E. coli MG1655 and R. sphaeroides 2.4.1 (ATCC, Manassas, VA, USA) were grown in 250 ml LB broth with shaking at 37°C and 30°C, respectively, to an OD600 of about 0.5. Each culture was divided into 50 ml aliquots. Cells were harvested by centrifugation at 4,000 × g for 10 minutes at room temperature. Pellets were resuspended in 25 ml of RNAlater (Ambion, Carlsbad, CA, USA). The tubes were agitated on a rotator at 4°C overnight, centrifuged at 4,000 × g for 10 minutes, placed in an ethanol/dry ice bath to flash freeze the pellet and stored at -80°C.
(MED4) cultures were a kind gift from Sallie Chisholm (MIT, Cambridge, MA, USA). Cells were grown in 8 L of PRO99 Sargasso sea water medium [33
] in a 24°C incubator with a simulated daily light/dark cycle, 13:11 light:dark cycle at 60 μmol Q m-2 s-1 [34
]. Cells were harvested at mid-log phase based on the fluorometric detection of bulk chlorophyll autofluorescence using a Turner Designs 10-AU Fluorometer. The culture was divided into 250 ml tubes and centrifuged at 15,000 × g for 10 minutes. After removal of the supernatant, RNAlater (50 ml) was added to each tube and the pellet resuspended. Half the cells in RNAlater (25 ml) were transferred into two 50 ml tubes and placed on a rotator overnight at 4°C. The tubes were centrifuged at 23,000 × g for 30 minutes, placed in an ethanol/dry ice bath to flash freeze the pellet and stored at -80°C.
RNA and DNA extraction
Bacterial cell pellets stored at -80°C in RNAlater (25 ml) were thawed on ice, resuspended and re-pelleted in 1 ml aliquots for 10 minutes at 4,000 × g in a microcentrifuge. The supernatant was removed and 200 μl bacterial lysis buffer (30 mM Tris·HCl, pH 8.0, 1 mM EDTA plus 15 mg/ml lysozyme (Sigma, St Louis, MO, USA) and 15 μl proteinase K (20mg/ml; QIAGEN, Valencia, CA, USA) were added to each tube. Samples were incubated at room temperature for 10 minutes, and vortexed for 10 s before and every 2 minutes during the incubation. QIAGEN RLT Plus buffer (750 μl) supplemented with 1 % v/v beta-mercaptoethanol (Sigma) was added to each tube and vortexed briefly to mix.
Two stool samples were collected 7 months apart from a single human donor (approved collection protocol by the Forsyth Institute Institutional Review Board, Assurance FWA00000398). Approximately 200 mg of stool were placed in approximately 2 ml RNAlater buffer, briefly mixed to disperse matter, and stored at room temperature during transport to the lab. The first sample (stool A) was extracted within 5 h of collection. The second sample (stool B) was stored at 4°C upon arrival for 24 h and then frozen and stored at -20°C for approximately 5 days until extracted. Prior to extraction, samples were vortexed briefly and centrifuged for 10 minutes at approximately 16,000 × g in a microcentrifuge at room temperature. Bacterial lysis buffer (100 μl) plus 10 μl proteinase K (20mg/ml) was added to half the stool sample (approximately 100 mg). Samples were incubated at room temperature for 10 minutes and vortexed for 10 s before and every 2 minutes during the incubation. QIAGEN RLT Plus buffer (1.2 ml) containing 1 % v/v beta-mercaptoethanol was added to each tube and vortexed briefly to mix. Samples were transferred into 2 ml sterile bead beating tubes (BioSpec Products Inc., Bartlesville, OK, USA) filled with 1 ml of 0. 1 mm glass beads (BioSpec Products), and placed in a bead beater (Mini Bead beater-8; BioSpec Products) for 3 minutes on 'homogenize' setting.
