Skip to main page content

• HOME
• Current Issue
• Archive
• Alerts
• About ASM
• Contact us
• Tech Support
• Journals.ASM.org

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Inoue, K. Articles by Kogure, K. Search for Related Content PubMed PubMed Citation Articles by Inoue, K. Articles by Kogure, K. Agricola Articles by Inoue, K. Articles by Kogure, K.

 Previous Article  |  Next Article 

Applied and Environmental Microbiology, February 2007, p. 1049-1053, Vol. 73, No. 4
0099-2240/07/$08.00+0     doi:10.1128/AEM.01158-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Separation of Marine Bacteria according to Buoyant Density by Use of the Density-Dependent Cell Sorting Method

Katsuyuki Inoue,^1* Masahiko Nishimura,¹ Binaya B. Nayak,² and Kazuhiro Kogure¹

Ocean Research Institute, The University of Tokyo, Nakano, Tokyo 164-8639, Japan,¹ Central Institute of Fisheries Education, Seven Bungalows, Versova, Mumbai 400 061, India²

Received 19 May 2006/ Accepted 1 December 2006

Top Abstract Introduction Materials and Methods Results Discussion References

 
The purpose of this study was to test whether some phylogenetic groups of natural marine bacteria have unique buoyant densities that allow them to be separated by the density-dependent cell sorting (DDCS) method. We first concentrated a natural bacterial assemblage to collect sufficient numbers of cells. They were separated into three fractions by DDCS, and the community structure in each was clarified by fluorescence in situ hybridization. The cells of Archaea tended to appear in the high-density fraction, whereas those of Cytophaga-Flavobacterium-Bacteroides were in the low-density fraction. We also calculated the sedimentation velocities of three typical marine bacteria (low density, middle density, and high density) using their buoyant density. The sedimentation velocities were approximately 10, 20, and 30 µm h^–1; these velocities have ecological implications when the heterogeneity of bacteria is considered at a microscale. To our knowledge, this is the first report on the buoyant density of natural marine bacteria.

Top Abstract Introduction Materials and Methods Results Discussion References

 
Buoyant density is a fundamental physical characteristic of aquatic unicellular organisms, such as phytoplankton and bacteria. For phytoplankton, this parameter must be accounted for when sedimentation velocities are considered. However, virtually no attention has been paid to this trait in natural communities of marine bacteria, because the implications of buoyant density of marine bacteria are less apparent.

Differences in buoyant densities can be used to separate particles, and the density-dependent cell sorting (DDCS) method has been applied to laboratory-cultured bacteria. For instance, cells in different physiological states have been successfully separated using this approach (see, e.g., references 9, 12, and 13), because physiological changes alter cellular components and the subsequent buoyant density. The DDCS method has been applied mostly to pure cultures, however, and only once to natural assemblages. To our knowledge, only one study has used DDCS to separate active from nonactive cells in natural seawater after short-term incubation with 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-phenyl-2H-tetrazolium chloride (INT) (22); the deposition of formazan on active cells increased their density, which made it possible to separate cells using the DDCS method.

The buoyant density of natural bacterial communities may have ecological implications when the behavior of bacteria is considered at a microscale. Natural bacterial communities are composed of physiologically and taxonomically different groups of cells, but it is not clear how these two factors affect the apparent buoyant density of each group. Because the nutrient level in natural seawater is generally low and relatively constant (about 1 mg of C liter^–1) (14), there should be no or very few cells in a physiological state corresponding to the exponential-growth phase achieved in the laboratory. In natural bacterial assemblages, the variation of physiological states may not be as large as that seen in batch culture, which contains cells in lag, exponential, and stationary phases. Therefore, we assumed that unique buoyant densities may be detectable among at least some phylogenetic groups and that those cells can be separated using the DDCS method.

