Sulfur bacteria promote dissolution of authigenic carbonates at marine methane seeps

Del Mar methane seep carbonate community analyses

The carbonate rock sample used for microbial community analyses was collected from Del Mar East Seep, Dive SO 177, (32°54.25456764 N, 117°46.9408327 W) at a depth of 1032.06 m. The site is in the northern portion of the San Diego Trough, about 50 km west of San Diego, California. Visible features of the seep include carbonate boulders and pavements colonized by orange and white bacterial mats, possible subsurface methane hydrate (large pits and craters), clam beds, and curtains of methane bubbles [20]. For a more in-depth discussion of the Del Mar Methane Seep, please see Grupe et al., (2015). A small chunk of the rock was sealed in a Mylar bag and shipped on dry ice to the University of Minnesota (Twin Cities). Upon arrival, the carbonate sample was temporarily stored at −80 °C until sampling for DNA extractions. All tools used for sampling biomass were autoclave-sterilized prior to use, and all work was performed in a LabConco A2 biological safety cabinet. Sterile aluminum foil was placed over ice packs to provide a cold and clean working environment. Three tubes were designated for the top of the rock, and three tubes were designated for the bottom of the rock. Biomass was scraped from the respective positions and placed in tubes. Top sample tubes had a biomass weight between 76.5–97.6 mg, and bottom sample tubes had a biomass weight between 30.7–59.3 mg. DNA was extracted from the seep samples using the ZymoBIOMICS DNA miniprep kit (Zymo, Irvine, CA). Nuclease-free water from the kit was processed alongside the samples as a negative control for iTag sequence analyses. The amplification of DNA and the generation iTag libraries of the V4 hypervariable region was performed by the University of Minnesota Genomics Center as previously described [21]. The samples and negative control were sequenced on ¼ of a lane of MiSeq for paired end 2×300 bp reads. Primers and adapters were removed with Cutadapt v. 2.10 [22]. The paired end reads were processed and assembled using DADA2 v1.16.0 [23]. The maximum expected error rate was 2 and reads detected as phiX were removed prior to error detection, merging of pairs and chimera detection. Taxonomic assignment was performed using the Silva database v. 138 [24]. Bioinformatic and statistical analyses was performed using tools in PhyloSeq 1.32.0 [25]. R package Decontam 1.8.0 [26] identified contaminating amplicon sequence variants (ASVs). Thus, ASVs of the genera Escherichia/Shigella, Haemophilus, Streptococcus, and the clade Chloroflexi S085 were bioinformatically removed from analyses. ASVs with statistically different abundances between the top and bottom surfaces were detected with DESeq2 v. 1.28.0.

Scanning electron microscopy of seep carbonates

The carbonate rock sample used for SEM imaging was collected from the Lasuen Knoll Seep (Dive SO 170, 33°23.57489996 N, 118°0.39814252 W) at a depth of 279.705 meters. A small chunk of carbonate rock with filtered bottom water and 25% vol/vol glutaraldehyde was stored at 4 °C, shipped on ice to the University of Minnesota (Twin Cities), and then stored at 4 °C. Several small pieces of the carbonate rock were broken off and placed in an 8-well plate. The carbonate pieces in the 8-well plate were rinsed, and subjected to an ethanol dehydration series as follows: 50% for 2 h, 70% over night, 80% for 15 min, 95% (x2) for 15 min, and 100% (x2) for 15 min. Carbonate samples were then subjected to critical point drying on a Tousimis Model 780 A Critical-Point-Dryer following standard procedures, followed by sputter coating 1–2 nm Iridium in a Leica ACE600 Sputter Coater. Lastly, carbonate samples were visualized at 1.0 kV on a Hitachi SU8230 Field Emission Gun Scanning Electron Microscope.