The lysed bacterial and stool samples were homogenized using QIAshredder spin columns (QIAGEN) and added to the AllPrep DNA spin columns for RNA and DNA isolation following the manufacturer's protocol. RNA integrity values (RIN values) [38
] were determined by running 1 μl aliquots on a Bioanalyzer 2100 (Agilent, Santa Clara, CA, USA). The RNA and DNA were stored in 5 to 20 μg aliquots at -80°C and -20°C, respectively.
Ribosomal RNA depletion methods performed at RNA level
An early access version of Meta-Bacteria Ribo-Zero rRNA removal kit (Epicentre, Madison, WI, USA) was used according to the manufacturer's instructions. RNA input amounts determined the amount of Ribo-Zero rRNA removal solution to add (10 µl rRNA removal solution for 2.5 to 5 µg or 8 µl for <2.5 µg total RNA per reaction). Samples in Ribo-Zero rRNA removal solution were incubated at 68°C for 10 minutes followed by a 15 minute incubation at room temperature. To remove the hybridized rRNA molecules from the mRNA, the RNA/rRNA solution reactions were incubated with the prepared microsphere beads, mixed well and placed at room temperature for 10 minutes (mixing every few minutes), then at 50°C for 10 minutes. The mRNAs were separated from the microspheres bound with rRNAs by a filter column provided in the kit. The final purification of eluted mRNA was performed using Agencourt RNAClean XP beads (2× the volume per mRNA volume; Beckman Coulter Genomics, Danvers, MA, USA).
MICROBExpress (Ambion/Applied Biosystems, Austin, TX, USA) was used according to the manufacturer's specifications. Briefly, total RNA (5 to 10 μg) was combined with binding buffer (200 μl) and capture oligonucleotide mix (4 μl). The RNA mix was heated to 70°C for 10 minutes then incubated at 37°C for 15 minutes to hybridize the capture oligos. The RNA/capture oligo mix was equilibrated with oligomag beads (50 μl, pre-warmed to 37°C) and incubated at 37°C for 15 minutes. Tubes were placed on a magnet to separate the supernatant containing the enriched mRNA from the oligomag beads. The enriched mRNA was purified and concentrated by ethanol precipitation according to the manual with precipitation at -80°C for 1 h.
mRNA-ONLY Prokaryotic mRNA Isolation Kit
Enzymatic reactions using the mRNA-ONLY Prokaryotic mRNA Isolation Kit (Epicentre) were performed according to the manufacturer's specifications. Briefly, total RNA (5 to 10 μg) was combined with 2 μl mRNA-ONLY 10× reaction buffer, 0.5 μl ScriptGuard RNase Inhibitor, 1 μl Terminator Exonuclease (1 U) and nuclease free water in a final volume of 20 μl and incubated at 30°C for 60 minutes. Reactions were terminated with the addition of 1 μl of mRNA-ONLY stop solution (100 mM EDTA). Agencourt RNAClean XP beads (0.8× of the reaction volume) were used to purify the reaction according to the manual.
The TURBO DNA-free kit (Ambion) was used for the DNase treatment. Total RNA or rRNA depleted RNA (following MICROBExpress, mRNA-ONLY, or Ribo-Zero treatment) was treated using a rigorous protocol that includes a second addition of DNase (2 to 4 units) and incubation at 37°C for 30 minutes according to the manufacturer's specifications. Reactions were terminated with the addition of the DNase inactivation reagent (0.2× the reaction volume) and purified using Agencourt RNAClean XP beads (0.8× of the reaction volume) according to the kit instructions. The presence of DNA contamination was assessed by PCR with 16S-specific primers. Each reaction included 1× AccuPrime PCR buffer II (10×), 0.75 U of AccuPrime Taq High Fidelity polymerase (5 U/μl; Invitrogen) and 200 nM of each primer (357F: 5'- CCTACGGGAGGCAGCAG -3' and 926R: 5'- CCGTCAATTCMTTTRAGT -3'). The DNase treated RNA (2 μl) was added to each reaction in a final reaction volume of 20 μl. Each reaction was run in parallel with a positive (E. coli DNA) and negative (nuclease free water) amplification control. The plates were sealed, centrifuged briefly, and placed in the thermal cycler (ABI 9700, Applied Biosystems, Foster City, CA, USA) with the following cycling conditions: 95°C for 2 minutes, 30 cycles of 95°C for 20 s, 50°C for 30 s, 72°C for 5 minutes. If an amplification product of approximately 600 bp was observed, the RNA was treated again with DNase.
