Abstract

   Hydrogen peroxide (H[2]O[2]) is a natural oxidant produced by aerobic
   organisms and gives rise to oxidative damage, including DNA mutations,
   protein inactivation and lipid damage. The genus Mycobacterium utilizes
   redox sensors and H[2]O[2] scavenging enzymes for the detoxification of
   H[2]O[2]. To date, the precise response to oxidative stress has not
   been fully elucidated. Here, we compared the effects of different
   levels of H[2]O[2] on transcription in M. smegmatis using
   RNA-sequencing. A 0.2 mM H[2]O[2] treatment had little effect on the
   growth and viability of M. smegmatis whereas 7 mM H[2]O[2] was lethal.
   Analysis of global transcription showed that 0.2 mM H[2]O[2] induced
   relatively few changes in gene expression, whereas a large proportion
   of the mycobacterial genome was found to be differentially expressed
   after treatment with 7 mM H[2]O[2]. Genes differentially expressed
   following treatment with 0.2 mM H[2]O[2] included those coding for
   proteins involved in glycolysis-gluconeogenesis and fatty acid
   metabolism pathways, and expression of most genes encoding ribosomal
   proteins was lower following treatment with 7 mM H[2]O[2]. Our analysis
   shows that M. smegmatis utilizes the sigma factor MSMEG_5214 in
   response to 0.2 mM H[2]O[2], and the RpoE1 sigma factors MSMEG_0573 and
   MSMEG_0574 in response to 7 mM H[2]O[2]. In addition, different
   transcriptional regulators responded to different levels of H[2]O[2]:
   MSMEG_1919 was induced by 0.2 mM H[2]O[2], while high-level induction
   of DevR occurred in response to 7 mM H[2]O[2]. We detected the
   induction of different detoxifying enzymes, including genes encoding
   KatG, AhpD, TrxB and Trx, at different levels of H[2]O[2] and the
   detoxifying enzymes were expressed at different levels of H[2]O[2]. In
   conclusion, our study reveals the changes in transcription that are
   induced in response to different levels of H[2]O[2] in M. smegmatis.

Introduction

   The genus Mycobacterium includes pathogens, such as Mycobacterium
   tuberculosis and non-pathogenic microorganisms, such as Mycobacterium
   smegmatis. Mycobacteria are able to respond to and survive under
   different stresses [[31]1]. Hydrogen peroxide (H[2]O[2]) is a natural
   stressor that is produced by aerobic organisms and leads to oxidative
   damage, such as DNA mutations, protein inactivation and lipid damage
   [[32]2]. In addition, when M. tuberculosis, the pathogen which causes
   human tuberculosis (TB), infects a host, the production of H[2]O[2] is
   an important innate defense mechanism against infection. As a
   successful pathogen, M. tuberculosis has evolved redox sensors and
   H[2]O[2] scavenging enzymes for the detoxification of H[2]O[2] damage
   [[33]3,[34]4], but the precise response to H[2]O[2] has not been fully
   elucidated. A number of studies have shown that M. tuberculosis
   contains several regulators that respond to H[2]O[2] and several
   enzymes that detoxify H[2]O[2] damage [[35]5–[36]7]. A recent study has
   reported different transcriptional profiles in M. tuberculosis in
   response to different H[2]O[2] concentrations [[37]7]. However, the
   transcriptional response of M. smegmatis to different concentrations of
   H[2]O[2] has yet to be explored. A greater understanding of the
   differences between pathogenic M. tuberculosis and nonpathogenic M.
   smegmatis in their response to H[2]O[2] will help us to understand the
   pathogenesis of M. tuberculosis.

   Transcriptional regulation in response to H[2]O[2] in the Mycobacteria
   is complex compared to that in Bacillus or Escherichia coli. M.
   tuberculosis has 13 sigma factors, and M. smegmatis has 28 sigma
   factors [[38]8,[39]9], of which SigE, SigH, SigL and SigF play
   important roles in oxidative stress [[40]3]. As classical
   transcriptional regulators such as OxyR, FNR and FixL are absent in M.
   tuberculosis, alternative transcriptional regulators have been
   suggested to be involved in oxidative stress, including FurA [[41]10],
   IdeR [[42]11], CarD [[43]12], and the WhiB proteins [[44]3]. In
   addition to transcriptional regulators involved in the response to
   H[2]O[2], the signal transduction network including two-component
   systems, one-component systems, and serine/threonine kinases, is also
   involved in relaying and orchestrating the response to H[2]O[2.] M.
   tuberculosis encodes 11 serine/threonine kinases (STKs), of which PknB,
   PknF, and PknG have been shown to be involved in the oxidative stress
   response [[45]13–[46]15]. Park et al. showed that PknB phosphorylates
   both SigH and its anti-sigma factor RshA and causes its release from
   the complex of SigH and RshA. The phosphorylated SigH then regulates
   the response to oxidative stress [[47]14]. Similar to PknB, PknD was
   shown to phosphorylate anti-anti-sigma factor Rv0516c and then to
   activate Rv0516c, which changes the expression of the SigF regulon
   [[48]13]. Moreover, M. tuberculosis produces many enzymes that scavenge
   H[2]O[2]. Mycobacterial KatG is a multifunctional heme-dependent
   catalase-peroxidase-peroxynitritase [[49]16] and efficiently protects
   Mycobacterium from reactive oxygen species damage [[50]17]. KatG is the
   target of the first-line drug isoniazid (INH) and is responsible for
   the conversion of the prodrug INH into active INH [[51]18]. Clinical
   strains with decreasing KatG activity showed higher levels of AhpC
   [[52]19], suggesting that AhpC contributes to defense against oxidative
   stress. The metabolic enzyme complex with Lpd, SucB, AhpC, and AhpD, is
   also involved in antioxidant defense [[53]20]. Thiol-dependent
   peroxidase Tpx is an antioxidant protein against oxidative stress
   [[54]21]. Voskuil et al. recently investigated whole genome expression
   in response to different levels of oxidative stress and showed that
   many genes related to oxidative stress are induced concurrently with
   the dormancy regulon at high concentrations of H[2]O[2] [[55]7].

