MATERIALS AND METHODS

Site Description

Samples for determination of bacterial populations and activity were taken from cores at three sites on the Blake Ridge (Paull, Matsumoto, Wallace, et al., 1996). These sites (Sites 994, 995, and 997) formed a transect from an area on the southern flank of the Blake Ridge where a BSR was not detectable to an area where an extremely well-developed BSR existed (Fig. 1).

No BSR was observed at Site 994, and thus Site 994 was originally intended as a control site without gas hydrates but with the same sediment types. However, downhole interstitial water chloride anomalies, temperature anomalies in recovered cores, and patterns in the downhole log data indicated that gas hydrates were present in the sediment, typically occupying between 1 and 3 vol% of the sediments between 220 and 430 mbsf, with several intervals containing up to 9.5% gas hydrate.

At Site 995 a modest BSR was detected (Fig. 1), and the interstitial water geochemistry and physical properties were very similar to those at Site 994. Although no gas hydrate was recovered from Site 995, downhole interstitial water chloride anomalies, anomalies in recovered core temperatures, and patterns in the downhole log data indicated that gas hydrates occupied between 1 and 4 vol% of the sediments between 195 and 450 mbsf, with several intervals containing up to 10% gas hydrate. The BSR was apparently associated with the first occurrence of free gas.

Site 997 was located on the topographic crest of the Blake Ridge, and possesses an extremely well-developed and distinct BSR. Interstitial water chloride anomalies, temperature anomalies in the recovered cores, and patterns in the downhole log data indicated that gas hydrate occupies between >1.5 and 4 vol% or more of the sediment from 195 to 450 mbsf. Several intervals contained as much as 11% gas hydrate. The BSR was associated with a zone of free gas, which was well developed for 30 m and extended for over 100 m below the BSR.

During Leg 164 the pressure core sampler (PCS) was deployed to allow samples of gassy sediment to be recovered at in situ pressure, releasing the gas in controlled laboratory conditions at the surface to allow accurate determination of gas quantity and composition. Results from the PCS indicated that substantial quantities of methane (~15 GT of carbon) existed in the form of solid gas hydrate at Blake Ridge (Dickens et al., 1997), and, equally important, that an equivalent or greater amount of free gas existed below the BSR.

Shipboard Sample Handling

Whole-Round Cores

All samples for measurement of potential bacterial activity and estimation of bacterial numbers were taken from either 20- or 25-cm-long whole-round cores (WRC). In total, 37 samples were obtained from Site 994 (15 20 cm samples, 0.1-40.81 mbsf), Site 995 (16 25 cm samples, 0.13-0.6 mbsf) and Site 997 (6 20 cm samples, 330.69-491.53 mbsf). The samples were taken from 1.5-m core sections using a specially constructed cutting rig (Parkes et al., 1995). The cut ends of the WRC were capped with sterile core end caps under a flow of sterile oxygen-free nitrogen (OFN) to maintain anaerobic conditions. Modified core end caps fitted with a one-way gas release valve were used for WRC samples containing gas hydrate to allow gas pressure to vent from the samples. Capped WRC were stored in gas-tight anaerobic bags (Cragg et al., 1992) in the ship's cold room at 4ºC and transported by air back to the laboratory in insulated trunks containing wet ice and ice packs. The samples remained cold throughout transportation.

Direct Bacterial Enumeration

For direct determination of bacterial numbers, 1-cm³ sediment samples were taken from the end of selected 1.5-m core sections immediately after the sections were cut on the catwalk. A thin layer of potentially contaminated sediment was removed from the core using a sterile scalpel. A 1-cm³ sample was then removed using a sterile (autoclaved) 5-mL syringe from which the luer end had been removed. The sample was ejected directly into a preweighed serum vial containing 9 mL of filter-sterilized (0.2 µm) 4% (v/v) formaldehyde in artificial seawater. Additionally, three samples of "pure hydrate" were taken from the massive gas hydrate deposit at Site 997 (331 mbsf), washed with filter-sterilized (0.1 µm) water to remove external sediment and stored in capped serum vials as before, allowing gas pressure to vent through a sterile needle.

