Arch Microbiol DOI 10.1007/s00203-016-1191-y
ORIGINAL PAPER
Understanding the host‑adapted state of Citrobacter rodentium by transcriptomic analysis Allen D. Smith1 · Xianghe Yan2 · Celine Chen1 · Harry D. Dawson1 · Arvind A. Bhagwat2
Received: 16 September 2015 / Revised: 17 December 2015 / Accepted: 12 January 2016 © Springer-Verlag Berlin Heidelberg (outside the USA) 2016
Abstract Citrobacter rodentium (Cr) is a mouse pathogen that mimics many aspects of enteropathogenic Escherichia coli infections including producing attaching and effacing (A/E) lesions. Host-adapted (HA) Cr cells that are shed at the peak of infection have been reported to be hyper-infective. The exact mechanism underlying this phenomenon has remained elusive since the pathogen loses its HA ‘status’ immediately upon subculturing in laboratory media. We sequenced the entire transcriptome of Cr directly from the feces of infected mice and analyzed the gene expression pattern. We observed that the entire transcriptional machinery as well as several transcriptional regulators to be differentially expressed when compared with the transcriptome of cells grown on laboratory media. Major adhesion and effector genes, tir and eae, were highly expressed in HA along with many genes located on all five loci of enterocyte effacement regions (LEE 1–5). Notable absent among the HA expressed genes were 19 fimbrial operons and non-fimbrial adhesions and several non-LEE encoded effectors. These results demonstrate that Communicated by Erko Stackebrandt. Electronic supplementary material The online version of this article (doi:10.1007/s00203-016-1191-y) contains supplementary material, which is available to authorized users. * Allen D. Smith [email protected] 1
Diet, Genomics and Immunology Laboratory, Beltsville Human Nutrition Research Center, USDA-ARS, 10300 Baltimore Ave., B307C, Rm. 228, BARC‑E, Beltsville, MD 20705, USA
2
Environmental, Microbial, and Food Safety Laboratory, Beltsville Agriculture Research Center, USDA-ARS, Beltsville, MD, USA
host-adapted Cr has a unique transcriptome that is associated with increased host transmission. Keywords Attaching–effacing pathogens · RNA-seq analysis · Host-adaptation · Enteropathogens
Introduction Bacteria must be able to grow in diverse environmental conditions. Pathogenic strains of bacteria including Escherichia coli, Salmonella typhimurium, and Vibrio cholera must adapt to growth in wastewater as well as the host gastrointestinal tract. Studies have demonstrated that these bacteria alter their gene expression profiles depending on growth conditions (Partridge et al. 2007; Arunasri et al. 2014; Le Bihan et al. 2015). Furthermore, it has been observed that E. coli and V. cholera undergo host adaption that results in a hyper-infectious state (Merrell et al. 2002; Brady et al. 2011). This, presumably, aids in the spread of these organisms. Characterizing gene expression under different growth conditions is important for understanding how these organisms achieve the host-adapted (HA) state. An RNAseq approach has been used to characterize gene expression in V. cholera grown in culture compared to two infant animal models (Mandlik et al. 2011). Directional RNAseq of E. coli K12 cultured under different conditions revealed condition-dependent changes to the genome (Li et al. 2013). There is, however, a lack of a good animal model for pathogenic strains of E. coli, especially in adult mice, making study of the transcriptome of host-adapted E. coli more difficult. Citrobacter rodentium (Cr) is a natural mouse pathogen that mimics many aspects of EPEC infections in humans
13
including formation of attaching and effacing lesions (Luperchio and Schauer 2001). In recent years, it has been studied extensively as a model of EPEC infections. We as well as others have shown that Cr is hyper-infectious after passage through the host (Wiles et al. 2005; Smith and Bhagwat 2013). It has been proposed that exposure to bicarbonate in the gastrointestinal tract may serve as a signal for induction of genes associated with colonization and virulence (Yang et al. 2010) and limited gene expression analysis has implicated components of the LEE as being important for hyper-infectivity (Bishop et al. 2007). Initial infection requires that the bacteria adhere to the epithelial cells lining the gastrointestinal tract and it has been demonstrated that host-adapted Cr (HA-Cr) has increased binding to human epithelial type 2 (HEp-2) cells compared to Luria–Bertani broth (LB)-grown Cr (Bishop et al. 2007). To gain a greater understanding into the factors contributing to the HA phenotype in Cr, we compared the transcriptomes of Cr gown under three conditions: traditional LB media (LB-Cr), fecal media (clarified 10 % fecal homogenate, FM-Cr), and HA-Cr obtained from feces 5–8 days post-infection.