   In this study, we compared the effects of different H[2]O[2] levels on
   transcription in M. smegmatis using RNA-sequencing. We show that low
   levels (0.2 mM) of H[2]O[2] have little effect on the growth and
   viability of M. smegmatis whereas high levels (7 mM) of H[2]O[2] are
   bactericidal. Relatively few changes in gene expression were observed
   on exposure to 0.2 mM H[2]O[2] while a large munber of differentially
   expressed genes were induced after treatment with 7 mM H[2]O[2]. Some
   differentially expressed genes involved in the
   glycolysis-gluconeogenesis and fatty acid metabolic pathways were
   induced by 0.2 mM H[2]O[2], and the expression of genes encoding
   ribosomal proteins was lower after treatment with 7 mM H[2]O[2.] Our
   analysis also identified differences in the sigma factors,
   transcriptional regulators, and detoxifying enzymes that are expressed
   in response to treatment with 0.2 mM and 7 mM H[2]O[2].

Materials and Methods

Bacterial strains and culture conditions

   Liquid cultures of the M. smegmatis mc^2155 strain were grown in
   Middlebrook 7H9 medium (Becton Dickinson) supplemented with 0.2% (v/v)
   glycerol (Beijing Modern Eastern Finechemical), 0.05% Tween 80 (v/v)
   (Sigma) and 10% ADS (albumin, dextrose, and saline). Middlebrook 7H10
   medium (Becton Dickinson) supplemented with 10% ADS and 0.2% (v/v)
   glycerol was used as the solid medium for M. smegmatis growth.

Response of the M. smegmatis mc^2155 strain to H[2]O[2] stress

   Log phase cultures (OD[600] of 0.8–1.0) of M. smegmatis mc^2155 were
   diluted 1:100 into 7H9 media and cultured for approximately 12 hours
   until the OD[600] reached 0.3. Re-inoculated cells were then treated
   with the indicated concentrations of H[2]O[2] (0, 0.2 and 7 mM) for
   periods of 30 min or 3 h, and surviving cells were grown on 7H10 media.
   Cells were collected after 30 min of 0.2 mM or 7 mM H[2]O[2] treatment,
   and total RNA was isolated from each sample and compared by
   RNA-sequencing to RNA from untreated cells that were prepared
   simultaneously.

RNA isolation for RNA-sequencing

   Fifty milliliters of bacterial culture (OD[600] of ~ 0.3) was collected
   and total RNA was isolated using FastPrep Purification kits (MP Bio)
   according to the manufacturer’s instructions. Construction and
   sequencing of the cDNA libraries of the various mycobacterial strains
   was performed by BGI-Shenzhen (China). Briefly, total RNA from treated
   mc^2155 strains was treated using a Ribominus Transcriptome Isolation
   Kit (Thermo Fisher Scientific) to remove rRNA contaminations. NEXTflex
   RNA Fragmentation Buffer (Bioo Scientific) was added to separate the
   mRNA into short fragments. Using these short fragments as templates,
   random hexamer-primers were used to synthesize the first strand of
   cDNA. The second strand of cDNA was synthesized using buffer, dNTPs,
   RNase H and DNA polymerase I. The short fragments were purified with a
   QIAQuick PCR extraction kit (QIAGEN) and resolved with EB buffer for
   end reparation and addition of poly(A). The short fragments were
   subsequently connected with sequencing adaptors. For amplification by
   PCR, we selected suitable fragments as templates, based on results of
   agarose gel electrophoresis. The library was then sequenced using
   Illumina HiSeq 2000. Clean reads were mapped to the reference genome
   and the gene sequences using SOAP2 [[56]22]. The RNA-sequencing dataset
   obtained has been submitted to ArrayExpress under the accession number
   E-MTAB-3594.

RNA-sequencing data analysis

   The raw data were filtered to 1) remove reads with adaptors, 2) remove
   reads with more than 10% of unknown nucleotides, 3) remove low quality
   reads (in which more than half of the base quality scores were less
   than 5). The resulting cleaned paired-end reads were mapped to the M.
   smegmatis mc^2155 reference genome (NC 008596.1) using SOAP2.
   Mismatches of no more than 5 bases were allowed in the alignment. We
   performed statistical analysis in read alignments on the genome and
   genes for each sample. Randomness of the mRNA/cDNA fragmentation was
   evaluated using the reads distribution of reference genes.

   Gene expression was calculated according to the RPKM method (reads per
   kilobase of exon model per million mapped reads) [[57]23], and
   eliminated the biases influence of different gene length and sequencing
   difference using the TPM method [[58]24]. We calculated the ratio of
   each gene between samples and identified genes differentially expressed
   between two samples using "the significance of digital gene expression
   profiles" [[59]25] based on the criteria FDR ≤ 0.001 and a fold change
   larger than 4. STRING (9.1) [[60]26] was used to analyze the
   interactions of differentially expressed genes and functional and
   pathway enrichment analysis. A p-value less than 0.05 was used as a
   threshold to indicate significant enrichment.