RNA fragmentation reactions were performed using fragmentation buffer (5×; GeneChip Sample Cleanup Module; Affymetrix, Santa Clara, CA) in a final concentration of 1× per reaction. A maximum of 5 μg of RNA was added to each 10 μl fragmentation reaction, incubated at 80°C for 4 minutes on a thermal cycler, and placed on ice. Agencourt RNAClean XP beads (2.0× of the reaction volume) were used to purify the reactions according to the manual. An RNA fragment size distribution with a mode of approximately 300 bases was achieved with these conditions.
cDNA synthesis was performed as previously described [29
]. Total or rRNA depleted RNA was combined with 3 μg of the random primers (Invitrogen; 3 μg/μl) in a final volume of 11 μl. The reaction was incubated at 70°C for 10 minutes and placed immediately on ice. The remaining reagents were added to the reaction in a final volume of 20 μl: 1× of first strand buffer (10×; Invitrogen), 10 mM of DTT (0.1 M; Invitrogen), 0.5 mM of dNTP mix (10 mM; Invitrogen), 20 U of SUPERase-In (20 U/μl; Ambion) and 200 U of SuperScript III (200 U/μl; Invitrogen). The first strand reaction was incubated at 25°C for 10 minutes followed by 55°C for 60 minutes and then placed on ice. The second strand was synthesized by adding 1× of second strand buffer (5×; Invitrogen), 0.2 mM of dNTPs (10 mM; Invitrogen), 40 U of E. coli
DNA polymerase I (10 U/μl; NEB, Ipswich, MA, USA), 10 U of E. coli
DNA ligase (10 U/μl; NEB), 5 U of RNase H (5 U/μl; Invitrogen) to the first strand reaction (150 μl total volume). After 2 h at 16°C, the reaction was stopped by adding 10 μl of 0.5 M EDTA and purified using MinElute PCR Clean up columns (QIAGEN) according to the manufacturer's instructions.
Strand-specific cDNAs were made by dUTP marking and degradation of second strand cDNA [26
] using a modification of the protocol by Levin et al
].. Total or rRNA depleted RNA was combined with 3 μg of the random primers (Invitrogen; 3 μg/μl) in a final volume of 7 μl, incubated at 70°C for 10 minutes and immediately placed on ice. The remaining reagents were added to the first strand synthesis reaction for a total volume of 20 μl: 1× of first strand buffer (5×; Invitrogen), 10 mM of DTT (0.1 M; Invitrogen), 0.5 mM of dNTP mix (10 mM; Invitrogen), 4 μg of Actinomycin D (USB, Cleveland, OH, USA), 20 U of SUPERase-In (20 U/μl; Ambion) and 200 U of SuperScript III (200 U/μl; Invitrogen). The first strand reaction was incubated at 25°C for 10 minutes followed by 55°C for 60 minutes and then placed on ice. The first strand reaction was purified with Agencourt RNAClean XP beads (2.0× of the reaction volume) to remove the Actinomycin D and dNTPs. The second strand synthesis reaction included 1× first strand buffer (5×), 1 mM of DTT (0.1 M; Invitrogen), 260 nM dNTPs (10 mM deoxynucleotide mix containing dUTP instead of dTTP; Roche Applied Science, Indianapolis, IN, USA), 1× second strand buffer (5×), 40 U of E. coli
DNA polymerase I (10 U/μl; NEB), 10 U of E. coli
DNA ligase (10 U/μl; NEB) and 5 U of RNase H (5 U/μl; Invitrogen) in a final volume of 150 μl. The second strand reaction was incubated at 16°C for 2 h. The reaction was stopped by adding 10 μl of 0.5 M EDTA and purified using MinElute PCR clean up columns (QIAGEN) according to the manufacturer's instructions or Agencourt AMPure XP beads (2.0× of the reaction volume).
