liter¹; Difco Corp.). Aliquots were taken from the exponential and stationary growth phases and preserved with formaldehyde (final concentration, 4%) overnight. Cells were collected by centrifugation (6,000 × g for 10 min), washed with phosphate-buffered saline (PBS; 130 mM sodium chloride, 10 mM sodium phosphate buffer; pH 7.5), and resuspended in TE buffer (10 mM Tris-HCl, 1 M EDTA; pH 7.5). One hundred microliters of cell suspension was mixed with 100 µl of TE-buffered dye solution. As nucleic acid stains, we used the blue fluorescent dye DAPI (Sigma, chemical no. D-9542) and the GFD SYBR Green I (10,000× in dimethyl sulfoxide [DMSO]; Molecular Probes, chemical no. S-7567), SYBR Green II (10,000× in DMSO; Molecular Probes, chemical no. S-7568), SYBR Gold (10,000× in DMSO; Molecular Probes, chemical no. S-11494), PicoGreen (dsDNA quantitation kit; Molecular Probes, chemical no. P-7581), and RiboGreen (RNA quantitation kit; Molecular Probes, chemical no. R-11490). The final concentration of DAPI ranged from 0.025 to 50 µg ml¹, and the final dilution of the GFD was between 1 × 10² and 5 × 10⁵ of the stocks provided by the manufacturer. Fluorescence was measured in duplicates with a Cytofluor 2350 fluorescence measurement system (Millipore) by using low-fluorescence 96-well microtiter plates (CytoPlate; Millipore) set at 360 nm (emission; bandwidth, ±20 nm) and 460 nm (excitation; bandwidth, ±12.5 nm) for DAPI and 485 nm (excitation; bandwidth, ±10 nm) and 530 nm (emission; bandwidth, ±12.5 nm) for GFD. As blanks, we used 0.2-µm-pore-size-filtered cell suspensions plus dye solution.

Staining and enumeration of soil and sediment bacteria. Samples were collected from a set of very diverse soil and sediment habitats (Table 1). For the analyses, 2 g of soil or sediment was preserved with 6 ml of 4% formaldehyde in PBS and stored at 4°C. Samples were vortexed, and 10- to 100-µl aliquots (depending on soil or sediment type) were removed before particles could sediment. Aliquots made up to 1 ml with 0.01 M sodium pyrophosphate, and 100 µl of this suspension was diluted 1:10 with 0.01 M sodium pyrophosphate, shaken for 45 min on a variable-speed vibration shaker equipped with an Eppendorf tube attachment, and sonicated for 1 min (4-mm needle diameter; Labsonic U 2000 set at 50 W and 0.5-s pulses). Two types of 0.2-µm-pore-size 25-mm membrane filters were used to collect bacteria from 1-ml samples: black polycarbonate filters (Nuclepore) and Anodisc aluminum oxide filters (Whatman). These filters were backed by 0.45-µm-pore-size cellulose nitrate membrane filters (Sartorius). With the polycarbonate filters, bacteria were first stained for 15 min within the filtration funnel (14) at the final dye concentrations shown below and then collected onto the filter by vacuum filtration (<10 kPa). With the Anodisc filters, cells were first collected onto the filter and processed by slightly modifying the protocol of Noble and Fuhrman (12). The filter sandwich was placed sample side up in a petri dish on a drop of TE buffer, and bacteria were stained by adding 200 µl of DAPI (typically 5 µg ml¹) or GFD solution (typically a 5 × 10⁴ dilution of the stock). After a staining period of 15 min in the dark, the filter sandwich was placed back on the filter holder and the dye solution was sucked off. The Anodisc filter was mounted on a glass slide with a drop of antifade solution (50% glycerol, 50% PBS, 0.5% ascorbic acid) and a 25-mm-square coverslip. Cells were enumerated under UV (DAPI) or blue excitation (GFD) by using a Zeiss epifluorescence microscope (Axiophot model 135TV). At least 200 cells from at least 10 eye fields were counted per filter, and cells were enumerated on two filters per sample (7).