Quantitative PCR of selected genes

   Log phase cultures (OD[600] = 0.8–1.0) of all the tested strains were
   diluted 1:100 in 7H9 media. The strains were cultured until the OD[600]
   reached 0.3 and then divided into control and treatment groups. In the
   treatment group, the cells were treated with 0.2 or 7 mM H[2]O[2] for
   30 min and then collected by centrifugation at 12,000 x g. Bacterial
   pellets were resuspended in TRIzol (Invitrogen, USA), and RNA was
   purified according to the manufacturer’s instructions. cDNA was
   synthesized using the SuperScript III First-Strand Synthesis System
   (Invitrogen, USA). Quantitative real-time PCR (qRT-PCR) was performed
   in a Bio-Rad iCycler using a 2x SYBR real-time PCR pre-mix (Takara
   Biotechnology Inc., Japan). The following cycling program was used:
   95°C for 1.5 min followed by 40 cycles of 95°C for 10 s, 60°C for 15 s,
   and 72°C for 15 s, followed by 72°C for 6 min. The M. smegmatis rpoD
   gene encoding the RNA polymerase sigma factor SigA was selected as a
   reference gene for normalizing gene expression. The 2^-ΔΔCT method was
   used [[61]27] to evaluate relative gene expression in the different
   strains and/or different treatments. All primers used are listed in
   [62]S1 Table.

Statistical analysis

   All statistical analyses were performed using GraphPad Prism 5.0c.
   Significant differences in the data were determined using t-tests.

Results and Discussion

Effects of H[2]O[2] on growth and viability

   Hydrogen peroxide (H[2]O[2]) is a natural oxidant produced by aerobic
   organisms and can lead to oxidative damage, such as DNA mutations,
   protein inactivation and drug resistance [[63]2]. In addition,
   increasing levels of toxic H[2]O[2] in the infected host is an
   important defensive mechanism against invading pathogens. Resistance to
   H[2]O[2] might increase bacterial survival in mycobacterial-infected
   macrophages. A previous study from our lab showed that increased
   resistance to H[2]O[2] in a mutant strain of M. smegmatis could lead to
   higher survival in infected macrophages [[64]28]. M. tuberculosis can
   persist in macrophages for decades, partly because it possesses many
   regulators that respond to H[2]O[2] and many enzymes that detoxify
   H[2]O[2] [[65]3,[66]7]. A recent study analyzed genome-wide changes in
   gene expression in response to different levels of oxidative and
   nitrosative stresses in M. tuberculosis [[67]7]. Responses to oxidative
   and nitrosative stresses were compared and their results revealed a
   common genetic response used by M. tuberculosis in response to these
   stresses. This study demonstrated that analyzing global transcription
   levels can help us understand the molecular mechanisms underlying the
   response of bacteria to H[2]O[2]. Here, we examined the global
   transcriptional response of M. smegmatis to different levels of
   H[2]O[2]using RNA-sequencing.

   Bacteria are most sensitive to environmental stresses at the early
   logarithmic phase [[68]29]. We therefore chose to treat the M.
   smegmatis strain mc^2155 with H[2]O[2] when bacteria reached the early
   logarithmic phase of growth (optical density, OD[600] of ~ 0.3). We
   have previously reported that, under experimental conditions in our
   laboratory, the MIC to H[2]O[2] in M. smegmatis is 0.039 mM [[69]28]
   and that of M. tuberculosis is 1mM. The ratio of the MIC to H[2]O[2] of
   M. smegmatis to M. tuberculosis is ~26. In order to compare the
   response of M. smegmatis and M. tuberculosis to H[2]O[2], we used
   concentrations of H[2]O[2] comparable to those used in Voskuil et al
   [[70]7]. Here we used an H[2]O[2] concentration of 7mM for M. smegmatis
   to correspond to the 200 mM H[2]O[2] treatment used by Voskuil et al.
   in M. tuberculosis. Similarly, the 0.2 mM H[2]O[2] treatment used here
   corresponded to the ~ 5 mM H[2]O[2] treatment used by Voskuil et al in
   M. tuberculosis. When the M. smegmatis mc^2155 strain reached the early
   logarithmic phase, bacterial cultures were exposed to two different
   H[2]O[2] concentrations, namely 0.2 mM or 7 mM for either 30 min or 3
   h, and cultures were then collected and spotted onto 7H10 media. As
   shown in [71]Fig 1, no growth differences were detected between the
   bacteria treated with 0.2 mM for 30 min or 3 h and untreated control
   bacteria, indicating that exposure to 0.2 mM H[2]O[2] had little effect
   on bacterial growth and viability. In contrast, when cultures were
   treated with 7 mM H[2]O[2], we observed that exposure to 7 mM H[2]O[2]
   resulted in a significant decrease in cell survival ([72]Fig 1),
   indicating that 7 mM H[2]O[2] has a bactericidal effect in M.
   smegmatis. Moreover, the genes induced in response to H[2]O[2] (5–10
   mM) in the Voskuil study were also found to be induced in M.
   tuberculosis during infection of activated macrophages[[73]30],
   indicating that a level of 5–10 mM H[2]O[2] is similar to that
   experienced by bacteria within infected macrophage. Here, the response
   of 7 mM of H[2]O[2] used to examine the M. smegmatis response might
   suggest the response of bacteria within infected macrophage.

Fig 1. The effect of H[2]O[2] stress on the survival of M. smegmatis.

   Fig 1
   [74]Open in a new tab

   The panel represents serial dilutions (1:10) of mc^2155 cultures
   treated with 0.2 mM or 7 mM H[2]O[2] for either 30 min or 3 hour. Three
   microliters of diluted M. smegmatis cultures were spotted onto solid
   7H10 medium. Images shown are representative of at least 3 experiments.

   The following experiments were performed to compare the transcriptional
   response of M. smegmatis exposed to 0.2 mM H[2]O[2] or 7 mM H[2]O[2]
   for 30 min when bacterial growth had reached an OD[600] of 0.3.

