Study site and experimental design—The experiment was carried out in Gossenko¨ llesee (GKS), a small alpine lake (area: 0.017 km2) situated at 2,417 m above sea level in the Austrian Alps (47u139N, 11u019E). GKS is a dimictic and holomictic lake covered by ice for ,7–8 months per year. The catchment area (0.3 km2) is composed of crystalline bedrock and covered with a poor soil layer (ca. 10% of its area) and sparse patches of alpine rankers. In the first week of August 2004 (i.e., 3 weeks after ice out), 50 liters of water were collected at 2-m depth and stored in acidcleaned carboys. The water was gently filtered first through a glass-fiber filter (AP 40, Millipore, 142-mm diameter) precombusted at 450uC for 4 h and subsequently twice through a 0.22-mm polycarbonate GSWP membrane (Millipore, 142-mm diameter). Filtered water was distributed among three sets of three replicates of 5-liter glass bottles (Schott). All glassware was soaked in 1 mol L21 hydrochloric acid (HCl) and subsequently rinsed several times with Milli-Q water and filtered lake water. The first set of bottles served as control of the experiments and did not receive any DOM addition. The second set of bottles received 300 mL of a soil extract to enrich the lake water with allochthonous DOM (hereafter, SOIL treatment). The soil extract was obtained according to Kablitz et al. (2003) using surface soil (upper 3–4 cm) collected from the catchment area of GKS. Finally, the third set of bottles received 200 mL of a filtrate from an algal lysate to enrich the autochthonous fraction of DOM (hereafter ALGAL treatment). The lysate was obtained from a culture of the planktonic green algae Chlorella minutissima grown in Woods Hole medium (at 17uC with 8:16 light:dark) until the early stationary phase was reached. The culture was then treated with a tip sonicator (2-mm diameter) for 15 min at 30 W to disrupt the cells and then filtrated onto a 0.22-mm polycarbonate membrane to eliminate algal rests and bacteria. The concentration of nutrients and DOC were both measured in the control and in the SOIL and ALGAL treatments. The DOC enrichment factor in the SOIL and ALGAL treatments was four-fold as compared to the natural lake concentration.
The lake bacterial assemblage was inoculated in all treatments following a 1:10 dilution obtained by filtering lake water collected from 2-m depth through a 0.8-mm polycarbonate membrane (ATTP, Millipore) to exclude bacterivores (microscopic examination at the end of the experiment revealed the absence of grazers). Bottles were incubated at 1-m depth under in situ temperature and light conditions for 6 days. Subsamples for bacterial production and abundance, DOC analysis, and absorbance measurements were collected at day 0, day 2, and thereafter every day. The first sample (i.e., day 0) was removed 2 h after the treatments were settled to allow bacteria to adapt to the experimental conditions. Additional subsamples were removed to perform micro-autoradiography combined with fluorescent in situ hybridization and signal amplification by catalyzed reporter deposition (CARD-FISH). Those samples were incubated with [3H]-L-leucine as described in the following section. At every sampling date, a water sample from GKS was collected at 2-m depth and the same analyses and measurements were done.
Incubation for micro-autoradiography—Subsamples from every treatment were incubated with [3H]-L-leucine (Amersham, specific activity 5 2,331 GBq mmol21; 20 nmol L21 final concentration) at in situ temperature for 6 h. The optimum incubation time was previously determined by monitoring disintegrations per minute (DPM) increase over time during a time series experiment. Control samples were killed with formaldehyde (2% final concentration) 20 min before adding the substrate and were incubated in parallel with the samples. Incubation ended by adding formaldehyde at a final concentration of 2%. Samples were fixed overnight at 4uC and filtered on the next day through 0.22- mm polycarbonate white filters that were subsequently rinsed twice with 5 mL of particle-free Milli-Q water. Afterward, filters were stored frozen (220uC) until further processing.
Bacterial production—Heterotrophic bacterial production was estimated from rates of protein synthesis with [14C]-L-leucine (Amersham, specific activity 5 11.3 GBq mmol21) (Simon and Azam 1989). Duplicate samples and one formaldehyde-killed control were incubated with 20 nmol L21 (final concentration) of [14C]-L-leucine. Samples (15–30 mL) were incubated at in situ temperature in the dark for 1 h. Incubations were terminated by adding formaldehyde at 2% final concentration. Subsequently, the samples were filtered through 0.22-mm Millipore GSWP filters and rinsed twice with 5 mL 5% trichloroacetic acid for 5 min. Filters were dissolved in 1-mL ethyl acetate (Riedel de Haen), and after 10 min, 6 mL of scintillation cocktail (Ready-safe, Beckman Coulter) were added, and the radioactivity was assessed after 15 h. The radioactivity of the filter was converted into bacterial carbon production using the formula given in Simon and Azam (1989).
Bacterial abundance—Bacterial numbers were assessed by flow cytometry. Subsamples of 450 mL were stained by adding 25 mL of a 50 mmol L21 SYTO 13 solution (Molecular Probes). Fluorescent microspheres (1-mm TransFluoSpheres 488/560, Molecular Probes) were added to a final concentration of 4.7 3 105 mL21 as a counting and internal fluorescence standard. The absolute concentration of the stock solution of the microspheres was assessed by flow cytometry combined with gravimetric volume measurement. Counts were made with a MoFlo (DakoCytomation) equipped with a water-cooled argonion laser tuned at 488 nm (200 mW). Bacteria were detected by their signatures in a plot of orthogonal side scatter versus green fluorescence.
