Experimental design
To enable simultaneous assessment of depth-resolved gross rates of light-driven sulfide consumption and O2 production, as well as the fate of freshly produced dissolved organic carbon (DOC), we sampled a cyanobacterial mat without the underlying sediment from the Frasassi sulfidic springs in September 2012 (Fig. S2). The mat was placed in a flow chamber that accommodated sufficient area for microsensor measurements and sub-sampling of the mat during defined conditions (Fig. S3) that are detailed in the following sections. The incubation started with exposure to darkness for 8 h. 13C-bicarbonate solution was added to the water column and to a spring water reservoir underneath the mat after ~5.5 h. During the following stepwise increase of light intensity (7, 19, 89, and 315 µmol photons m−2 s−1), net and gross rates of AP and OP were continuously monitored using microsensors in three replicate spots of the mat. Light intensity was only increased after a steady state had established for at least 30 min (determined from concentration depth profiles). Triplicate subsamples (1 cm2) of the mat were taken in regular intervals over the course of the experiment to (1) determine bulk rates of inorganic carbon assimilation, (2) identify the functional groups involved in this 13C assimilation based on fatty acids (FA), (3) follow the flow of assimilated carbon into the 13C-DOC pool, and (4) monitor changes in the active community based on 16S rRNA sequencing. To be able to differentiate between the effect of light intensity and photosynthetic O2 production, after exposure to 315 µmol photons m−2 s−1, DCMU (3-(3,4-dichlorophenyl)-1,1-dimethylurea; dissolved in ethanol), an inhibitor of OP [24], was added to the water column in the dark to a final concentration of ~10 µM. The mat was then again exposed to 315 µmol photons m−2 s−1 for 8 h. In a second incubation run with fresh mat material DCMU was added in the beginning, before addition of 13C-bicarbonate.
Sampling and setup
The cyanobacterial mat forms along the flow path of “Main Spring” that emerges from the Frasassi cave system (Fig. S2, 43°24′4″N, 12°57′56″E, [23]). The day before first mat sampling, water column samples for total sulfide determination were collected and conserved in 2% zinc acetate solution. Concentration was assessed on the same day according to Cline [25]. O2 concentration and pH were determined using microsensors (see below). Temperature at the mat surface was measured with a PT1000 mini-sensor (Umweltsensortechnik, Geschwenda, Germany). Spring water was collected from the outflow of main spring and transported to the laboratory facilities of the Osservatorio Geologico di Coldigioco (~45 min driving time) and immediately prepared for use in the flow chamber.
The flow chamber was a larger version of what is described in [26] (Fig. S3). Briefly, the upper part of the flow chamber was separated from a bottom chamber using fibrous web and GF/F filters. The bottom chamber was filled with HEPES-buffered (pH 7.2) spring water that was then purged with N2 using needles penetrating the rubber stoppers on the wall of the chamber. The upper flow chamber was connected with tubing via five inlets to a water pump in a thermostated 20 L recycle of freshly sampled N2-bubbled spring water.
The following day, a 30 × 40 cm piece of mat was carefully lifted off the sediment, transferred into a plastic container, and transported cooled and in the dark to the laboratory. A small subsection of the mat was flash-frozen for 16S rRNA analysis on site. Upon arrival in Coldigioco, the mat was immediately placed onto the GF/F filters in the flow chamber. Neutralized Na2S was slowly added to the 20 L recycle of the flow cell. After ~6 h of dark incubation, 12C- and 13C-sodium bicarbonate (13C-DIC final atom fraction of ≈6%) were injected into the bottom chamber and briefly stirred. Subsequently, 12C- and 13C-sodium bicarbonate (13C-DIC final atom fraction of ≈6%) was added to the recycle. To allow for homogeneous distribution of the label, the pumping speed was increased for 5–10 min. To minimize outgassing of H2S and exchange of 13CO2 with the atmosphere, the spring water in the 20 L recycle was covered with paraffin oil and the water column in the flow cell was covered with transparent plastic wrap. Small holes were kept in the wrap to allow microsensor measurements. Immediately after bicarbonate addition, the first mat and water column samples were taken. Homogenous illumination was achieved by using two large cold-white lamps (Envirolite), the distance of which to the mat was adjusted to change light conditions. Incident irradiance at the mat surface was determined using a cosine‐corrected quantum sensor connected to a LI‐250A light meter (both LI‐COR Biosciences GmbH, Germany).