Expression profiles of M. smegmatis in response to different levels of
H[2]O[2]

   Transcriptional reprogramming is a critical step in bacterial responses
   to various stress factors to ensure their survival. We therefore
   examined changes in mRNA expression following treatment with H[2]O[2]
   using RNA-sequencing. mRNA samples from of M. smegmatis mc^2155 with or
   without H[2]O[2] treatment were prepared as described in the “Materials
   and Methods”.

   RNA-sequencing mapping statistics showed that approximately 96% of the
   sequencing reads could be mapped to the M. smegmatis reference genome
   ([75]NC_008596.1) ([76]Table 1). The percentage of unique mapped reads
   for untreated mc^2155, mc^2155 treated with 0.2 mM H[2]O[2] and mc^2155
   treated with 7 mM H[2]O[2] were 92.34%, 98.28% and 98.17%, respectively
   and the number of reads mapped was 6089174, 6390446 and 6548357,
   respectively.

Table 1. RNA-sequencing mapping statistics.

   Sample name mc^2155 mc^2155 / 0.2 mM mc^2155/ 7 mM
   Reads number Percentage Reads number Percentage Reads number Percentage
   Total reads 6594442 100.00% 6502552 100.00% 6670090 100.00%
   Total base pairs 593499780 100.00% 585229680 100.00% 600308100 100.00%
   Total Mapped Reads 6089174 92.34% 6390446 98.28% 6548357 98.17%
   Perfect match 4382586 66.46% 5499445 84.57% 5525430 82.84%
   ≤ 5bp mismatch 1706588 25.88% 891001 13.70% 1022927 15.34%
   Unique match 5818976 88.24% 6065452 93.28% 6171788 92.53%
   Multi-position match 270198 4.10% 324994 5.00% 376569 5.65%
   Total unmapped reads 505268 7.66% 112106 1.72% 121733 1.83%
   [77]Open in a new tab

   Fully annotated data are presented in [78]S2–[79]S4 Tables. Genes were
   considered to be significantly differentially expressed if their
   changes in expression were > 4-fold greater compared to the non-treated
   wild type mc^2155 strain, with a false discovery rate (FDR) corrected
   P-value of < 0.01. To confirm the results obtained from the
   RNA-sequencing analysis ([80]Fig 2B), several induced genes were
   examined by quantitative RT-PCR (qRT-PCR). In three independent
   experiments, total RNA was isolated from M. smegmatis exposed to 0.2 mM
   or 7 mM H[2]O[2] for 30 min and relative levels of expression were
   analyzed by qRT-PCR. Results were consistent with those obtained from
   RNA-sequencing results, confirming the validity of our approach. For
   example, msmeg_0574 exhibited a 21.17 ± 1.11-fold enhancement when
   induced by 7 mM H[2]O[2], but little enhancement (1.61 ± 0.31-fold)
   when induced by 0.2 mM H[2]O[2] ([81]Fig 2A), consistent with
   RNA-sequencing results ([82]Fig 2B). In addition, msmeg_3242 exhibited
   a 3116.9 ± 182.8-fold enhancement when induced by 7 mM H[2]O[2] and a
   7.43 ± 0.11-fold enhancement when induced by 0.2 mM H[2]O[2] ([83]Fig
   2A)[.] The relative expression levels of the genes we chose to test
   under 0.2 mM and 7 mM H[2]O[2] were also consistent with RNA-sequencing
   results ([84]Fig 2). In summary, these results support the fidelity of
   the RNA-sequencing results for the analysis below.

Fig 2. Quantitative RT-PCR validation of RNA-sequencing results.

   Fig 2
   [85]Open in a new tab

   (A) Quantitative RT-PCR analysis of the mRNA expression of genes
   differentially expressed after treatment with different levels of
   H[2]O[2]. M. smegmatis cultures were treated with 2 mM or 7 mM H[2]O[2]
   for 30 min before extraction of RNA for qRT-PCR. The data represent 3
   independent experiments. (B) Fold changes of selected genes
   differentially expressed genes after treatment with 0.2 mM and 7 mM
   H[2]O[2] obtained by the RNA-sequencing.