Continuous flow bioreactor experiments

To investigate the potential for sulfur-oxidizing bacteria to dissolve carbonate minerals, we used flow-through biofilm reactors (herein referred to as bioreactors: CDC; CBR 90–3 CDC Biofilm Reactor). Bioreactors contained eight polypropylene coupon holder rods that each accommodate three 12.7 mm diameter coupons. The lid and coupon holder rods were mounted in a 1 L glass vessel with side-arm discharge port at ~400 mL. A liquid medium (described below and Supplementary Table 1) was circulated through the bioreactor, while mixing was generated by a magnetic stir bar at 80 RPM. Sampling of the coupons was conducted aseptically by removing individual coupon holders and harvesting the coupons, while replacing the removed coupon holder with a sterile rubber plug. Experiments were 21 days in duration. On the final day of bioreactor experiments, one coupon was transferred to a treatment imaging flow cell (herein referred to as flow cell: BioSurface Technologies; Model FC 310 Treatment Imaging Flow Cell) in order to visualize and measure the in vivo pH of the biofilm. The flow cell is an autoclavable polycarbonate plastic cell with a recessed inner circle for installing coupons with biofilms grown in CDC Biofilm Reactors, and is equipped with barbed influent and effluent connectors, and topped with a coverslip for in vivo imaging.

The bioreactor medium consisted of two solutions, salt solution 1 & 2, and six additional components: vitamin solution, trace element solution, yeast extract, sodium bicarbonate, sodium thiosulfate, and sodium metasilicate. One 10 L carboy was prepared by autoclaving 5 L of salt solution 1 and one 10 L carboy was prepared by autoclaving 4.1 L (with an additional 839 mL ddH₂O) of salt solution 2 at 121 °C for 80 min. The salt solutions were then aseptically pumped together, followed by the addition of sterile stocks of 10 mL 1000X vitamin solution, 10 mL 1000X trace element solution, 10 mL (1000 g/L) yeast extract, 11 mL 1 M NaHCO₃, 10 mL 1 M Na₂S₂O₃, and 20 mL 100 mM NaSiO₃ through a 0.2 μM filter.

The final medium had a pH of ~7.85 and contained approximately 1.1 mM HCO₃^–, 1 mM S₂O₃^2-, and 9.14 mM Ca²⁺, with a saturation state with respect to aragonite (Ω_aragonite) of ~0.5, similar to that measured near the sediment/water interface in certain seep environments where carbonate dissolution was observed to occur [7, 27]. The saturation state of aragonite was calculated using CO2SYS_v2.1 [28]. Ingredients for both salt solutions and the additional components added to the medium are presented in Supplementary Table 1. Measured saturation indices throughout bioreactor experiments are presented in Supplementary Table 2.

Mineral preparation for bioreactor experiments

Bulk mineral specimens of aragonite were obtained from D.J. Minerals (Butte, Montana). X-ray powder diffraction (XRD) of the aragonite sample, and peak matching indicates the mineral is nearly 100% aragonite. Samples used for the bioreactor were cored with a diamond coring bit on a drill press, cut to ~1–3 mm thickness and smoothed with a diamond saw. Coupons were then scored with a diamond tipped scribe pen to assign a reference code. After scoring, coupons were placed in a muffle furnace (NEY 2-525 Series II) at 500 °C for 4 h to remove residual organic matter, followed by weighing to 0.01 mg (CAHN 29 Automatic Electrobalance) and stored in a desiccator until being mounted in coupon holder rods.

Assembling the bioreactor

Weighed coupons were mounted in coupon holder rods, and fastened in place with plastic screws to avoid mineral chipping, as occurred with the stock metal screws. The bioreactor lid was equipped with a bubble trap on the media inflow port, a 0.45 μm airport, and a lure-lock covered in foil for inoculating via syringe. The entire bioreactor was autoclaved at 121 °C for 45 min.

Microbial inoculant

Celeribacter baekdonensis strain LH4 was used as the pure-culture inoculant for the bioreactor experiments. C. baekdonensis strain LH4 is a colorless, chemolithoheterotrophic, sulfur-oxidizing bacterium, belonging to the Rhodobacterales within the Alphaproteobacteria. It was isolated from marine sediments collected from a methane seep/brine pool at Green Canyon Block 233 [29] (Lat/Long: 27° 43.4392’ N, 91° 16.7638’ W, Depth: 648 m) in the Gulf of Mexico. Strain LH4 produces acid from thiosulfate oxidation, likely via the sox pathway (soxABCDXYZ) [30]. Strain LH4 also contains other genes related to sulfur oxidation, including four copies of sulfide quinone oxidoreductase ORFs and three copies that encode flavocytochrome c oxidases. Strain LH4 was grown up in the final medium, centrifuged at 10,000 G for 10 min, and washed 2X. OD₅₉₀ for experiments was on average ~0.07 (~2.3•10⁸ cells/mL based on cell counts and growth curves yielding the equation y = 3•10⁹(x) + 2•10⁷ where y equals cells/mL, and x equals OD₅₉₀) with an inoculum volume of 20 mL.