Selective cDNA synthesis: the Ovation Prokaryotic RNA-Seq System
The Ovation Prokaryotic RNA-Seq System (NuGEN Technologies, Inc., San Carlos, CA, USA) was used as follows. Intact RNA was DNase treated as described above and synthesized into cDNA according to the manufacturer's protocol. Briefly, the first strand primer was mixed with the intact RNA (500 ng/reaction), incubated at 65°C for 5 minutes, and placed on ice. First strand buffer and enzyme mix were added to each tube, mixed well, and incubated at 40°C for 30 minutes, 85°C for 5 minutes and a 4°C hold. Reaction Enhancement Enzyme mix was added to each tube and incubated at 37°C for 15 minutes with a 4°C hold. Second strand primer mix was added to the first strand reaction and incubated at 65°C for 5 minutes with a 4°C hold. The second strand master mix was added to each tube, mixed well and incubated at 25°C for 60 minutes with a 4°C hold. The cDNA was purified using a MinElute column (QIAGEN) and eluted in 1× low TE (10 mM Tris, 0.1 mM EDTA, pH 8.0). The cDNA was sheared using the Covaris S2 adaptive focused acoustics instrument (Covaris, Woburn, MA, USA) with the following conditions: duty cycle 5%, intensity 10, cycles/burst 200, time 6 minutes. The sheared products were purified and concentrated with Agencourt AMPure XP beads (2× the reaction volume).
Illumina sequencing libraries
Libraries for Illumina sequencing [39
] were made with NEB reagent kits and paired-end adapters using modified PCR amplification conditions to minimize base-composition bias [40
]. To simplify and streamline the process, especially for low input libraries, we transitioned to the 1 tube 'with bead' method [32
] in which all the steps (end repair, A-base addition and adaptor ligation ± indexing) were carried out in a single tube. Following adaptor ligation, the purified products were size selected on a gel (approximately 300 to 450 bp). cDNAs created with the second strand dUTP approach were treated with 1 U Uracil-Specific Excision Reagent enzyme mix (USER; NEB) at 37°C for 15 minutes followed by 95°C heat inactivation for 5 minutes. Samples were enriched with Illumina PE1.0 and PE2.0 primers (1 μM each), 1× of AccuPrime PCR buffer I (10×), 0.5 U of AccuPrime Taq High Fidelity polymerase (5 U/μl; Invitrogen) in a final volume of 25 μl. Enriched reactions were purified using Agencourt AMPure XP beads (0.8× the reaction volume).
Low C0t normalization of cDNA libraries using duplex-specific nuclease
The enrichment protocol following adaptor ligation was modified using 0.5 μM of each truncated paired end adaptor primer (SBS3_8 nt_F: 5'- TACACGACGCTCTTCCGATCT-3' and SBS8_7nt_R: 5' - CTGAACCGCTCTTCCGATCT-3'), 1× of AccuPrime PCR buffer I (10×), 0.5 U of AccuPrime Taq High Fidelity polymerase (5 U/μl; Invitrogen) in a final volume of 25 μl. Reactions were run on an ABI 9700 thermal cycler (Applied Biosystems) with the following cycling conditions: 98°C for 3 minutes, 20 cycles of 98°C for 30 s, 55°C for 30 s, 65°C for 1 minute with a final extension of 65°C for 10 minutes. Enriched reactions were purified using Agencourt AMPure XP beads (0.8× the reaction volume).