   Upon exposure to 0.2 mM H[2]O[2,] there were 303 up-regulated genes and
   331 down-regulated genes. Genes differentially expressed in the mc^2155
   strain on treatment with 0.2 mM H[2]O[2] were significantly enriched
   for several GO biological processes, including response to stress (p =
   1.44 x 10^−10), DNA repair (p = 2.87 x 10^−5), and ergothioneine
   biosynthesis (p = 1.18 x 10^−2) when compared to the untreated mc^2155
   strain ([86]Fig 3A). In GO molecular function categories, we found that
   genes differentially expressed after treatment with 0.2 mM H[2]O[2]
   were significantly enriched for nuclease activity (p = 2.49 x10^-4),
   helicase activity (p = 4.55 x 10^−2), and sulfur compound transmembrane
   transporter activity (p = 4.55 x 10^−2) when compared to the untreated
   mc^2155 strain ([87]Fig 3A). As H[2]O[2] causes DNA damage, genes
   involved in DNA repair (listed in [88]Table 2) were induced upon
   exposure to 0.2 mM H[2]O[2]. Induction of RecA, AlkA, and DNA helicase
   by H[2]O[2] was also found in the M. tuberculosis study [[89]7]. M.
   tuberculosis RecA is involved in nucleotide excision, recombination and
   the SOS response [[90]31]. In M. smegmatis, RecA is induced by DNA
   damage and is a key regulator element of the SOS response [[91]32]. In
   M. tuberculosis, dnaE2, which encodes an error–prone DNA polymerase,
   was shown to increase its expression in response to DNA damaging
   agents, suggesting that its role is involved in damage tolerance
   [[92]33,[93]34]. mRNA levels of M. smegmatis dnaE2 and recA were
   increased 46-fold and 12-fold, respectively, by 0.2 mM H[2]O[2], and
   5-fold and 7.8-fold, respectively, by 7 mM H[2]O[2] ([94]Table 2). The
   response profiles with high inductions of DNA repair genes in M.
   smegmatis by both low (0.2 mM) and high (7 mM) levels of H[2]O[2] were
   strikingly different to those in M. tuberculosis which showed high
   induction from mild levels of H[2]O[2] and no change in induction with
   bactericidal H[2]O[2] levels [[95]7][.] Future work should compare and
   investigate differences between M. smegmatis and M. tuberculosis in
   DNA-damage-mediated death caused by H[2]O[2] in order to provide
   greater insights into the pathogenicity of M. tuberculosis. The STRING
   database was used to establish protein interaction networks of physical
   and functional interactions among the differentially expressed genes
   identified ([96]Fig 3B). Interestingly, using the KEGG-User Data
   Mapping [[97]35] ([98]Fig 4B), seven genes involved in fatty acid
   metabolism (RM018 and RM020) were found and formed an interconnected
   cluster ([99]Table 2 and [100]Fig 4A). In addition, nine genes involved
   in glycolysis/gluconeogenesis (msm00010) were found in a partially
   interconnected cluster ([101]Table 2). Transcription of these genes was
   induced, suggesting that these differentially expressed genes are
   involved in the central carbon metabolism (CCM) switch and providing
   supporting evidence for a previous suggestion that the CCM of M.
   tuberculosis plays important roles in growth and pathogenicity
   [[102]36,[103]37]. An extensive transcriptional switch in M.
   tuberculosis CCM genes during host infection has been reported,
   indicating that there is a quick change in the metabolic pathway in
   response to various stresses [[104]38]. The gene pdhB involved in
   glycolysis/gluconeogenesis was repressed 9.7-fold by 0.2 mM H[2]O[2],
   and pdhA was repressed 11-fold. PdhA and PdhB are constituents of the
   mycobacterial pyruvate dehydrogenase complex which connects glycolysis
   and the tricarboxylic acid (TCA) cycle [[105]39]. In contrast to cells
   treated with 7 mM H[2]O[2], changes induced in pdhA and pdhB in the 0.2
   mM treatment were weak ([106]Table 2). Similarly, the seven
   differentially expressed genes induced by 0.2 mM were involved in fatty
   acid metabolism and decreased by 4 to 6-fold ([107]Table 2), whereas a
   decrease induction did not appear after the 7 mM H[2]O[2] treatment.
   Together, these results suggest that the metabolic switch of
   glycolysis/gluconeogenesis and fatty acid metabolism was specific to
   induction by 0.2 mM H[2]O[2].

Fig 3. Overview of the differential expression profiles in response to 0.2 mM
H[2]O[2] in M. smegmatis.

   [108]Fig 3
   [109]Open in a new tab

   (A) Enrichment analysis. The differently colored bars indicate the gene
   number for the enrichment of the annotations. (B) Interaction network
   of the differentially expressed genes of M. smegmatis induced by 0.2 mM
   H[2]O[2] using STRING (9.1) at confidence scores ≥ 0.4. The network is
   enriched among the 634 differentially expressed genes and 111
   interactions were observed (p value = 0).

Table 2. Fold changes of genes differentially expressed after treatment with
0.2 mM and 7 mM H[2]O[2] (treated vs untreated).