Arch Microbiol
culture was concentrated by centrifugation and resuspended in LB medium to a concentration of 5.0 × 1010 cfu/ml and 0.2 ml administered (1.0 × 1010 cfu) by oral gavage to male 6- to 8-week-old C57Bl/6 mice. The exact viable count was determined by retrospective plating on LB agar media containing 50 µg/ml nalidixic acid. HA-Cr cells were prepared using a 10 % fecal homogenate from mice infected with Cr (5–8 days post-infection) and were suspended either in RNA stabilization reagent (Bhagwat et al. 2003, 2013) for RNA isolation (see below) or in PBS buffer for estimating viable cell counts. RNA extraction and quantitative PCR for determining gene expression RNA was isolated from Cr cells from either LB broth, FM medium, or HA-state that were treated with RNA stabilization reagent (Bhagwat et al. 2003). Cells were suspended in hot TRIzol RNA isolation reagent (LifeTechnologies, Carlsbad, CA) and processed for RNA isolation as described (Bhagwat et al. 2013). RNA was treated with DNase three times (twice on column and once in solution), and DNA contamination was checked by PCR using primers against rpoD and 16S RNA genes of Cr.
Materials and methods Bacterial growth conditions The Cr strain used is a nalidixic acid (nalr)-resistant spontaneous mutant of ATCC strain 51459. Cultures of Cr, nalr were streaked on LB agar plates from freezer stocks, and a single colony was inoculated in LB broth (Difco Chemicals, Detroit, MI) (Bhagwat et al. 2005). For aerobic growth, a single colony of Cr was inoculated in 10 ml LB medium in 125-ml flask and incubated in shaker incubator at 210 rpm at 37 °C for 18–20 h (Bhagwat et al. 2005) and then diluted and grown to an OD600 of 1.5. For anaerobic growth, cells were grown in anaerobic bags which achieve 4 % CO2 within 2 h of incubation at 37 °C (Becton–Dickinson, Sparks, MD) (Bhagwat and Bhagwat 2004; Smith and Bhagwat 2013). Mouse fecal medium was prepared from fresh fecal pellets obtained from uninoculated mice. Mouse fecal pellets (pH 8.0 ± 0.1) were homogenized (10 %, w/v) in 1 mM sodium phosphate pH 7.0 and 154 mM NaCl and filter sterilized using 0.22-µm filter and used immediately as fecal medium.
Construction and sequencing of illumina RNA‑Seq libraries The messenger RNA population from all RNA samples was enriched by depleting rRNA sequences, and strandspecific libraries were constructed as described earlier (Bhagwat et al. 2014). Briefly, starting with 8–50 ng rRNAdepleted RNA, random-primed cDNA synthesis was done using ScriptSeq v2 RNA-Seq library preparation kit (Epicenter, WI). cDNA was purified using Agencourt AMPure XP system (BeckmanCoulter, NJ). Libraries were amplified using FailSafe PCR enzyme kit (Epicenter, WI). Typically, 12 (50–100 ng starting RNA) or 15 (8 ng starting RNA) PCR cycles were used and reverse primer from the kit was replaced with one of the ScriptSeq Index primers. After PCR amplification, libraries were purified and size selected (~280 bp) using the Agencourt AMPure XP system (BeckmanCoulter, NJ). Profiles of library insert sizes were verified on the Experion microfluidic platform using 1K DNA chip (Bio-Rad Laboratories, CA). Libraries were sequenced on an Illumina Hi-Seq2500 (100 cycle pairedend mode for HA samples) or MiSeq instrument (150 cycle paired-end mode for LB and FM samples).
