Use of the [¹⁴C]Leucine Incorporation Technique To Measure Bacterial Production in River Sediments and the Epiphyton

Abstract

Bacterial production is a key parameter for the understanding of carbon cycling in aquatic ecosystems, yet it remains difficult to measure in many aquatic habitats. We therefore tested the applicability of the [¹⁴C]leucine incorporation technique for the measurement of bulk bacterial production in various habitats of a lowland river ecosystem. To evaluate the method, we determined (i) extraction efficiencies of bacterial protein from the sediments, (ii) substrate saturation of leucine in sediments, the biofilms on aquatic plants (epiphyton), and the pelagic zone, (iii) bacterial activities at different leucine concentrations, (iv) specificity of leucine uptake by bacteria, and (v) the effect of the incubation technique (perfused-core incubation versus slurry incubation) on leucine incorporation into protein. Bacterial protein was best extracted from sediments and precipitated by hot trichloroacetic acid treatment following ultrasonication. For epiphyton, an alkaline-extraction procedure was most efficient. Leucine incorporation saturation occurred at 1 μM in epiphyton and 100 nM in the pelagic zone. Saturation curves in sediments were difficult to model but showed the first level of leucine saturation at 50 μM. Increased uptake at higher leucine concentrations could be partly attributed to eukaryotes. Addition of micromolar concentrations of leucine did not enhance bacterial electron transport activity or DNA replication activity. Similar rates of leucine incorporation into protein calculated for whole sediment cores were observed after slurry and perfused-core incubations, but the rates exhibited strong vertical gradients after the core incubation. We conclude that the leucine incorporation method can measure bacterial production in a wide range of aquatic habitats, including fluvial sediments, if substrate saturation and isotope dilution are determined.

Most organic carbon metabolism in running waters occurs on or in sediments (38, 42). Bacteria play a key role in organic carbon processing (10, 64) and influence many aspects of the chemistry and biology of river ecosystems (43). The quantification of bacterial production in sediments is therefore important for holistic studies of these ecosystems. However, methodological problems make it difficult to measure production rates of intact bacterial communities in many aquatic habitats.

Several methods have been suggested for the measurement of bacterial production in natural aquatic systems (e.g., see references 19, 20, 46, and 62). One of the most promising approaches consists in the measurement of leucine incorporation into bacterial protein (24, 50). This technique provides more-direct results for bacterial carbon production than the more widely used thymidine method (18, 22), because it measures an increase of a major biomass fraction. It is also one order of magnitude more sensitive, because over time bacterial cells incorporate about 10 times more leucine than thymidine (50). In addition, the method has potential to measure production of anaerobic bacteria (8, 32). Leucine incorporation into bacteria has been tested extensively in pelagic systems (e.g., see references 21, 24, 44, 50, and 60) and is now commonly used for the measurement of bacterial production in these environments (23). Recently, adaptations for other habitats (soil [4], epiphyton [54], leaf litter [51], and sediments [30, 58]) have been tested.

However, the leucine incorporation technique is still far from being a routine assay and relies on assumptions that have not been thoroughly tested in many aquatic habitats. In this study we examined several of these critical questions about the application of the [¹⁴C]leucine method in a variety of aquatic habitats. (i) Can free and incorporated leucine be completely recovered from sediments? (ii) Can leucine incorporation into protein be saturated, and can internal and external isotope dilution be excluded? (iii) Do high leucine concentrations stimulate bacterial activity? (iv) Is leucine taken up solely by bacteria, even if high concentrations are used? (v) Does the incubation technique—vial incubation versus perfused-core incubation—affect bacterial production estimates?

Our aim was to develop methodologies for routine measurements of bacterial production in aquatic habitats that have received little attention, most notably sediments. We also compared our bacterial production estimates with those determined in other riverine habitats, such as the water column (pelagic zone) and the biofilm on aquatic plants (epiphyton).

MATERIALS AND METHODS

General methods.

Sediment was sampled by using a sediment corer from shifting sands in the 6th order River Spree, approximately 40 km upstream of the city of Berlin in Germany. These sediments were characterized by low organic matter content (loss on ignition, <1% ash-free dry mass), a homogeneous particle-size distribution with a median particle size of 0.5 mm, high hydraulic conductivity (k = 0.001 to 0.004 m/s), and the presence of manganese and iron oxide coatings. Further information on the River Spree and the study site is found in references 26 and 52. The upper 4-cm-thick strata of five sediment cores were pooled, using fresh samples taken on the day of each experiment. Sediments were mixed gently with a spatula, and subsamples with wet weight of 0.5 g were weighed into sterile 10-ml centrifuge vials containing 4 ml (2 ml in the respiratory-activity experiment) of fresh, sterile-filtered (pore size, 0.2 μm) river water. The vials were stored at 4°C for up to 5 h prior to experiments.

l-[U-¹⁴C]leucine (Amersham Ltd.) at a specific activity of 11.6 GBq/mmol was diluted with cold l-leucine to achieve the specific activities given for the various experiments. In order to simulate in-stream conditions, samples were incubated at in situ temperatures with gentle shaking. The incubation was terminated by the addition of formaldehyde (final concentration, 3.2%).

