Proteomic Delineation of the ArcA Regulon in Salmonella Typhimurium During Anaerobiosis

Abstract

Salmonella enterica serovar Typhimurium (S. Typhimurium) is one of the most used models for bacterial pathogenesis and successful infection requires its adaptation to the low oxygen environment in host gastrointestinal tracts. Central to this process is the Arc (aerobic respiratory control) two-component regulatory system that contains a sensor kinase ArcB and a response regulator ArcA. Nevertheless, a comprehensive profile of the ArcA regulon on the proteome level is still lacking in S. Typhimurium. Here we quantitatively profiled Salmonella proteome during anaerobiosis in an arcA-deleting mutant compared with its parental strain. In addition to known processes under its control, notably we found that ArcA represses ethanolamine utilization by directly binding to the promoter region of the eut operon. Furthermore, we found opposing changes of several bacterial genes on the protein and transcript levels in the arcA-deleting mutant including the virulence genes of Salmonella pathogenicity island 1 (SPI-1), thereby indicating potentially prevalent post-transcriptional regulatory mechanisms. Altogether, our study provides important new insights into ArcA-dependent bacterial physiology and virulence during Salmonella anaerobiosis.

Salmonella enterica serovar Typhimurium (S. Typhimurium)¹ is a Gram-negative, facultative intracellular bacterial pathogen that causes acute gastroenteritis via oral ingestion of contaminated food or water. During infection, S. Typhimurium survives gastric acidity to gain access to the intestinal epithelium (1). The bacterial virulence is highly dependent on two distinct type III secretion systems (T3SSs) encoded on Salmonella pathogenicity islands 1 and 2 (SPI-1 and -2). On adherence to the apical surface of host cells, S. Typhimurium injects a set of SPI-1 virulence factors (called effectors) to facilitate initial invasion. After bacterial internalization into host cells, SPI-2 T3SS is activated and plays a major role in promoting intracellular survival and replication (1–3). Additionally, to achieve successful infection S. Typhimurium must cope with the drastic shift from ambient conditions in vitro to the low oxygen environment in host gastrointestinal tracts (4).

An important signal transduction system that mediates the adaptation of S. Typhimurium to different respiratory conditions is the Arc two-component system (TCS). It comprises the transmembrane sensor kinase ArcB and the cytosolic cognate response regulator ArcA (5, 6). Under microaerobic or anaerobic conditions, ArcB undergoes autophosphorylation and catalyzes the phosphorylation of ArcA (7). Phosphorylated ArcA is then activated as a transcription factor and turns on/off the expression of a wide spectrum of downstream genes, thereby reshuffling bacterial metabolic pathways to optimize energy conversion (6, 7). When bacteria encounter aerobic conditions, oxidized quinone electron carriers in the membrane inhibit the autophosphorylation of ArcB and hence the phosphorylation/activation of ArcA (5, 8). Thus, the Arc TCS permits bacterial adaptation to changing oxygen levels.

The regulatory role of the Arc TCS has been extensively studied in E. coli by using various techniques, especially high-throughput transcriptomic profiling (5–7, 9–13). In S. Typhimurium, however, few studies have been carried out to define its ArcA regulon. Evans et al. examined the transcriptome of S. Typhimurium ATCC 14028s wild-type (WT) and its isogenic ΔarcA mutant under anaerobic conditions (14). In addition to the central metabolism regulated similarly as in E. coli, Salmonella ArcA also controls the expression of distinct pathways such as repression of ethanolamine utilization and activation propanediol metabolism. Nevertheless, a proteomic landscape of the ArcA-regulated pathways during Salmonella anaerobiosis has been lacking.

