Sample collection and storage: Subsurface sediment samples from Hydrate Ridge (IODP Leg 204 Site 1244a; 44° 35′ 17″ N, 125° 07′ 19″ W), Peru Margin (IODP Leg 201 Site 1227a; 79° 57′ 349″ W, 08° 59′ 463″ S), and Eastern Equatorial Pacific (IODP Leg 201 Site 1225a; 110° 43′ 289″ W, 02° 46′ 247″ N) were obtained from the Gulf Coast Core Repository (University of Texas A&M). Gravity core subsurface samples from North Pond near the Mid-Atlantic Ridge (22° 48′ 04″ E, 46° 06′ 30″ N) and Benguela Upwelling System (14° 15′ 04″ E, 27° 44′ 40″ S) were collected on March 3rd 2009 onboard the R/V Maria Merian and April 21st 2008 onboard the R/V Meteor, respectively and were provided by Andreas Teske (University of North Carolina, Chapel Hill, NC). Careful precautions were taken during sampling to avoid contamination during the sampling process. For IODP cores, contamination tests were performed using Perfluorocarbon tracers and fluorescent microspheres. Sediment samples were immediately frozen at −80 degrees C after sampling and stored at −80 degrees C until RNA was extracted. Sediment samples at a sediment depth of 0.01 and 0.08 mbsf from Little Sippewissett Salt Marsh were taken November 13th 2011 using a sterile syringe. Sulfide was detectable in both samples and thus samples were presumed anoxic. No specific permits were required for the described field studies. The locations sampled are not privately owned or protected and field studies did not involve endangered or protected species.

RNA extraction and purification: RNA was extracted from 25 grams of sediment using the FastRNA Pro Soil-Direct Kit in a laminar flow hood to reduce contamination from aerosols. Extractions were performed at Woods Hole Oceanographic Institution. Several modifications were made to the protocol provided with the kit to increase RNA yield from low biomass subseafloor samples. It was necessary to scale up the volume of sediment that is typically extracted with the kit (~0.5 grams) due to the expected low biomass of subsurface eukaryotes. Four 15 ml Lysing Matrix E tubes (MP Biomedicals, Solon, OH) were filled with 5 g sediment and 5 ml of Soil Lysis Solution. Tubes were vortexed to suspend the sediment and Soil Lysis Solution was added to the tube leaving 1 ml of headspace. Tubes were then homogenized for 60 seconds at a setting of 4.5 on the FastPrep-24 homogenizer. Contents of the 15 ml tubes were combined into two RNAse free 50 ml falcon tubes and centrifuged for 30 minutes at 4000 RPM. The supernatants were combined in a new 50 ml RNAse-free falcon tube and 1/10 volume of 2M Sodium Acetate (pH 4.0) was added. An equal volume of phenol-chloroform (pH 6.5) was added and vortexed for 30 seconds, incubated for 5 minutes at room temperature, and centrifuged at 4000 RPM for 20 minutes at 4 degrees C. The top phase was carefully transferred to a new 50 ml falcon tube and 2.5x volumes 100% ethanol and 1/10 volume 3M Sodium Acetate were added and incubated overnight at −80 degrees C. After incubation, tubes were centrifuged at 4000 RPM for 60 minutes at 4 degrees C and the supernatant removed. Pellets were washed with 70% ethanol, centrifuged for 15 minutes at 4 degrees C, and air-dried. Dried pellets were resuspended with 0.25 ml RNAse-free sterile water and combined into a new 1.5 ml RNAse-free tube. 1/10 volume of 2M Sodium Acetate (pH 4.0) and an equal volume of phenol:chloroform (pH 6.5) were added, the tube was vortexed for 1 minute, and incubated for 5 minutes at room temperature. The tube was then centrifuged for 10 minutes at 4 degrees C, the top phase removed into a new RNAse free 1.5 ml tube, and 0.7 volumes of 100% isopropanol was added and incubated for 1hour at −20 degrees C. After incubation tubes were centrifuged for 20 minutes at 14000 RPM at 4 degrees C and the supernatant was removed. Pellets were washed with 70% ethanol and centrifuged at 14000 RPM for 5 minutes at 4 degrees C. Ethanol was removed and the pellets air-dried. Pellets were resuspended with 200 ul of RNAse free sterile water and DNA was removed using the Turbo DNA-free kit (Life Technologies, Grand Island, NY). DNAse incubation times were increased to 1 hour to ensure removal of contaminating DNA. Samples were then taken through the protocol supplied with the FastRNA Pro Soil-Direct kit to the end (starting at the RNA Matrix and RNA Slurry addition step), including the optional column purification step to remove residual humic acids. To further purify the RNA, we used the MEGA-Clear RNA Purification Kit. Extraction blanks were performed (adding sterile water instead of sample) to identify aerosolized contaminants that may have entered sample and reagent tubes during the extraction process. To reduce contamination, all RNA extractions were performed in a laminar flow hood.

