SAMPLING THE DEEP BIOSPHERE

Shipboard Sampling
Water samples were collected to determine background microbial populations introduced in the drilling process. Undisturbed water in Hole 801C was viewed as a potentially excellent in situ microbial culture, and was thus sampled using the water-sampling temperature probe (WSTP) before the hole was disturbed; ~300 mL of water was recovered from a depth of 540 m in the hole. Water samples were also collected from the drill pipe and from the sea surface (from the z-boat) upwind of the JOIDES Resolution.
A series of samples of different rock types were collected for culturing, DNA and Adenosine triphosphate (ATP) analyses, and total cell counts. Figure 17 shows an example of a bacterial population contaminated with fluorescent microspheres. Whole-round samples of basement rocks, including glass samples, were (1) taken and transported to an anaerobic chamber where they were cracked open and interior pieces were used to inoculate cultures and (2) also used for shore-based microscopic analyses and DNA extraction. Cultures have also been started at in situ pressure, and some samples have been stored under pressure for shore-based analyses.
Hole 801C, and to a lesser extent volcanic rocks from Site 1149 preserve small amounts (several grams to a few grains) of fresh basaltic glass. These samples were studied microscopically for traces of microbial activity. Several samples preserve filament like textures that are identical to textures attributed to microbes in deep-sea basalt from young oceanic pillow basalt. Whether these textures are trace fossils of past microbial activity or signs of extant microbes living in this high-pressure seafloor environment remains a fascinating subject of shore-based study.

Tracer Tests
The issue of core contamination often makes interpretation of observations in deep subsurface microbiology difficult because nonindigenous microorganisms can be introduced during the drilling and sampling processes. This concern is amplified for extremely deep environments where the abundance of microorganisms is generally very low and the potential for contamination is great. Because the Ocean Drilling Program ventures into the realm of deep biosphere research, the issue of sample contamination, especially as it pertains to the drilling process on board the JOIDES Resolution, must be addressed.
The drilling fluids (mud and/or surface seawater), the drill string itself, and sample handling procedures on deck are all possible sources of contamination. Quantitative contamination tests have been conducted in other drilling operations using a variety of chemical, microbiological, and particulate tracers (see review by Griffin et al., 1997). An ideal chemical tracer is inert, not found naturally in the environment, and easily detected at extremely low concentrations. One of the goals during Leg 185 was to establish shipboard procedures on the JOIDES Resolution to deliver perfluorocarbon tracers (PFTs) and fluorescent microspheres 0.5-1.0 µm in size during drilling and to detect the tracers in recovered cores. Sediment and igneous rock require different drilling techniques, and tracer experiments were performed in both types of formations.
PFT data, along with the measured concentration of microbial cells in surface waters, can be used to estimate the extent of potential microbial contamination. The measured number of cells in the surface water at Site 801 was 4.2 x 10² cells/µL. Based on this value, microbial contamination is less than one cell per gram when there is less than ~0.0025 µL of drilling fluid contamination per gram of sample. For interior samples, this is the case for five of the 12 unconsolidated sediments cored by APC, one of the five basalt analyses, and one of the four consolidated sediments analyzed that were cored by RCB.
Caution must be used in the interpretation of the PFT contamination tests. Although low PFT values can be taken as proof of minimal contamination, high PFT levels do not unequivocally prove microbial contamination. There are two reasons for this. First, the PFT is able to penetrate much smaller pores than contaminating microorganisms. Second, perfluoro (methylcyclohexane) is volatile at room temperature and can cause contamination during handling via gas-phase transfer of the tracer from drilling fluid on the exterior sediment and rock surfaces to the interior samples.
The microspheres are not a perfect mimic of microorganisms. Whereas microspheres are similar in size to microorganisms, they have different surface properties and, therefore, may behave differently as they migrate through the formation or attach to mineral surfaces. Microspheres are dispersed in the core barrel when rock passes through the core catcher, so the rocks are exposed to the tracer only after it has been saturated with drilling fluid and the physical disturbance of the formation by drilling. Thus, microspheres are less likely to be introduced into cracks than PFTs.
The rapid analysis and high sensitivity of PFTs makes it possible to quickly check samples for the possibility of contamination before performing the microbiological manipulations and, therefore, avoid wasting resources on contaminated samples. In addition, if samples are routinely taken, it may be possible to build up a database correlating the probability of contamination with different types of sediment and rock formation characteristics (e.g., porosity, percent veins, and mineralogy) and type of drilling. As little as 10^-12g of PFT is detectable. Higher sensitivity may possibly be achieved by using a smaller bore column on the gas chromatograph. This should sharpen the peaks and improve the signal to noise ratio. The use of a less volatile PFT may also improve the reliability of the method as an indicator of microbial contamination.
The processing of hard rock relies on the use of water as a lubricant and cooling fluid, both during subseafloor drilling and sample preparation in the laboratory. Water can introduce microbial contamination to the core as well as transfer surface contaminants into the interior of a sample during the processing of wet samples. For this reason, to sample core interiors core surfaces should be dried before handling and broken with a rock splitter, which does not require water, instead of a rock saw.
Because rock saws and polishing grits are required to make thin sections, it is critical that contamination is controlled during these processes as well. Establishing a protocol for the cutting and polishing of thin sections, without the introduction of microbial contaminants, is essential.
On average, the sediments cored with the APC showed less susceptibility to contamination, and several core interiors were entirely free of contaminants. RCB coring resulted in the presence of chemical, but not particulate, tracers in the interior of the cores. This difference is probably, though not unequivocally, because of the nature of the coring. The APC core barrel is fired into the sediment in less than a second, whereas the RCB can take several hours to cut a 9.5-m core with drilling fluid continuously flowing through the bit.
As more research on microbial activity in the deep biosphere is conducted, it is important to remain cognizant of the issue of contamination. Although the absolute quantity of the contaminant seems very small, unlike chemical contamination, microbial contamination can be amplified after the sample has been taken. This is a concern with nonindigenous microorganisms growing in cultures inoculated with material from the deep biosphere, as well as in samples where the DNA of nonindigenous microbes may be amplified using the polymerase chain reaction. Therefore, continued, routine testing for microbial contamination during drilling on board the JOIDES Resolution will provide the scientific community with quality data on deep biosphere microbiology.

To 185 Summary