Herein we performed the first large-scale proteomic profiling of S. Typhimurium and its isogenic ΔarcA mutant during anaerobiosis. Among ∼1700 protein identifications, 260 proteins were differentially expressed in the strain lacking arcA compared with the wild-type. In contrast to the previous report, we found that ArcA represses ethanolamine utilization by directly binding to the promoter region of the eut operon and importantly such regulation is physiologically relevant during bacterial infection in vivo. Furthermore, we provide evidence that the regulation of the TCA cycle by ArcA is partially mediated by its repression of the transcription factor YdcI. Interestingly, our data also reveal some ArcA-regulated proteins (e.g. SPI-1 virulence factors) exhibit changes that differed between protein and transcript levels, indicating potential post-transcriptional regulatory mechanisms.

MATERIALS AND METHODS

Bacterial Strains and Culture Conditions

The Salmonella enterica serovar Typhimurium wild-type strain SL1344 and its isogenic arcA mutant were maintained frozen at −80 °C in the peptone solution (2% peptone, 25% glycerol). Bacteria were routinely grown on LB plates with 1.5% agar and 30 μg/ml streptomycin at 37 °C. A single colony was inoculated into 3 ml of MOPS (morpholinepropanesulfonic acid)-buffered (100 mm, pH 7.4) LB broth supplemented with 20 mm d-xylose (LB-MOPS-X, to avoid potential indirect effects of pH and catabolite repression (14)) (10, 15) and grown overnight at 37 °C with shaking. The overnight culture was then diluted 1:20 into 5 ml of LB-MOPS-X broth in a centrifugal tube (5 ml). The tube was tightly sealed with parafilm to maintain anaerobic conditions during bacterial culturing. The bacteria were harvested when OD₆₀₀ reached ∼0.3.

Construction of Bacterial Mutants

The S. Typhimurium arcA deletion mutant was generated by using the standard homologous recombination method as previously described (16). Briefly, a PCR fragment containing the 5′ and 3′-flanking sequences of the target gene was cloned into a suicide vector pSR47s. The resulting vector was transferred into the wild-type strain through E. coli DH5α (λpir)-mediated conjugation. The desirable transconjugants were selected on LB agar plates containing 50 μg/ml kanamycin and 30 μg/ml streptomycin. These colonies were further screened for marker-less in-frame deletion by growth on LB agar plates containing 15% sucrose without NaCl, and the deletion of the target gene was confirmed by both sequencing and PCR analyses.

The S. Typhimurium eut* mutant was constructed by replacing the promoter region (–250 to −1 relative to the start codon of the eutS gene) with a constitutive promoter J23119 (http://partsregistry.org/Part:BBa_J23119). Approximately 400 ng of PCR products containing the J23119 sequence and the chloramphenicol resistance gene cat was electroporated into S. Typhimurium SL1344 containing pKD46. The desirable construct was selected on LB agar plates containing 25 μg/ml chloramphenicol and further verified by colony PCR as well as DNA sequencing.

Proteomic Sample Preparation and Peptide Dimethyl Labeling

The bacteria were harvested when OD₆₀₀ reached ∼0.3 and bacterial pellets were washed twice with PBS before being boiled in the SDS sample buffer. Approximately 100 μg of proteins were loaded onto SDS-PAGE and fractionated into eight fractions. In-gel protein digestion was performed as previously reported (17). Briefly, gel slices were cut into 1 mm³ cubes and destained with 50% acetonitrile (ACN) in 50 mm triethyl ammonium bicarbonate (TEAB) and dehydrated with pure ACN. After dehydration, in-gel trypsin digestion was performed in a buffer containing 1.2 ng/μl trypsin, 50 mm TEAB and 10% (v/v) ACN. The enzymatic reaction was allowed to proceed overnight at 37 °C. The resulting tryptic peptides were extracted from gel cubes twice with 50% ACN and 5% formic acid (FA) for 20 min at 37 °C with constant shaking. Finally, extracted peptides were pooled and vacuum dried for further dimethyl labeling.

The isotopic labeling experiments were performed as described previously (18). The protein digest was resuspended in 100 μl of 100 mm TEAB, and then 4 μl of 0.6 m sodium cyanoborohydride (NaBH₃CN) were added. Peptides from WT and ΔarcA strains were labeled with 4 μl of 4% formaldehyde (CH₂O) and deuterated formaldehyde (CD₂O), respectively. The reaction mixture was vortexed and incubated at room temperature for 1 h. The reaction was quenched by sequential addition of 16 μl of 1% (v/v) ammonia solution and 8 μl of formic acid. Finally, light- and heavy-labeled peptides were mixed, and vacuum dried immediately before further mass spectrometric analyses.