RT-PCR amplification of eukaryotic rRNA: To amplify the V4 hypervariable region of eukaryotic rRNA, we used PCR primers targeting this region: EukV4F (5′ – CGTATCGCCTCCCTCGCGCCATCAGxxxxxxxxxxCCAGCASCYGCGGTAATTCC – 3′) and EukV4R (5′ – CTATGCGCCTTGCCAGCCCGCTCAGACTTTCGTTCTTGATYRA – 3′), where the x region represents the unique MID barcode used for each sample, the linker primer sequence is underlined, and the 18S rRNA eukaryotic primer is bold. These primers were chosen because they target a wide range of eukaryotic taxa. RT-PCR was performed using the SuperScript One-Step RT-PCR with Platinum Taq kit. Individual reactions consisted of 2 ul RNA template, 25 ul buffer, 1 ul of forward Primer, 1 ul of reverse primer, 2 ul of the Platinum RT-Taq enzyme mix, and 18 ul RNAse free sterile water. The cDNA step was performed at 55 degrees C and cDNA was amplified in 40 cycles of PCR with an annealing temperature of 65 degrees C (55 degrees C for 30 minutes, 95 degrees C for 5 minutes, [95 degrees C for 15 seconds, 65 degrees C for 30 seconds, 68 degrees C for 1 minute]x40, 68 degrees C for 5 minutes). To check for DNA carryover during the RNA extraction protocol, a separate PCR reaction (at the same number of cycles) was included in which Taq polymerase was substituted for the reverse-transcriptase/platinum Taq enzyme mix. For each sample, 5–10 RT-PCR reactions were performed and extracted using the Zymo Research Gel Extraction Kit. A gel volume of 100% isopropanol was added to each dissolved gel slice before addition to the DNA collection column. Dissolved gel slices from each sample were pooled by centrifuging them all through the same DNA collection column. cDNA was quantified fluorometrically prior to 454 sequencing using the Qubit 2.0. To identify contaminants we performed additional RT-PCR amplifications at 55 cycles using RNAse free sterile water and RNA extraction blanks (resulting from RNA extractions in which no sample was added) as template. Contaminants were amplified with primers containing a unique MID in 55 cycles of PCR.

Quality control, clustering, and taxonomic assignment of 454 data: cDNA amplicons were sequenced on a GS-FLX Titanium 454 sequencer at EnGenCore (University of South Carolina, Columbia, SC), which resulted in ~37000 reads. To reduce homopolymer errors inherent to 454 sequencing, the dataset was put through the denoise protocol as described in the QIIME software package using the denoise_wrapper.py command. After denoising, chimeric sequences were identified and removed using ChimeraSlayer with the blast_fragments method in QIIME. The data were subjected to quality score filtering using the split_libraries.py command and clustered at various levels of sequence identity (80%, 85%, 90%, 93%, 95%, 97%) in QIIME using the uclust method of all-to-all pair-wise comparisons via the pick_otus.py command.

The QIIME taxonomy classification pipeline was not able to accurately classify the majority of eukaryotic OTUs. Thus, we used Jaguc, a program developed specifically for classification of eukaryotic rRNA sequence data, to classify our sequence reads. 90% of eukaryotic OTUs were classified to genus using this approach. OTU tables were created using the make_otu_table.py command in QIIME and the Jaguc taxonomy for each OTU was amended onto this table using a custom perl script developed by the authors for this purpose. This perl script is available from the authors upon request.

Terminal Restriction Fragment Length Polymorphism (TRFLP) analysis of fungal rRNA: To further investigate the fungal diversity in our samples, we used a TRFLP approach using PCR primers specific to fungal 18S rRNA. The fungal primers used were EF3 (5′ – TCCTCTAAATGACCAAGTTTG – 3′) and Fung5 (5′ – GTAAAAGTCCTGGTTCCCC – 3′). The forward primer, EF3, was labeled with the phosphoramidite dye 6-Carboxyfluorescein (6-FAM) at the 5′-end (Integrated DNA Technologies, Coralville, Iowa). Fungal rRNA was amplified using a cDNA incubation step at 50 degrees C followed by 40 cycles of PCR with an annealing temperature of 53 degrees C. Three RT-PCR reactions were performed for each sample, gel extracted, and pooled using the same protocol as above. Fungal rRNA amplicons were digested with three different restriction enzymes: MspI, RsaI, and HhaI (New England Biolabs, Ipswich, MA), for 1 hr at 37 degrees C. These restriction enzymes were chosen because they have been shown to provide statistically significant TRFLP data for interpreting fungal community structure across different samples. Digests were mixed with the Applied Biosystems size marker GS600LIZ and HiDi Formamide in the ratio 1:1:9 and run on an Applied Biosystems 3730 DNA analyzer. Electropherograms were analyzed using the PeakScanner software package (Applied Biosystems, Carlsbad, CA) to identify the size, height, and peak area of each T-RF. T-REX was used to filter out noise from true peaks and to align peaks.

Related references:
The manuscript is at http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0056335