Comparison of fluorescence yields. The highest fluorescence yield of DAPI for E. coli cells in the stationary growth phase was found at a dye concentration of 1 to 10 µg ml¹ (Fig. 1), which is similar to DAPI concentrations (average, 2.4 µg ml¹) typically used for enumerating soil and sediment bacteria (6, 15). The highest fluorescence yield for cells of the stains SYBR I, SYBR II, and SYBR Gold was detected at a dilution of 1 × 10³ to 2 × 10⁴ compared to a dilution of 1 × 10² to 4 × 10³ for PicoGreen and RiboGreen. These dye concentrations were within the range of optimal dye concentrations of SYBR I, SYBR II, and PicoGreen reported for flow cytometry counts of bacteria (8-10, 16). The fluorescence yield of cells was at least an order of magnitude higher for the GFD than for DAPI, pointing to the great potential of the GFD for enumerating bacteria. The highest fluorescence yield was obtained with SYBR II. Data similar to those shown in Fig. 1 were also obtained with E. coli cells from the exponential growth phase (17).

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Variation of the fluorescence intensity of E. coli cells for DAPI, SYBR I, SYBR II, SYBR Gold, PicoGreen, and RiboGreen as a function of dye concentration. GFD concentration is given as microliters of stock solution per milliliter of dye solution.

Comparison of different fluorescent dyes and filter types. In 45% of the samples, the DAPI-polycarbonate filter method could not be used for the enumeration of bacteria, since background fluorescence was high (sandy soils at Wittenberg) or distinction between small coccoid cells and other particles was not possible due to similar staining intensities (aquifer at Bitterfeld and sandy deep-sea sediment [Table 1]). We tested whether GFD would improve the visualizing of bacteria in the problematic sandy soils and found that with all GFD the background fluorescence was reduced sufficiently so that the enumeration of cells was possible. Moreover, although detritus and inorganic particles were occasionally faintly stained with GFD, the signal of the stained cells was much brighter, thus allowing for a distinction between cells and detritus or inorganic particles. Brief centrifugation after the sonication step was a means to reduce the background fluorescence with DAPI in several (but not all) samples. However, centrifugation resulted in losses of cells and thus in an underestimation of the bacterial counts and did not improve the distinction between small cells and inorganic particles with DAPI.

View larger version (7K): [in this window] [in a new window]   FIG. 2.   GFD-stained samples collected onto Anodisc aluminum oxide membrane filters. (A) Sandy Wittenberg soil spiked with Pseudomonas sp. strain SN45(p111). Double exposure shows all bacteria as green SYBR I-stained cells, and spiked cells can be distinguished from the indigenous bacteria by extensive yellow immunofluorescence (yellow-stained spiked cells shifted slightly beside the SYBR I-stained version of spiked cells). Note that we could not obtain cell counts in this sample with the DAPI-polycarbonate filter method (Table 1). (B) Sandy River Enns sediment. Cells were stained with SYBR II.

Conclusion. Our data indicate that all GFD can be used for staining bacteria in soils and sediments. SYBR I, SYBR II, and SYBR Gold can be used at higher dilutions than those of PicoGreen and RiboGreen. From a 1-ml stock solution of SYBR I, SYBR II, and SYBR Gold provided by the manufacturer, 10,000 samples can be stained when a 5 × 10⁴ dilution of the stock solution is used. Since the fluorescence yield was ca. 1.5 times higher for SYBR II than for SYBR I and SYBR Gold, this dye is the most promising candidate to stain bacteria for epifluorescence enumeration. In addition to the enumeration of cells, the new method can also be used to estimate size, shape, and biomass of cells, which are important parameters in soil and sediment microbiology (13). Preliminary data indicate that due to the bright staining of the cells and the low fluorescence of other particles the new method is very promising for automated image analysis (17), which is a major improvement in the investigation of soil bacteria (1). Overall, the use of SYBR II and Anodisc filters represents a rapid, precise, and inexpensive method for counting bacteria in soils and sediments.