   Gene Name Gene Product log2 Ratio (0.2mM/mc^2155) P-value FDR log2
   Ratio (7mM/mc^2155) P-value FDR
   DNA repair
   msmeg_2839 transcriptional accessory protein -1.7523338 2.02E-243
   5.7E-242 -1.6975845 9.0323E-80 5.4177E-79
   alkA methylated-DNA—protein-cysteine methyltransferase -2.106759
   8.54E-09 2.73E-08 -1.1090648 0.00020253 0.00030498
   msmeg_1383 endonuclease IV -2.8849149 0 0 -0.4418784 0 0
   msmeg_1756 endonuclease VIII and DNA N-glycosylase with an AP lyase
   activity 3.01822392 2.08E-105 3.04E-104 2.96577039 9.82E-104 7.22E-103
   dnaE2 error-prone DNA polymerase 5.52502439 0 0 2.44964676 1.32E-277
   2.48E-276
   lig ATP-dependent DNA ligase 2.37386688 1.03E-184 2.21E-183 1.60677746
   5.66E-69 3E-68
   ligA NAD-dependent DNA ligase LigA 2.32734432 0 0 2.47575753 0 0
   recA recombinase A 3.59091497 0 0 2.96402929 0 0
   uvrB excinuclease ABC subunit B 3.25788081 0 0 4.50884994 0 0
   lexA LexA repressor 3.10981267 0 0 1.01678782 4.81E-103 3.51E-102
   radA DNA repair protein RadA 2.82617738 0 0 1.71244303 0 0
   dinP DNA polymerase IV 4.26301055 1.85E-174 3.84E-173 4.18732816
   2.79E-169 3.17E-168
   msmeg_1622 putative DNA repair polymerase 4.79317887 8.44E-277
   2.71E-275 -0.1339395 0.658812 0.69126208
   recD exodeoxyribonuclease V, alpha subunit 3.50057135 3.24E-216
   8.07E-215 2.03125644 2.04E-46 8.42E-46
   msmeg_1756 endonuclease VIII and DNA N-glycosylase with an AP lyase
   activity 3.01822392 2.08E-105 3.04E-104 2.96577039 9.82E-104 7.22E-103
   helicase ATP-dependent DNA helicase 4.16679804 0 0 4.74283172 0 0
   tag DNA-3-methyladenine glycosylase I 2.81384991 2.09E-184 4.47E-183
   2.86173411 2.34E-201 3.15E-200
   Glycolysis/gluconeogenesis
   msmeg_1543 eptc-inducible aldehyde dehydrogenase -2.6859775 8.91E-139
   4.21E-148 -1.2892582 5.64E-56 2.60E-55
   msmeg_1762 piperideine-6-carboxylic acid dehydrogenase 2.93235592
   2.30E-149 4.21E-148 -1.6876011 2.28E-12 4.83E-12
   pfkB 6-phosphofructokinase isozyme 2 -5.7653785 3.88E-196 8.90E-195
   -2.8501465 6.61E-126 5.73E-125
   pdhB pyruvate dehydrogenase E1 component subunit beta -3.2722769
   3.01E-30 1.92E-29 0.71595693 5.59E-07 9.61E-07
   pdhA pyruvate dehydrogenase E1 component, alpha subunit -3.4705711
   1.84E-89 2.39E-88 -0.3895867 0.00010596 0.00016157
   adhB alcohol dehydrogenase B -4.0478653 2.97E-105 4.33E-104 -2.1612025
   4.92E-60 2.38E-59
   msmeg_6616 S-(hydroxymethyl)glutathione dehydrogenase -2.956082
   0.00E+00 0.00E+00 -1.1422205 1.36E-159 1.45E-158
   msmeg_6687 aldehyde dehydrogenase, thermostable -3.706221 1.64E-29
   1.03E-28 -1.5049935 9.36E-12 1.94E-11
   msmeg_6834 alcohol dehydrogenase -2.7848309 2.21E-05 5.51E-05
   0.08733237 0.840712 0.85579383
   fadA acetyl-CoA acetyltransferase -2.7273404 0.00E+00 0.00E+00
   0.33348045 5.36E-35 1.86E-34
   fadB putative 3-hydroxyacyl-CoA dehydrogenase -2.6949142 0.00E+00
   0.00E+00 1.39743214 0 0
   msmeg_5199 putative acyl-CoA dehydrogenase -2.6174798 7.78E-91 1.02E-89
   -0.4351779 5.88E-07 1.01E-06
   msmeg_4832 acyl-CoA dehydrogenase -2.2726247 1.84E-34 1.30E-33
   -1.4448556 2.45E-20 6.42E-20
   msmeg_4833 putative acyl-CoA dehydrogenase -2.176661 8.50E-49 7.39E-48
   -2.3488919 3.50E-57 1.65E-56
   echA4 enoyl-CoA hydratase -2.5812975 1.27E-32 8.59E-32 -0.1876747
   0.1805012 0.20801109
   fadE13 putative acyl-CoA dehydrogenase -2.4369076 3.84E-11 1.38E-10
   -0.1310912 0.579156 0.6139111
   Sigma factors
   sigG RNA polymerase factor sigma-70 2.93398738 0 0 3.88123043 0 0
   msmeg_0573 putative ECG sigma factor RpoE1 1.38509413 0.0569946
   0.08901435 7.09387521 1.35E-209 1.91E-208
   msmeg_0574 putative ECG sigma factor RpoE1 0.18840222 0.422278
   0.5067336 4.08443636 3.1E-276 5.78E-275
   msmeg_1348 RNA polymerase ECF-subfamily protein sigma factor -0.0465176
   0.804554 0.84797939 -2.3586787 1E-17 2.48E-17
   sigL RNA polymerase sigma factor SigL 0.52825986 0.00083713 0.00176811
   -2.2355973 4.27E-20 1.12E-19
   msmeg_1970 sigma factor -1.6126499 4.2715E-20 1.1152E-19 -4.1999183
   4.27E-20 1.12E-19
   mysB RNA polymerase sigma factor SigB -0.4921846 4.74E-104 6.85E-103
   2.50111799 0 0
   msmeg_3008 putative sigma 54 type regulator -1.8964762 1.61E-11
   5.87E-11 -2.0197975 3.44E-13 7.46E-13
   msmeg_5214 RNA polymerase sigma-70 factor 4.65312922 2.09E-184
   4.47E-183 2.81376729 1.63E-13 3.59E-13
   msmeg_5444 RNA polymerase sigma-70 factor protein 0.03717083 0.894318
   0.91679688 -2.0919913 9.72E-09 1.8E-08
   Transcriptional Regulators
   devR two-component system response regulator 1.01514452 0.00136283
   0.00280756 6.64431111 0 0
   furA (msmeg_6383) transcription regulator FurA 2.0393827 1.57E-38
   1.19E-37 4.16725591 0 0
   phoP DNA-binding response regulator PhoP 2.90191537 4.67E-220 1.19E-218
   -0.5325639 0.00089158 0.00128064
   msmeg_4517 TetR-type transcriptional regulator of sulfur metabolism
   3.21516913 0.00000169 0.00000457 4.13040109 7.31E-14 1.62E-13
   msmeg_4925 transcriptional regulator 0.69321643 0.00330216 0.00644626
   2.93688979 1.02E-78 6.03E-78
   msmeg_1919 Transcription factor WhiB 2.52297686 0 0 1.07820136
   2.53E-126 2.21E-125
   msmeg_4025 transcriptional regulator, LysR family protein 1.30263197
   0.00055968 0.00120459 5.38130323 1.6E-210 2.28E-209
   msmeg_6253 fur family protein transcriptional regulator 1.24070422
   0.0469844 0.07479434 3.90800867 1.22E-25 3.59E-25
   Detoxification enzymes
   trxB thioredoxin-disulfide reductase -0.0780198 0.1478068 0.20755437
   2.71109242 0 0
   trx thioredoxin 0.12205973 0.0602136 0.09347541 3.84889994 0 0
   msmeg_6884 NADP oxidoreductase 0.87052096 0.1166366 0.1680318
   4.68101209 4.52E-68 2.38E-67
   KatG catalase/peroxidase HPI 3.82587233 0 0 4.95397834 0 0
   msmeg_4890 alkylhydroperoxidase 3.42735043 6.73E-19 3.21E-18 4.22476909
   0.0465926 0.05777446
   msmeg_3448 two-component system sensor kinase 0.55015338 0.1588256
   0.22100457 5.07235098 4.09E-202 5.55E-201
   ahpD alkylhydroperoxidase AhpD core -2.3697934 0.0210456 0.03618406
   3.85029317 4.71E-31 1.53E-30
   msmeg_3708 catalase 2.47422002 5.45E-56 5.11E-55 -3.0001305 5.73E-15
   1.31E-14
   [110]Open in a new tab

