Sediment samples for ATP analysis were obtained from the end of core sections immediately after the sections were cut on the ship catwalk. Potentially contaminated sediment was removed first, using a sterile scapula. Then, using a 10-cm3 plastic cylinder, a minicore sample was taken, sealed in a plastic bag, and dropped into liquid nitrogen. The samples were stored at -20ºC and shipped frozen. Samples that had thawed during transport were discarded.
All glassware was cleaned prior to use by soaking for 24 hr in 0.1-M HCl, rinsed in distilled water, and dried at 100ºC. All other equipment was stored and cleaned with alcohol. The use of gloves was mandatory. ATP was extracted using the extraction mixture developed by Webster et al. (1984), but omitting Ethylene-dinitrilo-tetra-acetate (EDTA) and Zwittergent 3,10. In detail, a 2-cm3 subcore was taken from the center of the 10-cm3 minicore by means of a steel tube and extruded into a 15-cm3 preweighed centrifuge tube. The exact mass of the subcore was determined by weighing. The sample was homogenized with 5 cm3 of the extraction mixture by means of a hand-held homogenizer. Addition of the strongly acidic reagent mixture led to vigorous CO2 degassing from the CaCO3-rich sediments (see "Results and Discussion" section, this chapter). The centrifuge tubes were placed in an ultrasonic bath for 60 s, an additional volume of 3-cm3 extraction mixture was added, and the tubes were left covered, but not sealed, for 10 min to degas. The centrifuge tubes were capped and placed in a frame, to prevent further degassing of CO2 and blowing the caps off. The frame was placed on a rotating wheel, and further extraction continued for 20 min. Following 20 min centrifugation at 4000 rpm, 1 cm3 of the supernatant was transferred to a test tube containing 1 cm3 strongly acidic cation exchanger (Dowex C-500), which had been cleansed previously with the extraction mixture. In the original method (Webster et al. 1984) EDTA was added to complex cations that might interfere with the Luciferin-Luciferase system. However, initial tests with the original extraction mixture showed very low luminescence, presumably because the amount of EDTA added was insufficient to sequestrate the high concentration of Ca2+ (0.4-1.8 M) in the extracts. Atomic absorption spectroscopic analysis showed that treatment with the cation exchanger removed more than 99% of the Ca2+. Hence, EDTA was omitted from the extraction mixture to minimize possible contamination.
The content of ATP in the Ca-free extract was determined based on the reaction with Luciferin-Luciferase (Webster et al., 1984) by the means of a portable Lumac Biocounter M1500, and nucleotide releasing reagent (NRM) and Luciferin-Luciferase reagent from Lumac. The role of the NRM is similar to the role of Zwittergent 3,10 in the original extraction mixture developed by Webster et al. (1984); hence Zwittergent 3,10 was omitted from the extraction mixture used here. The Lumac reagents have been developed for hygienic tests. During initial tests on sediment samples from Leg 164, these reagents were found to yield to low luminescence for quantitative work. Addition of extra Luciferase prepared by dissolving 5-mg Luciferase (Fluka) in 2-mL deionized water gave satisfactory results. Increasing the concentration of Luciferin had little effect. The acidic Ca-free extract (0.1 cm3) was neutralized to pH 7.7 with 0.5 cm3 of 1 M Hepes (4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic acid) solution adjusted to pH 8.2 with NaOH(s). The neutralized sample (0.1 cm3) was mixed with 0.1 cm3 NRM, 0.1 cm3 of the Lumac Luciferin-Luciferase reagent and 0.025 cm3 of the prepared Luciferase solution in a optical cell and the bioluminescence was measured and integrated for 10 s. Three standard additions of 0.010 cm3 2000 pg cm-3 ATP prepared from Adenosine-5-Triphosphate Disodium Salt hydrate (Fluka) were used to determine the content of ATP in the sample. Recording of the standard-addition curve took about 65 s, during which the bioluminescence decreased by 5%, as determined by repeated analyses of a sample spiked with 0.030 cm3 of the ATP standard. Corrections were made for this effect. Repeated (n = 5) analysis of a sample prepared by mixing sediment from several samples (mean ATP 456 pg g-1) gave a standard deviation of 54 pg g-1. Reagent grade CaCO3 that had been ignited at 550ºC for 6 hr was run as a procedure blank. All measurements were corrected for the blank (51 ± 7 pg g-1 ATP, n = 6).
Estimates of bacteria cell numbers (cells/gram) from ATP concentration data (pg g-1) were obtained by the use of literature data on C/ATP and C/bacteria cell ratios (Karl, 1980; Gerlach, 1978; Findlay et al., 1986). The bacteria cell numbers were further converted to cells per cubic centimeter of wet sediment by means of the porosity data (Table 1) and a mean particle density of 2.71 g cm-3 (Paull, Matsumoto, Wallace, et al., 1996), for easy comparison with data from direct bacteria enumeration (Wellsbury et al., Chap. 36, this volume).
Suspensions of 10 mg sediment in 3 cm3 0.5% sodium-pyrophosphate (dispersant) and 1 cm3 1% sodium azide (biocide) were left for one week and then dispersed for 60 s in an ultrasonic bath. A subsample was further diluted to an absorbance of 0.7-0.9 with 0.5% sodium-pyrophosphate before the particle-size distribution was determined by means of a HORIBA CAPA-500 centrifugal automatic particle analyzer. Particles in the size range 65-6.5 µm were analyzed in gravitational mode in intervals of 6.5 µm, and particles with diameters less than 6.5 µm were analyzed in centrifugal mode with intervals of 0.5 µm.
Total carbon, nitrogen, and organic carbon were analyzed by the means of a Carlo Erba 1100 elemental analyzer on freeze-dried and homogenized sediments. Inorganic carbon was removed prior to determination of organic carbon by addition of 10-cm3 1-M HCl to 0.2 g sediment. The suspensions were centrifuged after 48 hr reaction time, and the particles were washed with 10-cm3 deionized water, centrifuged, and freeze dried.
Calculation of principal components (PCs) were made by means of the software package SIRIUS (Pattern Recognition Systems, High Technology Center in Bergen, N-5007 Bergen, Norway) after normalization of the variables by division by their standard deviation.