Nanoflow LC-MS/MS Analyses

Nanoflow reversed-phase LC separation was carried out on an EASY-nLC 1200 System (Thermo Scientific). The capillary column (75 μm × 150 mm) with a laser-pulled electrospray tip (Model P-2000, Sutter Instruments) was home-packed with 4 μm, 100 Å Magic C18AQ silica-based particles (Michrom BioResources Inc., Auburn, CA). Labeled peptides were dissolved in solvent A (described below) and ∼200 ng of sample was loaded onto the analytical column in a single LC-MS/MS analysis. The mobile phase was comprised of solvent A (97% H₂O, 3% ACN, and 0.1% FA) and solvent B (100% ACN and 0.1% FA). The LC separation was carried out with the following gradient: solvent B was started at 7% for 3 min, and then raised to 35% over 40 min. Subsequently, solvent B was rapidly increased to 90% in 2 min and maintained for 10 min before 100% solvent A was used for column equilibration. Peptides eluted from the capillary column were electrosprayed directly onto a hybrid linear ion trap-Orbitrap mass spectrometer (LTQ Orbitrap Velos, Thermo Scientific) for MS/MS analyses in a data-dependent mode. One full MS scan (m/z 350–1500) was acquired by the Orbitrap mass analyzer with r = 60,000 and subsequently the ten most intense ions were selected and fragmented by collision-induced dissociation (CID) in the ion trap with the following parameters: ≥ +2 precursor ion charge, 2 Da precursor ion isolation window, and 35% normalized collision energy. Dynamic exclusion was set with repeat duration of 30 s and exclusion duration of 12 s.

Experimental Design and Statistical Rationale

To globally define the regulon of the ArcAB two-component system, we performed proteomic analyses of the ΔarcA mutant compared with the wild-type strain under anaerobic conditions. In total, we analyzed three biological replicates of paired ΔarcA and wild-type samples in 48 LC-MS/MS experiments. The raw MS files were analyzed by MaxQuant (http://maxquant.org/, version 1.5.4.1). MS/MS spectra were searched against the S. Typhimurium LT2 protein database (better annotated than SL1344 with 5199 sequences downloaded from UniProt modified at November 30, 2016) using the Andromeda search engine embedded in MaxQuant. The precursor mass tolerance was set at 20 ppm and the fragment mass tolerance was set at 0.8 Da. The digestion enzyme was set as trypsin with a maximum of two missed cleavages. Dimethyl (K, N-term) and dimethyl (D₄K, D₄N-term) were set as fixed modifications for light (L)- and heavy (H)-labeled samples, respectively. Oxidation (M) was set as a variable modification. Both peptide and protein assignments were filtered to achieve a false discovery rate (FDR) < 1%. Only the proteins having at least two unique peptides were quantified. The MaxQuant software was used to calculate the intensity of H- and L-labeled proteins. The intensity values from MaxQuant were normalized and further processed by using the Perseus software (version 1.5.4.1). We removed those protein hits that were only assigned with modifications or matched to the reverse database as well as common contaminants. Logarithmic values (log₂) of the H- and L-labeled protein intensity were calculated. Missing values (that refer to the scenario where a peptide signal is absent or not detected in one of the two-paired samples) were replaced with random numbers from a normal distribution with the default parameters (width = 0.3, shift = 1.8) by using the imputation method in Perseus. All altered proteins with missing values were listed in Table S2 sheet 4. We have performed additional data analyses that demonstrate the applicability of this imputation method for our dataset (see supplemental Fig. S1). The p values were obtained by using the two-tailed Student's t test. Proteins with ratios (H/L) > 2.0 or < 0.5 and p values <0.05 were considered as significant difference between the WT and ΔarcA strains. To have sufficiently more protein hits for subsequent enrichment analyses, the multiple hypothesis testing was not performed. Functional annotation and clustering enrichment of differentially expressed proteins were conducted by using the clusters of orthologous groups (COGs) as defined by the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/COG). For the analysis of protein-protein interactions and/or their functional association, differentially expressed proteins were searched against the STRING database (http://string-db.org/) with the highest confidence score (score >0.9). The interaction network only showed the proteins with connections, and the unconnected proteins were not presented.