Fig 4. Connected network of the enriched differentially expressed genes
following exposure to 0.2 mM H[2]O[2].

   Fig 4
   [111]Open in a new tab

   (A) Connected network of enriched differentially expressed genes
   involved in fatty acid metabolism (RM018 and RM020). (B) Partial fatty
   acid metabolism in M. smegmatis. Genes expressed differentially after
   0.2 mM H[2]O[2] treatment assigned to RM018 and RM020 are marked in
   red.

   Compared to the down-regulation of 331 genes under the 0.2 mM H[2]O[2]
   treatment, 1671 genes were down-regulated under the 7 mM H[2]O[2]
   treatment and 343 genes were up-regulated. In contrast to the small
   proportion of genes in the genome that responded to the 0.2 mM H[2]O[2]
   treatment (663 differentially-expressed genes, ~10% of the genes in the
   M. smegmatis genome), 2002 genes were induced in response to 7 mM
   H[2]O[2], representing 29.3% of the genes in the M. smegmatis genome.
   In contrast to the interaction networks obtained among genes which
   showed differential expression at the mRNA level in response to the
   0.2mM H[2]O[2] treatment, we did not find enrichment in specific
   metabolic pathways among genes that were differentially expressed in
   response to the 7 mM H[2]O[2] treatment. This might be due to the fact
   that the 7 mM H[2]O[2] treatment had more global effects on
   transcription, which were not limited to specific metabolic pathways.
   We also conducted an analysis of GO biological processes and identified
   enrichment in processes including gene expression (p = 2.47 x 10^−1),
   macromolecule biosynthetic processes (p = 2.47 x 10^−1) and regulation
   of gene expression (p = 2.84 x 10^−1) ([112]S1 Fig). Ribosome
   biogenesis was also enriched, though the P-value was 5.48 x 10^−1,
   slightly higher than the cutoff value (P < 0.5). Oxidative stress
   results in the rapid inhibition of protein synthesis as well as in the
   reprogramming of gene expression, resulting in growth reduction as an
   adaption to oxidative stress [[113]40].

Responses of specific genes to exposure to low and high levels of H[2]O[2]

   Mycobacterium utilizes diverse antioxidant systems to combat H[2]O[2]
   stress, including sigma factors, transcriptional regulators, STKs and
   detoxifying enzymes. To further clarify the responses to different
   levels of H[2]O[2] stress, we compared the biological categories
   involved in oxidative stress. In M. tuberculosis, sigma factors (such
   as SigE, SigH, SigL and SigF) are important for initiating the H[2]O[2]
   detoxification pathway [[114]7], and so we examined the response of
   sigma factors to different levels of H[2]O[2] in M. smegmatis. We did
   not find any differences in the expression of sigF in the 0, 0.2 mM,
   and 7 mM H[2]O[2] treatments. SigF was first described as a
   stationary-phase stress response sigma factor in M. tuberculosis
   [[115]41], and we have previously shown that SigF is involved in the
   oxidative stress response in mycobacteria [[116]41,[117]42]. Lack of
   SigF induction here may have been due to the fact that SigF is a
   stationary-phase stress response sigma factor and does not function at
   the early logarithmic phase, which was used in this study. It will be
   interesting to compare the global transcriptional response to H[2]O[2]
   at different growth phases. While Voskuil et al. conducted studies
   showing that M. tuberculosis SigH is highly induced upon exposure to
   high levels of H[2]O[2] [[118]7], here, M. smegmatis SigH was
   moderately induced at both low and high stresses (data not shown). The
   gene msmeg_0573, which encodes the ECF sigma factor RpoE1, was the most
   highly induced sigma factor, exhibiting a 136-fold increase in
   expression following the 7 mM H[2]O[2] treatment. No change in its
   expression was observed following the 0.2 mM H[2]O[2] treatment. Its
   paralog, msmeg_0574, which encodes the ECF sigma factor RpoE1, was also
   specifically induced (~17-fold) by 7 mM H[2]O[2]. These results showed
   that both MSMEG_0573 and MSMEG_0574 are specifically induced by 7 mM
   H[2]O[2]. In contrast, msmeg_5214, which encodes the RNA polymerase
   sigma-70 factor, was specifically up-regulated in 0.2 mM H[2]O[2]
   ([119]Table 2). Together, these data indicate that distinct sigma
   factor genes respond to different levels of H[2]O[2]. In addition to
   sigma factors, transcriptional regulators also have important roles in
   the regulation of the oxidative stress response. As M. tuberculosis
   does not have a functional OxyR, the transcription regulators FurA,
   IdeR, CarD and WhiB play important roles in the oxidative stress
   response. FurA was induced 4-fold at 0.2 mM H[2]O[2] and 18-fold at 7
   mM H[2]O[2]. Furthermore, no significant changes were found in IdeR and
   CarD expression levels at either 0.2 mM or 7 mM H[2]O[2.] The M.
   smegmatis genome contains six whiB genes, and only MSMEG_1919 was
   induced by 0.2 mM H[2]O[2] ([120]Table 2). DevR has previously been
   shown to be highly induced in M. tuberculosis only when exposed to high
   concentrations of H[2]O[2] [[121]7]. Consistent with this report, our
   results showed that in M. smegmatis, devR expression was mildly
   increased by 0.2 mM and strongly increased (100-fold) after treatment
   with 7 mM H[2]O[2] treatment ([122]Table 2). We next compared the
   transcription responses of genes coding for serine/threonine-protein
   kinases (SPTKs) to H[2]O[2] exposure. Of the STPKs, we found that
   following 7 mM H[2]O[2] treatment, only pknK was up-regulated. M.
   tuberculosis PknK has been shown to regulate the slow growth of
   mycobacteria in response to various stresses and during persistence in
   infected mice [[123]43]. Just as PknK is involved in the oxidative
   stress response pathway in M. tuberculosis, PknK plays an important
   role in M. smegmatis in regulating the response to high levels of
   H[2]O[2]. The functions of PknK and its involvement in the response to
   H[2]O[2] require further exploration.

