Mycobacterium tuberculosis strains were isolated from lung tissue specimens of four patients, hospitalized with TB, diagnosed at the Beijing Ditan Hospital (Beijing, China), between April 2010 and October 2012. Lung tissue was obtained by thoracoscopic lung biopsy. TB was diagnosed based on the Chinese Pulmonary Tuberculosis Diagnostic Criteria (WS288-2008) and the Chinese TB Volume of Clinical Diagnosis and Treatment Guidelines (9).

The present study was approved by the Ethics Committee of Beijing Ditan Hospital, Capital Medical University (Beijing, China), according to the Declaration of Helsinki (10). Written informed consent was obtained from all of the patients.

Sputum/bronchoalveolar lavage fluid, pleural fluid and tissue samples were collected from the patients with TB for myco-bacterial assessment, by culturing the bacteria with either Lowenstein-Jensen (L-J) culture medium (Roche Molecular Systems, Inc., Branchburg, NJ, USA) or in a BACTEC 960 system (Roche Molecular Systems, Inc.). Blood and sputum samples, were initially cultured using the BACTEC 9120 Blood Culture system (Roche Molecular Systems, Inc.), and the positive samples were then cultured in L-J media at 37°C for 30 days. The Mycobacterium strains were identified using multi-locus polymerase chain reaction (PCR; Roche Molecular Systems, Inc.) (11,12). The total volume of the PCR reaction mixture was 25 µl (13 µl 2X PCR mix; 1 µl forward primer; 1 µl reverse primer; 1 µl DNA template; 9 µl DNase free water). The PCR amplification procedure was as follows: Pre-degeneration at 94°C for 5 min; 35 cycles of degeneration at 94°C for 1 min, annealing at 60°C for 1 min and extension at 72°C for 1 min; followed by a final extension step at 72°C for 10 min. The Mastercycler Nexus (Eppendorf, Hamburg, Germany) was used to conduct PCR. The sequences of the primers (synthesized by Invitrogen Life Technologies, Carlsbad, CA, USA) were as follows (lower case, plasmid; upper case, TB): Forward (F)1, tgtaaaacgacggccagtCGGATMACCGCTTTCGCCG, reverse (R)1, caggaaacagctatgaccGACATGTGTGAGCTGTTTGC; F2, tgtaaaacgacggccagtGAAGGCGGTATTCAAGC, R2, caggaaacagctatgaccGAGTCACCCTCCACAATGTA; F3, tgtaaaacgacggccagtGAAACCATTTCAACGGGTTC, R3, caggaaacagctatgaccCCATTGTAGCTGTACCAAGCACCC; F4, tgtaaaacgacggccagtTGGCCATAACGACATTCTG, R4, caggaaacagctatgaccGAGCACCAACGTGTTTAGC; F5, tgtaaaacgacggccagtACGGCTACGCAAAAGAAATG; R5, caggaaacagctatgaccTTGAGGCTGAGCCGATACTT; F6, tgtaaaacgacggccagtAGCAACCGGTAAAATTGTCG, R6, caggaaacagctatgaccCAGTGTAAGAACCGGCACAA; and F7, tgtaaaacgacggccagtTGTACGAAATTGCCACCAAA, and R7, caggaaacagctatgaccAATATTTTCGCCGCATCAAC. Drug sensitivity testing (DST) of the Mycobacterium tuberculosis strains was performed using the proportion method with four first-line anti-tuberculosis drugs: Isoniazid, rifampicin, streptomycin and ethambutol. Briefly, the tested bacteria liquid (20 mg/l) was prepared, and then 0.01 ml bacteria liquid was inoculated in L-J medium, with was 0.1 µg and 0.001 µg of each drug, respectively. After inoculation in 37°C culture medium for 4 weeks, the colony numbers were counted. If the percentage of colony numbers compared to control is ≤1% the strain is considered sensitive, and if it is >1% the strain is considered resistant.

RNA was extracted from the Mycobacterium tuberculosis using the RNA isolation reagent TRIzol^® (Invitrogen Life Technologies). Initially, 30 mg cultured Mycobacterium tuberculosis was added to a methanol/chloroform (Sigma-Aldrich, St. Louis, MO, USA) suspension (1:3). The suspension was then agitated with 5 ml TRIzol^®, centrifuged at 1,000 × g for 15 min at 4°C, and the colorless upper phase was collected. An equivalent quantity of isopropanol (0.5 ml) was mixed with this upper phase, centrifuged and the collected sediment at the bottom of the tube was mixed with ddH₂O. To prevent DNA contamination, the total RNA was treated with 20 µl RNase-free DNase II (Invitrogen Life Technologies).

Small RNA fractions, with a length ≤50 nt, were subjected to hybridization and ligation using Adaptor mix (Agilent Technologies, Santa Clara, CA, USA). Subsequently, the RNA samples were reverse transcribed and sequenced using miRNA sequencing on an Illumina HiSeq 2000 platform (Illumina, San Diego, CA, USA).