Pore Waters

Pore-water samples were filter sterilized (0.1 µm) and stored, frozen, in 2.5-mL capped vials before transportation back to the laboratory for determination of acetate concentrations by high performance liquid chromatography (HPLC).

Laboratory Sample Handling

Whole-Round Cores

On arrival at the laboratory in Bristol the WRC samples were stored in a constant-temperature room at 4ºC prior to further handling. All samples were processed within three weeks of arrival.

All sample handling was performed under aseptic, anaerobic conditions (Parkes et al., 1995). WRC were cut into 5-cm sections, from each of which ten 5-cm³ syringe subcores were removed for radiotracer potential activity measurements, and one 5-cm³ syringe subcore for most probable number (MPN) viable counts. Syringe subcores were taken from the center of the WRC, avoiding sediment near the core liner to avoid the possibility of contamination.

The 20-cm WRC from Sites 994 and 997 were sliced into four 5-cm sections as described above, and potential rates of (1) methanogenesis from bicarbonate, (2) sulfate reduction, (3) methane oxidation and (4) thymidine incorporation were determined. Additionally, at Site 995, the fifth 5-cm section was used to determine rates of acetate turnover.

Direct Microscopy

Total numbers of bacteria were determined using acridine orange staining and epifluorescence microscopy (Fry, 1988; Fry, 1990). Formaldehyde-fixed samples were vortex mixed, and 5 µL was added to 10 mL of 2% filter-sterilized (0.1-µm filter) formaldehyde, along with 50 µL of filter-sterilized, 1 g L^-1 acridine orange solution. After 3 min incubation, the solution was filtered through a 25-mm Nucleopore (0.2-µm pore size) black polycarbonate membrane (Costar, UK), which was rinsed with a further 10 mL of 2% filter-sterilized (0.1-µm filter) formaldehyde and mounted on a glass slide in a minimal amount of paraffin oil under a cover slip.

Mounted membranes were viewed under epifluorescent illumination (Zeiss Axioscop, 50W mercury vapor lamp, blue excitation, 100 Plan Neofluor oil-immersion objective, and 10x eyepiece). Fluorescent bacteria were enumerated; cells were recorded as "on" or "off" particles, doubling the number of cells on particles in the final calculations to account for masking. Dividing cells (with a clear invagination) and divided cells (pairs of cells with identical morphology) were also counted. Triplicate membranes were prepared and counted for each sample, with a minimum number of 200 fields of view examined for each membrane. Where replicate log₁₀ counts differed by more than 0.5, a fourth membrane was prepared. This gives a detection limit of 1 10⁵ cells mL^-1 (Cragg and Parkes, 1994). Periodic blank membranes were also counted to check for potential contamination.

Enumeration of Viable Bacteria

A most probable number (MPN) technique (Hurley and Roscoe, 1983) was used to estimate numbers of viable fermentative heterotrophs; nitrate- and sulfate-reducing bacteria; and methane-oxidizing, sulfate-reducing bacteria. This procedure involved 9-11 dilution levels from the original sediment, descending serially in quadruplicate one in five dilutions. All MPN enrichments were performed in 7-mL serum vials of anaerobic media, sealed with butyl rubber septa and aluminum crimp tops (Phase Separations Ltd., Deeside, UK). Compositions of the growth media were as follows:

Fermentative heterotroph medium: 0.2 g L^-1 KH₂PO₄; 0.25 g L^-1 NH₄Cl; 30.0 g L^-1 NaCl; 6.0 g L^-1 MgCl₂·6H₂O; 0.5 g L^-1 KCl; 0.15 g CaCl₂·2H₂O; 1.0 mg L^-1 Resazurin; 0.5 g L^-1 casamino acids; 0.1 g L^-1 yeast extract. The pH was adjusted to 7.5 with NaOH and the medium autoclaved. Once autoclaved, the following were added from sterile, stock solutions: 3.0 mL L^-1 of "Combined Vitamin Solution" (40.0 mg L^-1 4-aminobenzoic acid; 10.0 mg L^-1 D[+] biotin; 100.0 mg L^-1 thiamine-HCl; 20.0 mg L^-1 folic acid; 100.0 mg L^-1 pyridoxine-HCl; 50.0 mg L^-1 riboflavin; 50.0 mg L^-1 nicotinic acid; 50.0 mg L^-1 DL calcium pantothenate; 50.0 mg L^-1 lipoic acid; 50.0 mg L^-1 cyanocobalmine); 3.0 mL L^-1 of "Trace Elements 1" (190.0 mg L^-1 CoCl₂·6H₂O; 100.0 mg L^-1 MnCl₂·4H₂O; 70.0 mg L^-1 ZnCl₂; 62.0 mg L^-1 H₃BO₃; 36.0 mg L^-1 Na₂MoO₄·2H₂O; 24.0 mg L^-1 NiCl₂·6H₂O; 17.0 mg L^-1 CuCl₂·2H₂O; 1.5 g L^-1 FeCl₂·4H₂O [dissolved first in 10 mL 25% {v/v} HCl]); 3.0 mL L^-1 of selenite/tungstate (3.0 mg L^-1 NaSeO₃·5H₂O; 4.0 mg L^-1 NaWO₄·2H₂O); 30.0 mL L^-1 of 84.0 g L^-1 saturated NaHCO₃ solution, 3.0 mL L^-1 of 12.0 g L^-1 Na₂S solution; 10.0 mL L^-1 of a mixed solution of 0.5 g glucose and 0.49 g glycerol in 10 mL water. The pH of the medium was readjusted to 7.5 and dispensed under a N₂:CO₂ headspace into sterile 7-mL vials containing 2.5 mg chitin and 2.5 mg cellulose. Each vial was sealed with a butyl rubber septum and aluminum crimp top. Positive growth in each of the vials was determined by phase-contrast microscopy.

Nitrate-reducing bacteria medium: 25.0 g L^-1 Nutrient Broth No. 2 (Merck UK); 1.01 g L^-1 KNO₃; 1.0 mg L^-1 Resazurin; 25.8 g L^-1 NaCl; 5.2 g L^-1 MgCl₂·6H₂O. The pH was adjusted to 7.5 with NaOH, and the medium autoclaved and dispensed under OFN. Positive growth in each of the vials was determined by presumptive color change (orange to yellow).

Sulfate-reducing bacteria medium: 0.5 g L^-1 KH₂PO₄; 1.0 g L^-1 NH₄Cl; 1.0 g L^-1 CaSO₄; 2.0 g L^-1 MgSO₄·7H₂O; 0.875 g L^-1 sodium lactate; 0.45 g L^-1 sodium acetate; 0.1 g L yeast extract; 0.1 g L^-1 ascorbic acid; 0.1 g L^-1 thioglycollic acid; 0.5 g L^-1 FeSO₄·7H₂O; 26.6 g L^-1 NaCl; 5.4 g L^-1 MgCl₂·6H₂O; 1.0 mg L^-1 resazurin. The pH was adjusted to 7.2 with NaOH under OFN, 30.0 mL L^-1 of 84.0 g L^-1 saturated NaHCO₃ solution added, and the medium transferred to an anaerobic cabinet (Forma Scientific, UK, gas composition 80% N₂, 10% CO₂, 10% H₂). The medium was dispensed and crimp-sealed (butyl septa) in the anaerobic cabinet and autoclaved. Positive growth in each of the vials was determined by production of a black FeS precipitate.

Methane-oxidizing sulfate-reducing bacteria medium: 0.5 g L^-1 KH₂PO₄; 1.0 g L^-1 NH₄Cl; 1.0 g L^-1 CaSO₄; 2.0 g L^-1 MgSO₄·7H₂O; 0.1 g L^-1 yeast extract; 0.1 g L^-1 ascorbic acid; 0.1 g L^-1 thioglycollic acid; 0.5 g L^-1 FeSO₄·7H₂O; 26.6 g L^-1 NaCl; 5.4 g L^-1 MgCl₂·6H₂O; 1.0 mg L^-1 resazurin. The pH was adjusted to 7.2 with NaOH under OFN, and 30.0 mL L^-1 of 84.0 g L^-1 saturated NaHCO₃ solution added. The mixture was transferred to an anaerobic cabinet, and the medium was dispensed, crimp-sealed (butyl septa), and autoclaved. Once injected with diluted sample, the headspace was pressurized to +5 p.s.i. with sterile (0.1-µm filter) methane. Positive growth in each of the vials was determined by increase in headspace CO₂ and decrease in headspace methane in each vial by GC.