In this study, we first concentrated a natural bacterial assemblage in order to collect sufficient numbers of cells. They were separated into three fractions using DDCS, and the community structure in each fraction was clarified by fluorescence in situ hybridization (FISH). The cells of Archaea tended to appear in the high-density fraction, whereas those of Cytophaga-Flavobacterium-Bacteroides (CFB) were in the low-density fraction. To our knowledge, this is the first report on the buoyant density of natural marine bacteria.

Top Abstract Introduction Materials and Methods Results Discussion References

 
Using a sterile plastic bucket, 4 liters of surface seawater was collected in Aburatsubo Inlet (35°09.5'N, 139°36.5'E; Sagami Bay, Japan) on 28 July 2004, 25 October 2004, 24 January 2005, and 18 April 2005. The samples were prefiltered using a Nuclepore polycarbonate filter (diameter, 47 mm; pore size, 3.0 µm; type PC MB 111112; Whatman, Middlesex, United Kingdom). Next, samples were passed through a Nuclepore polycarbonate filter (diameter, 90 mm; pore size, 0.2 µm; type PC MB 111706; Whatman) in an ultrafiltration system (UHP-90K; Advantec, Tokyo, Japan). Nitrogen pressurizing below 1 MPa was applied, and particles were concentrated to 133 to 200 times their original concentrations. The samples were kept below 4°C, and the filtering and concentration steps were completed within 4 h.

Three subsamples were taken from the concentrate and treated in parallel. Each subsample was fractionated using DDCS according to the method of Nishino et al. (13), with slight modifications. In brief, we used a Percoll (Amersham Biosciences, Uppsala, Sweden) gradient working solution containing 61 to 63% Percoll, 10% 4 M NaCl, 10% 10x phosphate-buffered saline (pH 7.4), and 17 to 19% distilled water. Density marker beads (Amersham Biosciences) were added to one tube to measure the density of the working solution with a density gradient. One milliliter of sample was layered on top of 9 ml of the working solution, followed by ultracentrifugation for 20 min at 50,512 x g and 4°C with an SW40 Ti rotor (Beckman, CA) in a Beckman Optima XL-90 centrifuge (13). After ultracentrifugation, the samples were divided into three fractions (top, middle, and bottom) using a Piston Gradient Fractionator (Biocomp, Fredericton, Canada). The three fractions contained particles for which the buoyant densities were <1.064, 1.064 to 1.074, and >1.074 g cm^–3, respectively.

Each fraction was diluted with autoclaved, 0.2-µm-filtered artificial seawater and fixed with paraformaldehyde (final concentration, 2%). After staining with DAPI (4',6'-diamidino-2-phenylindole), total cell counts were obtained by epifluorescence microscopy (BH-2; Olympus, Tokyo, Japan) (15). For each sample, at least 20 fields and more than 400 cells were counted.

Each fractionated sample was filtered through Isopore polycarbonate filters (diameter, 25 mm; pore size, 0.2 µm; type GTTP 02500; Millipore, Eschborn, Germany) by applying a vacuum of <100 kPa. The filters and samples were stored below –20°C until treatment. Cy3-labeled oligonucleotides were purchased from QIAGEN (Tokyo, Japan). The probe names, target positions, sequences, and references are given in Table 1. Probes BET42a and GAM42a were used with competitor oligonucleotides (11). Each filter was cut into eight sections, placed on glass slides covered with Parafilm (American National Can, Chicago, IL), and covered with 30 µl of hybridization solution containing 29.4 µl of hybridization buffer (Table 2) and 0.6 µl of 250-ng µl^–1 Cy3-labeled oligonucleotide probe. After incubation at 46°C for 90 min in a moist chamber (4), the filters were transferred to a vial containing 20 ml of prewarmed (48°C) washing solution (Table 2) and incubated at 48°C for 15 min (4). The filter sections were dried on Toyo filter paper (Advantec), placed on Parafilm, and covered with 50 µl of DAPI solution (1 µg ml^–1 in distilled water run through a 0.2-µm-pore-size filter) for 5 min at room temperature in the dark. Next, they were gently washed in 50 ml of 0.2-µm-filtered distilled water, dried on Toyo filter paper, and mounted on glass slides with Citifluor AF1 (Citifluor Ltd., Canterbury, United Kingdom) (4). After hybridization, the bacteria on the filter sections were enumerated by epifluorescence microscopy (BH-2; Olympus) with a Cy3 filter set. For each sample and probe, at least 20 fields and more than 400 cells were counted. All counts were corrected by subtracting the count obtained with the NON338 probe (negative control).

