- Email: [email protected]
A species-specific bacterial productivity method using immunomagnetic separation and radiotracer experiments
Journal of Microbiological Methods 28 (1997) 207–219
Journal of Microbiological Methods
A species-specific bacterial productivity method using immunomagnetic separation and radiotracer experiments D.G. Bard, B.B. Ward* Ocean Sciences Department, A316 EMS, University of California, Santa Cruz, Santa Cruz, CA 95064, USA Received 27 August 1996; revised 17 January 1997; accepted 23 January 1997
Abstract A species-specific bacterial productivity assay, combining radiotracer methods with immunomagnetic bead separation, was developed and tested in the marine environment. The capture method was optimized using cultures of the marine denitrifying strain, Pseudomonas stutzeri (ATCC 14 405). Immunocapture was optimal at a bead to target cell ratio of 10:1 using an indirect antibody technique, in which the target cell is first incubated with specific (primary) polyclonal antiserum and then with the secondary antibody-coated beads. Primary antibody concentration was less important than target cell concentration in determining the efficiency of target cell recovery. Reproducible recovery efficiencies of 75% could be obtained using cultures, but at natural seawater abundance levels, efficiency was much lower, around 20%. Estimates of total heterotrophic bacterial production and P. stutzeri production, based on radiotracer incorporation, were obtained for seawater samples from Monterey Bay, CA. To measure species-specific production, samples were incubated with radiotracers (methyl-[ 3 H]thymidine and [ 14 C]-leucine), fixed, and concentrated. After separating P. stutzeri cells from the bacterial assemblage using immunomagnetic separation, target cell fraction radioactivity was measured. P. stutzeri abundance, estimated by immunofluorescence, represented less than 0.1% of the total bacterial abundance, whereas radiotracer incorporation by the target fraction represented 1–3% of the total assemblage tracer incorporation. 1997 Elsevier Science Ireland Ltd.
1. Introduction Measurement of total bacterial secondary production using radioactively labeled organic substrates [1–4] has become the standard method for evaluating the role of bacteria in aquatic food webs. This approach can be used to estimate the productivity of the total heterotrophic bacterial assemblage in seawater, but it cannot provide information on the contribution of individual species to the overall rates.
*Corresponding author. Tel.: 11 408 4593171; fax: 11 408 4594882; [email protected]
Metabolic rates of bacterial species have been measured in pure cultures, but it is not realistic to extrapolate those results to natural systems when that same species is part of a larger bacterial assemblage. The simplest approach might be to enumerate individual strains present in the natural assemblage and assume that their contribution to production is proportional to their numbers. Even this approach is not usually possible with currently available methods. Thus, with minor exceptions, it has not been possible to assess the contribution of individual species to total bacterial abundances nor to bacterially mediated processes in the environment. Antibodies which are specific for individual strains or species offer a basis for species-specific identifica-
0167-7012 / 97 / $17.00 1997 Elsevier Science Ireland Ltd. All rights reserved PII S0167-7012( 97 )00987-1
208
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
tion, enumeration and, in conjunction with immunomagnetic beads, species-specific productivity estimates. Immunofluorescence has been used to enumerate several kinds of biogeochemically important bacteria in terrestrial, marine and freshwater systems [5–8]. The same antibodies can be used for bacterial cell separation, for example, using affinity columns [9] or antibody coated magnetic beads. Originally developed in the medical field for the fractionation of specific human blood cells [10], immunomagnetic separation (IMS) has been applied in microbiological research for the detection and enrichment of selected bacterial species [11–14]. Mammalian antibodies confer a high degree of specificity and sensitivity on this approach. Antibodies developed against a wholecell antigen (primary antibodies) added to a heterogeneous cell sample should bind selectively to target cells. Antibodies against the primary antibody (i.e., secondary antibodies), covalently bound to the outside of magnetic microspheres (known as immunomagnetic beads, IMB), can then bind the target cells to the beads. A magnet is then used to separate and isolate target cells from the sample. A general description of bacterial activity in the water column can be obtained from production estimates, based on methods such as [ 3 H]thymidine incorporation into DNA [15] and radiolabeled leucine incorporation into protein [16]. The present study employed the dual-label method [17], which utilizes the simultaneous incorporation of [ 3 H]thymidine and [ 14 C]leucine to obtain two different measures of production. Pseudomonas stutzeri (ATCC 14 405) was chosen as the target bacterium for methods development because its metabolism is not atypical of marine bacteria – it is a heterotrophic, facultative anaerobe – and because of the availability of a well characterized P. stutzeri-specific antiserum [18]. To estimate P. stutzeri production, the entire bacterial assemblage was exposed to radiolabeled substrates and then target cells were separated from the assemblage using IMS. This study describes the development and evaluation of a method to separate P. stutzeri cells from seawater, and the use of this technique in conjunction with a standard production assay to estimate P. stutzeri production in situ. Initial results from a field study of P. stutzeri relative abundance and relative production are also presented.
2. Materials and Methods
2.1. Bacterial strains Pseudomonas stutzeri (ATCC 14 405), formerly P. perfectomarina [19], was obtained from the ATCC. TBD-8b, an unidentified marine denitrifier, was originally isolated from intertidal sediment from Tomales Bay, California [18]. P. stutzeri and TBD8b were grown in a seawater medium containing 5 g carbon (as peptone) l 21 (CP medium; [20]). To produce radiolabeled cells for IMS optimization experiments (not for productivity measurements), cells were labeled with radioisotopes during exponential growth. P. stutzeri cultures were labeled by incubating cells in a solution of diluted CP medium (1:100) with either 10 nM methyl- 3 H-thymidine (specific activity, 107.8 Ci mmol 21 ; New England Nuclear, Dupont, Wilmington, DE, USA) for 2 h, or 100 nM 14 C(U)-leucine (specific activity, 321.6 mCi mmol 21 ; New England Nuclear) for 30 min. After incubation, cells were washed several times with phosphate buffered saline (PBS; per liter of distilled water: 8 g NaCl, 0.2 g KCl, 1.15 g Na 2 HPO 4 , and 0.2 g KH 2 PO 4 ), and then preserved with formalin (37% formaldehyde, buffered with saturated sodium borate, filtered prior to use, 2% v / v final concentration).
2.2. Immunomagnetic separation The primary antibody solution was polyclonal antiserum specific for P. stutzeri [18] which was absorbed with Escherichia coli (to remove nonspecific interactions) prior to use. IMB (Dynabeads M-280, 2.8 mm in diameter, 6–7310 8 beads ml 21 , with covalently linked sheep anti-rabbit IgG (secondary antibody); Dynal A / S, Oslo, Norway) were used to perform the immunomagnetic separation. In most experiments, samples were first incubated with the primary antibody solution (allowing the antiserum to bind to target cells), and then the beads were added, allowing the antibodies, with cells attached, to bind to the secondary antibody-coated beads (the indirect technique). A bacterial culture or a concentrated seawater sample was incubated with the primary antiserum (1:100 final dilution) in 1.0 ml of a 1:1 mixture of PBS–gelatin for 3 h. Gelatin (2%
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
in water, base hydrolyzed at pH 11 and 1208C for 15 min, neutralized) was used to reduce nonspecific binding. All incubations of cells with antiserum, and subsequent incubations with IMB, were performed in volumes of 1.0 ml or less in 1.5 ml microcentrifuge tubes at 158C with continuous, end-over-end rotation. After incubation with the primary antibody, excess antiserum was removed by washing twice with PBS and resuspending cells in 1.0 ml PBS–gel (1:1) and 20 ml of IMB. After an overnight incubation with the beads, the sample was ready for the magnetic separation procedure. In a few experiments, the primary antiserum was first bound to the secondary antibody-coated beads (the direct technique) and then the bead-antibody complex was allowed to bind to target cells. In this case, IMB were incubated in 1.0 ml of the P. stutzeri antiserum (1:100 final dilution) at 158C for 3 h. The antiserum-coated beads were then washed twice with PBS and resuspended in PBS–gel (1:1). Coated beads (20 ml) were then added to a sample and incubated overnight. The sample was then ready for the separation. Separation of P. stutzeri cells (i.e., obtaining a target fraction) was performed by placing the sample epitube in a magnetic particle concentrator (MPC-E: Dynal) for 5 min. Bead-cell complexes (BCC) were drawn against the wall of the epitube and the residual suspension was removed by aspiration. To reduce the loss of BCC by aspiration, ends of pipette tips were cut to enlarge the tip inner diameter and the magnet was used to recover the complexes from the aspirated fluid. The beads were washed twice with 1.0 ml PBS, resuspended in PBS, and filtered onto a 25 mm, 0.2 mm polycarbonate filter using a ten-place filtration manifold (Hoefer Scientific Instruments, San Francisco, CA). Filtered cells could be inspected by microscopy after staining with DAPI or a fluorescently labeled secondary antibody, or if the cells were radioisotopically labeled, the filters were radioassayed (see below).