If not stated otherwise, the following procedure was used for the further processing of the samples. The fixed samples were vortexed, sonicated (10 min, 60% power) in a sonication bath (Elma T 710 DH) and vortexed again. After this step, a subsample could be taken in order to determine bacterial abundance. To the remaining samples, including the sediments, trichloroacetic acid (TCA) was added to a final concentration of 5%. In order to dissolve the nonprotein fraction of the cells, the samples were then incubated at 95°C for 30 min. After cooling on ice, the remaining precipitate was filtered onto 0.2-μm-pore-size membrane filters (polycarbonate filters [Nuclepore] or RoTrac polyester filters [Oxyphen GmbH, Grosserkmannsdorf, Germany]). Filters were thoroughly rinsed with deionized water to eliminate unincorporated leucine. The amount of radioactivity in the filtrate was recorded. Filters were then put into 4-ml scintillation picovials, completely dissolved in 0.5 ml of solvent (Soluene; Packard), and mixed with 2.5 ml of scintillation fluid (Hionic Fluor; Packard). Radioactive decays were measured in a Canberra Packard 1900 scintillation counter. Controls were fixed with formaldehyde (final concentration, 3.2%) immediately at the start of the incubation and generally contributed less than 10% of the total leucine incorporation.

Additionally, an alkaline extraction procedure (14, 30) was tested. Briefly, protein was extracted from sediments at 25°C for 20 h (experiment one) and 40 h (experiment two and epiphyton experiments) with a solution of 0.3 M NaOH, 25 mM EDTA, and 0.1% sodium dodecyl sulfate (SDS). Protein was precipitated with TCA, washed with TCA, and redissolved in NaOH as described by Marxsen (30). An aliquot of 0.8-ml volume was mixed with 4.2 ml of scintillation cocktail (Ultima Gold; Packard) in a 6-ml scintillation minivial and measured as described above.

The results are given as rates of leucine incorporation into protein. Bacterial carbon production can be calculated from these values by using conversion factors given by Simon and Azam (50). Statistical analyses were performed using the software SPSS (Release 6.0; SPSS Inc.).

Substrate saturation.

A wide range of leucine concentrations was tested in three saturation experiments conducted under the following different incubation conditions. (i) For the experiment conducted in March 1997, leucine concentrations ranged from 0.05 to 200 μM, incubation temperature was 13°C, incubation time was 3 h, and the specific activity of leucine was 296 Bq/nmol of leucine. (ii) For the experiment conducted in May 1997, these conditions were 0.5 to 200 μM, 17°C, 1.5 h, and 148 Bq/nmol of leucine, respectively. For that conducted in February 1998, they were 3 to 300 μM, 5°C, 2 h, and 92.5 Bq/nmol of leucine, respectively. The resulting incorporation velocities were iteratively fitted to the hyperbola function of the Michaelis-Menten enzyme kinetics by using nonlinear regression (Origin 4.0; Microcal Software Inc.) (28, 44, 60). The plot was used for calculations of the theoretical maximum uptake velocity (V_max) and of the sum of the half saturation constant and the natural leucine concentration [(K_t + S_n)] as measures of substrate affinity.

Stimulation of bacterial metabolism by added leucine.

The electron transport system of metabolically active bacteria reduces the vital stain 5-cyano-2,3-ditolyl tetrazolium chloride (CTC), thus building up the red fluorescent compound CTC-formazan. This compound accumulates in active bacteria and can be microscopically viewed (45). In order to test whether added leucine has an effect on bacterial respiratory activity, sediments were incubated with 4 nM (final concentration) CTC and cold leucine at final concentrations of 0, 2, 50, and 200 μM. After 2 h of incubation at 17°C, samples were fixed and processed within 2 days. Bacteria were counterstained with DAPI (4′,6-diamidino-2-phenylindole) and enumerated immediately after filtration onto black polycarbonate filters (pore size, 0.2 μm) (Nuclepore), using a Nikon FXA photomicroscope and the filter sets EX 330-380, DM 400, and BA 400 for cells stained with DAPI and EX 420-490, DM 510, and BA 580 for CTC-stained bacteria (16, 41).