Bioreactor experiments

Experiments were run to measure the dissolution of aragonite in a medium with a saturation state of 0.5 with respect to aragonite, similar to that measured near the sediment/water interface in seep environments [27]. Unlike in the sediments where AOM-generated alkalinity is high and carbonate actively precipitates, at the sediment/water interface previous studies have measured a notable drop in alkalinity [11, 31,32,33,34,35,36] and carbonate dissolution was observed to occur [7,8,9, 11, 27]. In total, six continuous -flow bioreactor experiments were run: three biotic experiments under identical conditions at 10 °C, pH ~7.85, 1.1 mM HCO₃^–, 1 mM S₂O₃^2-, and yeast extract; two uninoculated controls run under the same conditions as above to obtain an abiotic dissolution rate in undersaturated conditions; and one biotic experiment was run under identical conditions, minus the addition of S₂O₃^2- to elucidate the impact of heterotrophy on mineral stability.

10 L carboys containing the final medium and sterile bioreactors were assembled aseptically in a class A2 biological safety bench. Approximately 325 mL of the media was pumped into the bioreactor, followed by inoculating with 20 mL of pure culture and ~600 μL 1 M HCO₃^–. The luer-lock was removed and a 0.2 μM filter was added in its place. Bioreactors were then placed in a 10 °C refrigerator on a stir plate at 80 RPM.

Inoculated bioreactors were run in batch phase for 4 h to promote ample attachment of cells to aragonite coupon surfaces. Following the batch phase, the bioreactor was moved back to the clean hood where one coupon holder rod was removed and replaced with a sterile rubber stopper. Two coupons were placed in the incinerator at 500 °C for 4 h, followed by weighing for mass loss. The additional coupon was fixed in 4% PFA for 2 h, washed, and stained with DAPI [4,6-diamidino-2-phenylindole] for 30 min. Stained coupons were washed and mounted on coverslips (0.17 mm thickness) with DPX mountant for cell counting. After batch phase harvesting, the pump was turned on to the max flow rate (31 mL/min) and the first 100 mL out of the reactor was collected to measure pH, alkalinity, [Ca²⁺], and OD₅₉₀. pH and alkalinity were measured using a Hanna Instruments total alkalinity mini titrator (HI-84431) and pH meter (HI 1131B) following the manufacturers protocol (Hanna Instruments, Woonsocket, RI, USA). [Ca²⁺] was measured using a Hach hardness test kit (product #2063900) following to the manufacturer’s protocol (Hach, Loveland, CO, USA). OD₅₉₀ was measured on a Thermo Scientific Spectronic 20D + (Thermo Fisher Scientific, Waltham, MA, USA), where 1 mL of sterile medium was added to a cuvette to blank the spectrophotometer, and then 1 mL of the collected outflow was placed in a cuvette and measured. The pump was kept at the max flow rate for approximately 3–4 h after batch phase to remove planktonic cells and the additional HCO₃^– added during batch phase. Directly after flushing the bioreactor, water chemistry measurements were taken as they were above, and then the flow rate was dropped to 10 mL/min. Henceforth, the flow rate was only adjusted to keep the bulk fluid at approximately the pH and alkalinity of the starting conditions (max flow rate of 21 mL/min during week 3). To ensure consistency of the bulk fluid chemistry, pH and alkalinity measurements were taken twice daily (8–12 h intervals), just before changing carboys. Carboys were changed about every 8–12 h, depending on the current flow rate. To change carboys, the following steps were taken: (1) The pump was briefly turned off, and sterile aluminum foil was wrapped around the outflow tubing. (2) The bioreactor and currently connected carboy were moved to the biological safety bench on a roll cart. (3) Tubing that connected the carboy to the bioreactor was disconnected, and the male-port of the tubing still connected to the bioreactor was submerged in sterile 70% EtOH. (4) A new 10 L carboy containing the final medium was removed from the 10 °C refrigerator. (5) Aluminum foil covering the female-port of the tubing on the carboy was removed, and sprayed with sterile 70% EtOH. (6) Tubing from the bioreactor and carboy were then connected and placed on the roll cart, and moved back to the 10 °C refrigerator. (7) The pump was turned back on and flow resumed. One coupon holder rod was harvested every four days for 21 days, where two coupons were incinerated and weighed for mass loss and one coupon was fixed and stained for cell counting as described above. Surface area of aragonite coupons were estimated using the diameter of the coupon, the density of aragonite (2.93 g/cm³), and the mass of the coupon before and after bioreactor experiments. Density and mass were used to estimate the volume of the coupon, followed by deriving an estimated surface area. Using the mass of coupons before and after experiments, we calculated the total number of moles lost per unit area. Averaging these data out over time we were able to determine a dissolution rate in units of μmol CaCO₃ • cm⁻² • hr⁻¹, which we then converted to μmol CaCO₃ • cm⁻² • yr⁻¹ (8,760 h/year).