2.3. Evaluation of immunomagnetic separation method The effect of pH on IMS was tested on pure cultures of P. stutzeri. Triplicate samples of [ 14 C]leucine labeled P. stutzeri cells (9310 6 cells
209
ml 21 ) were incubated in PBS–Gel under different pH conditions (pH 4.0, 7.0, and 10.0). Recovery efficiency was not determined in the pH experiment, but the effect of pH was assessed by simply comparing radioactivity (normalized to cell number) in the target fraction of each treatment. The effect of bead to cell ratio on P. stutzeri cell recovery in pure culture conditions was tested by incubating duplicate samples of [ 14 C]leucine labeled P. stutzeri cells (3–7310 6 cells) in 1.0 ml PBS–gel (1:1) with antiserum (1:100 final dilution) for 3 h. After incubation, excess antiserum was removed by centrifugation with two washes of PBS–gel. The antiserum-coated P. stutzeri cells were resuspended in 1.0 ml PBS–Gel (1:1). To one of the duplicate samples, the bead sample, the appropriate concentration of IMB was added to yield bead to cell ratios between 0.5 and 75. After an overnight incubation, a target fraction was obtained, using the magnetic particle concentrator as previously described, and radioassayed. The other of the duplicate samples, the pre-bead sample, was incubated without the beads. After incubation, the pre-bead sample was filtered and radioassayed. Target cell recovery was calculated by dividing the bead activity by the pre-bead activity. The effect of antiserum concentration was investigated by incubating triplicate samples of [ 14 C]leucine-labeled P. stutzeri cells (9310 6 –3.63 10 7 cells ml 21 ) with a range of primary antibody concentrations, at the optimum bead to cell ratio determined above (10). Efficiency of cell recovery was determined by comparing the radioactivity recovered in the bead fraction with that in an aliquot containing the same initial number of filtered cells (i.e., a pre-bead fraction). A sensitivity experiment was performed using different concentrations of [ 14 C]leucine- labeled cultured P. stutzeri in 900 ml of sterile filtered seawater. P. stutzeri concentrations ranged from 80– 1300 cells ml 21 . The samples were then concentrated by filtering the sample through 47 mm, 0.2 mm pore size polycarbonate filters (Poretics, Livermore, CA, USA) under low vacuum pressure. Filters were placed in 5 ml PBS and the cells resuspended by vortexing for 30 s. The resuspended cells were drawn off and the filters washed twice with 4 ml PBS and these washes added to the cells. The
210
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
resuspended cells were transferred to 2.0 ml epitubes and centrifuged in a microcentrifuge (Eppendorf 5415) at 10 0003g for 4 min. Centrifugation and resuspension continued (3–4 times) until the sample volume was reduced to 1.0 ml. The concentrated samples were then processed for cell separation and recovery using the IMS method. Target fractions were filtered and the filtrate was collected. The target and nontarget fractions and the filtrates were radioassayed as described above. To estimate cell recovery, the target fraction activity was divided by the sum of target, nontarget and filtrate activities. A target fraction purity experiment was performed using seawater with an intact natural seawater bacterial population (SWB). For each treatment, 4.5 l of natural seawater were concentrated using an Omega 300K tangential filtration device (Filtron Technology, MA, USA) and a Masterflex 60 600 peristaltic pump (Cole-Parmer, Chicago) to a final volume of approximately 400 ml. Triplicate subsamples (40 ml each) from the concentrated sample were used in each of three treatments and a single 40 ml subsample was used for a dead control. Treatments included: (A) the dead control (SWB preserved with formalin); (B and D) SWB with added P. stutzeri cells (approximately 10 7 cells); (C) SWB without added P. stutzeri cells. All treatments were incubated overnight with 10 nM (final concentration) [ 14 C]leucine, and then treatments B, C, and D were preserved with formalin. Treatments were then processed using IMS, as previously described, except that treatment B was incubated in a PBS–gel solution without antiserum. To compare the efficiency and specificity of the direct and indirect techniques of binding P. stutzeri to the beads, 200 ml of dual-labeled seawater were concentrated by centrifugation to reduce the final volume to 1.0 ml. Triplicate samples were tested with each technique using the indirect and direct procedures described above.
2.4. Sampling sites Water column samples were collected in Monterey Bay California during September 1993 and November 1994, and in the Southern California Bight during April 1994. Samples were collected using 5 l Niskin or Go-Flo bottles and sampling
locations are stated in the results of each experiment. Radiotracer incubations of the Monterey Bay 1993 and the Southern California Bight 1994 samples were performed immediately after collection at in situ temperatures using a flowing seawater incubator on deck. The Monterey Bay 1994 samples were obtained aboard the R.V. Point Lobos and 2 l water samples were kept cool and dark while transported to the laboratory. Incubations were performed at in situ temperatures within 12 h of sample collection.
2.5. Total bacterial production The dual-label method [17] was performed with slight modifications. Immediately before tracer additions, water samples were filtered through a 210 mm nylon screen mesh to remove larger particulates and zooplankton. For each depth, one or two 20 ml subsamples were incubated with 5 nM (final concentration) methyl-[ 3 H]-thymidine, and 10 nM (final concentration) [ 14 ]C(U)-leucine. TCA-insoluble material was extracted using a slight modification of the original method [17]. Samples were filtered through 25 mm, 0.2 mm poresize polycarbonate filters (Poretics) using the filtration manifold (Hoefer) at low pressure (#10 cm Hg pressure differential). Filtered samples from 1993 were counted with a Packard PRIAS model PL liquid scintillation counter. Filtered samples from 1994 were counted with a Beckman LS 6500 liquid scintillation spectrometer. Quenching was corrected with the external standard in the Compton edge shift mode (H[). In both cases, dual isotope counting modes and the use of quench curves maximized the resolution of the two isotopes.
2.6. Seawater concentration Seawater samples were concentrated in order to increase the concentration of target cells. For 1993 Monterey Bay and 1994 Southern California Bight samples, 1 l of seawater was concentrated to a final volume of approximately 100 ml using the Filtron casette (see above). Concentrated samples were incubated with radiotracers, as described above, preserved with formalin, and then kept cool in the dark during transit to the laboratory. In the lab, samples were stored in the dark at 48C and processed
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
211
within one month from time of collection. (While storage for this length of time probably resulted in loss of cells, it was unavoidable for this experiment. The least biased assumption is that cell loss affected all cell types equally, thus not biasing the resulting strain specific comparison.) Samples were centrifuged in 40 ml round-bottom centrifuge bottles at 10 0003g for 30 min and the pellet was resuspended in PBS. Samples were transferred to 2.0 ml epitubes and centrifuged at 10 0003g for 4 min. Centrifugation and resuspension continued (3–4 times) until the sample volume was reduced to 1.0 ml. For 1994 Monterey Bay samples, 1 l of seawater was incubated with radiotracers as described above and preserved with formalin after the incubation. Concentration was performed after incubation by filtering the sample through 47 mm, 0.2 mm pore size Poretics polycarbonate filters using low-vacuum pressure. Filters were placed in 5 ml PBS and the cells resuspended by vortexing and repeated PBS– gel washes. The resuspended cells were transferred to 2.0 ml epitubes and centrifuged to complete the concentration step as described above. The concentrated samples were used for measurement of total bacterial radiotracer incorporation or for IMS of P. stutzeri to estimate the contribution of P. stutzeri to bacterial radiotracer uptake using the protocols described above. A radiotracer experiment was conducted on cultured P. stutzeri cells to determine the magnitude of the correction for the difference between whole cells and cold TCA insoluble material. Fifteen replicate subsamples of exponentially growing P. stutzeri cells were incubated in 1:100 diluted CP medium containing 10 nM methyl-[ 3 H]-Thymidine. After 2 h, five replicates were filtered onto 0.2 mm pore-size Gelman Supore filters. Five replicates were preserved with 2% formalin for 30 min, then filtered, and five replicates were treated with cold TCA as in the natural seawater experiments. Filters were assayed by scintillation counting. In an identical duplicate experiment, the 15 replicate samples were incubated in the presence of 20 nM [ 14 ]C-Leucine.
DAPI staining method [21] using a Zeiss Axioskop microscope with 100 W Hg illumination, a UV G365 filter for excitation, and a LP 420 barrier filter (Zeiss filter set 487 902) at 1600 power. Sample volumes of 1 ml or less of the culture and 10 to 20 ml of field samples were diluted or filtered to a volume of about 1 ml and DAPI was added at 0.35 mg ml 21 . Abundances were calculated from single filter counts or duplicates were averaged for mean abundance. A total of over 200 cells were counted for each field sample. For most field samples, DAPI and immunofluorescence (see below) enumeration were done on the same preparation, by switching between optical filter sets. The indirect immunofluorescence (IIF) staining method [8] was used to enumerate P. stutzeri in field samples preserved with formalin, using the primary antibody for P. stutzeri [18] and fluorescein isothiocyanate-conjugated sheep anti-rabbit immunoglobulin G (IgG; Miles Laboratories, Elkhardt, IN, USA) as the secondary antibody. Poretics polycarbonate filters (25 mm, 0.2 mm pore size) were stained in Irgalan Black to reduce background fluorescence. Stained samples were viewed within one day using epifluorescence microscopy. Cell counts were made at 1600 power using a Zeiss Axioskop microscope (Carl Zeiss, Germany) with 100 W Hg illumination with a BP 450 / 490 filter for excitation and a LP 520 barrier filter. P. stutzeri cells were distinguished by a staining reaction of 13 or 14 (compared to the homologous reaction of 14) and the characteristic rod shape. A total of 350 fields were inspected for each count of natural samples. Calculations and simple statistical tests for optimization of methods and comparisons of experimental treatments were performed using the spreadsheet program Excel 5.0.
2.7. Enumeration of bacteria
Several parameters which might influence recovery of target cells were tested in order to optimize the IMS method. The effects of pH, antiserum concentration, and bead to cell ratio were tested on
Enumeration of pure cultures of P. stutzeri and the total bacterial assemblage was performed using the
3. Results
3.1. Optimization of immunomagnetic separation method
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
212
Table 1 Parameters optimized for immunomagnetic separation Cell recovery(% activity)a
[ 14 C]Leucine activity a (DPM)
[3H]Thymidine activity a (DPM)
ND ND ND
5868 6669 6466
NA NA NA
Antiserum concentration c 1:20 1:50 1:100 1:150 1:300 1:800 1:1000
75615 85613 746 3 746 2 806 3 686 5 706 6
NA NA NA NA NA NA NA
NA NA NA NA NA NA NA
Binding technique d Direct Indirect
ND ND
961 5463
1062 44618
Parameter varied pH b 4.0 7.0 10.0
a
Mean6(S.D.). [ 14 C]leucine labeled P. stutzeri cells separated using indirect technique. c 14 [ C]leucine labeled P. stutzeri cells separated using indirect technique. Antiserum concentration expressed as dilution of antiserum stock concentration. d Performed on natural seawater bacterial assemblage with dual-label radiotracer incubation. Sample was collected from SCB station 202 (338299N, 117845.59W, total depth554 m). NA, not applicable; ND, not determined. b
pure cultures of P. stutzeri. No significant differences were detected among the three pH levels tested (Table 1). Bead to cell ratio had a significant effect on recovery efficiency; cell recovery varied from a low of 25% at both extremes of bead to cell ratios, to a maximum of 55% around the bead to cell ratio of 10:1 (Fig. 1). At the optimum bead to cell ratio, cell
recoveries averaged 75% (6S.D. 5.8%) and did not vary as a function of antiserum dilutions ranging from 1:20 to 1:1000 (Table 1). Indirect and direct methods were compared using replicate subsamples of a concentrated dual-labeled natural bacterial population from seawater. Activities based on both radiotracers were much higher for the indirect technique than for the direct technique (Table 1), and no further experiments using the direct technique were performed.
3.2. Target fraction purity
Fig. 1. Cell recovery as a function of bead to cell ratio (note log scale). [ 14 C]leucine labeled P. stutzeri cells were recovered by indirect IMS.
Purity of the target cell fraction was assessed indirectly by comparing the target fraction activity of [ 14 C]leucine-labeled P. stutzeri cells with the target fraction activity of [ 14 C]leucine-labeled P. stutzeri cells plus an approximately ten-fold excess of nonlabeled TBD-8b cells. The target fraction activities did not differ (data not shown) suggesting that the TBD-8b cells did not interfere with target cells binding to the beads. Target fraction purity was also examined by
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
213
the range of concentrations tested were approximately 20% (data not shown). The target fraction activity was plotted against P. stutzeri concentration, and a best-fit line was calculated (Fig. 3). Using the regression function, and setting the lower limit of significant detectable activity as twice background activity ( y-intercept), the limit of detection was 324 cells ml 21 .