Reverse Transcription-Quantitative PCR (RT-qPCR)

RT-qPCR measurements of gene expression were performed as previously described (19). S. Typhimurium was grown anaerobically to an OD₆₀₀ of ∼0.3 as described above. The total RNA was extracted using the illustra RNAspin Mini Kit (GE Healthcare Life Sciences, Chicago, IL) following the manufacturer's instructions. The extracted RNA samples were subjected to DNase I treatment using the turbo DNA Free Kit (Ambion) to remove any genomic DNA contaminants. Reverse transcription was performed by using Super Script II reverse transcriptase (Invitrogen, Carlsbad, CA) and random hexamer (Invitrogen) to generate cDNA. Quantitative PCR was performed by using the SYBR Green PCR master mix (Applied Biosystems, Foster City, CA) and specific primers on a StepOnePlus Real-time PCR system (ABI). The house-keeping gene rrsA (encoding 16S RNA) was used as an internal control. Primers are listed in supplemental Table S1.

β-galactosidase Activity Assays

β-galactosidase activity assays were performed according to the Miller method (1972). Briefly, bacteria were grown anaerobically to an OD₆₀₀ of ∼0.3. Tetracycline (with a final concentration of 10 μg/ml) was added to terminate protein synthesis and bacterial cultures were placed on ice. Appropriate amounts of cell suspension were mixed with Z-buffer (60 mm Na₂HPO₄, 40 mm NaH₂PO₄, 10 mm KCl, 1 mm MgSO₄, 50 mm β-mercaptoethanol, pH 7.0), and bacteria were lysed with chloroform and SDS. The lysate samples were incubated at 28 °C for 5 min, followed by the addition of 4 mg/ml ortho-nitrophenyl-β -galactoside (ONPG). Color development of the reaction was monitored and measured at 420 nm by using a spectrophotometer. Promoter activities were calculated as the Miller units.

Purification of Recombinant His₆-ArcA Proteins

Purification of His₆-ArcA proteins and subsequent in vitro phosphorylation were carried out as previously described (20). Briefly, E. coli strain BL21 transformed with pET-His₆-ArcA was grown in LB to an OD₆₀₀ of 0.4 before the addition of 1 mm isopropyl-β-d-thiogalactopyranoside (IPTG) to induce the expression of His₆-ArcA. After additional 3–6 h of growth, bacteria were harvested by centrifugation. Bacterial pellets were then resuspended in 2 ml of the lysis buffer (0.5 m NaCl, 50 mm Tris-HCl, pH 7.2, 40% glycerol, 20 mm imidazole, 150 mm PMSF, and 500 mm DTT) followed by extensive sonication. Cell lysates were cleared by centrifugation at 15,000 × g for 30 min and then run through a 1.5 ml Ni-nitrilotriacetic acid-agarose column (Qiagen) equilibrated with the binding buffer (0.5 m NaCl, 50 mm Tris-HCl, pH 7.2, 40% glycerol, and 40 mm imidazole). Following three rounds of washing with the binding buffer, His₆-ArcA was eluted with the elution buffer (0.5 m NaCl, 50 mm Tris-HCl, pH 7.2, 40% glycerol, and 250 mm imidazole).