   The M. smegmatis genome encodes several enzymes involved in the
   detoxification of H[2]O[2] [[124]4,[125]8]. Our analysis showed that
   mRNA levels of katG were up-regulated 12-fold and 30-fold following
   exposure to 0.2 and 7 mM H[2]O[2] respectively ([126]Table 2). Notably,
   trxB (encoding thioredoxin-disulfide reductase) and trx (encoding
   thioredoxin) expression was induced strongly by 7 mM H[2]O[2] treatment
   ([127]Table 2), but not by 0.2 mM H[2]O[2]. Similarily, ahpD (encoding
   alkylhydroperoxidase) and msmeg_6884 (encoding NADP oxidoreductase)
   responded to 7 mM H[2]O[2] but not to 0.2 mM H[2]O[2]. msmeg_3708
   (encoding catalase) exhibited a 5-fold increase in mRNA expression only
   after exposure to 0.2 mM H[2]O[2], indicating that it is specifically
   induced in response to low levels of H[2]O[2] exposure. It will be
   interesting to investigate the distinct biological roles of enzymes
   that scavenge different levels of H[2]O[2] in mycobacteria. Such
   studies will lead to greater understanding of the basic roles of these
   enzymes.

Conclusion

   In this study, we have shown that, in M. smegmatis, different genes are
   induced in response to low and high levels of H[2]O[2]. A notable
   difference in the response to low-level (0.2 mM) H[2]O[2] and
   high-level (7 mM) H[2]O[2] was observed. When exposed to 0.2 mM
   H[2]O[2], expression of approximately 10% of the genes in the M.
   smegmatis genome was significantly changed. In contrast, 29.3% of M.
   smegmatis genes were significantly changed in response to 7 mM
   H[2]O[2.]Transcriptional analysis suggested that a metabolic switch in
   glycolysis/gluconeogenesis and fatty acid metabolism was potentially
   involved in the response to the 0.2 mM H[2]O[2] treatment but not to
   the 7 mM H[2]O[2] treatment. We also observed that transcriptional
   levels of genes encoding ribosomes decreased when bacterial cells were
   treated with 7 mM H[2]O[2]. This result suggests that 7 mM H[2]O[2]
   treatment affected the protein synthesis apparatus and thus reduced
   protein synthesis, resulting in reduced bacterial growth. The
   expression level of gene msmeg_5214 (encoding the RNA polymerase
   sigma-70 factor) was induced in response to 0.2 mM H[2]O[2], and the
   rpoE1s (msmeg_0573 and msmeg_0574) were induced specifically in
   response to 7 mM H[2]O[2]. In addition, different regulators were
   observed to respond to different levels of H[2]O[2]. MSMEG_1919 was
   induced following exposure to 0.2 mM H[2]O[2], while DevR was highly
   induced by the 7 mM H[2]O[2] treatment. Our results show that pknK, a
   gene encoding a STPK, is involved in the 7 mM H[2]O[2] treatment
   response and that different genes encoding detoxifying enzymes,
   including the genes encoding KatG, AhpD, TrxB and Trx, were expressed
   in response to different levels of H[2]O[2]. In summary, this study of
   global transcriptional changes that occur after exposure to different
   levels of H[2]O[2] documents changes in transcriptional regulation in
   response to exposure to low and high level H[2]O[2] treatments,
   including the use of different sigma factors, regulators,
   serine/threonine kinases and differences in the transcriptional levels
   of detoxifying enzymes used to combat H[2]O[2] stress. Further study of
   these genes will aid our understanding of the mechanisms underlying the
   precise regulation and scavenging of H[2]O[2].

Supporting Information

   S1 Fig. Overview of differential gene expression in response to 7 mM
   H[2]O[2] in M. smegmatis.

   (JPG)
   [128]Click here for additional data file.^ (84.4KB, jpg)
   S1 Table. Oligonucleotide primers used in this study.

   (DOC)
   [129]Click here for additional data file.^ (48.5KB, doc)
   S2 Table. RNA-sequencing expression data for the control of M.
   smegmatis mc^2155 without H[2]O[2] treatment.

   (XLSX)
   [130]Click here for additional data file.^ (362.7KB, xlsx)
   S3 Table. RNA-sequencing expression data for the M. smegmatis mc^2155
   strain treated with 0.2 mM H[2]O[2.]

   (XLSX)
   [131]Click here for additional data file.^ (364.7KB, xlsx)
   S4 Table. RNA-sequencing expression data for the M. smegmatis mc^2155
   strain treated with 7 mM H[2]O[2.]

   (XLSX)
   [132]Click here for additional data file.^ (355.1KB, xlsx)

Acknowledgments