Bacterial DNA replication activity was tested by adding [³H]thymidine (final concentration, 100 nM; specific activity, 925 kBq/nmol; incubation temperature, 17°C; incubation time, 30 min) to samples containing cold leucine at concentrations of 0, 2, 50, and 200 μM. We extracted labeled DNA as described by Moran and Hodson (36) for measuring bacterial production on plant detritus, with the following modifications: fixed samples were cooled on ice immediately, and ice-cold TCA was added to a final concentration of 5%. Samples were then sonicated for 10 min in a cooled sonication bath and vortexed, and half of the volume of each sample was filtered onto polycarbonate filters (pore size, 0.2 μm) (Nuclepore). Filters were rinsed four times with ice cold TCA (concentration, 5%) to remove unincorporated label. The filtration apparatus was kept cool during this procedure. In order to hydrolyze DNA, filters were then incubated for 30 min at 100°C in capped vials containing 2 ml of TCA (concentration, 5%). The components that were insoluble in hot TCA (mainly proteins) were then filtered onto a second polycarbonate filter. An 0.8-ml subsample of the filtrate containing the dissolved DNA was mixed with 4 ml of scintillation cocktail (Ultima Gold; Packard) in a 6-ml scintillation minivial for counting.

Eukaryotic organism leucine uptake.

The proportion of eukaryotic organism leucine uptake was assessed at leucine concentrations of 2, 50, and 200 μM. A mixture of colchicine and cycloheximide (final concentrations, 0.01 and 0.02%, respectively) was added to five vials for each treatment. In combination, these antibiotics effectively inhibit eukaryotic organism reproduction and feeding and have no direct effect on bacterial growth (49). After 1 h, leucine was added (specific activity, 148 Bq/nmol of leucine; temperature, 15°C; n, 5). Short incubation times (50 min) were applied to reduce indirect effects of dead eukaryotic cells on bacterial growth. A second experiment was conducted in winter by using leucine at only the 200-μM concentration (specific activity, 92.5 Bq/nmol of leucine; temperature, 5°C; n, 7).

Effects of incubation technique.

We compared bacterial protein production measured in sediment slurries in incubation vials with results obtained with a perfused-sediment-core technique (30, 31). Sediment cores of 7.6-cm length and 2-cm inner diameter were taken from the river bed and perfused with prefiltered river water in a once-through mode at a rate of 18.2 ml h⁻¹ (residence time, 30 min) for 24 h at 20°C. We obtained this rate as a rough estimate of in situ interstitial flow by model calculations (53) and by dye experiments conducted in laboratory flumes. Four cores were perfused with the flow directed from the top sediment layer toward the deeper sediment layer (“top-down perfusion”), and four additional cores were perfused in the opposite direction (“bottom-up perfusion”). Subsequently, leucine was added to the stock of perfusion water to a final concentration of 50 μM and a specific activity of 7 Bq/nmol of leucine. A ninth core was sterilized with 5% formaldehyde and served as a control. Leucine perfusion lasted for another 12 h and was terminated by perfusing a solution of 5% formaldehyde for 4 h. Sediment cores were then cut into four sections corresponding to depths of 0 to 1.9, 1.9 to 3.8, 3.8 to 5.7, and 5.8 to 7.6 cm. Protein was extracted from 0.5-cm³ aliquots of the sediments, and its activity was measured as described above. The sediment cores for the slurry incubation were cut in the same way before incubation, and aliquots of 0.5 cm³ from each depth were incubated in vials and processed as described above.

Measurements of production by bacteria in epiphyton and the pelagic zone.

We tested two protein extraction methods for epiphyton and assessed the substrate saturation concentration of leucine in this habitat. In July 1997, we randomly cut leaves located at 10 to 40 cm below the water surface of one of the most common species (Sagittaria sagittifolia L., Alismataceae Vent.) and dissected discs of 1 cm² with a corkborer. Five leaf discs each were then pooled and allotted to 20-ml scintillation vials containing 5 ml of prefiltered river water. Before we determined leucine saturation for epiphytic bacteria, the two protein extraction methods were tested (leucine concentration, 1,500 nM; specific activity, 2,960 Bq/nmol; incubation time, 90 min; incubation temperature, 21°C). The incubation was terminated by adding formaldehyde to final concentration of 3.2%, and leaf discs were then sonicated at 100% power for 10 min. By the alkaline-extraction method, protein was extracted at 25°C for 40 h in a solution of 0.3 M NaOH, 25 mM EDTA, and 0.1% SDS. Protein was precipitated with TCA, washed three times, and redissolved in NaOH as described by Marxsen (30) for sediment samples. A 0.8-ml aliquot of sample and 4.2 ml of scintillation cocktail (Ultima Gold; Packard) were mixed in a scintillation minivial and measured. For acidic extraction, TCA was added to the samples after sonication at final concentration of 5%. Samples were then incubated for 30 min at 95°C. Subsequently, aliquots were filtered and washed as described above for the sediment samples.