The hybridization reaction was prepared on ice in a 96-well plate with 100 ng enriched cDNA plus 1× hybridization buffer (50 mM HEPES, pH 7.3, USB; and 0.5 M NaCl, Ambion) in a final volume of 18 μl. The plate was incubated in a thermal cycler (ABI 9700) at 98°C for 10 minutes and 68°C for 4 h. A 68°C pre-heated mix of 2× DSN buffer (20 μl) and 2 μl DSN enzyme (Evrogen, Moscow, Russia) was added to each reaction (40 μl final volume) and incubated for another 25 minutes at 68°C. The reaction was stopped by adding 40 μl of the 2× DSN stop solution (10 mM EDTA) and purified with Agencourt AMPure XP beads (1.6× of the reaction volume). Samples were enriched with full-length Illumina PE1.0 and PE2.0 primers (1 μM each), 1× of AccuPrime PCR buffer I (10×), 0.5 U of AccuPrime Taq High Fidelity polymerase (5 U/μl; Invitrogen) in a final volume of 25 μl. Enriched reactions were purified using Agencourt AMPure XP beads (0.8× of the reaction volume) and sequenced.
Illumina sequencing and data analysis pipeline
Libraries were sequenced on either Illumina GAII or Hi-Seq instruments. Sequencing mode (single or paired end) and read lengths for each experiment are available in Additional file 6
. The raw reads of RNA-seq and DNA-seq data were processed using the Picard pipeline [41
]. Briefly, the reads were aligned and assigned to the reference genomes using the program BWA [42
], version 5.9, with parameters: -q 5 -l 32 -k 2 -t 4 -o 1. Sequence data for the PER mock community (Prochlorococcus marinus
str. CCMP1986 (MED4), Escherichia coli
, K12 substr
, Rhodobacter sphaeroides 2.4.1
) were aligned to the respective genome sequences. BWA-aligned reads were then analyzed and assigned to individual genes according to the genome annotations provided by GenBank (E. coli: NC_000913.gff
; P. marinus
; R. sphaeroides
). The normalized read counts for each gene, RPKM, was calculated by 1,000 × (The sum of reads/Gene length) × (106
/Total mappable reads).
DNA-seq reads were aligned to a reference set of 649 selected bacterial genomes (Additional file 2
). The initial list of the reference genomes and their annotations were downloaded from the Human Microbiome Project [6
], which included 1,700+ organisms from GenBank and the draft genomes sequenced by the Human Microbiome Project. To reduce the misalignment and crosstalk between similar genomes, the reference genomes were aligned using all against all pairwise whole genome alignments in MUMMER [43
] and clustered based on their MUMi values [44
]. One representative from each genome cluster, sharing at minimum a MUMi value of 0.3, was selected for the final reference set. To further reduce the size of the reference set of 4 gigabases, BWA's upper limit, we removed genomes that had not been previously observed in healthy human gut microbiomes. Genomes to be eliminated were determined empirically from examination of whole genome shotgun sequencing data from hundreds of Human Microbiome Project samples, representing various body sites from 100 healthy individuals. To reduce the possibility of spurious alignment, the BWA-aligned reads were post-filtered at a minimum sequence identity of 97% to the best aligned reference genome. Since the human gut microbiome is often dominated by a handful of organisms, we chose the top 19 most abundant bacterial organisms observed in the meta-genomic data that had sufficient sequence coverage and depth to analyze the consequences of rRNA depletion from the same sample. Accession numbers for the reference sequences of these 649 species are listed in Additional file 2
. Draft genomes lacking annotated rRNA genes were annotated in-house using the program RNAmmer [45
]. The RPKM value for each gene was calculated as described above.
To measure strand specificity, we calculated the normalized abundance values for ORFs (RPKMO) as described [46
]. Briefly, RPKMO values for regions corresponding to the sense and antisense strands of ORFs correspond to the number of reads aligning to these regions divided by the length of the region (in kb) and by the total number of reads aligning to the sense strand of all annotated ORFs in that sample (in millions). Annotations of protein-encoding genes were based on RefSeq NC_000913.gff, NC_005072.gff, NC_007493.gff, and NC_007494.gff for E. coli
, P. marinus
, and R. sphaeroides
chromosome 1 and 2, respectively.
The sequencing data have been submitted to the Sequence Read Archive, and accession numbers are listed in Additional File 6