3.4. Total bacterial and target fraction radiotracer incorporation Fig. 2. Activity of target cell fractions from treatments using natural bacterial assemblage from seawater (SWB) and P. stutzeri cells. Triplicate samples for treatments B–D; error bars represent standard deviation: Treatment A, dead control: SWB, P. stutzeri cells (preserved with formalin) 118 (primary) antibody added during IMS; treatment B, live SWB1P. stutzeri cells, no primary antiserum added during IMS; treatment C, live SWB118 antibody added during IMS; treatment D, live SWB1P. stutzeri cells118 antibody added during IMS.
comparing the target fraction activity of four different treatments using a natural seawater sample and P. stutzeri cell additions (Fig. 2). The dead control (treatment A) showed no significant activity. Some activity was associated with the treatment B, even though no antiserum had been added, showing that minor nonspecific binding occurred. The activity in treatment C was not significantly higher than in treatment B (P.0.20). These results imply that the natural population of P. stutzeri was small and that non-specific antibody binding was negligible. The activity in treatment D was much higher than the other three treatments (P,0.10), reflecting the addition of cultured target cells to the resident population. The recovery efficiency of this experiment cannot be determined from radioactivity alone because the added P. stutzeri cells acquired their radioactivity as part of the experiment and could not be independently assayed.
Radiotracer incubations using concentrated and unconcentrated samples were compared to determine whether the concentration procedure influenced bacterial production estimates. Samples were concentrated using a tangential flow filtration device and incubated alongside unconcentrated samples using the same radiotracer procedures. In most cases, concentration resulted in rates which were disproportionately lower than predicted from the concentration factor (data not shown). This is consistent with the observed decrease in cell number associated with incubation of concentrated samples (see above). To avoid preconcentration artefacts, the natural seawater experiments in May 1994 were incubated with radiotracers and then concentrated. Radioisotope incorporation estimates from Monterey Bay in 1994 were obtained using a dual-label method. For total bacterial incorporation, radioactivity of the cold TCA-insoluble cell material was
3.3. Limit of detection The sensitivity of this method was investigated by performing the IMS procedure on different concentrations of [ 3 H]thymidine-labeled P. stutzeri cells in 1.0 l of filtered seawater. Cell recoveries across
Fig. 3. Limit of detection for IMS recovery of cultured P. stutzeri cells from natural seawater determined from y-intercept of linear regression ( y50.112x13.759, r 2 50.875).
214
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
measured. For target fraction incorporation, whole cell radioactivity was measured. TCA-insoluble radiotracer incorporation determined from experiments with cultured P. stutzeri was 46% for thymidine and 80% for leucine; target fraction incorporation rates were corrected for these factors. A second correction was made for estimated cell recovery efficiency of 20% at the typical P. stutzeri cell concentration encountered in natural samples (50– 100 cells ml 21 ). Corrected radiotracer incorporation rates for the target fraction and total rates were plotted against depth using a log scale to allow display of the large range of magnitudes (Fig. 4). Similar trends with depth were observed for both total and P. stutzeri rates. Highest incorporation occurred near the surface and the rates decreased with depth. Incorporation estimates for the target fraction at most depths were approximately two orders of magnitude less than incorporation by the total bacterial fraction. The contribution of P. stutzeri target cells to the total bacterial incorporation decreased with depth ranging from a minimum of 1.2% to a maximum of 3.3% (Table 2).
Table 2 Target fraction contribution to total bacterial production a Depth (m)
10 20 100 200
% Total tracer incorporation due to target fraction b Thymidine
Leucine
2.8 2.8 1.5 1.3
3.3 2.7 2.2 1.2
a
Monterey Bay in November 1994, central station, total depth5 260 m. b Mean % contribution (6S.D.)52.1460.82 for thymidine, 2.3660.89 for leucine.
3.5. Enumeration of bacteria Total bacteria and P. stutzeri in Monterey Bay were enumerated using DAPI and indirect immunofluorescence, respectively (Table 3). For every sample examined, P. stutzeri cells were present, usually at less than 110 cells ml 21 , and accounted for much less than 1% (usually less than 0.1%) of the total bacterial assemblage. With two exceptions, relative abundance of P. stutzeri was similar at all nine depths, even though P. stutzeri abundance in the water column decreased with increasing distance from shore and with increasing depth. Cell loss during radiotracer incubations was examined by enumerating total bacterial cells before and after incubation in samples of concentrated and unconcentrated seawater which were incubated alongside production samples (Table 4). There was little cell loss observed in unconcentrated samples, but a significant decrease in cell number occurred in concentrated samples.
4. Discussion
4.1. Evaluation of immunomagnetic separation Fig. 4. Radiotracer incorporation by target fraction and total bacterial assemblage. Samples were collected from the Monterey Bay water column at a central station (maximum depth at 260 m). Target fraction estimates (open symbols) and total bacterial incorporation (closed symbols) based on radiotracer uptake of thymidine (h, j) and leucine (x, ♦).
Initial characterization and optimization of the immunomagnetic separation method were performed using pure cultures of P. stutzeri. Bead and antiserum concentrations relative to target and non-target cell concentration were expected to be the most im-
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
215
Table 3 Relative abundance of P. stutzeri in Monterey Bay seawater samples a Total bacteria b (cells ml 21 )
P. stutzeri c (cells ml 21 )
Relative abundance (% total)
Station 1156: 36854.979 N, 121856.439 W (total depth526 m) 1 2.2310 5 10 1.0310 5 24.5 1.4310 5
110 60 55
0.05 0.06 0.04
Station 1116: 36843.519 N, 121856.289 W (total depth597 m) 5 1.4310 5 15 3.0310 5 95 3.0310 4
70 50 50
0.05 0.02 0.20
Station H3: 36846.259 N, 122800.929 W (total depth5911 m) 5 8.0310 4 150 2.0310 4 908 3.0310 4
30 40 20
0.04 0.20 0.07
Depth (m)
a
Samples collected in September, 1993. Enumeration by DAPI staining method [21]. c Enumeration by IIF staining method [8]. b
portant variables in determining the efficiency of the separation procedure. A wide range of antiserum concentrations worked equally well, with an average activity recovery of 75%. The antiserum dilution used in these experiments, 1:100, was that found previously to optimize strength of the immunoflourescence staining reaction with this serum, without causing non-specific staining of heterologous cells [18]. In a separate set of experiments designed to
optimize the bead to cell ratio, performed using an antiserum dilution of 1:100, the maximum activity recovery was only 55%. The difference between average recovery in the bead to cell ratio experiment and the antiserum concentration experiment is probably due to the introduction of an additional wash and recovery step in the latter. In the bead to cell ratio experiment, beads removed during the washes were not recovered by a secondary magnetic recovery step, which was incorporated into the protocol for
Table 4 Total bacterial counts before and after productivity incubation a Sample b
Depth (m)
Total cell count before (cells ml 21 )
Total cell count after (cells ml 21 )
% Cell loss
1 500 820
5.4310 5 6.9310 4 4.0310 4
4.5310 5 6.9310 4 5.3310 4
17 0 0
1 500 820
1.4310 6 4.1310 5 5.2310 5
2.4310 5 2.4310 5 2.0310 5
83 42 62
Unconcentrated
Concentrated c
a
Enumeration by DAPI staining method [21]. Southern California Bight, Station 205 (33818.79N, 118809.69W, total depth5905 m) April, 1994. c Concentration was performed using a tangential filtration device (Filtron). b
216
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
the antiserum concentration experiment and all subsequent experiments. The effect of absolute cell number or concentration of target cells was not addressed in these experiments, which were all performed with target cell concentrations in the range 10 6 –10 7 cells ml 21 . Subsequent experiments with P. stutzeri cells suspended in filtered or natural seawater at concentrations less than 1000 cells ml 21 (in the limit of detection experiments) yielded recoveries on the order of 20%, indicating that regardless of bead and antiserum concentration, recovery efficiency decreases with decreasing target cell concentration, due to low encounter frequencies. For a given target cell concentration, the separation method provides reproducible recovery, and a relationship between cell concentration and recovery efficiency can be used to normalize activity measurements. Because the antiserum had been previously characterized [18], those tests were not repeated here. However, the tests of target fraction purity implied that the specificity of the immunomagnetic separation was consistent with the specificity of the antiserum itself. When challenged with a mixture of cultured cells that contained only 10% target cells, the presence of nontarget cells did not reduce binding of target cells. This suggests that nonspecific binding is too low to interfere with the specific binding capacity of the beads. Similarly, when cultured target cells were resuspended in a natural seawater sample (at a final added target cell concentration of approximately 10 5 cells ml 21 ), the level of activity associated with nonspecific binding was low and easily distinguished from that associated with labeled target cells. The optimization experiments were designed to develop a reproducible protocol that could be applied to natural samples. These experiments illustrated that the principle of immunomagnetic separation can be applied to specific separation of bacterial cells and that under controlled conditions of antiserum, bead and cell concentration, a reproducible separation and concentration of target cells can be achieved. The major complication in applying this method to field samples is the low concentration of target cells. While there are few data available on the absolute abundance of individual strains of non-photosynthetic bacteria in seawater, previous estimates
from immunofluorescence enumeration of nitrifying and denitrifying bacteria indicate that individual serotypes are usually present at very low levels, usually less than 0.1% of the total cell abundance. At these levels, low recovery efficiencies would be expected, and even low levels of nonspecific binding could become relatively more important. The low abundance of P. stutzeri detected in the samples used in this study was consistent with previous enumeration in this environment [18] and was below the detection limit determined in Fig. 3. Such low abundances presented challenges to target fraction activity measurement; when samples were first tested using IMS, activity in the target fraction was too low to differentiate P. stutzeri production from nonspecific binding (see Fig. 2). Sample concentration was investigated as a means to increase the number of target cells in the sample. Concentration itself proved problematic. Concentration prior to radiotracer incubation has the advantage of requiring less radioisotope use and generating less isotope waste, but significant cell losses, accompanied by low productivity rates, occurred during incubation of concentrated samples (Table 4). This decrease in cell number probably was due to increased bacterial grazing by microzooplankton which survived in the concentrated sample. While this study was not intended to investigate concentration procedures exhaustively, we did note significant differences in recovery due to concentration method alone. Efficiency of concentration can be quantified, whichever procedure is used, so inefficiency at this step does not prevent quantification of species-specific activity. In the future, it may be more effective to use radiotracers with higher specific activities and thus avoid the need for sample concentration, but if concentration is required prior to IMS, it should be performed after the tracer incubation step. Low target cell abundance not only affects efficiency of target recovery, but it also affects target fraction purity. Streifel [22] reported that low target cell abundance reduced target fraction purity in IMS with eukaryotic cells. Lund et al. [23] reported similar findings using IMS of bacterial cells. The nonspecific binding observed in separations of natural samples in this study might have been caused by nontarget cells trapped amongst the beads. Although washing should reduce this nonspecific entrapment,
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
we observed bead agglutination even in the absence of bound target cells, which suggests that mechanical trapping of nontarget cells in the bead-cell complex may occur independently of the specificity of the antibody reaction. Target fraction contamination may have accounted for part of the activity differences observed between the direct and indirect immunocapture techniques. The increased activity associated with the indirect method could have been due to a greater number of target cells captured or it could have resulted from a relatively greater amount of nonspecific binding. In future applications of this method, target fraction purity and target cell recovery might be directly assessed by enumerating target and nontarget cells by immunofluorescence in all fractions and washes. This is not very practical, however, because the cell number remaining is below the limit of detection and statistically significant counts cannot be obtained. The opposite problem, clumping and three-dimensional overlapping of cells bound to beads, means that the number of cells recovered is also not easily determined. Radioisotopically labeled cells were used in these optimization experiments in order to increase the sensitivity and ease of assaying recovery and to circumvent the error and sensitivity problems associated with direct microscopic cell counts, but microscopy may also be useful.