The saturation experiment was conducted by using the alkaline-extraction method and leucine concentrations of 30, 120, 400, 1,000, 2,000, and 4,000 nM. Incubation time was 90 min, at 21°C and at a specific activity of 2,960 Bq/nmol of leucine. For pelagic-zone samples, 0.2 ml of leucine was added to 4.8 ml of river water, resulting in final leucine concentrations of 10, 20, 40, 80, 140, and 200 nM. Incubation time was 50 min, temperature was 17°C, and specific activity was 11,562 Bq/nmol of leucine. Incubation was terminated by adding formaldehyde to final concentration of 3.2%. Samples were processed as described above for the sediment samples, but with the sonication step excluded.

RESULTS

Extraction of protein from sediment samples.

TCA extraction in combination with sonication and filtration of the precipitate maximized extraction of bacterial protein from sediments. In addition, polycarbonate filters (0.2-μm pore size (Nuclepore) produced significantly lower control values than the more-commonly used cellulose nitrate filters over a wide range of leucine concentrations (paired t-test, P = 0.03, n = 8). However, we found that the stability of the polycarbonate filters was not sufficient in all cases. Damage occurred during filtration to approximately every tenth filter. RoTrac polyester filters (Oxyphen GmbH) needed longer solubilization times but produced both low leucine adsorption and high resistance to TCA.

A total of 102.6% ± 3.6% (mean ± standard deviation) (n = 6) of the added label was recovered after 2 h of incubation, subsequent sonication, and TCA extraction. On the filter remained 2.73% ± 0.43% of the recovered label, of which 93.0% was incorporated into bacterial protein and 7.0% was abiotically adsorbed to sediment particles and to the filter itself. Of the total added label 97.3% ± 4.5% was recovered in the filtrate. We were unable to routinely recover bacterial protein from sediments by the alkaline-extraction method. In the first experiment (20 h of alkaline extraction), the radioactive label recovered in protein amounted to only 1.5% of the protein determined with the TCA extraction method. In the second experiment (40 h of alkaline extraction), 10.8% of the protein determined with the TCA extraction method was found.

Substrate saturation.

None of our substrate saturation experiments with sediments revealed typical Michaelis-Menten kinetics. In general, there seemed to be a depression or a plateau at the leucine concentration of 50 μM, followed by increased leucine uptake at higher concentrations (Fig. 1a, b, and c). In order to account for a possible multiphasic leucine uptake, we performed fitting for data obtained at leucine concentrations up to 50 μM as well as for the complete data set (concentrations up to 200 or 300 μM). The former was much closer to the measured data, thus exhibiting a significantly lower chi-square value. We therefore calculated the parametric values of (K_t + S_n) and V_max separately for concentration ranges of up to 50 μM and up to 200 (or 300) μM. At up to 50 μM (K_t + S_n) for leucine ranged from 6.0 to 12.6 μM, with V_max of 652 pmol of leucine cm⁻³ h⁻¹ in the February experiment and 1,105 and 1,525 pmol of leucine cm⁻³ h⁻¹ in the May and March experiments, respectively. At up to 200 μM V_max values were 2,088 and 2,261 pmol of leucine cm⁻³ h⁻¹ for the March and May experiments, respectively (Table 1).

FIG. 1.

Kinetics of uptake of leucine for sediment bacteria with fitting by using the Michaelis-Menten equation. Fittings are shown for leucine concentrations up to 50 μM (solid lines), up to 200 or 300 μM (dashed lines), and 50 to 200 μM or 100 to 300 μM (dotted lines). Experiments were conducted in March 1997, at incubation temperature of 13°C (a), in May 1997, at incubation temperature of 17°C (b), and in February 1998, at incubation temperature of 5°C (c).

TABLE 1.

Parametric values for leucine incorporation into protein by bacteria in sediments^a

Date	Leucine concn (μM)	Vmeasb (pmol cm−3 h−1) ± SD	Kt + Sn (μM) ± SE	Vmax (pmol cm−3 h−1) ± SE	Vmeas/Vmax ratio	Chi-squares, r2, P (n)c
March 1997	Up to 50	1,387 ± 74	6.0 ± 0.7	1,525 ± 50	1.10	4,797, 0.99,*** (18)
	Up to 200		14.1 ± 2.3	2,088 ± 88	1.51	26,730, 0.98,*** (24)
	50–200		19.2 ± 4.7	2,313 ± 124	1.67	21,031, 0.96,*** (9)
May 1997	Up to 50	850 ± 156	6.7 ± 2.8	1,105 ± 130	1.30	27,062, 0.91,*** (15)
	Up to 200		40.6 ± 13.0	2,261 ± 253	2.66	70,496, 0.86,*** (21)
	50–200		37.8 ± 13.5	2,460 ± 281	2.89	57,533, 0.70,** (9)
February 1998	Up to 50	507 ± 154	12.6 ± 5.8	652 ± 104	1.29	7,738, 0.78,*** (12)
	Up to 300		776 ± 697	8,444 ± 5834	16.7	92,618, 0.40,** (21)
	100–300		78.2 ± 44.3	3,155 ± 633	6.22	129,389, 0.14 N.S. (9)

^aCalculations were performed by non-linear regression of saturation curves using the Michaelis-Menten equation. Results are given for all data of each experiment (up to 200 and 300 μM), for data at leucine concentration up to 50 μM, and for data at leucine concentration exceeding 25 or 50 μM. 