View this table: [in this window] [in a new window]

TABLE 1. Oligonucleotide probes used in this study

View this table: [in this window] [in a new window]

TABLE 2. Composition of hybridization and wash buffers used in this study

Top Abstract Introduction Materials and Methods Results Discussion References

 
Cell concentration.
The concentration efficiency (the ratio of absolute cell numbers in the seawater before and after concentration, expressed as a percentage) of DAPI-positive cells ranged from 88% to 96% (Table 3). With regard to the probe-specific cells, the concentration efficiency ranged from 66% (25 October 2004, CF319a) to 100% (24 January 2005, GAM42a; mean, 91%; standard deviation [SD], 9.2%; coefficient of variation [CV], 10%; n = 24) (Table 3). During this concentration process, there were no changes in the relative abundance of each phylogenetic group (P < 0.05 by one-way analysis of variance; P < 0.05 by the Bartlett test).

View this table: [in this window] [in a new window]

TABLE 3. Concentration efficiencies of DAPI-positive and probe-specific cells in each sample

Total (DAPI-positive) cell number in each fraction.
Each subsample was divided into three fractions: top (buoyant density, <1.064 g cm^–3), middle (1.064 to 1.074 g cm^–3), and bottom (>1.074 g cm^–3). The bacteria enumerated after staining with DAPI in each fraction are referred to as "total bacteria." For the four sample periods, the sum of total bacteria in the three fractions ranged from 49% to 71% (mean, 61%; SD, 9.1%; CV, 15%; n = 4) of the numbers before the fractionation. We were unable to recover any larger fraction of bacteria under the present conditions. We did not notice any morphological or size differences among the bacteria in the three fractions. The relative proportions of total bacteria in each fraction were 26 to 48% (mean, 40%; SD, 9.5%; CV, 24%; n = 4), 16 to 54% (mean, 32%; SD, 16%; CV, 50%; n = 4), and 20 to 41% (mean, 29%; SD, 9.2%; CV, 31%; n = 4) (Fig. 1) in the top, middle, and bottom, respectively. The CVs among the three subsamples were small (0.4 to 6.7%, 12 sample sets).

View larger version (38K): [in this window] [in a new window]

FIG. 1. Relative proportions of bacteria stained with DAPI or FISH probes. Ju, July; O, October; Ja, January; A, April.

Probe-specific cell number in each fraction.
We collected samples during four different seasons. The percentages of probe-specific cell numbers to total bacteria are shown in Fig. 1. The CV was 0.24 to 13% (72 sample sets; mean, 3.8%; SD, 3.2%) in three tubes of all probe-specific cells. There was a general trend in the relative proportion of each phylogenetic group (Fig. 1). The cells belonging to CFB were more abundant in the top fraction than in the bottom fraction. Conversely, cells belonging to Archaea were more abundant in the bottom fraction than in the top fraction (Fig. 1).

Top Abstract Introduction Materials and Methods Results Discussion References

 
Concentrated cells from natural seawater were separated into three fractions by DDCS, and the community structure in each was clarified by FISH. The cells of Archaea tended to appear in the high-density fraction, whereas CFB cells were in the low-density fraction. This result shows that some taxonomic groups have unique buoyant densities.