4.2. Species-specific radiotracer incorporation rates In situ incorporation rates estimated for the total heterotrophic bacterial community in Monterey Bay are consistent with previously published rates (on the order of 1 mg C l 21 h 21 in surface waters [18]). The objective of the species specific productivity measurement using IMS was to determine what fraction of this total productivity could be attributed to individual strains present in the population. This comparison is made on the basis of radiotracer incorporation (dpm l 21 h 21 ) rather than on the basis of carbon production to avoid unnecessary assumptions about conversion factors. If all cells and all strains are equally active, then the proportion of activity due to any strain parallels its abundance. If some strains are more or less active than average,
217
then abundance and activity might be differentially distributed. In these results from Monterey Bay, the distribution of P. stutzeri tracer incorporation parallels that of the total bacterial incorporation. In the absence of selection favoring the growth of P. stutzeri (e.g., low oxygen conditions which induce denitrification), P. stutzeri functions as an aerobic heterotrophic generalist, so that its abundance and activity are simply subsets of the total heterotrophic bacterial population. The mean relative production attributed to P. stutzeri was approximately an order of magnitude larger than the mean relative abundance of P. stutzeri observed elsewhere in the bay. Because we did not quantify target purity or target cell recovery by immunofluorescence in the exact samples used for productivity measurements, spatial or temporal variation in P. stutzeri abundance may account for this difference. Alternatively, P. stutzeri may be relatively more active than some other members of the microbial community. Recent findings that not all cells are active [24,25] could be due to differences in activity among strains. It is also possible that mechanical agglutination of the beads may cause some trapping of non-target cells, such that not all of the production associated with the beads was due to P. stutzeri. More measurements of this kind are necessary to assess the significance of this species specific production estimate. Previous studies using immunological methods to investigate species-specific activity have reported variation among bacterial species within habitats. Fliermans and Schmidt [26] reported distinct differences in activity of two species of Nitrobacter in mixed culture and Ward [27] observed differences in relative activity of two species of ammonia-oxidizing bacteria in the Northeast Pacific Ocean off the Washington coast, both using immunofluorescence in combination with autoradiography. IMS could be an improvement over autoradiography because, although verification of target fraction purity and independent enumeration of target cells require microscopy, it is somewhat less tedious than counting grains on autoradiograms. IMS in combination with standard radioisotopic productivity methods should yield a direct measurement of production, rather than a relative activity index. Accurate measurements of species activity de-
218
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
pends, in large part, on assessing the effectiveness of target cell separation from the overall bacterial assemblage. In the future, using radiotracers with higher specific activity for target cells with low abundance and performing direct counts of target species using immunofluorescence to assess IMS efficiency are likely to improve species-specific production estimates using this method. This method could be used to test the hypothesis that denitrifiers contribute relatively more to bacterial production in oxygen minimum zones, i.e., where its abundance and activity, should be favored.
Acknowledgments This research was funded by an ONR contract to BBW. General assistance from Mary Hogan and Margaret Geissler is gratefully acknowledged.
References [1] Joint, I.R. and Morris, R.M. (1982) The role of bacteria in the turnover of organic matter in the sea. Oceanogr. Mar. Biol. Annu. Rev. 20, 65–118. [2] Azam, F. and Furhman, J.A. (1984) Measurement of bacterioplankton growth in the sea and its regulation by environmental conditions. In: Heterotrophic Activity in the Sea (Eds. J.E. Hobbie and P.J.L. Williams). Plenum Press, New York, pp. 179–196. [3] Karl, D.M. and Winn, C.D. (1984) Adenine metabolism and nucleic acid synthesis: Applications to microbiological oceanography. In: Heterotrophic Activity in the Sea (Eds. J.E. Hobbie and P.J.L. Williams). Plenum Press, New York, pp. 196–216. [4] Ducklow, H.W. and Carlson, C.A. (1992) Oceanic Bacterial Production. Adv. Microb. Ecol. 12, 113–181. [5] Stanley, P.M., Gage, M.A. and Schmidt, E.L. (1979) Enumeration of specific populations by immunofluorescence. In: Native Aquatic Bacteria: Enumeration, Activity, and Ecology, ASTM STP 695. (Eds. J.W. Costerton and R.R. Colwell) American Society for Testing and Materials, pp. 46–55. [6] Josserand, A.G., Gay, G. and Faurie, G. (1981) Ecological study of two Nitrobacter serotypes coexisting in the same soil. Microbiol. Ecol. 7, 275–280. [7] Dahle A.B. and Laake M. (1982) Diversity dynamics of marine bacteria studied by immunofluorescent staining on membrane filters. Appl. Environ. Microbiol. 43, 169–176. [8] Ward, B.B. and Carlucci, A.F. (1985) Marine ammonia- and nitrite-oxidizing bacteria: Serological diversity determined by immunofluorescence in culture and in the environment. Appl. Environ. Microbiol. 50, 194–201.
[9] Heyman, U., Heyman, B. and Ward, B.B. (1988) Cell affinity chromatography for a marine nitrifying bacterium. In: Immunochemical Approaches to Estuarine, Coastal and Oceanographic Questions, (Eds. C.M. Yentsch, F.C. Mague and P.K. Horan). Springer-Verlag, New York, pp. 100–116. [10] Lea, T., Vartdal, F., Davies, C. and Ugelstad, J. (1985) Magnetic monosized polymer particles for fast and specific fractionation of human mononuclear cells. Scand. J. Immunol. 22, 207–216. [11] Skjerve, E., Rørvik, L.M. and Olsvik, O. (1990) Detection of Listeria monocytogenes in foods by immunomagnetic separation. Appl. Environ. Microbiol. 56, 3478–3481. [12] Morgan, J.A., Winstanley, C., Pickup, R. and Saunders, J.R. (1991) Rapid immunocapture of Pseudomonas putida cells from lake water by using bacterial flagella. Appl. Environ. Microbiol. 57, 503–509. [13] Christensen, R., Torsvik, T. and Lien, T. (1992) Immunomagnetically captured thermophilic sulfate-reducing bacteria from North Sea oil field waters. Appl. Environ. Microbiol. 58, 1244–1248. [14] Bifulco, J.M. and Schaefer III, F.W. (1993) Antibody– magnetite method for selective concentration of Giardia lambdia cysts from water samples. Appl. Environ. Microbiol. 59, 772–776. [15] Furhman, J.A. and Azam, F. (1980) Bacterioplankton secondary production estimates for coastal waters of British Columbia, Australia and California. Appl. Environ. Microbiol. 39, 1085–1095. [16] Kirchman, D., K’nees, E. and Hodson R. (1985) Leucine incorporation and its potential as a measure of protein synthesis by bacteria in natural aquatic systems. Appl. Environ. Microbiol. 49, 599–607. [17] Chin-Leo, G. and Kirchman, D.L. (1988) Estimating bacterial production in marine waters from the simultaneous incorporation of thymidine and leucine. Appl. Environ. Microbiol. 54, 1934–1939. [18] Ward, B.B. and Cockcroft, A.R. (1993) Immunofluorescence detection of the denitrifying strain Pseudomonas stutzeri (ATCC 14 405) in seawater and intertidal sediments. Microb. Ecol. 25, 233–246. [19] Dohler, K., Huss, V.A.R. and Zumft, W.G. (1987) Transfer of Pseudomonas stutzeri (Zobell) Baumann, Bowditch, and Beaman 1983 to Pseudomonas stutzeri (Lehman and Neuman 1896) Sijderius 1946. Int. J. Syst. Bacteriol. 37, 1–3. [20] Carlucci, A.F. and Pramer. D. (1957) Factors influencing the plate method for determining abundance of bacteria in sea water. Proc. Soc. Exp. Biol. Med. 96, 392–394. [21] Porter, K.G. and Fieg, Y.S. (1980) The use of DAPI for identifying and counting aquatic microflora. Limnol. Oceanog. 25, 943–948. [22] Streifel, J. (1988) Microspheres and cell separation. In: Microspheres: medical and biological approaches (Eds. A. Rembaum and Z.A. Tokes). CRC Press, Boca Raton, FL, pp. 77–88. [23] Lund, A., Helleman, A.L. and Vartdal, F. (1988) Rapid isolation of K88 1 Escherichia coli by using immunomagnetic particles. J. Clin. Microbiol. 26, 2572–2575.
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219 [24] Zweifel, U.L. and Hagstrom, A. (1995) Total counts of marine bacteria include a large fraction of non-nucleoidcontaining bacteria (ghosts). Appl. Environ. Microbiol. 61, 2180–2185. [25] Choi, U.W., Sherr, E.B. and Sherr, B.R. (1996) Relation between presence–absence of a visible nucleoid and metabolic activity in bacterioplankton cells. Limnol. Oceanogr. 41, 1161–1168.
219
[26] Fliermans, C.B. and Schmidt, E.L. (1975) Autoradiography and immunofluorescence combined for autecological study of single cell activity with Nitrobacter as a model system. Appl. Environ. Microbiol. 30, 676–684. [27] Ward, B.B. (1984) Combined autoradiography and immunofluorescence for estimation of single cell activity by ammonium-oxidizing bacteria. Limnol. Oceanogr. 29, 402–410.