Microsensors
O2, H2S, and pH microsensors with a tip diameter of 10, 20, and 50 µm, respectively, and response time of <2 s were constructed and calibrated as described previously [22, 27,28,29]. All sensors were mounted on a multi-sensor holder and the tips were separated by less than 1 cm. The motorized positioner for vertical microprofiling was mounted on a horizontal motorized positioner, which allowed automated and reproducible repositioning of the sensors in three replicate spots during the incubation. At each light condition H2S, pH, and O2 depth profiles were measured in the three spots. After correction for the measurement angle, depth resolution of profiling was ~450 µm in the water column and depths greater than 4 mm, and ~180 µm in the uppermost 4 mm of the mat. Total sulfide (Stot) concentration (∑[S2−, HS−, H2S]) was calculated from H2S concentration and pH. When steady state was reached at each light intensity, gross photosynthesis rates over depth in one of the replicate spots was measured using the previously described O2– and H2S-based light-dark shift methods for OP and AP, respectively [22]. Fluxes and local volumetric net rates of production/consumption were calculated from concentration depth profiles using Fick’s first and second law of diffusion, respectively, using diffusion coefficients corrected for temperature and salinity (1.35 × 10−5 cm2 s−1 for sulfide and 1.78 × 10−5 cm2 s−1 for O2).
To be able to compare rates of OP, AP, and chemosynthetic sulfide oxidation to rates derived from the SIP assays, we calculated potential C-fixation rates. For OP we multiplied the depth-integrated gross rates of O2 production with a factor 1 assuming the stoichiometry:
$${mathrm{H}}_{mathrm{2}}{mathrm{O}} + {mathrm{CO}}_2 to {mathrm{O}}_2 + {mathrm{CH}}_{mathrm{2}}{mathrm{O}}.$$
(1)
For AP we took a similar approach and multiplied the depth-integrated gross rates of light-dependent sulfide consumption by a factor 2 assuming sulfide oxidation to zero-valent sulfur according to:
$${mathrm{2H}}_{mathrm{2}}{mathrm{S}} + {mathrm{CO}}_2 to 2{mathrm{S}}^{mathrm{0}} + {mathrm{CH}}_{mathrm{2}}{mathrm{O}} + {mathrm{H}}_{mathrm{2}}{mathrm{O}},$$
(2)
as previously described in [13, 22]. The rate of predicted CO2 fixation by sulfur-oxidizing bacteria (SOB) was estimated based on the fluxes of sulfide and O2 into the zone of aerobic sulfide oxidation, and the previously determined energy conservation efficiency of 16.9% for autotrophic aerobic sulfide oxidation in Frasassi mats [22, 30]. The end member stoichiometries for predominant oxidation of sulfide to S0 and SO42− under the incubation conditions follow, respectively:
$${mathrm{H}}_{mathrm{2}}{mathrm{S}} + 0.4{mathrm{O}}_{mathrm{2}} + 0.1{mathrm{CO}}_2 to 1{mathrm{S}}^0 + 0.1{mathrm{CH}}_2{mathrm{O}} + 0.9{mathrm{H}}_{mathrm{2}}{mathrm{O}},$$
(3)
and
$${mathrm{H}}_{mathrm{2}}{mathrm{S}} + 1.5{mathrm{O}}_{mathrm{2}} + 0.5{mathrm{CO}}_2 + 0.5{mathrm{H}}_{mathrm{2}}{mathrm{O}} to 1{mathrm{SO}}_4^{2 – } + 0.5{mathrm{CH}}_2{mathrm{O}} + 2{mathrm{H}}^ +.$$
(4)
The stoichiometry was adjusted according to the concentrations of H2S and O2, and pH for each time point assuming a constant thermodynamic efficiency but variable products of sulfide oxidation (Table S1).
CO2 assimilation rates
To determine the 13C/12C of the dissolved inorganic carbon (DIC) pool, water column, and bottom chamber water samples were taken in regular intervals during the incubation and preserved by addition of HgCl2 and ZnCl2 in Exetainers (Labco, UK) without headspace. The 13C/12C ratio was determined by isotope-ratio-monitoring gas chromatography–mass spectrometry (GC-MS) (VG Optima; Micromass, Manchester, UK) [31].
The 13C/12C in the mat sampled during the incubation was determined using an automated elemental analyzer (FlashEA, 1112 series) coupled to a Delta Plus Advantage mass spectrometer (Finnigan DeltaplusXP, both from Thermo Scientific) after freeze-drying and decalcification with ortho-phosphoric acid. The leftovers of freeze-dried samples were pooled and used for FA-SIP and DOC extraction. Total CO2 fluxes were calculated as the rate of increase in the isotopic labeling of the mat, considering the average areal weight of mat and correcting for the labeling of the DIC pool.