Mouse inoculation protocols Mapping and assembly of raw reads For mouse inoculation, an overnight culture of Cr in LB media was diluted the following morning into fresh LB medium and grown to an OD600 of 1.5. To obtain HA-Cr, the
13
Sequences were initially analyzed using CLC Genomics software v6.0. Sequences with a quality score of
ORIGINAL PAPER
Understanding the host‑adapted state of Citrobacter rodentium by transcriptomic analysis Allen D. Smith1 · Xianghe Yan2 · Celine Chen1 · Harry D. Dawson1 · Arvind A. Bhagwat2
Received: 16 September 2015 / Revised: 17 December 2015 / Accepted: 12 January 2016 © Springer-Verlag Berlin Heidelberg (outside the USA) 2016
Abstract Citrobacter rodentium (Cr) is a mouse pathogen that mimics many aspects of enteropathogenic Escherichia coli infections including producing attaching and effacing (A/E) lesions. Host-adapted (HA) Cr cells that are shed at the peak of infection have been reported to be hyper-infective. The exact mechanism underlying this phenomenon has remained elusive since the pathogen loses its HA ‘status’ immediately upon subculturing in laboratory media. We sequenced the entire transcriptome of Cr directly from the feces of infected mice and analyzed the gene expression pattern. We observed that the entire transcriptional machinery as well as several transcriptional regulators to be differentially expressed when compared with the transcriptome of cells grown on laboratory media. Major adhesion and effector genes, tir and eae, were highly expressed in HA along with many genes located on all five loci of enterocyte effacement regions (LEE 1–5). Notable absent among the HA expressed genes were 19 fimbrial operons and non-fimbrial adhesions and several non-LEE encoded effectors. These results demonstrate that Communicated by Erko Stackebrandt. Electronic supplementary material The online version of this article (doi:10.1007/s00203-016-1191-y) contains supplementary material, which is available to authorized users. * Allen D. Smith [email protected] 1
Diet, Genomics and Immunology Laboratory, Beltsville Human Nutrition Research Center, USDA-ARS, 10300 Baltimore Ave., B307C, Rm. 228, BARC‑E, Beltsville, MD 20705, USA
2
Environmental, Microbial, and Food Safety Laboratory, Beltsville Agriculture Research Center, USDA-ARS, Beltsville, MD, USA
host-adapted Cr has a unique transcriptome that is associated with increased host transmission. Keywords Attaching–effacing pathogens · RNA-seq analysis · Host-adaptation · Enteropathogens
Introduction Bacteria must be able to grow in diverse environmental conditions. Pathogenic strains of bacteria including Escherichia coli, Salmonella typhimurium, and Vibrio cholera must adapt to growth in wastewater as well as the host gastrointestinal tract. Studies have demonstrated that these bacteria alter their gene expression profiles depending on growth conditions (Partridge et al. 2007; Arunasri et al. 2014; Le Bihan et al. 2015). Furthermore, it has been observed that E. coli and V. cholera undergo host adaption that results in a hyper-infectious state (Merrell et al. 2002; Brady et al. 2011). This, presumably, aids in the spread of these organisms. Characterizing gene expression under different growth conditions is important for understanding how these organisms achieve the host-adapted (HA) state. An RNAseq approach has been used to characterize gene expression in V. cholera grown in culture compared to two infant animal models (Mandlik et al. 2011). Directional RNAseq of E. coli K12 cultured under different conditions revealed condition-dependent changes to the genome (Li et al. 2013). There is, however, a lack of a good animal model for pathogenic strains of E. coli, especially in adult mice, making study of the transcriptome of host-adapted E. coli more difficult. Citrobacter rodentium (Cr) is a natural mouse pathogen that mimics many aspects of EPEC infections in humans
13
including formation of attaching and effacing lesions (Luperchio and Schauer 2001). In recent years, it has been studied extensively as a model of EPEC infections. We as well as others have shown that Cr is hyper-infectious after passage through the host (Wiles et al. 2005; Smith and Bhagwat 2013). It has been proposed that exposure to bicarbonate in the gastrointestinal tract may serve as a signal for induction of genes associated with colonization and virulence (Yang et al. 2010) and limited gene expression analysis has implicated components of the LEE as being important for hyper-infectivity (Bishop et al. 2007). Initial infection requires that the bacteria adhere to the epithelial cells lining the gastrointestinal tract and it has been demonstrated that host-adapted Cr (HA-Cr) has increased binding to human epithelial type 2 (HEp-2) cells compared to Luria–Bertani broth (LB)-grown Cr (Bishop et al. 2007). To gain a greater understanding into the factors contributing to the HA phenotype in Cr, we compared the transcriptomes of Cr gown under three conditions: traditional LB media (LB-Cr), fecal media (clarified 10 % fecal homogenate, FM-Cr), and HA-Cr obtained from feces 5–8 days post-infection.