Journal of Microbiological Methods
A species-specific bacterial productivity method using immunomagnetic separation and radiotracer experiments D.G. Bard, B.B. Ward* Ocean Sciences Department, A316 EMS, University of California, Santa Cruz, Santa Cruz, CA 95064, USA Received 27 August 1996; revised 17 January 1997; accepted 23 January 1997
Abstract A species-specific bacterial productivity assay, combining radiotracer methods with immunomagnetic bead separation, was developed and tested in the marine environment. The capture method was optimized using cultures of the marine denitrifying strain, Pseudomonas stutzeri (ATCC 14 405). Immunocapture was optimal at a bead to target cell ratio of 10:1 using an indirect antibody technique, in which the target cell is first incubated with specific (primary) polyclonal antiserum and then with the secondary antibody-coated beads. Primary antibody concentration was less important than target cell concentration in determining the efficiency of target cell recovery. Reproducible recovery efficiencies of 75% could be obtained using cultures, but at natural seawater abundance levels, efficiency was much lower, around 20%. Estimates of total heterotrophic bacterial production and P. stutzeri production, based on radiotracer incorporation, were obtained for seawater samples from Monterey Bay, CA. To measure species-specific production, samples were incubated with radiotracers (methyl-[ 3 H]thymidine and [ 14 C]-leucine), fixed, and concentrated. After separating P. stutzeri cells from the bacterial assemblage using immunomagnetic separation, target cell fraction radioactivity was measured. P. stutzeri abundance, estimated by immunofluorescence, represented less than 0.1% of the total bacterial abundance, whereas radiotracer incorporation by the target fraction represented 1–3% of the total assemblage tracer incorporation. 1997 Elsevier Science Ireland Ltd.
1. Introduction Measurement of total bacterial secondary production using radioactively labeled organic substrates [1–4] has become the standard method for evaluating the role of bacteria in aquatic food webs. This approach can be used to estimate the productivity of the total heterotrophic bacterial assemblage in seawater, but it cannot provide information on the contribution of individual species to the overall rates.
*Corresponding author. Tel.: 11 408 4593171; fax: 11 408 4594882; [email protected]
Metabolic rates of bacterial species have been measured in pure cultures, but it is not realistic to extrapolate those results to natural systems when that same species is part of a larger bacterial assemblage. The simplest approach might be to enumerate individual strains present in the natural assemblage and assume that their contribution to production is proportional to their numbers. Even this approach is not usually possible with currently available methods. Thus, with minor exceptions, it has not been possible to assess the contribution of individual species to total bacterial abundances nor to bacterially mediated processes in the environment. Antibodies which are specific for individual strains or species offer a basis for species-specific identifica-
0167-7012 / 97 / $17.00 1997 Elsevier Science Ireland Ltd. All rights reserved PII S0167-7012( 97 )00987-1
208
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
tion, enumeration and, in conjunction with immunomagnetic beads, species-specific productivity estimates. Immunofluorescence has been used to enumerate several kinds of biogeochemically important bacteria in terrestrial, marine and freshwater systems [5–8]. The same antibodies can be used for bacterial cell separation, for example, using affinity columns [9] or antibody coated magnetic beads. Originally developed in the medical field for the fractionation of specific human blood cells [10], immunomagnetic separation (IMS) has been applied in microbiological research for the detection and enrichment of selected bacterial species [11–14]. Mammalian antibodies confer a high degree of specificity and sensitivity on this approach. Antibodies developed against a wholecell antigen (primary antibodies) added to a heterogeneous cell sample should bind selectively to target cells. Antibodies against the primary antibody (i.e., secondary antibodies), covalently bound to the outside of magnetic microspheres (known as immunomagnetic beads, IMB), can then bind the target cells to the beads. A magnet is then used to separate and isolate target cells from the sample. A general description of bacterial activity in the water column can be obtained from production estimates, based on methods such as [ 3 H]thymidine incorporation into DNA [15] and radiolabeled leucine incorporation into protein [16]. The present study employed the dual-label method [17], which utilizes the simultaneous incorporation of [ 3 H]thymidine and [ 14 C]leucine to obtain two different measures of production. Pseudomonas stutzeri (ATCC 14 405) was chosen as the target bacterium for methods development because its metabolism is not atypical of marine bacteria – it is a heterotrophic, facultative anaerobe – and because of the availability of a well characterized P. stutzeri-specific antiserum [18]. To estimate P. stutzeri production, the entire bacterial assemblage was exposed to radiolabeled substrates and then target cells were separated from the assemblage using IMS. This study describes the development and evaluation of a method to separate P. stutzeri cells from seawater, and the use of this technique in conjunction with a standard production assay to estimate P. stutzeri production in situ. Initial results from a field study of P. stutzeri relative abundance and relative production are also presented.
2. Materials and Methods
2.1. Bacterial strains Pseudomonas stutzeri (ATCC 14 405), formerly P. perfectomarina [19], was obtained from the ATCC. TBD-8b, an unidentified marine denitrifier, was originally isolated from intertidal sediment from Tomales Bay, California [18]. P. stutzeri and TBD8b were grown in a seawater medium containing 5 g carbon (as peptone) l 21 (CP medium; [20]). To produce radiolabeled cells for IMS optimization experiments (not for productivity measurements), cells were labeled with radioisotopes during exponential growth. P. stutzeri cultures were labeled by incubating cells in a solution of diluted CP medium (1:100) with either 10 nM methyl- 3 H-thymidine (specific activity, 107.8 Ci mmol 21 ; New England Nuclear, Dupont, Wilmington, DE, USA) for 2 h, or 100 nM 14 C(U)-leucine (specific activity, 321.6 mCi mmol 21 ; New England Nuclear) for 30 min. After incubation, cells were washed several times with phosphate buffered saline (PBS; per liter of distilled water: 8 g NaCl, 0.2 g KCl, 1.15 g Na 2 HPO 4 , and 0.2 g KH 2 PO 4 ), and then preserved with formalin (37% formaldehyde, buffered with saturated sodium borate, filtered prior to use, 2% v / v final concentration).
2.2. Immunomagnetic separation The primary antibody solution was polyclonal antiserum specific for P. stutzeri [18] which was absorbed with Escherichia coli (to remove nonspecific interactions) prior to use. IMB (Dynabeads M-280, 2.8 mm in diameter, 6–7310 8 beads ml 21 , with covalently linked sheep anti-rabbit IgG (secondary antibody); Dynal A / S, Oslo, Norway) were used to perform the immunomagnetic separation. In most experiments, samples were first incubated with the primary antibody solution (allowing the antiserum to bind to target cells), and then the beads were added, allowing the antibodies, with cells attached, to bind to the secondary antibody-coated beads (the indirect technique). A bacterial culture or a concentrated seawater sample was incubated with the primary antiserum (1:100 final dilution) in 1.0 ml of a 1:1 mixture of PBS–gelatin for 3 h. Gelatin (2%
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
in water, base hydrolyzed at pH 11 and 1208C for 15 min, neutralized) was used to reduce nonspecific binding. All incubations of cells with antiserum, and subsequent incubations with IMB, were performed in volumes of 1.0 ml or less in 1.5 ml microcentrifuge tubes at 158C with continuous, end-over-end rotation. After incubation with the primary antibody, excess antiserum was removed by washing twice with PBS and resuspending cells in 1.0 ml PBS–gel (1:1) and 20 ml of IMB. After an overnight incubation with the beads, the sample was ready for the magnetic separation procedure. In a few experiments, the primary antiserum was first bound to the secondary antibody-coated beads (the direct technique) and then the bead-antibody complex was allowed to bind to target cells. In this case, IMB were incubated in 1.0 ml of the P. stutzeri antiserum (1:100 final dilution) at 158C for 3 h. The antiserum-coated beads were then washed twice with PBS and resuspended in PBS–gel (1:1). Coated beads (20 ml) were then added to a sample and incubated overnight. The sample was then ready for the separation. Separation of P. stutzeri cells (i.e., obtaining a target fraction) was performed by placing the sample epitube in a magnetic particle concentrator (MPC-E: Dynal) for 5 min. Bead-cell complexes (BCC) were drawn against the wall of the epitube and the residual suspension was removed by aspiration. To reduce the loss of BCC by aspiration, ends of pipette tips were cut to enlarge the tip inner diameter and the magnet was used to recover the complexes from the aspirated fluid. The beads were washed twice with 1.0 ml PBS, resuspended in PBS, and filtered onto a 25 mm, 0.2 mm polycarbonate filter using a ten-place filtration manifold (Hoefer Scientific Instruments, San Francisco, CA). Filtered cells could be inspected by microscopy after staining with DAPI or a fluorescently labeled secondary antibody, or if the cells were radioisotopically labeled, the filters were radioassayed (see below).