13C-DOC
To estimate the 13C/12C of the DOC pool, the remaining freeze-dried and decalcified mat material was pooled for each time point and 1.5 mL of ultrapure water were added to each sample. The re-suspended mat material was vigorously shaken. After centrifugation, the supernatant was filtered through 0.45 µm PES syringe filters into 2 mL septum vials (Zinsser). To convert the DOC into CO2, we followed the approach of Menzel and Vaccaro [32, 33] by adding 30 mg potassium persulfate and 60 µL 3% ortho-phosphoric acid before autoclaving for 1.5 h. The 13C/12C ratio in the resultant CO2 pool in the headspace was determined with isotope-ratio-monitoring GC-MS (VG Optima; Micromass, Manchester, UK). 13C-Glucose was used as a standard to assess conversion efficiency. As the efficiency of conversion into CO2, however, likely varies amongst different compounds of the DOC pool, we did not aim to quantify DOC but only report the relative changes of the 13C/12C-DOC.
FA-SIP
The total lipids extracts (TLE) of freeze-dried mat samples were obtained following the procedure in [34], with modifications (see Supplementary material). Elemental sulfur was removed from the TLE using copper powder (Sigma-Aldrich), activated with 4 N HCl as explained in [35]. An aliquot of the TLE was saponified according to [36]. Prior to analysis, FAs were derivatized using boron trifluoride (BF3) in methanol (Merck), leading to FA methylesters.
FAs were identified by coupled GC-MS (Agilent 6890N GC with Agilent 5973N mass selective detector). Quantification was done by GC coupled to a flame ionization detector using squalene as injection standard. The carbon isotopic compositions were determined by GC-isotopic ratio-MS using a Thermo Scientific Trace GC Ultra coupled to a Thermo Scientific Delta V Plus IRMS. The carbon isotope ratios were expressed in the delta notation (δ13C), based on which the relative increase of label (Δδ13C) was calculated by subtracting the δ13C of each FA at the first time point [37].
FA were classified into cyanobacterial-FA, sulfur reducing bacterial (SRB)-FA and SOB-FA according to the literature. Cyanobacterial-FA included the even-numbered monounsaturated C16:1ω9 and C18:1ω9, and the polyunsaturated FA C16:2 and C18:2 [37,38,39,40,41]. SRB-FA included aiC15:0, iC15:0, 10Me-C16:0, aiC17:0, iC17:0, and iC17:1, as well as C15:1, C17:0, and C17:1 [42,43,44,45]. SOB-FA included the even-numbered monounsaturated C16:1ω7 and C18:1ω7 [46,47,48].
Relative 13C uptake rate contribution of FAs (FA-RUR) for each interval between time points were calculated as the ratio of the rate of increase in the isotopic labeling of each SRB-FA (δ13C) over time (t) and the total rate of increase of all SRB-FAs over time, as:
$$mathrm{FA -RUR} = frac{{frac{{Delta delta ^{13}mathrm{C}}}{{Delta t}}}}{{Sigma frac{{Delta delta ^{13}mathrm{C}}}{{Delta t}}}}.$$
(5)
We then clustered sub-groups of SRB-FAs according to the patterns of increase/decrease in rate (examples in Fig. S4a, b), with group 1 (SRB-FA1) comprising 10Me-C16:0 and iC17:1, group 2 (SRB-FA2) comprising aiC17:0, iC17:0, and C17:1 and group 3 (SRB-FA3) comprising aiC15:0, iC15:0, C15:1, and C17:0.
16S rRNA extraction and sequencing
The mat sample taken directly in main spring, and six of the subsamples taken during the flow chamber incubations were chosen for sequencing. RNA was extracted from these seven RNAlater stabilized subsamples with FAST RNA Pro Soil direct kit (MP Bio). Pyrosequencing libraries were constructed as described previously [49] with modifications (see Supplementary information). Emulsion PCR, emulsion braking and sequencing were performed applying the GS FLX Titanium chemistry following the supplier’s protocols (Roche).
16S rRNA analysis
The raw sequencing data sets were initially processed with the next-generation sequencing analysis pipeline of the SILVA project (available at www.arb-silva.de/ngs) [50] to obtain sequence and alignment quality-based filtering of the amplicons, aligned sequences, and a taxonomic classification (see Supplementary information). Based on quality filtering, a subset of cyanobacterial and deltaproteobacterial sequencing reads were selected for further oligotyping analysis using Oligotyping version 2.1 (available from https://github.com/merenlab/oligotyping). Oligotype representatives were then added to the SILVA RefNR 132 guide phylogenetic tree using the ARB-parsimony addition tool [51] (Table S2). Further processing of oligotype data was performed in R environment for statistical computing (https://www.R-project.org/), using package phyloseq (version: 1.19.1) [52] (see Supplementary information). Sequence data has been deposited in the European Nucleotide Archive (ENA) at EMBL-EBI under accession number PRJEB38493 [53] (see Supplementary information).
Source: Ecology - nature.com