Arch Microbiol
culture was concentrated by centrifugation and resuspended in LB medium to a concentration of 5.0 × 1010 cfu/ml and 0.2 ml administered (1.0 × 1010 cfu) by oral gavage to male 6- to 8-week-old C57Bl/6 mice. The exact viable count was determined by retrospective plating on LB agar media containing 50 µg/ml nalidixic acid. HA-Cr cells were prepared using a 10 % fecal homogenate from mice infected with Cr (5–8 days post-infection) and were suspended either in RNA stabilization reagent (Bhagwat et al. 2003, 2013) for RNA isolation (see below) or in PBS buffer for estimating viable cell counts. RNA extraction and quantitative PCR for determining gene expression RNA was isolated from Cr cells from either LB broth, FM medium, or HA-state that were treated with RNA stabilization reagent (Bhagwat et al. 2003). Cells were suspended in hot TRIzol RNA isolation reagent (LifeTechnologies, Carlsbad, CA) and processed for RNA isolation as described (Bhagwat et al. 2013). RNA was treated with DNase three times (twice on column and once in solution), and DNA contamination was checked by PCR using primers against rpoD and 16S RNA genes of Cr.
Materials and methods Bacterial growth conditions The Cr strain used is a nalidixic acid (nalr)-resistant spontaneous mutant of ATCC strain 51459. Cultures of Cr, nalr were streaked on LB agar plates from freezer stocks, and a single colony was inoculated in LB broth (Difco Chemicals, Detroit, MI) (Bhagwat et al. 2005). For aerobic growth, a single colony of Cr was inoculated in 10 ml LB medium in 125-ml flask and incubated in shaker incubator at 210 rpm at 37 °C for 18–20 h (Bhagwat et al. 2005) and then diluted and grown to an OD600 of 1.5. For anaerobic growth, cells were grown in anaerobic bags which achieve 4 % CO2 within 2 h of incubation at 37 °C (Becton–Dickinson, Sparks, MD) (Bhagwat and Bhagwat 2004; Smith and Bhagwat 2013). Mouse fecal medium was prepared from fresh fecal pellets obtained from uninoculated mice. Mouse fecal pellets (pH 8.0 ± 0.1) were homogenized (10 %, w/v) in 1 mM sodium phosphate pH 7.0 and 154 mM NaCl and filter sterilized using 0.22-µm filter and used immediately as fecal medium.
Construction and sequencing of illumina RNA‑Seq libraries The messenger RNA population from all RNA samples was enriched by depleting rRNA sequences, and strandspecific libraries were constructed as described earlier (Bhagwat et al. 2014). Briefly, starting with 8–50 ng rRNAdepleted RNA, random-primed cDNA synthesis was done using ScriptSeq v2 RNA-Seq library preparation kit (Epicenter, WI). cDNA was purified using Agencourt AMPure XP system (BeckmanCoulter, NJ). Libraries were amplified using FailSafe PCR enzyme kit (Epicenter, WI). Typically, 12 (50–100 ng starting RNA) or 15 (8 ng starting RNA) PCR cycles were used and reverse primer from the kit was replaced with one of the ScriptSeq Index primers. After PCR amplification, libraries were purified and size selected (~280 bp) using the Agencourt AMPure XP system (BeckmanCoulter, NJ). Profiles of library insert sizes were verified on the Experion microfluidic platform using 1K DNA chip (Bio-Rad Laboratories, CA). Libraries were sequenced on an Illumina Hi-Seq2500 (100 cycle pairedend mode for HA samples) or MiSeq instrument (150 cycle paired-end mode for LB and FM samples).
Mouse inoculation protocols Mapping and assembly of raw reads For mouse inoculation, an overnight culture of Cr in LB media was diluted the following morning into fresh LB medium and grown to an OD600 of 1.5. To obtain HA-Cr, the
13
Sequences were initially analyzed using CLC Genomics software v6.0. Sequences with a quality score of