^bRate of incorporation of leucine into protein by bacteria in sediments at added leucine concentration of 50 μM. 

^cChi-square values are given for nonlinear regression, and r² and P values are for modified Lineweaver-Burk plots (65). ***, P ≤ 0.0001; **, P ≤ 0.01; N.S., P > 0.05. 

Isotope dilution.

The ratio of V_max to the incorporation rates measured by using a specific leucine concentration represents the isotope dilution at that leucine concentration (44, 60). Isotope dilution for the leucine concentration of 50 μM varied from 1.1 to 1.3 for saturation curves calculated for leucine concentrations up to the 50 μM. However, isotope dilution would be higher if all available data (acquired at concentrations up to 200 or 300 μM) were incorporated into the calculation of V_max (Table 1).

Stimulation of bacterial metabolism by added leucine.

The sediments were colonized by 1.9 × 10⁹ ± 0.19 × 10⁹ bacteria cm⁻³, with a mean proportion of respiratorily active cells of 26.8% ± 2.9%. The mean rate of [³H]thymidine incorporation into DNA was 3.65 ± 0.91 pmol cm⁻³ h⁻¹. The CTC and thymidine experiments revealed no significant effects of added leucine on bacterial electron transport activity (i.e., respiratory activity) (analysis of variance [ANOVA], P = 0.42, n = 20) and DNA synthesis (ANOVA, P = 0.89, n = 20). Using the data of the experiment, the minimum difference in the percentage of active bacteria calculated with 90% confidence of detection (see equation 10.36 in reference 66) was 3.5%. The minimum detectable difference in the rate of thymidine incorporation was 0.98 pmol cm⁻³ h⁻¹.

Eukaryotic organism leucine uptake.

We found clear effects of leucine concentration on leucine incorporation rate (two-way ANOVA, F = 225, P < 0.001, n = 30). However, we also detected significant effects of the antibiotics on leucine incorporation (F = 5.02, P = 0.035) and two-way interactions between the use of antibiotics and leucine concentration (F = 4.12, P = 0.029). Differences between treatments with and without antibiotics were significant only at the leucine concentration of 200 μM (t-test, P = 0.05, n = 5). At that leucine concentration, leucine incorporation in samples with antibiotics was 20% lower than that in samples without antibiotics, with a 95% confidence interval of 0 to 40%. A second experiment supported these results; samples containing antibiotics had 35% lower leucine incorporation than those without antibiotics (t-test, P = 0.013, n = 7), with a 95% confidence interval of 9 to 61%.

Effects of incubation technique.

Incorporation of leucine into protein measured in sediment cores significantly differed from that measured in incubation vials. In vials, incorporation rates were equal at all sediment depths (ANOVA, F = 0.53, P = 0.68, n = 16). In contrast, leucine incorporation in perfused sediment cores significantly declined from the inflow layer to the outflow layers (ANOVA, F = 101, P < 0.0001, n = 32). Leucine incorporation was highest in the 0-to-1.9-cm-deep layer of the top-down-perfused cores and in the 5.7-to-7.6-cm-deep layer of the bottom-up-perfused cores. Incorporation rates measured in cores clearly exceeded those measured in vials in the inflow layers, and incorporation rates were lower in the outflow layers (Fig. 2). Leucine incorporation calculated per whole sediment core did not differ significantly among the methods: it amounted to 3.23 ± 0.32 nmol cm⁻³ h⁻¹ for the vial incubation, 3.19 ± 0.10 nmol cm⁻³ h⁻¹ for the top-down-perfused core incubation, and 3.36 ± 0.16 nmol cm⁻³ h⁻¹ for the bottom-up-perfused core incubation. Thus, we found no significant effect of the applied method alone on leucine incorporation (two-way ANOVA, F = 0.74, P = 0.49, n = 48), but we found a highly significant effect of the sediment depth (F = 14.5, P < 0.001, n = 48). This effect was strongly modified by the method, as two-way interactions between the sediment depth and method were strong (F = 40, P < 0.001, n = 48).

FIG. 2.