2.3. Evaluation of immunomagnetic separation method The effect of pH on IMS was tested on pure cultures of P. stutzeri. Triplicate samples of [ 14 C]leucine labeled P. stutzeri cells (9310 6 cells
209
ml 21 ) were incubated in PBS–Gel under different pH conditions (pH 4.0, 7.0, and 10.0). Recovery efficiency was not determined in the pH experiment, but the effect of pH was assessed by simply comparing radioactivity (normalized to cell number) in the target fraction of each treatment. The effect of bead to cell ratio on P. stutzeri cell recovery in pure culture conditions was tested by incubating duplicate samples of [ 14 C]leucine labeled P. stutzeri cells (3–7310 6 cells) in 1.0 ml PBS–gel (1:1) with antiserum (1:100 final dilution) for 3 h. After incubation, excess antiserum was removed by centrifugation with two washes of PBS–gel. The antiserum-coated P. stutzeri cells were resuspended in 1.0 ml PBS–Gel (1:1). To one of the duplicate samples, the bead sample, the appropriate concentration of IMB was added to yield bead to cell ratios between 0.5 and 75. After an overnight incubation, a target fraction was obtained, using the magnetic particle concentrator as previously described, and radioassayed. The other of the duplicate samples, the pre-bead sample, was incubated without the beads. After incubation, the pre-bead sample was filtered and radioassayed. Target cell recovery was calculated by dividing the bead activity by the pre-bead activity. The effect of antiserum concentration was investigated by incubating triplicate samples of [ 14 C]leucine-labeled P. stutzeri cells (9310 6 –3.63 10 7 cells ml 21 ) with a range of primary antibody concentrations, at the optimum bead to cell ratio determined above (10). Efficiency of cell recovery was determined by comparing the radioactivity recovered in the bead fraction with that in an aliquot containing the same initial number of filtered cells (i.e., a pre-bead fraction). A sensitivity experiment was performed using different concentrations of [ 14 C]leucine- labeled cultured P. stutzeri in 900 ml of sterile filtered seawater. P. stutzeri concentrations ranged from 80– 1300 cells ml 21 . The samples were then concentrated by filtering the sample through 47 mm, 0.2 mm pore size polycarbonate filters (Poretics, Livermore, CA, USA) under low vacuum pressure. Filters were placed in 5 ml PBS and the cells resuspended by vortexing for 30 s. The resuspended cells were drawn off and the filters washed twice with 4 ml PBS and these washes added to the cells. The
210
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
resuspended cells were transferred to 2.0 ml epitubes and centrifuged in a microcentrifuge (Eppendorf 5415) at 10 0003g for 4 min. Centrifugation and resuspension continued (3–4 times) until the sample volume was reduced to 1.0 ml. The concentrated samples were then processed for cell separation and recovery using the IMS method. Target fractions were filtered and the filtrate was collected. The target and nontarget fractions and the filtrates were radioassayed as described above. To estimate cell recovery, the target fraction activity was divided by the sum of target, nontarget and filtrate activities. A target fraction purity experiment was performed using seawater with an intact natural seawater bacterial population (SWB). For each treatment, 4.5 l of natural seawater were concentrated using an Omega 300K tangential filtration device (Filtron Technology, MA, USA) and a Masterflex 60 600 peristaltic pump (Cole-Parmer, Chicago) to a final volume of approximately 400 ml. Triplicate subsamples (40 ml each) from the concentrated sample were used in each of three treatments and a single 40 ml subsample was used for a dead control. Treatments included: (A) the dead control (SWB preserved with formalin); (B and D) SWB with added P. stutzeri cells (approximately 10 7 cells); (C) SWB without added P. stutzeri cells. All treatments were incubated overnight with 10 nM (final concentration) [ 14 C]leucine, and then treatments B, C, and D were preserved with formalin. Treatments were then processed using IMS, as previously described, except that treatment B was incubated in a PBS–gel solution without antiserum. To compare the efficiency and specificity of the direct and indirect techniques of binding P. stutzeri to the beads, 200 ml of dual-labeled seawater were concentrated by centrifugation to reduce the final volume to 1.0 ml. Triplicate samples were tested with each technique using the indirect and direct procedures described above.
2.4. Sampling sites Water column samples were collected in Monterey Bay California during September 1993 and November 1994, and in the Southern California Bight during April 1994. Samples were collected using 5 l Niskin or Go-Flo bottles and sampling
locations are stated in the results of each experiment. Radiotracer incubations of the Monterey Bay 1993 and the Southern California Bight 1994 samples were performed immediately after collection at in situ temperatures using a flowing seawater incubator on deck. The Monterey Bay 1994 samples were obtained aboard the R.V. Point Lobos and 2 l water samples were kept cool and dark while transported to the laboratory. Incubations were performed at in situ temperatures within 12 h of sample collection.
2.5. Total bacterial production The dual-label method [17] was performed with slight modifications. Immediately before tracer additions, water samples were filtered through a 210 mm nylon screen mesh to remove larger particulates and zooplankton. For each depth, one or two 20 ml subsamples were incubated with 5 nM (final concentration) methyl-[ 3 H]-thymidine, and 10 nM (final concentration) [ 14 ]C(U)-leucine. TCA-insoluble material was extracted using a slight modification of the original method [17]. Samples were filtered through 25 mm, 0.2 mm poresize polycarbonate filters (Poretics) using the filtration manifold (Hoefer) at low pressure (#10 cm Hg pressure differential). Filtered samples from 1993 were counted with a Packard PRIAS model PL liquid scintillation counter. Filtered samples from 1994 were counted with a Beckman LS 6500 liquid scintillation spectrometer. Quenching was corrected with the external standard in the Compton edge shift mode (H[). In both cases, dual isotope counting modes and the use of quench curves maximized the resolution of the two isotopes.
2.6. Seawater concentration Seawater samples were concentrated in order to increase the concentration of target cells. For 1993 Monterey Bay and 1994 Southern California Bight samples, 1 l of seawater was concentrated to a final volume of approximately 100 ml using the Filtron casette (see above). Concentrated samples were incubated with radiotracers, as described above, preserved with formalin, and then kept cool in the dark during transit to the laboratory. In the lab, samples were stored in the dark at 48C and processed
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
211
within one month from time of collection. (While storage for this length of time probably resulted in loss of cells, it was unavoidable for this experiment. The least biased assumption is that cell loss affected all cell types equally, thus not biasing the resulting strain specific comparison.) Samples were centrifuged in 40 ml round-bottom centrifuge bottles at 10 0003g for 30 min and the pellet was resuspended in PBS. Samples were transferred to 2.0 ml epitubes and centrifuged at 10 0003g for 4 min. Centrifugation and resuspension continued (3–4 times) until the sample volume was reduced to 1.0 ml. For 1994 Monterey Bay samples, 1 l of seawater was incubated with radiotracers as described above and preserved with formalin after the incubation. Concentration was performed after incubation by filtering the sample through 47 mm, 0.2 mm pore size Poretics polycarbonate filters using low-vacuum pressure. Filters were placed in 5 ml PBS and the cells resuspended by vortexing and repeated PBS– gel washes. The resuspended cells were transferred to 2.0 ml epitubes and centrifuged to complete the concentration step as described above. The concentrated samples were used for measurement of total bacterial radiotracer incorporation or for IMS of P. stutzeri to estimate the contribution of P. stutzeri to bacterial radiotracer uptake using the protocols described above. A radiotracer experiment was conducted on cultured P. stutzeri cells to determine the magnitude of the correction for the difference between whole cells and cold TCA insoluble material. Fifteen replicate subsamples of exponentially growing P. stutzeri cells were incubated in 1:100 diluted CP medium containing 10 nM methyl-[ 3 H]-Thymidine. After 2 h, five replicates were filtered onto 0.2 mm pore-size Gelman Supore filters. Five replicates were preserved with 2% formalin for 30 min, then filtered, and five replicates were treated with cold TCA as in the natural seawater experiments. Filters were assayed by scintillation counting. In an identical duplicate experiment, the 15 replicate samples were incubated in the presence of 20 nM [ 14 ]C-Leucine.
DAPI staining method [21] using a Zeiss Axioskop microscope with 100 W Hg illumination, a UV G365 filter for excitation, and a LP 420 barrier filter (Zeiss filter set 487 902) at 1600 power. Sample volumes of 1 ml or less of the culture and 10 to 20 ml of field samples were diluted or filtered to a volume of about 1 ml and DAPI was added at 0.35 mg ml 21 . Abundances were calculated from single filter counts or duplicates were averaged for mean abundance. A total of over 200 cells were counted for each field sample. For most field samples, DAPI and immunofluorescence (see below) enumeration were done on the same preparation, by switching between optical filter sets. The indirect immunofluorescence (IIF) staining method [8] was used to enumerate P. stutzeri in field samples preserved with formalin, using the primary antibody for P. stutzeri [18] and fluorescein isothiocyanate-conjugated sheep anti-rabbit immunoglobulin G (IgG; Miles Laboratories, Elkhardt, IN, USA) as the secondary antibody. Poretics polycarbonate filters (25 mm, 0.2 mm pore size) were stained in Irgalan Black to reduce background fluorescence. Stained samples were viewed within one day using epifluorescence microscopy. Cell counts were made at 1600 power using a Zeiss Axioskop microscope (Carl Zeiss, Germany) with 100 W Hg illumination with a BP 450 / 490 filter for excitation and a LP 520 barrier filter. P. stutzeri cells were distinguished by a staining reaction of 13 or 14 (compared to the homologous reaction of 14) and the characteristic rod shape. A total of 350 fields were inspected for each count of natural samples. Calculations and simple statistical tests for optimization of methods and comparisons of experimental treatments were performed using the spreadsheet program Excel 5.0.
2.7. Enumeration of bacteria
Several parameters which might influence recovery of target cells were tested in order to optimize the IMS method. The effects of pH, antiserum concentration, and bead to cell ratio were tested on
Enumeration of pure cultures of P. stutzeri and the total bacterial assemblage was performed using the
3. Results
3.1. Optimization of immunomagnetic separation method
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
212
Table 1 Parameters optimized for immunomagnetic separation Cell recovery(% activity)a
[ 14 C]Leucine activity a (DPM)
[3H]Thymidine activity a (DPM)
ND ND ND
5868 6669 6466
NA NA NA
Antiserum concentration c 1:20 1:50 1:100 1:150 1:300 1:800 1:1000
75615 85613 746 3 746 2 806 3 686 5 706 6
NA NA NA NA NA NA NA
NA NA NA NA NA NA NA
Binding technique d Direct Indirect
ND ND
961 5463
1062 44618
Parameter varied pH b 4.0 7.0 10.0
a
Mean6(S.D.). [ 14 C]leucine labeled P. stutzeri cells separated using indirect technique. c 14 [ C]leucine labeled P. stutzeri cells separated using indirect technique. Antiserum concentration expressed as dilution of antiserum stock concentration. d Performed on natural seawater bacterial assemblage with dual-label radiotracer incubation. Sample was collected from SCB station 202 (338299N, 117845.59W, total depth554 m). NA, not applicable; ND, not determined. b
pure cultures of P. stutzeri. No significant differences were detected among the three pH levels tested (Table 1). Bead to cell ratio had a significant effect on recovery efficiency; cell recovery varied from a low of 25% at both extremes of bead to cell ratios, to a maximum of 55% around the bead to cell ratio of 10:1 (Fig. 1). At the optimum bead to cell ratio, cell
recoveries averaged 75% (6S.D. 5.8%) and did not vary as a function of antiserum dilutions ranging from 1:20 to 1:1000 (Table 1). Indirect and direct methods were compared using replicate subsamples of a concentrated dual-labeled natural bacterial population from seawater. Activities based on both radiotracers were much higher for the indirect technique than for the direct technique (Table 1), and no further experiments using the direct technique were performed.
3.2. Target fraction purity
Fig. 1. Cell recovery as a function of bead to cell ratio (note log scale). [ 14 C]leucine labeled P. stutzeri cells were recovered by indirect IMS.
Purity of the target cell fraction was assessed indirectly by comparing the target fraction activity of [ 14 C]leucine-labeled P. stutzeri cells with the target fraction activity of [ 14 C]leucine-labeled P. stutzeri cells plus an approximately ten-fold excess of nonlabeled TBD-8b cells. The target fraction activities did not differ (data not shown) suggesting that the TBD-8b cells did not interfere with target cells binding to the beads. Target fraction purity was also examined by
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
213
the range of concentrations tested were approximately 20% (data not shown). The target fraction activity was plotted against P. stutzeri concentration, and a best-fit line was calculated (Fig. 3). Using the regression function, and setting the lower limit of significant detectable activity as twice background activity ( y-intercept), the limit of detection was 324 cells ml 21 .