Confocal microscopy

We utilized the flow cell in conjunction with the ratiometric dye C-SNAFL-1 [5-[6]-carboxyseminaphthofluorescein] and Hoechst 33342 for measuring the in vivo pH of 21-day-old biofilms, and visualizing the biofilm, respectively. Coupons transferred to the flow cell were connected to a peristaltic pump at 3 mL/min until imaging occurred.

Spectrofluorometric assays

To calibrate C-SNAFL-1 photon excitations and to evaluate the dye’s potential use in measuring pH in biofilms, 1 mL aliquots of the final media with 10 mM HEPES buffer were adjusted to pH 5.0, 5.2, 5.6, 6.0, 6.4, 6.8, 7.2, 7.6, and 8.0. The final probe concentration was ~1.09 μM of C-SNAFL-1. Additional calibrations were performed with the addition of 4% (vol/vol) C. baekdonensis strain LH4.

Confocal scanning laser microscopy

All biofilm images, pH and ratio calibrations in the flow cell were collected using a Nikon TiE inverted microscope equipped with an A1Rsi confocal scan head (Nikon). Hoechst 33342 fluorescence was excited with a 405 nm laser and collected between 425 and 475 nm using a regular PMT to visualize the biofilm. SNAFL emission was collected between 570 and 620 nm with a GaAsP PMT alternating excitation with the 488 and 561 nm lasers to obtain a ratiometric image. This image was used to calculate the pH across the field (below). 10X/0.45 PlanApo and 20X/0.75 PlanApo VC objectives were used to collect 512- by 512-bit resolution z-stacks. Depths in this paper are the distance from the substratum (base of biofilm) to the focal plane (distance furthest from the biofilm).

To visualize the biofilms, flow was suspended and coupons were stained with Hoechst 33342 (2 drops/mL medium) for 15 min and then flushed. C-SNAFL-1 was added to the medium at a final concentration of 20 μM. Flow resumed with C-SNAFL-1 in the medium and then was suspended for no more than 15 min to minimize the accumulation of acidic metabolites, which could artificially modify the local pH. Image planes were set up to visualize the coupon-attached biofilm and the bulk media adjacent to the coupon in the same image. Biofilms were excited at 405 nm, and emission was detected between 425 and 475 nm. pH microenvironment images were captured using dual excitation at 488 nm and 561 nm, and single emission at 600 nm. To determine the z-stack image parameters, we first excited the sample at 405 nm to visualize the base of the biofilm, and then switched to the dual excitation channels to determine the end of the focal plane (e.g. the coverslip; point furthest away from the biofilm in the z-dimension where the signal diminishes). After determining the z-distance to image, we imaged the biofilm at 405 nm, and then the pH of the microenvironment at 488/561 nm. The pH of the microenvironment was determined by calculating the ratio of excitation intensities (i.e., pixel values) between the two channels (Ex_{488 nm}/ Ex_{561 nm}). All intensities were determined using NIS Elements software.