Leucine incorporation of sediment bacteria incubated in vials (■), top-down-perfused cores (▾), and bottom-up-perfused cores (▴). Experiments were performed in August 1998, at incubation temperature of 20°C. Data are expressed as mean ± 1 standard deviation (n = 4).

Bacterial production in epiphyton and the pelagic zone.

The alkaline- and TCA extraction methods recovered similar amounts of leucine incorporated into protein from epiphytic biofilms (t-test, P = 0.8, n = 10). Although Levene’s test showed no significant differences between the variances for the treatments (P = 0.18), we noticed that the samples that were treated with the TCA extraction method contained large undisrupted leaf particles which made subsampling more difficult and probably increased the variability between measurements. Apparently, the label was not completely extracted from the leaves, so that subsamples containing larger leaf particles exhibited higher activities. We therefore used the alkaline-extraction method for further investigations. However, adding a homogenization step to the TCA extraction procedure should probably make both methods equivalent.

Leucine incorporation into bacterial protein of the epiphytic biofilm was close to saturation at concentrations above 1,000 nM (Fig. 3). We calculated an isotope dilution of 1.3 for a leucine concentration of 2,000 nM. (K_t + S_n) for leucine was 621 nM in the epiphyton, with a V_max of 70 pmol cm⁻² h⁻¹ (Table 2). In the pelagic zone, leucine incorporation was close to saturation at concentrations above 80 nM (Fig. 4). The calculated isotope dilution was 1.2 at that concentration (it was 1.1 at 140 nM). (K_t + S_n) of pelagic bacteria was 20 nM and thus much lower than those for bacteria in sediments and epiphyton. The V_max was 1,718 pmol liter⁻¹ h⁻¹ (Table 2).

FIG. 3.

Kinetics of uptake of leucine for epiphytic bacteria with fitting by using the Michaelis-Menten equation. The experiment was performed in July 1997, at incubation temperature of 21°C.

TABLE 2.

Parametric values for leucine incorporation into protein by bacteria in epiphyton and in the pelagic zone^a

Habitat	Vmeasbc ± SD	(Kt + Sn) (nM) ± SE	Vmaxc ± SE	Vmax/Vmeas ratio	Chi-square, r2, P (n)d
Epiphyton	69.8 ± 11.4	621 ± 155	90.1 ± 6.8	1.29	65.7, 0.97,*** (18)
Pelagic zone	1,417 ± 91	20.4 ± 2.8	1,718 ± 61	1.21	11,236, 0.99,*** (18)

^aCalculations were performed by nonlinear regression of the saturation curves by using the Michaelis-Menten equation. 

^bRate of incorporation of leucine into protein by bacteria in epiphyton at an added leucine concentration of 2,000 nM and by bacteria in the pelagic zone at 80 nM. 

^cThe values are reported as picomoles of leucine cm⁻² h⁻¹ for bacteria in epiphyton and picomoles of leucine liter⁻¹ h⁻¹ for bacteria in the pelagic zone. 

^dSee Table 1, footnote c, for an explanation of the statistical values. 

FIG. 4.

Kinetics of uptake of leucine for pelagic-zone bacteria with fitting by using the Michaelis-Menten equation. Experiments were performed in May 1997, at incubation temperature of 17°C.

DISCUSSION

Extraction methods.

The alkaline-extraction method, which was successfully applied by Bååth (4) and by Marxsen (30), was not suitable for protein extraction from River Spree sediments. Possibly, the iron and manganese coating on the sediment particles interfered with the complex-forming EDTA used in this technique. The combustion method of Tuominen (58) measures the total leucine uptake into bacteria. However, the intracellular pool of leucine not incorporated into proteins may be high (50), and bacterial production thus might be overestimated with the combustion method. In our study, hot TCA extraction combined with filtration of the precipitate yielded good results, was less time-consuming than the alkaline extraction, and does not require an oxidizer like the combustion method does. However, it is more expensive in terms of filters and solubilizer needed.

Substrate saturation of leucine.

Saturation of nutrient uptake rates occurs if enzyme-dependent or transporter-dependent steps limit nutrient influx into the cells (7). In pure cultures, leucine is taken up via active transport systems that produce a biphasic saturation curve (1). However, it seems obvious that kinetic diversity of a natural bacterial population will not yield simple one-enzyme–one-substrate saturation curves of the Michaelis-Menten type. Kinetic variability can be caused by several factors, e.g., diffusion gradients of enzymes and substrates in the biofilm matrix (29, 40), the physiological heterogeneity of natural bacterial populations (27, 63), and the influence of additional limiting factors (1, 7). A bi- or multiphasic mode of substrate uptake has therefore been postulated for marine pelagic environments (3, 59, 63).