Cell concentration.
For DDCS, 1 ml of sample was loaded on top of the working solution. To obtain a sufficient number of cells for FISH with different types of probes, it was necessary to concentrate the cells prior to the procedures. After a preliminary investigation (data not shown), we decided to obtain a suspension with approximately 10⁸ cells ml^–1, which requires 100- to 200-fold concentration of natural seawater samples. There have been only a few reports of the concentration of natural aquatic bacteria. The concentration efficiency of the present method was about 90%, which is comparable to that for cultured Escherichia coli cells using the tangential flow system (20). We found little difference in concentration efficiencies between probe-specific groups (mean, 87%; SD, 11%; CV, 12%; n = 18), indicating that there was little bias during the concentration process (Table 3). In this study, after samples were divided by DDCS, the FISH method was used, because it gives the most reliable quantitative data among culture-independent techniques. If another method requiring fewer cells were applied, however, the cell concentration step might not be necessary.

Application of DDCS to a natural marine bacterial assemblage.
In previous studies, DDCS was applied exclusively to cultured bacterial cells under experimental conditions (see, e.g., references 9, 12, and 13). The application of DDCS to a natural bacterial population by using INT was specifically intended to increase the density of actively respiring bacteria (22); thus, it was not based on the differences in buoyant densities inherent to natural communities. When applied to natural assemblages, the DDCS method should separate the cells solely on the basis of their buoyant densities and not on the basis of other factors such as cell size or clumping. If the sedimentation equilibrium had not been established under the present ultracentrifugation condition, those factors might have affected the apparent results. Assuming that the cell radius is 0.3 µm and the buoyant density is 1.087 g cm^–3, the cell may move 163 mm under the present ultracentrifugation condition. Since density marker beads with a density of 1.087 g cm^–3 locate at a depth of 65 mm from the surface in the test tube, the equilibrium should be established within the ultracentrifugation time period. The effect of the cell shape also does not cause a problem, because a prolate ellipsoid with an axial ratio of 20 may move 82 mm in our centrifugation condition. Therefore, factors such as cell size, morphology, and clumping should hardly affect the present results. Microscopic observation supports this idea.

Using our present protocol, about 61% of the natural marine bacterial assemblage was recovered; part of the remaining 39% may have undergone cellular lysis during the procedure. In order to separate cells according to their original buoyant densities, cells were fixed after the DDCS step. Some ghost cells, if any existed, might not be able to withstand ultracentrifugation with Percoll. Another possibility is that there were some "heavy bacteria" in our samples. After ultracentrifugation with Percoll, condensed materials remained in the bottom of the centrifugation tube. Because we could not recover cells from these materials, the possible presence of heavy bacteria remains to be investigated. Such heavy bacterial cells may be associated with condensed elements or metals (5).

Buoyant density of marine Archaea and Eubacteria.
Using DDCS, we revealed that free-living cells belonging to CFB had lower buoyant densities (distributed in the top fractions) than did free-living cells of Archaea (bottom fraction) (Fig. 1). This finding coincides with the vertical distribution of these groups in natural seawater. CFB bacteria tend to live near the surface, especially in association with phytoplankton (8), whereas Archaea comprise half or more of the total bacteria in the deep ocean, especially below 1,000 m (7, 19). It is not clear which cellular characteristics account for this difference in buoyant density, although the membranes of archaeal cells are typically composed of archaeol and caldarchaeol, whereas the membranes of eubacterial cells are ester lipid. Thus, biochemical investigations may clarify the factors involved in buoyant density. We were unable to find trends in the buoyant densities of other phylogenetic groups in this study, perhaps because they are mixtures of physiologically different cells and/or because each group comprises subgroups possessing a wide range of unique densities.

Calculation of bacterial sedimentation velocity.
Because this work provides the first quantitative data on the buoyant density of marine bacterial communities, we also estimated their sedimentation velocity using Stokes' law (6). We assumed the following for marine bacteria and seawater: the cell is a sphere with a radius of 0.3 µm; the buoyant densities of low-, middle-, and high-density bacteria and seawater are 1.047, 1.067, 1.087, and 1.027 g cm^–3, respectively; and the temperature of the seawater is 5°C. Based on these assumptions, the sedimentation velocities of low-, middle-, and high-density bacteria are about 10, 20, and 30 µm h^–1, respectively. These values increase by 1.5 times at 20°C.