Electrophoretic Mobility Shift Assays (EMSA)

Purified His₆-ArcA proteins were further phosphorylated by incubation in the TEGD buffer (50 mm Tris-HCl, pH 7.5, 0.5 mm EDTA, and 10% glycerol) supplemented with 5 mm MgCl₂ and 50 mm carbamoyl phosphate. The reaction mixture was incubated at 25 °C for 90 min, and then the phosphorylated ArcA was used immediately for subsequent DNA binding. To initiate DNA binding, various concentrations of DNA probes and phosphorylated proteins were mixed and incubated at 37 °C for 20 min in the EMSA binding buffer (pH 7.2, 20 mm Tris, 50 mm NaCl, 1 mm EDTA, 20 mm DTT, 10% glycerol, and 0.5 mg/ml BSA). The reaction mixtures were then subjected to 6% non-denaturing polyacrylamide gel electrophoresis in the 0.5×TBE buffer. The polyacrylamide gel was visualized under UV light (254–366 nm) following the staining in the 0.5×TBE buffer containing 0.5 μg/ml ethidium bromide (EB) for 10 min.

Western Blot Analysis

The WT and ΔarcA strains expressing chromosomally 3×FLAG-tagged SipA were constructed by using the homologous recombination method as described above. Bacteria were grown in LB-MOPS-X medium anaerobically to an OD₆₀₀ of ∼0.3. Equivalent volumes of bacterial cultures were clarified by centrifugation at 10,000 × g for 10 min and the supernatants were filtered prior to protein precipitation with 20% (final concentration) of trichloroacetic acid (TCA). After washing with cold acetone, the precipitated proteins were resuspended in the SDS sample buffer. To probe the level of proteins in cell lysates, bacterial pellets were directly resuspended in the SDS sample buffer prior to SDS-PAGE fractionation. Gel-separated proteins were then transferred to the PVDF membrane followed by immunoblotting detection with primary antibodies specific for FLAG (Sigma, St. Louis, MO) (1:5000) and anti-mouse HRP-conjugated secondary antibodies (Sigma) (1:5000). As a loading control, GroEL was also probed by using polyclonal anti-GroEL (Sigma) (1:80,000) and anti-rabbit HRP-conjugated secondary antibodies (Sigma) (1:5000).

Bacterial Growth Competition Assays

Overnight cultures of WT or ΔarcA strains were co-inoculated in LB-MOPS-X medium with a final cell density of 1 × 10⁶ cells/ml under anaerobic conditions. At selected time points, bacterial cultures were diluted and plated on LB agar plates with and without the addition of 20 μg/ml kanamycin, respectively. The plates were incubated at 37 °C for 24 h. The number of viable bacteria in the original cultures was determined by colony-forming unit (CFU) assays. The competitive index is defined as the number of viable WT bacteria divided by that of ΔarcA bacteria. Results are presented as the mean of three independent experiments.

In Vivo Competition Assays In C. elegans Infection

In vivo competition assays in C. elegans were performed following the description by Portal-Celhay et al. with slight modifications (21). Briefly, C. elegans wild-type (N2) embryos were cultured on NGM agar plates (0.25% peptone, 0.3% NaCl, 2% agar, 5 μg/ml cholesterol, 1 mm MgSO₄, and 25 mm KH₂PO₄, pH 6) at 25 °C and fed with E. coli OP50 as described (22). At day 2, synchronized nematodes (L4 stage) were transferred to LB+MOPS+X plates containing 100 μl of mixed bacterial lawns of the wild-type S. Typhimurium and the eut* mutant strains at a ratio of 1:1, and worms were allowed to feed on the mixed lawns at 25 °C. After 24 hpi or 48 hpi, 12 worms were transferred to 500 μl of M9 worm buffer (0.5% NaCl, 0.3% KH₂PO₄, 0.6% Na₂HPO₄, and 1 mm MgSO₄) and washed in M9 worm buffer for 6 times to remove external bacteria. The worms were finally suspended in ddH₂O and disrupted mechanically. S. Typhimurium that survived in the C. elegans intestine were then recovered and quantified by dilution-plating on selective MacConkey plates. CFUs were determined following incubation of the plates at 37 °C for 24 h. The competitive index (CI) was calculated as follows: CI = (the number of WT recovered/the number of eut* recovered)/(the number of WT inoculated/the number of eut* inoculated).