Leucine incorporation of bacterial communities in sediments of the River Spree also seems to be at least biphasic with a close fit to the Michaelis-Menten saturation curve for leucine concentrations up to 50 μM. We therefore intend to conduct saturation experiments prior to measuring bacterial production in sediments, with 50 μM as a leucine concentration for guidance. This concentration has been shown to saturate leucine incorporation in streambed sediments (30), and incorporation was close to saturation in lake sediments (58) and in the hyporheic zone of a prealpine stream (6). Values for bacterial production calculated from uptake of leucine at 50 μM concentration therefore represent conservative estimates of bacterial activity. For epiphytic bacteria on leaves of S. sagittifolia from the River Spree, leucine incorporation was saturated at leucine concentrations above 1,000 nM. This is slightly higher than the saturation concentration range (400 to 800 nM) found for epiphytic bacteria present on detritus of Juncus effusus from the Talladega Wetland (54) and on eelgrass (Zostera marina) leaves (56). The saturation leucine concentration of 100 nM we found for pelagic-zone bacterial communities of the River Spree is within the range of those found for eutrophic lakes (21, 44).

Why are substrate affinities so different for pelagic, epiphytic, and sediment bacteria? Differences in the assimilation of leucine by bacterial communities among these habitats can be explained by specific nutritional conditions in their environments: as leucine uptake is enhanced by the presence of glucose (1), bacteria living in biofilms that effectively trap dissolved organic carbon (12, 17, 33) might be physiologically adapted to these conditions. However, Bright and Fletcher (5) and van Loosdrecht et al. (61) argued that the observed variability of K_t values is an indirect effect due to the differences in the environments of the cells, unaccompanied by change of the bacterial assimilation behavior. Mass transfer by diffusion through the biofilm should then be the limiting step, and the observed (K_t + S_n) and V_max values of biofilm bacteria should not reflect physiological differences among the bacterial communities.

Isotope dilution.

The calculation of V_max yields a potential maximum leucine incorporation rate at which the value for external isotope dilution is one (no isotope dilution) and internal isotope dilution is minimized due to feedback inhibition of de novo synthesis of leucine (25, 44). Considering that V_max is reached only at an infinite substrate concentration, the values calculated for isotope dilution at the leucine concentration of 50 μM are low (range, 1.1 to 1.3) and imply external leucine concentrations of 5 to 15 μM.

These calculated external leucine concentrations do not always accord well with the calculated data for (K_t + S_n). For the fitting at 50 μM, data seem to be realistic: for the March experiment, with a theoretical isotope dilution of 1.1 and (K_t + S_n) value of 6.0 μM (Table 1), the natural substrate (leucine) concentration, S_n, should be 5 μM, whereas K_t should be approximately 1 μM. For the May and February experiments, with a theoretical isotope dilution of 1.3, K_t should approach negative values if S_n is calculated via isotope dilution. However, values for (K_t + S_n) as well as for V_max show larger variations in these experiments, and a K_t of approximately 1 μM would well fit into the range of the standard errors of these data. In contrast, the theoretical isotope dilutions for the complete data set as well as for the data acquired in the 50 to 200 μM range are in a realistic range only for the March experiment and far too high for the May and February experiments. For the latter, other processes, which cannot be described by the Michaelis-Menten equation, seem to have been important.

The calculated isotope dilution can also be verified by comparing it to the amino acid content of the interstitial water: dissolved amino acids are immobilized rapidly in sediment biofilms by a combination of microbial and abiotic mechanisms (11, 12). The existence of a large pool of dissolved amino acids within the sediments is therefore improbable. In various interstitial waters, median concentrations ranging from 5 to 100 μM were found (55); this range is of the same order of magnitude as the natural leucine concentrations that can be calculated with our isotope dilution data (5 to 15 μM for concentrations up to 50 μM, and 25 and 75 μM for those up to 200 or 300 μM). These dissolved amino acids are most often combined rather than free (55), but after they are hydrolyzed with extracellular enzymes, their uptake might compete with that of dissolved free amino acids (47) like the added [¹⁴C]leucine. The concentration of dissolved free amino acids in aquatic environments is generally in the lower nanomolar range (55) and should therefore not contribute significantly to isotope dilution.

However, on a small (microliter) scale, microzones containing much higher ambient leucine concentrations are likely to occur in the sediments. Pelagic algae and other organic matter might be trapped here and release exudates in close proximity to the bacteria. Therefore, in these “hot spots” of heterotrophic metabolism, local isotope dilution might be higher. Multiphasic uptake kinetics with K_t values in the higher micromolar range and high V_max values might here be of evolutionary use for sediment bacteria.

Changes in bacterial activity.