The total RNA was isolated from each sample using TRIzol^®, and the degradation and contamination of RNA was assessed using agarose gel (Abcam, Cambridge, MA, USA) electrophoresis. The RNA was purified using a KingFisher™ Pure RNA Tissue kit (Thermo Fisher Scientific, Inc., Waltham, MA, USA), and the ratio of the absorbance at 260 and 280 nm (A260/280) was determined to confirm purity. The RNA integrity was assessed using an RNA Nano 6000 Assay kit on a Bioanalyzer 2100 system (Agilent Technologies). A total of 3 g RNA per sample was used as input material for the RNA sample preparations, and all samples had RNA integrity number values >8. The samples of three individuals were then pooled within each group in equal quantities to generate two mixed samples. The pooled samples were then used to prepare six separate Illumina sequencing libraries, containing three technical replicates for each condition. cDNA libraries were generated using an Illumina TruSeq™ RNA Sample Preparation kit (Illumina), according to the manufacturer's instructions, and index codes were added to attribute sequences to each sample. Briefly, the mRNA was purified from the total RNA using poly-T oligo-attached magnetic beads (Thermo Fisher Scientific, Inc., Rockford, IL, USA). Fragmentation was performed using divalent cations under elevated temperature (55°C) in Illumina proprietary fragmentation buffer (Illumina). First strand cDNA was synthesized using random oligonucleotides and SuperScript II (Abcam). Second strand cDNA synthesis was subsequently performed using DNA Polymerase I and RNase H (Abcam). The remaining overhangs were converted into blunt ends via exonuclease/polymerase activities and the enzymes were removed. Following adenylation of the 3′ ends of the DNA fragments, the Illumina PE adapter oligo-nucleotides were ligated to prepare for hybridization. In order to select cDNA fragments of 200 bp in length, the library fragments were purified using an AMPure XP system (Beckman Coulter Genomics, Danvers, MA, USA). DNA fragments with ligated adaptor molecules on each end were selectively enriched using Illumina PCR Primer Cocktail (Illumina) in a 10 cycle PCR reaction. Primersequences were as follows: Forward: 5′-AATGATACGGCGACCACCGAGA-3′ and reverse: 5′-CAAGCAGAAGACGGCATACGAGT-3′). Products were purified using AMPure XP system (Beckman Coulter, Beverly, MA, USA) and quantified using the Agilent high sensitivity DNA assay on the Agilent Bioanalyzer 2100 system (Agilent Technologies). Prior to sequencing, all the individual libraries were normalized and pooled together in a single lane on an Illumina HiSeq 2000 platform, and 0–100 bp paired-end reads were generated.

To detect standard Mycobacterium tuberculosis strains in PubMed, particularly CDC1551 (http://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?id=83331), a library of reference sequences was prepared by identifying the sequences of mature miRNAs, together with five flanking nucleotides, within the hairpins deposited in miRBase version 19 (http://www.mirbase.org/) (13). The Illumina HiSeq platform requires a minimum read length of 50 nt, therefore, all small RNAs were extended using specific adapters (forward, 5′-AATGATACGGCGACCACCGAGA-3′; reverse, 5′-CAAGCAGAAGACGGCATACGAGT-3′), which were annealed to their 3′ ends during library preparation. Removal of the adapters was performed in silico on the raw, 50 nt Illumina sequence reads, using Cutadapt software (14). The resulting 15–30 nt long sequences were subjected to further analysis as potential miRNAs. The sequences were mapped on the prepared reference library using Bowtie version 0.12.7 (15), with the requirement of perfect matching. The numbers of mapped reads were subsequently calculated for each miRNA and provided as a number of each of the unique reads mapped to each reference sequence, and as a number of all the reads mapped to each reference sequence. Data obtained for each sample was normalized using reads per million (RPM) normalization, according to the following formula: RPM = (N_ref/N_all) × 10⁶. N_ref indicates the number of reads mapped to the miRNA reference, and N_all indicates the total number of reads mapped in the sample (16).

Selection of miRNAs and isomiRs deregulated between the analyzed groups was performed using a Welch t-test, paired for comparison between sensitive MTB and MDR MTB samples. False discovery rate was used to assess multiple testing errors. Statistical analyses were conducted using SPSS version 13.0 (SPSS Inc., Chicago, IL, USA) Hierarchical clustering of samples, based on the expression profiles of the selected miRNAs, was performed using Ward's agglomeration method, operated on Euclidean distance measures. Identification of target genes for each miRNA seed sequence, with significantly deregulated expression between the isoniazid-sensitive MTB and MDR MTB samples was performed using Target Rank version 3.2 software (16,17).

A total of 5, 785 and 195, and 6, 290 and 595 qualified Illumina reads were obtained from the two MDR MTB strains; and 6, 673 and 665, and 7, 210 and 217 qualified Illumina reads were obtained from the two sensitive MTB strains, respectively (Table I and Fig. 1). According to the statistical results, the average quality of >99% of the reads was >20 in each sample, indicating the quality of the sequencing was suitable (Table II and Fig. 2).

The adapter sequence (3′ adapter: AGATCGGAAGAGCACACGTCT) was filtered from the NGS raw data (Fig. 3) using the fastx_cliper program. The low quality reads were removed using the fastq_quality_filter program (http://seqanswers.com/forums/showthread.php?goto=nextoldest&t=24679), to ensure the quality score of at least 95% of the bases was >20. Clustering was then performed and identical base sequences were recorded as one tag. The length of the mature miRNA sequences was between 18 and 30 nt, which were further analyzed (Table III). The tags are shown in Fig. 4.

The tags were mapped to the genome sequence of the Mycobacterium tuberculosis strain CDC1551 in PubMed, using match software Bowtie version 0.12.7. The genome-mapped tags and reads, which were calculated based on 18–30 nt reads or tags, are shown in Table IV. The genome map rates of the samples are shown in Fig. 5. The RNA family database, Rfam (version 11.0; http://rfam.xfam.org/) was used to analyze the variety of RNAs in the samples. The Rfam reads and tags are shown in Tables V and VI.

Mirdeep (version 2; http://www.mdc-berlin.de/8551903/en/research/research_teams/systems_biology_of_gene_regulatory_elements/projects/miRDeep) was used to predict the miRNAs in the samples. The data are shown in Table VII. The target genes of miRNAs were predicted using Miranda software (http://www.miranda-im.org/; score >150; energy <−15).