Determination of Cellular ATP Levels

Measurements of cellular ATP levels were performed by using the CellTiter-Glo luminescent cell viability assay (Promega) according to the manufacturer's instructions. Approximately 10⁷ cells were pelleted and resuspended in 100 μl of PBS. Cell suspensions or ATP standards diluted in PBS (100 μl) were mixed 1:1 with 100 μl of CellTiter-Glo reagents in standard opaque-walled 96-well plates. The plates were incubated at room temperature for 15 min before luminescence measurements (Synergy HTX Plate Reader). Cell-free controls were included to account for background readings.

RESULTS

Comparative Proteomics Analysis of S. Typhimurium and Its Isogenic ΔarcA Strain During Anaerobiosis

To globally define the ArcAB regulon, we quantitatively examined the proteome of wild-type S. Typhimurium and its isogenic mutant lacking arcA (ΔarcA) cultured under anaerobic conditions. Upon enzymatic digestion of bacterial proteins, the resulting peptides from WT and ΔarcA samples were isotopically labeled and mixed in equal amounts before LC-MS/MS analyses. In total, we identified 1551 bacterial proteins (FDR<1%) from three biological replicates. By using the criteria described in the method section, 149 and 102 proteins were up- and down-regulated respectively in the ΔarcA mutant compared with its parental strain. Therefore, these altered proteins are likely to be regulated either directly or indirectly by the ArcAB two-component system. Specifically, proteins with higher or lower levels in the deletion mutant are the candidates being repressed or activated by ArcA. A full list of all protein assignments and differentially expressed proteins is provided as supplemental Table S2.

To obtain a global view of ArcA-controlled biological processes, we classified proteins with altered expression levels in S. Typhimurium ΔarcA cells into Clusters of Orthologous Groups (COGs) (Table I). The largest category contains mostly metabolic enzymes, accounting for 49 and 41% of ArcA-repressed and activated proteins respectively. In fact, many ArcA-regulated proteins are associated with energy production and conversion as well as amino acid metabolism and transport. Among the ArcA-repressed proteins, we also found several proteins required for signal transduction, cell wall/envelop biogenesis as well as inorganic ion transport and metabolism. Furthermore, ArcA seems to activate a set of proteins associated with protein translation. Last, many ArcA-regulated proteins (approximately one fourth) are of unknown functions or poorly characterized, suggesting that our current understanding of the ArcAB regulatory system is still far from being complete.

Table I. Classification of differentially expressed Salmonella proteins in the arcA mutant according to Clusters of Orthologous Groups (COGs).

Functional Groups	Number of Proteins
	ArcA-repressed	ArcA-activated
Energy production and conversion	34	10
Amino acid metabolism and transport	20	12
Nucleotide metabolism and transport	2	8
Carbohydrate metabolism and transport	10	6
Coenzyme metabolism	0	4
Lipid metabolism	8	4
Metabolisma	74	44
Translation	0	14
Transcription	4	5
Replication and repair	0	2
Cell wall/membrane/envelop biogenesis	8	3
Cell motility	6	1
Post-translational modification, protein turnover, chaperone functions	4	0
Inorganic ion transport and metabolism	7	6
Secondary structure	4	1
Signal transduction	11	3
Intracellular trafficking and secretion	3	1
General functional prediction only	8	7
Function unknown	41	28

^a The bolded functional category contains a summary of the unbolded COG functional groups above associated with metabolism.