In marine, pelagic environments, the addition of nanomolar concentrations of leucine repressed biosynthesis of leucine and did not alter the rate of protein synthesis (24, 25). In the eutrophic environment of the River Spree even micromolar additions of leucine did not significantly alter the activity of sediment bacteria. However, this might not be true for oligotrophic systems. Leucine might here be respired by bacteria to a greater extent, and additionally protein turnover might occur (25), leading to an overestimation of bacterial production.

Eukaryotic organism leucine uptake.

A variety of eukaryotic organisms are capable of osmotrophic uptake of organic compounds. Whereas heterotrophic protozoa can only feed osmotrophically in highly organically enriched environments (48), heterotrophy is well known for a multitude of marine and freshwater algal species (reviewed in reference 57). Facultative heterotrophy can be seen as a selective advantage of benthic algae in rivers, which are often transported into deeper, dark sediment layers (personal observation). Compared to bacteria, however, protozoa and most benthic algae probably are poor competitors for dissolved organic carbon, due to their lower surface-to-volume ratio and lower substrate affinity. This concept is in accordance with our finding of substantial leucine uptake by eukaryotes (20 to 35%) only at concentrations of 200 μM.

Effects of incubation technique.

Sampling and laboratory incubation of sediments from natural habitats always imply a disturbance of the in situ gradients of reduced carbon and electron acceptors. This procedure might therefore enhance bacterial activity (13, 35, 39), although in other studies (9, 30, 37) little effect of sediment disruption on bacterial production has been revealed. Disturbance effects can be minimized by incubating the sediments as whole cores that are perfused with river water or groundwater at natural rates (30, 31). In our study, differences of bacterial activity due to sediment disturbance are unlikely, because in the River Spree these sands are prone to steady movement caused by the current and thus form a very dynamic habitat. We attribute the higher bacterial production in the surface layer of the sediment core, and the sharp decrease with depth, to an enhanced oxygen and nutrient supply of the uppermost layer caused by the steady water flow. The flow in perfused sediment cores reduces diffusion limitation and therefore enhances leucine incorporation.

Loss of the added leucine by bacterial incorporation in upper sediment layers and abiotic adsorption is unlikely to account for the lower activity in deeper layers of the perfused cores, because 80.6% ± 1.3% (n = 8) of the added tracer was recovered at the outflow and a gradual rather than a sudden decrease should have occurred in consecutive layers if it were due to adsorption. Rather than leucine, oxygen is probably the key factor responsible for the observed pattern: estimated bacterial production was as high as 8.7 ± 0.9 μg of C cm⁻³ h⁻¹ in the upper layer (to 1.9 cm depth) of the perfused cores. If a bacterial growth efficiency of 30% (34) is assumed, bacteria in this layer should consume 3.9 mg of O₂ within a 12-h incubation time. By perfusion of oxygen-saturated water at a rate of 18.2 ml h⁻¹, 2 mg of oxygen was supplied to the core during the 12-h incubation. This means that at approximately 1-cm sediment depth oxygen depletion occurs and anaerobic metabolism should prevail. The same calculations can be performed by using data from the bottom-up-perfused cores, where leucine incorporation rates exhibited a striking symmetry with those for the top-down-perfused cores (Fig. 2). Oxygen should be consumed in the 7.6-to-5.7-cm-deep layer of the bottom-up-perfused cores, and anaerobic metabolism should prevail in the upper sediment layers. Bacterial abundance as well as organic matter was roughly equally distributed within the upper 10 cm of the shifting Spree River sands (15). Therefore, oxygen, and to a lesser extent leucine and nutrients, from the flowing water limit bacterial protein production in deeper sediment layers.

By maintaining a steady water flow, the perfused-core method reflects the natural conditions occurring in a river bed, where water frequently infiltrates and convectively flows within the sediments (53). If experiments aim at the estimation of bacterial production in the natural environment, the dissolved organic carbon and oxygen contents of the water that is used for perfusion should approach natural conditions. The perfused-core method makes possible long incubation times (30), which allow for the use of a low-specific-activity tracer, resulting in extremely low control values. However, as the perfused water does not circulate, a greater volume of radioactive waste (with low specific activity) is produced than by vial incubation methods. For a study of the basic principles of bacterial production measurements, the incubation in vials seems appropriate, because comparable nutrient and oxygen conditions can here be maintained in parallel.

Conclusions.

Bacterial production can be measured by the leucine incorporation method in a wide range of aquatic habitats, including fluvial sediments. In the sediments, substrate saturation was achieved only at high leucine concentrations, which have rarely been applied in previous studies. Therefore, bacterial production measurements can be interpreted correctly only if substrate saturation experiments involving large concentration ranges are performed concomitantly and if isotope dilution is estimated. The perfused-core incubation technique simulates in situ conditions more closely than vial incubations, as vertical concentration gradients, which influence bacterial metabolism substantially, develop.