Injection Schedule

The 5-cm³ subcores were allowed to equilibrate overnight before injection with radiotracers using a specially designed injection rig (Parkes et al., 1995) that allowed an even distribution of isotopes along the center line of the subcore. Each group of ten subcores was divided into one time-zero control and triplicate samples in each of three incubation periods. Time-zero subcores were prechilled at 4ºC, injected and immediately frozen, and stored in anaerobic bags at -20ºC. Incubated samples were sealed in anaerobic bags after injection and incubated at the approximate mean downhole temperature of 12ºC for varying periods (specified in text below). Incubations were terminated by freezing at -20ºC, and subcores were stored frozen prior to analysis. In all cases, time-zero control results were subtracted from experimental data before calculation of potential activity rates.

Potential Methane Oxidation

The ¹⁴C methane (supplied by Amersham, U.K.) was removed from an adapted break-seal ampule using a sterile gas syringe. Fifty µL of gas were injected into each subcore. Equivalent volumes of sterile 0.1 M NaOH were added to the gas reservoir to maintain pressure before further removal of gas samples. Methane oxidation rates were determined from the amounts of ¹⁴CO₂ produced during incubation periods varying from 18 hr to 48 days, depending on sample depth. Frozen syringe subcores were ejected into 30-mL serum vials containing 10 mL of 0.6 M NaOH and a small magnetic stirrer. The vial was crimp sealed, shaken vigorously to disperse the sediment, and flushed with OFN at 60 mL min^-1 for 30 min. The methane content of the vial was then determined after oxidation to ¹⁴CO₂ (Cragg et al., 1995), and once all residual methane had been stripped from the vial it was transferred to a second flushing rig including two acid traps (1 M HCl) to intercept aerosol droplets. The vial contents were acidified (2 mL of 2 M H₂SO₄) and stirred, and the headspace flushed with OFN for 40 min at 80 mL min^-1. Labeled CO₂ was trapped in three sequential vials containing scintillation cocktail designed to trap CO₂ (800 mL toluene, 70 mL -phenylethylamine, 80 mL methanol, 5 g PPO, 0.1 g POPOP, [Cragg et al., 1990]) and ¹⁴C activity determined by liquid scintillation counting (LSC, LKB-Wallace 1414).

Potential Methanogenesis from Bicarbonate

Sample subcores were injected with 7.2 µL (4.8 µCi) of sodium ¹⁴C-bicarbonate solution (supplied by Amersham UK, diluted with filter-sterilized [0.2 µm], degassed distilled water), and incubated for varying time periods (from 18 hr to 48 days) before termination by freezing. Rates of methanogenesis from bicarbonate were determined from the amount of ¹⁴CH₄ produced, calculating pore-water carbon dioxide from alkalinity data. Frozen subcores were ejected into 10 mL of 0.6 M NaOH in a 30-mL serum vial, which was crimp sealed with a butyl rubber septum. ¹⁴CH₄ was stripped from the vial by flushing with OFN (60 mL min^-1 for 40 min), oxidized to ¹⁴CO₂, and trapped in -phenylethylamine scintillation cocktail for LSC as described above.

Potential Sulfate Reduction

Rates of sulfate reduction were determined from the proportion of ³⁵S-labeled sulfide produced. Subcores were injected with 7.2 µL (3.6 µCi) of ³⁵-S sodium sulfate solution (Amersham UK, diluted with filter-sterilized [0.2 µm], autoclaved distilled water) and incubated for varying time periods (from 18 hr to 48 days) before termination by freezing.