Calibration standards for CSLM were performed in the flow cell with a sterile aragonite coupon and media with 20 μM C-SNAFL-1 and 10 mM HEPES buffer at pH 8.0, 7.8, 7.7, 7.55, 7.2, 7.0, and 6.8. Calibrations were performed with identical microscope settings as in the experiment above. Titration curves of pH versus ratios were calculated using NIS Elements software according to Eq. IV and were used to convert C-SNAFL-1 excitation ratios to pH values.

Image analysis

pH gradients within the biofilm and surrounding microenvironment were evaluated with NIS Elements software by using kymographs on the ratio images (Ex_{488 nm}/ Ex_{561 nm}). Each kymograph was ~500 μm in length (xy) and ~100 μm in depth (z). The resulting kymograph image had 4–5 regions of interest selected, including 2–3 spanning the biofilm into the bulk fluid directly above, and 2-3 in the bulk fluid adjacent to the biofilm. Regions of interest varied in z-depth depending on the thickness of the biofilm, though all were 13 μm in width (xy).

Estimation of CaCO₃ dissolved from various seep locations

Seep locations

PDC1 – Dive, SO 162; Site, Point Dume; Feature, Chimney complex 1; Latitude, 33 56.45753693; Longitude, 118 50.70879299; Depth (m), 729.522

PDC2 – Dive, SO 164; Site, Point Dume; Feature, Chimney complex 2; Latitude, 33 56.50074983; Longitude, 118 50.64165898; Depth (m), 725.723

PDC3 – Dive, SO 164; Site, Point Dume; Feature, Chimney complex 3; Latitude, 33 56.45417243; Longitude, 118 50.70728425; Depth (m), 728.966

Determining total surface area

Each image taken by submersible at Point Dume had two scale bars generated at the time of imaging, one in the background and one in the foreground, with each representing 10 cm. These two scale bars, in conjunction with a designed macro in Fiji/ImageJ were used to interpolate the total surface area within each image. The macro is based on the equation y-y₁ = m(x-x₁). The slope, m, was determined using the equation ((y₂-y₁)/(x₂-x₁)), where y₂ = 10 cm/(length in pixels of the foreground scale bar), y₁ = 10 cm/(length in pixels of the distant scale bar), x₂ = (y-coordinate of the foreground scale bar), x₁ = (y-coordinate of the distant scale bar). Each image was imported into Fiji/ImageJ, and scaled to 1333 × 750 pixels. All scales were removed to put all measurements in pixel values. Centroid and integrated density were added to set measurements. Each image was duplicated and the duplicated images were changed to 32-bit images. The macro was then run on each respective image. A polygon was then drawn around the visible area within each image, and added to the ROI manager. The ROIs were added to their respective duplicate and measured. These data gave us the total visible surface area present in each image.

Determining surface area of exposed carbonates and carbonated-attached bacterial mats

Polygons were drawn around all exposed carbonates within the visible region. These ROIs were then added to their respective duplicated image, and the total area of exposed carbonates was measured. Carbonate ROIs that contained small regions in which bacterial mats were not visible had “negative” ROIs drawn around them. When applicable, these negative ROIs were subtracted from the total exposed carbonate measurement, resulting in the total mat-on-rock surface area measurement. It is likely that bacteria are also colonizing the surfaces on which no obvious mat can be observed, but removal of these areas from consideration was done to provide a conservative lower estimate of bacterial coverage.

Moles of carbonate rock dissolved was calculated as follows

Experimentally-derived carbonate dissolution rate = 1773.97 μmol CaCO₃ • cm⁻² • yr⁻¹.

Apply dissolution rate to seep surface area covered by rock (e.g. Hydrate Ridge) to obtain the annual dissolution; 17.7397 mol CaCO₃ • m⁻² • yr⁻¹ * 3.00E + 05 m² = 5.32E + 06 mol CaCO₃ • yr⁻¹.

Apply 92.77% average carbonate-attached bacterial mat percent coverage (Supplementary Table 3) to annual dissolution; (5.32E + 06 mol CaCO₃ • yr⁻¹) • (0.9277) = 4.94E + 06 mol CaCO₃ • yr⁻¹.

We then applied these calculations to the remaining five seep sites in Table 1 to obtain the annual amount of carbonate dissolved per seep.

Source: Ecology - nature.com