This sedimentation velocity value is small, and marine bacteria may not actually be able to move through water masses or currents with physicochemical factors and densities different from those assumed in our calculation. However, a sedimentation velocity of about 10 to 30 µm h^–1 would indicate that the cells can be affected by the downward force of gravity, and this may result in their vertical movements. Recent works have indicated that there is a heterogeneous distribution of bacterial cells at a microscale (2, 16, 17, 18). Bacterial sedimentation velocity depending on buoyant density may contribute to such distribution patterns and may offer new insight into this heterogeneity. Bacterial variability at this scale, such as microzones or hot spots, will lead to localized biogeochemical cycling (18). Therefore, the buoyant density of marine bacteria is important in considering the distribution at such scales.

In conclusion, DDCS was applied to natural marine bacterial assemblages, and the community structure in each fraction was examined by FISH. CFB bacteria tended to have low densities, whereas Archaea tended to have high densities. The apparent sedimentation velocity calculated according to Stokes' law was roughly 10 to 30 µm h^–1. To our knowledge, this is the first report of the fractionation of marine bacteria based on their buoyant densities. We are currently investigating the relationship between buoyant density and bacterial functions in natural seawater.

 
We are grateful to the technicians and staff of the Misaki Marine Biological Station, The University of Tokyo, for support during samplings. We also thank Hiroshi Ogawa for excellent technical advice on the bacterial concentration techniques and Hideaki Nomura for valuable comments and discussion.

This work was partly supported by the Sasakawa Scientific Research Grant from The Japan Science Society and by a Research Project Grant-in-Aid for Scientific Research (A) (18201003), funded by the Ministry of Education, Science, Sports and Culture of Japan.

 
* Corresponding author. Mailing address: Ocean Research Institute, The University of Tokyo, 1-15-1 Minamidai, Nakano, Tokyo 164-8639, Japan. Phone: 81 3 5351 6834. Fax: 81 3 5351 6482. E-mail: kix{at}ori.u-tokyo.ac.jp.

Published ahead of print on 8 December 2006.

Top Abstract Introduction Materials and Methods Results Discussion References

 