A sequential distillation regime was used to separate labeled sulfides into three fractions (Allen and Parkes, 1995). Frozen syringe subcores were ejected into 10 mL of 20% (w/v) zinc acetate solution and allowed to thaw with occasional mixing. Subsample aliquots (1 mL) were removed from the thawed samples after vortex mixing and added to a conical flask containing 10 mL of 35 NaCl and a magnetic stirrer. The flask was attached to a distillation rig, and the headspace flushed with OFN for 15 min at 80 mL min^-1 before addition of 2 mL of 6 M HCl. The flasks were heated to 80ºC and distilled for 40 min, trapping labeled acid-volatile sulfide in 10 mL of 10% (w/v) zinc acetate. The flasks were then allowed to cool before addition of 5 mL 95% (v/v) ethanol, 25 mL CrCl₂ and 5 mL of concentrated HCl. This cold chromous chloride distillation (pyritic sulfide fraction) was allowed to proceed for 40 min, after which the flask was heated to 80ºC for a further 40 min of hot chromous chloride distillation (elemental sulfur fraction). At the end of each distillation period the zinc acetate traps were changed and aliquots (1 mL) removed for determination of ³⁵S activity by LSC and total sulfide by spectrophotometry (Cline, 1969). Total reduced inorganic sulfide (TRIS) was calculated as the sum of the three sulfide fractions.

Thymidine Incorporation

Syringe subcores were injected with 25 µL (18 µCi) of methyl[³H]thymidine (83 Ci mmol^-1, Amersham UK) and incubated for varying time periods (from 30 min to 36 hr) before termination of the incubations by freezing. Frozen samples were transferred to a 13-mL centrifuge tube containing 5 mL of 20% (w/v) aqueous trichloroacetic acid (TCA) solution and allowed to thaw, mixing thoroughly at intervals. Labeled DNA was extracted from the sediment by an acid-base hydrolysis method (Wellsbury et al., 1993; Wellsbury et al., 1994; Wellsbury et al., 1996). Rinsed, dried sediment was hydrolyzed with 1 M NaOH at 37ºC for 1 hr, and centrifuged (2500 g for 10 min at 4ºC). The supernatant was removed, cooled to 4ºC, and DNA reprecipitated by addition of 1.5 mL of 20% (w/v) TCA in 3.6 M HCl, Kieselguhr, and 50 µL of a saturated solution of unlabeled "carrier" DNA. Following centrifugation (2500 g for 10 min at 4ºC) and rinsing (once each with 5% (w/v) TCA, 95% (v/v) ethanol, all at 4ºC), labeled DNA was extracted from the pellet in 5% TCA at 100ºC for 30 min, and activity determined by LSC.

Potential Acetate Turnover

Replicate syringe subcores were injected with 7.4 µL (7.4 µCi) of [1-(2)¹⁴C] acetate (Amersham UK) and incubated for varying time periods (from 1 hr to 13 days) before termination by freezing. Frozen subcores were ejected into 30-mL serum vials containing 10 mL of 0.6 M NaOH, crimp sealed with butyl septa, and allowed to thaw before extraction of labeled ¹⁴CH₄ and ¹⁴CO₂ as described above. Rates of methane and carbon dioxide production were calculated based on the proportion of labeled gas produced and the bioavailable pore-water acetate pool determined by HPLC.

Pore-Water Acetate Determination by HPLC

Pore-water acetate concentrations were determined using an enzymatic method (King, 1991). Aliquots (20 µL) of each of (1) bovine serum albumin (BSA, 200 µg mL^-1); (2) disodium ATP (10 mM); (3) coenzyme A (sodium salt, from yeast, 10 mM); and (4) acetyl coenzyme A syntheses (20 U mL^-1, ~4.9 U mg^-1 protein) were added to 1 mL thawed pore water in a screw-cap 2.5-mL vial with an integral O-ring seal (Sarstedt, Leicester, UK). The samples were mixed thoroughly and incubated at 37ºC for 1 hr before termination by immersion in a boiling water bath for 2 min. After cooling, 800-µL aliquots were transferred to glass autosampler vials and stored in a cooled autos ampler at 4ºC before separation by HPLC (Waters UK). Samples (10-µL injections) were injected onto an analytical column (Supelco LC-18-T, 25 cm 4.6 mm) with a mobile phase of 0.1 M KH₂PO₄ (pH 6.0) at 30ºC at 0.8 mL min^-1. Detection was by UV/vis detector at 254 nm and quantification by peak-area integration (Waters data module). Deeper samples containing high concentrations were diluted appropriately such that the detector response was 100-µM acetate.