3.4. Total bacterial and target fraction radiotracer incorporation Fig. 2. Activity of target cell fractions from treatments using natural bacterial assemblage from seawater (SWB) and P. stutzeri cells. Triplicate samples for treatments B–D; error bars represent standard deviation: Treatment A, dead control: SWB, P. stutzeri cells (preserved with formalin) 118 (primary) antibody added during IMS; treatment B, live SWB1P. stutzeri cells, no primary antiserum added during IMS; treatment C, live SWB118 antibody added during IMS; treatment D, live SWB1P. stutzeri cells118 antibody added during IMS.
comparing the target fraction activity of four different treatments using a natural seawater sample and P. stutzeri cell additions (Fig. 2). The dead control (treatment A) showed no significant activity. Some activity was associated with the treatment B, even though no antiserum had been added, showing that minor nonspecific binding occurred. The activity in treatment C was not significantly higher than in treatment B (P.0.20). These results imply that the natural population of P. stutzeri was small and that non-specific antibody binding was negligible. The activity in treatment D was much higher than the other three treatments (P,0.10), reflecting the addition of cultured target cells to the resident population. The recovery efficiency of this experiment cannot be determined from radioactivity alone because the added P. stutzeri cells acquired their radioactivity as part of the experiment and could not be independently assayed.
Radiotracer incubations using concentrated and unconcentrated samples were compared to determine whether the concentration procedure influenced bacterial production estimates. Samples were concentrated using a tangential flow filtration device and incubated alongside unconcentrated samples using the same radiotracer procedures. In most cases, concentration resulted in rates which were disproportionately lower than predicted from the concentration factor (data not shown). This is consistent with the observed decrease in cell number associated with incubation of concentrated samples (see above). To avoid preconcentration artefacts, the natural seawater experiments in May 1994 were incubated with radiotracers and then concentrated. Radioisotope incorporation estimates from Monterey Bay in 1994 were obtained using a dual-label method. For total bacterial incorporation, radioactivity of the cold TCA-insoluble cell material was
3.3. Limit of detection The sensitivity of this method was investigated by performing the IMS procedure on different concentrations of [ 3 H]thymidine-labeled P. stutzeri cells in 1.0 l of filtered seawater. Cell recoveries across
Fig. 3. Limit of detection for IMS recovery of cultured P. stutzeri cells from natural seawater determined from y-intercept of linear regression ( y50.112x13.759, r 2 50.875).
214
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
measured. For target fraction incorporation, whole cell radioactivity was measured. TCA-insoluble radiotracer incorporation determined from experiments with cultured P. stutzeri was 46% for thymidine and 80% for leucine; target fraction incorporation rates were corrected for these factors. A second correction was made for estimated cell recovery efficiency of 20% at the typical P. stutzeri cell concentration encountered in natural samples (50– 100 cells ml 21 ). Corrected radiotracer incorporation rates for the target fraction and total rates were plotted against depth using a log scale to allow display of the large range of magnitudes (Fig. 4). Similar trends with depth were observed for both total and P. stutzeri rates. Highest incorporation occurred near the surface and the rates decreased with depth. Incorporation estimates for the target fraction at most depths were approximately two orders of magnitude less than incorporation by the total bacterial fraction. The contribution of P. stutzeri target cells to the total bacterial incorporation decreased with depth ranging from a minimum of 1.2% to a maximum of 3.3% (Table 2).
Table 2 Target fraction contribution to total bacterial production a Depth (m)
10 20 100 200
% Total tracer incorporation due to target fraction b Thymidine
Leucine
2.8 2.8 1.5 1.3
3.3 2.7 2.2 1.2
a
Monterey Bay in November 1994, central station, total depth5 260 m. b Mean % contribution (6S.D.)52.1460.82 for thymidine, 2.3660.89 for leucine.
3.5. Enumeration of bacteria Total bacteria and P. stutzeri in Monterey Bay were enumerated using DAPI and indirect immunofluorescence, respectively (Table 3). For every sample examined, P. stutzeri cells were present, usually at less than 110 cells ml 21 , and accounted for much less than 1% (usually less than 0.1%) of the total bacterial assemblage. With two exceptions, relative abundance of P. stutzeri was similar at all nine depths, even though P. stutzeri abundance in the water column decreased with increasing distance from shore and with increasing depth. Cell loss during radiotracer incubations was examined by enumerating total bacterial cells before and after incubation in samples of concentrated and unconcentrated seawater which were incubated alongside production samples (Table 4). There was little cell loss observed in unconcentrated samples, but a significant decrease in cell number occurred in concentrated samples.
4. Discussion
4.1. Evaluation of immunomagnetic separation Fig. 4. Radiotracer incorporation by target fraction and total bacterial assemblage. Samples were collected from the Monterey Bay water column at a central station (maximum depth at 260 m). Target fraction estimates (open symbols) and total bacterial incorporation (closed symbols) based on radiotracer uptake of thymidine (h, j) and leucine (x, ♦).
Initial characterization and optimization of the immunomagnetic separation method were performed using pure cultures of P. stutzeri. Bead and antiserum concentrations relative to target and non-target cell concentration were expected to be the most im-
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
215
Table 3 Relative abundance of P. stutzeri in Monterey Bay seawater samples a Total bacteria b (cells ml 21 )
P. stutzeri c (cells ml 21 )
Relative abundance (% total)
Station 1156: 36854.979 N, 121856.439 W (total depth526 m) 1 2.2310 5 10 1.0310 5 24.5 1.4310 5
110 60 55
0.05 0.06 0.04
Station 1116: 36843.519 N, 121856.289 W (total depth597 m) 5 1.4310 5 15 3.0310 5 95 3.0310 4
70 50 50
0.05 0.02 0.20
Station H3: 36846.259 N, 122800.929 W (total depth5911 m) 5 8.0310 4 150 2.0310 4 908 3.0310 4
30 40 20
0.04 0.20 0.07
Depth (m)
a
Samples collected in September, 1993. Enumeration by DAPI staining method [21]. c Enumeration by IIF staining method [8]. b
portant variables in determining the efficiency of the separation procedure. A wide range of antiserum concentrations worked equally well, with an average activity recovery of 75%. The antiserum dilution used in these experiments, 1:100, was that found previously to optimize strength of the immunoflourescence staining reaction with this serum, without causing non-specific staining of heterologous cells [18]. In a separate set of experiments designed to
optimize the bead to cell ratio, performed using an antiserum dilution of 1:100, the maximum activity recovery was only 55%. The difference between average recovery in the bead to cell ratio experiment and the antiserum concentration experiment is probably due to the introduction of an additional wash and recovery step in the latter. In the bead to cell ratio experiment, beads removed during the washes were not recovered by a secondary magnetic recovery step, which was incorporated into the protocol for
Table 4 Total bacterial counts before and after productivity incubation a Sample b
Depth (m)
Total cell count before (cells ml 21 )
Total cell count after (cells ml 21 )
% Cell loss
1 500 820
5.4310 5 6.9310 4 4.0310 4
4.5310 5 6.9310 4 5.3310 4
17 0 0
1 500 820
1.4310 6 4.1310 5 5.2310 5
2.4310 5 2.4310 5 2.0310 5
83 42 62
Unconcentrated
Concentrated c
a
Enumeration by DAPI staining method [21]. Southern California Bight, Station 205 (33818.79N, 118809.69W, total depth5905 m) April, 1994. c Concentration was performed using a tangential filtration device (Filtron). b
216
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
the antiserum concentration experiment and all subsequent experiments. The effect of absolute cell number or concentration of target cells was not addressed in these experiments, which were all performed with target cell concentrations in the range 10 6 –10 7 cells ml 21 . Subsequent experiments with P. stutzeri cells suspended in filtered or natural seawater at concentrations less than 1000 cells ml 21 (in the limit of detection experiments) yielded recoveries on the order of 20%, indicating that regardless of bead and antiserum concentration, recovery efficiency decreases with decreasing target cell concentration, due to low encounter frequencies. For a given target cell concentration, the separation method provides reproducible recovery, and a relationship between cell concentration and recovery efficiency can be used to normalize activity measurements. Because the antiserum had been previously characterized [18], those tests were not repeated here. However, the tests of target fraction purity implied that the specificity of the immunomagnetic separation was consistent with the specificity of the antiserum itself. When challenged with a mixture of cultured cells that contained only 10% target cells, the presence of nontarget cells did not reduce binding of target cells. This suggests that nonspecific binding is too low to interfere with the specific binding capacity of the beads. Similarly, when cultured target cells were resuspended in a natural seawater sample (at a final added target cell concentration of approximately 10 5 cells ml 21 ), the level of activity associated with nonspecific binding was low and easily distinguished from that associated with labeled target cells. The optimization experiments were designed to develop a reproducible protocol that could be applied to natural samples. These experiments illustrated that the principle of immunomagnetic separation can be applied to specific separation of bacterial cells and that under controlled conditions of antiserum, bead and cell concentration, a reproducible separation and concentration of target cells can be achieved. The major complication in applying this method to field samples is the low concentration of target cells. While there are few data available on the absolute abundance of individual strains of non-photosynthetic bacteria in seawater, previous estimates
from immunofluorescence enumeration of nitrifying and denitrifying bacteria indicate that individual serotypes are usually present at very low levels, usually less than 0.1% of the total cell abundance. At these levels, low recovery efficiencies would be expected, and even low levels of nonspecific binding could become relatively more important. The low abundance of P. stutzeri detected in the samples used in this study was consistent with previous enumeration in this environment [18] and was below the detection limit determined in Fig. 3. Such low abundances presented challenges to target fraction activity measurement; when samples were first tested using IMS, activity in the target fraction was too low to differentiate P. stutzeri production from nonspecific binding (see Fig. 2). Sample concentration was investigated as a means to increase the number of target cells in the sample. Concentration itself proved problematic. Concentration prior to radiotracer incubation has the advantage of requiring less radioisotope use and generating less isotope waste, but significant cell losses, accompanied by low productivity rates, occurred during incubation of concentrated samples (Table 4). This decrease in cell number probably was due to increased bacterial grazing by microzooplankton which survived in the concentrated sample. While this study was not intended to investigate concentration procedures exhaustively, we did note significant differences in recovery due to concentration method alone. Efficiency of concentration can be quantified, whichever procedure is used, so inefficiency at this step does not prevent quantification of species-specific activity. In the future, it may be more effective to use radiotracers with higher specific activities and thus avoid the need for sample concentration, but if concentration is required prior to IMS, it should be performed after the tracer incubation step. Low target cell abundance not only affects efficiency of target recovery, but it also affects target fraction purity. Streifel [22] reported that low target cell abundance reduced target fraction purity in IMS with eukaryotic cells. Lund et al. [23] reported similar findings using IMS of bacterial cells. The nonspecific binding observed in separations of natural samples in this study might have been caused by nontarget cells trapped amongst the beads. Although washing should reduce this nonspecific entrapment,
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
we observed bead agglutination even in the absence of bound target cells, which suggests that mechanical trapping of nontarget cells in the bead-cell complex may occur independently of the specificity of the antibody reaction. Target fraction contamination may have accounted for part of the activity differences observed between the direct and indirect immunocapture techniques. The increased activity associated with the indirect method could have been due to a greater number of target cells captured or it could have resulted from a relatively greater amount of nonspecific binding. In future applications of this method, target fraction purity and target cell recovery might be directly assessed by enumerating target and nontarget cells by immunofluorescence in all fractions and washes. This is not very practical, however, because the cell number remaining is below the limit of detection and statistically significant counts cannot be obtained. The opposite problem, clumping and three-dimensional overlapping of cells bound to beads, means that the number of cells recovered is also not easily determined. Radioisotopically labeled cells were used in these optimization experiments in order to increase the sensitivity and ease of assaying recovery and to circumvent the error and sensitivity problems associated with direct microscopic cell counts, but microscopy may also be useful.