1
1. Amann, R. I., B. J. Binder, R. J. Olson, S. W. Chisholm, R. Devereux, and D. A. Stahl. 1990. Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analyzing mixed microbial populations. Appl. Environ. Microbiol. 56:1919-1925.[Abstract/Free Full Text]
2
2. Azam, F., and A. Z. Worden. 2004. Microbes, molecules, and marine ecosystems. Science 303:1622-1624.[Abstract/Free Full Text]
3
3. Brosius, J., T. J. Dull, D. D. Sleeter, and H. F. Noller. 1981. Gene organization and primary structure of a ribosomal RNA operon from Escherichia coli. J. Mol. Biol. 148:107-127.[CrossRef][Medline]
4
4. Glöckner, F. O., B. M. Fuchs, and R. Amann. 1999. Bacterioplankton compositions of lakes and oceans: a first comparison based on fluorescence in situ hybridization. Appl. Environ. Microbiol. 65:3721-3726.[Abstract/Free Full Text]
5
5. Heldal, M., K. M. Fagerbakke, P. Tuomi, and G. Bratbak. 1996. Abundant populations of iron- and manganese-sequestering bacteria in coastal water. Aquat. Microb. Ecol. 11:127-133.[CrossRef]
6
6. Johnson, B. D., and P. E. Kepkay. 1992. Colloid transport and bacterial utilization of oceanic DOC. Deep-Sea Res. 39:855-869.
7
7. Karner, M. B., E. F. DeLong, and D. M. Karl. 2001. Archaeal dominance in the mesopelagic zone of the Pacific Ocean. Nature 409:507-510.[CrossRef][Medline]
8
8. Kirchman, D. L. 2001. The ecology of cytophaga-flavobacteria in aquatic environments. FEMS Microbiol. Ecol. 39:91-100.
9
9. Makinoshima, H., A. Nishimura, and A. Ishihama. 2002. Fractionation of Escherichia coli cell populations at different stages during growth transition to stationary phase. Mol. Microbiol. 43:269-279.[CrossRef][Medline]
10
10. Manz, W., R. Amann, W. Ludwig, M. Vancanneyt, and K.-H. Schleifer. 1996. Application of a suite of 16S rRNA-specific oligonucleotide probes designed to investigate bacteria from the phylum Cytophaga-Flavobacter-Bacteroides in the natural environment. Microbiology 142:1097-1106.[Abstract/Free Full Text]
11
11. Manz, W., R. Amann, W. Ludwig, M. Wagner, and K.-H. Schleifer. 1992. Phylogenic oligodeoxynucleotide probes for the major subclass of Proteobacteria: problems and solutions. Syst. Appl. Microbiol. 15:593-600.
12
12. Nayak, B. B., E. Kamiya, T. Nishino, M. Wada, M. Nishimura, and K. Kogure. 2005. Separation of active and inactive fractions from starved culture of Vibrio parahaemolyticus by density-dependent cell sorting. FEMS Microbiol. Ecol. 51:179-186.[CrossRef][Medline]
13
13. Nishino, T., B. B. Nayak, and K. Kogure. 2003. Density-dependent sorting of physiologically different cells of Vibrio parahaemolyticus. Appl. Environ. Microbiol. 69:3569-3572.[Abstract/Free Full Text]
14
14. Ogawa, H., and E. Tanoue. 2003. Dissolved organic matter in oceanic waters. J. Oceanogr. 59:129-147.[CrossRef]
15
15. Porter, K. G., and Y. S. Feig. 1980. The use of DAPI for identifying and counting aquatic microflora. Limnol. Oceanogr. 25:943-948.
16
16. Seymour, J. R., J. G. Mitchell, L. Pearson, and R. L. Waters. 2000. Heterogeneity in bacterioplankton abundance from 4.5-millimeter resolution sampling. Aquat. Microb. Ecol. 22:143-153.[CrossRef]
17
17. Seymour, J. R., J. G. Mitchell, and L. Seuront. 2004. Microscale heterogeneity in the activity of coastal bacterioplankton communities. Aquat. Microb. Ecol. 35:1-16.
18
18. Seymour, J. R., L. Seuront, and J. G. Mitchell. 2005. Microscale and small-scale temporal dynamics of a coastal planktonic microbial community. Mar. Ecol. Prog. Ser. 300:21-37.[CrossRef]
19
19. Teira, E., T. Reinthaler, A. Pernthaler, J. Pernthaler, and G. J. Herndl. 2004. Combining catalyzed reporter deposition-fluorescence in situ hybridization and microautoradiography to detect substrate utilization by bacteria and archaea in the deep ocean. Appl. Environ. Microbiol. 70:4411-4414.[Abstract/Free Full Text]
20
20. Trinel, P. A., and H. Leclerc. 1976. Concentration of bacteria in water using the ultrafiltration method. Ann. Microbiol. 127:201-212. (In French.)
21
21. Wallner, M., R. Amann, and W. Beisker. 1993. Optimizing fluorescent in situ hybridization with rRNA-targeted oligonucleotide for flow cytometric identification of microorganisms. Cytometry 14:136-143.[CrossRef][Medline]
22
22. Whiteley, A. S., M. R. Barer, and A. G. O'Donnell. 2000. Density gradient separation of active and non-active cells from natural environments. Antonie Leeuwenhoek 77:173-177.[CrossRef][Medline]

---

Applied and Environmental Microbiology, February 2007, p. 1049-1053, Vol. 73, No. 4
0099-2240/07/$08.00+0     doi:10.1128/AEM.01158-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Inoue, K. Articles by Kogure, K. Search for Related Content PubMed PubMed Citation Articles by Inoue, K. Articles by Kogure, K. Agricola Articles by Inoue, K. Articles by Kogure, K.