Dissolved organic carbon—Subsamples for DOC analysis were filtered immediately after sampling through a precombusted (4 h at 450uC) GF/F filter (Whatman) placed on a stainless steel syringe holder. Filters were rinsed first with Milli-Q water and then with the sample. The filtrate was collected in precombusted glass bottles (Schott), acidified with HCl to pH 2, and stored in the dark at 4uC until further analysis within 48 h. DOC was measured by hightemperature catalytic oxidation with a Shimadzu TOC analyzer Model 5000. The instrument was equipped with a Shimadzu platinized-quartz catalyst for high sensitivity analysis. Three to five injections were analyzed for every sample and blanks (Milli-Q water).
DOM absorption—Samples for DOM absorbance measurements were filtered as described for DOC analysis. Upon arrival at the laboratory, samples were scanned in a spectrophotometer (double-beam Hitachi U-2000) from 250 nm to 750 nm using a 10-cm quartz cuvette (SUPRASIL I). The measurements were referenced against Milli-Q water. Apparent absorption coefficients at specific wavelengths (absl) were calculated as absl 5 (Dl 3 ln 10)/L, where Dl is the absorbance at the wavelength considered, and L is the path length (m) of the cuvette. True absorption coefficients (al) were corrected for the effect of scattering by colloids using a long reference wavelength (740 nm) applying the formula: al 5 absl 2 (abs740 3 740/l). The ratio a250:a365 was used to provide information about the relative size of DOM molecules (Strome and Miller 1978).
Hybridization and tyramide signal amplification— CARD-FISH was carried out on filter sections embedded in low gelling point agarose (0.2%). Sections were permeabilized with lysozyme and achromopeptidase according to Sekar et al. (2003). Hybridization with 59- horseradish peroxidase (HPR)-labeled oligonucleotide probes, was carried out according to the protocol of Pernthaler et al. (2002) in 0.5-mL reaction vials filled up with 300 mL of hybridization buffer and the 59-HPRlabeled probe at a final concentration of 0.5 ng mL21. Five different group–specific oligonucleotide probes (Thermo- Hybrid, Germany) were targeted to the domain Bacteria (EUB338), to ß-Proteobacteria (BET42a) and its subgroup R-BT (R-BT065), to Cytophaga-like bacteria (CF319a), and to the class Actinobacteria (HGC69a). The proportion of formamide in the hybridization buffer was always 55% except for probe HGC69a (35%). Cells were hybridized at 35uC for at least 2 h and up to 4 h. After hybridization, the filter sections were washed at 37uC for 15 min in a prewarmed washing buffer containing 5 mmol L21 ethylenediaminetetraacetic acid (pH 8), 20 mmol L21 Tris-(Hydroxymethyl)-aminomethane hydrochloride (pH 7.5), 0.01% sodium dodecyl sulfate, and the appropriate amount of sodium chloride. The sections were then transferred to phosphate-buffered saline amended with Triton X-100 (PBS-T) for 10 min at room temperature.
Amplification was carried out in the dark at 37uC for at least 15 min and up to 30 min in a reaction vial containing amplification buffer and tyramide-Alexa 488 in a 1:100 ratio. Thereafter, filters were washed for 10 min in PBS-T in the dark at room temperature, then rinsed with Milli-Q water, and finally dipped in 96% ethanol followed by air drying. Filters were stored frozen until they were processed for micro-autoradiography.
Micro-autoradiography—The microautoradiographic procedure we used is described in detail elsewhere (Teira et al. 2004). Briefly, hybridized filter sections were transferred onto slides coated with a molten Kodak NTB-2 photographic emulsion. Subsequently, slides were placed on a cold plate for a few minutes until the emulsion hardened. Slides exposure was carried out at 4uC for 24 h in light-tight boxes containing a drying agent. Optimum exposure time was determined empirically in a preliminary experiment. Development and fixation of the slides were done according to the specifications of the manufacturer. After fixation, the slides were rinsed in Milli-Q water for at least 2 min, then the filters were peeled off, and finally the cells were stained with an anti-fading solution containing 49,6-diamidino-2-phenylindole (DAPI) (final concentration of 1 mg mL21).
Microscopy—The slides were examined using a Zeiss Axioplan microscope equipped with a 100-W Hg lamp and Zeiss filter sets number 1 for DAPI and number 9 for Alexa 488. Silver grains around bacterial cells were observed using the transmission mode of the instrument. In the control samples, ,0.6% of the cells were associated with three or more silver grains. Cells were counted in at least 20 randomly selected microscopic fields and for every field four different counts were recorded: (1) DAPI-positive cells, (2) probe-specific positive cells, (3) DAPI + autoradiography positive cells, and (4) probe specific + autoradiography positive cells. At least 350 DAPI-stained cells were counted per sample.
Statistical analysis—A two-way analysis of variance (ANOVA) was used to compare changes in the proportions of the bacterial groups detected by the respective oligonucleotide probes (expressed as a percentage of DAPI stained cells) at different times in the control and in the ALGAL and SOIL treatments. The pertinent post-hoc comparisons were made by the Holm-Sidak method with an overall significance level of 0.05.