4.2. Species-specific radiotracer incorporation rates In situ incorporation rates estimated for the total heterotrophic bacterial community in Monterey Bay are consistent with previously published rates (on the order of 1 mg C l 21 h 21 in surface waters [18]). The objective of the species specific productivity measurement using IMS was to determine what fraction of this total productivity could be attributed to individual strains present in the population. This comparison is made on the basis of radiotracer incorporation (dpm l 21 h 21 ) rather than on the basis of carbon production to avoid unnecessary assumptions about conversion factors. If all cells and all strains are equally active, then the proportion of activity due to any strain parallels its abundance. If some strains are more or less active than average,
217
then abundance and activity might be differentially distributed. In these results from Monterey Bay, the distribution of P. stutzeri tracer incorporation parallels that of the total bacterial incorporation. In the absence of selection favoring the growth of P. stutzeri (e.g., low oxygen conditions which induce denitrification), P. stutzeri functions as an aerobic heterotrophic generalist, so that its abundance and activity are simply subsets of the total heterotrophic bacterial population. The mean relative production attributed to P. stutzeri was approximately an order of magnitude larger than the mean relative abundance of P. stutzeri observed elsewhere in the bay. Because we did not quantify target purity or target cell recovery by immunofluorescence in the exact samples used for productivity measurements, spatial or temporal variation in P. stutzeri abundance may account for this difference. Alternatively, P. stutzeri may be relatively more active than some other members of the microbial community. Recent findings that not all cells are active [24,25] could be due to differences in activity among strains. It is also possible that mechanical agglutination of the beads may cause some trapping of non-target cells, such that not all of the production associated with the beads was due to P. stutzeri. More measurements of this kind are necessary to assess the significance of this species specific production estimate. Previous studies using immunological methods to investigate species-specific activity have reported variation among bacterial species within habitats. Fliermans and Schmidt [26] reported distinct differences in activity of two species of Nitrobacter in mixed culture and Ward [27] observed differences in relative activity of two species of ammonia-oxidizing bacteria in the Northeast Pacific Ocean off the Washington coast, both using immunofluorescence in combination with autoradiography. IMS could be an improvement over autoradiography because, although verification of target fraction purity and independent enumeration of target cells require microscopy, it is somewhat less tedious than counting grains on autoradiograms. IMS in combination with standard radioisotopic productivity methods should yield a direct measurement of production, rather than a relative activity index. Accurate measurements of species activity de-
218
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219
pends, in large part, on assessing the effectiveness of target cell separation from the overall bacterial assemblage. In the future, using radiotracers with higher specific activity for target cells with low abundance and performing direct counts of target species using immunofluorescence to assess IMS efficiency are likely to improve species-specific production estimates using this method. This method could be used to test the hypothesis that denitrifiers contribute relatively more to bacterial production in oxygen minimum zones, i.e., where its abundance and activity, should be favored.
Acknowledgments This research was funded by an ONR contract to BBW. General assistance from Mary Hogan and Margaret Geissler is gratefully acknowledged.
References [1] Joint, I.R. and Morris, R.M. (1982) The role of bacteria in the turnover of organic matter in the sea. Oceanogr. Mar. Biol. Annu. Rev. 20, 65–118. [2] Azam, F. and Furhman, J.A. (1984) Measurement of bacterioplankton growth in the sea and its regulation by environmental conditions. In: Heterotrophic Activity in the Sea (Eds. J.E. Hobbie and P.J.L. Williams). Plenum Press, New York, pp. 179–196. [3] Karl, D.M. and Winn, C.D. (1984) Adenine metabolism and nucleic acid synthesis: Applications to microbiological oceanography. In: Heterotrophic Activity in the Sea (Eds. J.E. Hobbie and P.J.L. Williams). Plenum Press, New York, pp. 196–216. [4] Ducklow, H.W. and Carlson, C.A. (1992) Oceanic Bacterial Production. Adv. Microb. Ecol. 12, 113–181. [5] Stanley, P.M., Gage, M.A. and Schmidt, E.L. (1979) Enumeration of specific populations by immunofluorescence. In: Native Aquatic Bacteria: Enumeration, Activity, and Ecology, ASTM STP 695. (Eds. J.W. Costerton and R.R. Colwell) American Society for Testing and Materials, pp. 46–55. [6] Josserand, A.G., Gay, G. and Faurie, G. (1981) Ecological study of two Nitrobacter serotypes coexisting in the same soil. Microbiol. Ecol. 7, 275–280. [7] Dahle A.B. and Laake M. (1982) Diversity dynamics of marine bacteria studied by immunofluorescent staining on membrane filters. Appl. Environ. Microbiol. 43, 169–176. [8] Ward, B.B. and Carlucci, A.F. (1985) Marine ammonia- and nitrite-oxidizing bacteria: Serological diversity determined by immunofluorescence in culture and in the environment. Appl. Environ. Microbiol. 50, 194–201.
[9] Heyman, U., Heyman, B. and Ward, B.B. (1988) Cell affinity chromatography for a marine nitrifying bacterium. In: Immunochemical Approaches to Estuarine, Coastal and Oceanographic Questions, (Eds. C.M. Yentsch, F.C. Mague and P.K. Horan). Springer-Verlag, New York, pp. 100–116. [10] Lea, T., Vartdal, F., Davies, C. and Ugelstad, J. (1985) Magnetic monosized polymer particles for fast and specific fractionation of human mononuclear cells. Scand. J. Immunol. 22, 207–216. [11] Skjerve, E., Rørvik, L.M. and Olsvik, O. (1990) Detection of Listeria monocytogenes in foods by immunomagnetic separation. Appl. Environ. Microbiol. 56, 3478–3481. [12] Morgan, J.A., Winstanley, C., Pickup, R. and Saunders, J.R. (1991) Rapid immunocapture of Pseudomonas putida cells from lake water by using bacterial flagella. Appl. Environ. Microbiol. 57, 503–509. [13] Christensen, R., Torsvik, T. and Lien, T. (1992) Immunomagnetically captured thermophilic sulfate-reducing bacteria from North Sea oil field waters. Appl. Environ. Microbiol. 58, 1244–1248. [14] Bifulco, J.M. and Schaefer III, F.W. (1993) Antibody– magnetite method for selective concentration of Giardia lambdia cysts from water samples. Appl. Environ. Microbiol. 59, 772–776. [15] Furhman, J.A. and Azam, F. (1980) Bacterioplankton secondary production estimates for coastal waters of British Columbia, Australia and California. Appl. Environ. Microbiol. 39, 1085–1095. [16] Kirchman, D., K’nees, E. and Hodson R. (1985) Leucine incorporation and its potential as a measure of protein synthesis by bacteria in natural aquatic systems. Appl. Environ. Microbiol. 49, 599–607. [17] Chin-Leo, G. and Kirchman, D.L. (1988) Estimating bacterial production in marine waters from the simultaneous incorporation of thymidine and leucine. Appl. Environ. Microbiol. 54, 1934–1939. [18] Ward, B.B. and Cockcroft, A.R. (1993) Immunofluorescence detection of the denitrifying strain Pseudomonas stutzeri (ATCC 14 405) in seawater and intertidal sediments. Microb. Ecol. 25, 233–246. [19] Dohler, K., Huss, V.A.R. and Zumft, W.G. (1987) Transfer of Pseudomonas stutzeri (Zobell) Baumann, Bowditch, and Beaman 1983 to Pseudomonas stutzeri (Lehman and Neuman 1896) Sijderius 1946. Int. J. Syst. Bacteriol. 37, 1–3. [20] Carlucci, A.F. and Pramer. D. (1957) Factors influencing the plate method for determining abundance of bacteria in sea water. Proc. Soc. Exp. Biol. Med. 96, 392–394. [21] Porter, K.G. and Fieg, Y.S. (1980) The use of DAPI for identifying and counting aquatic microflora. Limnol. Oceanog. 25, 943–948. [22] Streifel, J. (1988) Microspheres and cell separation. In: Microspheres: medical and biological approaches (Eds. A. Rembaum and Z.A. Tokes). CRC Press, Boca Raton, FL, pp. 77–88. [23] Lund, A., Helleman, A.L. and Vartdal, F. (1988) Rapid isolation of K88 1 Escherichia coli by using immunomagnetic particles. J. Clin. Microbiol. 26, 2572–2575.
D.G. Bard, B.B. Ward / Journal of Microbiological Methods 28 (1997) 207 – 219 [24] Zweifel, U.L. and Hagstrom, A. (1995) Total counts of marine bacteria include a large fraction of non-nucleoidcontaining bacteria (ghosts). Appl. Environ. Microbiol. 61, 2180–2185. [25] Choi, U.W., Sherr, E.B. and Sherr, B.R. (1996) Relation between presence–absence of a visible nucleoid and metabolic activity in bacterioplankton cells. Limnol. Oceanogr. 41, 1161–1168.
219
[26] Fliermans, C.B. and Schmidt, E.L. (1975) Autoradiography and immunofluorescence combined for autecological study of single cell activity with Nitrobacter as a model system. Appl. Environ. Microbiol. 30, 676–684. [27] Ward, B.B. (1984) Combined autoradiography and immunofluorescence for estimation of single cell activity by ammonium-oxidizing bacteria. Limnol. Oceanogr. 29, 402–410.