Network Analyses of Functional Associations and/or Interactions Among ArcA-Regulated Proteins

To further understand functional associations or interactions of those proteins regulated by ArcA, next we performed network analysis using the STRING database. Among the proteins repressed by ArcA (i.e. those proteins with higher levels in the arcA mutant), there are several distinct networks with the most salient one associated with the tricarboxylic acid (TCA) cycle (Fig. 1A). In addition, we also found the presence of a few small clusters including those required for amino acid and lipid metabolism, suggesting their anaerobic repression by ArcA (Fig. 1A). In general, these proteomic findings (enclosed in black circles) are consistent with the previous transcriptomic data (14). Notably, our data also reveal a notable network comprised of EutB, EutC, EutE, and EutQ under anaerobic conditions, which are encoded by the ethanolamine utilization (eut) operon. (Fig. 1A). Intriguingly, the repression of the eut operon contrasts with the previous report of activation of this operon by ArcA on the transcript level (14). Furthermore, our network analyses reveal a small cluster of proteins that are associated with trehalose biosynthesis (OtsA, OtsB) and glycogen catabolism (STM1558, STM1559 and STM1560) (shown in the red circle, Fig. 1A). The repression of these proteins by ArcA has not been reported thus far in S. Typhimurium.

Fig. 1.

Network analyses of protein-protein interactions and functional associations of differentially expressed proteins in the arcA mutant. Different clusters of associated proteins were identified by using the STRING software with a highest confidence score. A, ArcA-repressed proteins. B, ArcA-activated proteins. Black or blue circles indicate those ArcA-regulated proteins that are consistent or inconsistent with previous transcriptome data, respectively. Red circles indicate the ArcA-regulated proteins that were newly identified in this study. The network nodes represent differentially expressed proteins and color-coded lines linking different nodes represent the types of evidence used in prediction (red line: fusion evidence; green line: neighborhood evidence; blue line: co-occurrence evidence; purple line: experimental evidence; yellow line: text-mining evidence; light blue line: database evidence; black line; co-expression evidence). The color of nodes is used only as a visual aid to identify which node goes with which description in the list of input. Empty nodes represent proteins of unknown 3D structures and filled nodes represent those with known or predicted 3D structures. The different size of nodes indicates the availability of structural information (i.e. it is larger to fit a thumbnail picture).

Next we focus our attention on the ArcA-activated proteins (i.e. those proteins of lower levels in the arcA mutant, Fig. 1B). In addition to the pdu operon encoding proteins for propanediol utilization, notably four new clusters of proteins were identified including several ribosomal proteins, maltose utilization proteins (MalE, MalM, and LamE), sulfate assimilation proteins (CysP, CysN, CysH, CysI, and Sbp) and several hydrogenases encoded by the hyb operon (enclosed in the red circles, Fig. 1B). Furthermore, a cluster of down-regulated proteins are associated with S. Typhimurium SPI-1 T3SS including SipA, SptP, PrgH, and HilA (Fig. 2B) and previously the mRNA levels of SPI-1 genes were mostly unaltered in the arcA mutant (14). A list of all discussed proteins in the text is provided in supplemental Table S2 (sheet 3).

Fig. 2.

Repression of the ethanolamine utilization pathway by ArcA. A, A schematic diagram of the eut operon. B, Representative mass spectra of dimethyl labeled peptides from EutQ, EutE, EutB and EutC. Peptides in the WT and ΔarcA samples with light and heavy labels are indicated by open and filled triangles respectively. C, RT-qPCR analyses of eutA, eutE, and eutQ mRNA levels in the WT and ΔarcA strains grown under anaerobic conditions. D, β-galactosidase activities of Peut-lacZ in the WT and ΔarcA strains grown under anaerobic conditions.

Repression of the Ethanolamine Utilization Operon by ArcA

Several proteins of the ethanolamine utilization operon, including EutB, EutC, EutE and EutQ, are of significantly higher (23–49 fold) levels in ΔarcA cells, suggesting their repression by ArcA (Fig. 2A and 2B). Previous microarray analysis by Evans et al. showed the activation of the eut operon by ArcA in S. Typhimurium 14028s strain (14). Given the importance of ethanolamine utilization for the physiology of S. Typhimurium at the host-pathogen interface (23, 24), we set out to investigate its regulation by using both in vivo and in vitro approaches. RT-qPCR analysis found that the mRNA levels of several genes (eutA, eutE, and eutQ) were orders of magnitude higher in ΔarcA cells than those in the WT (Fig. 2C). Next we constructed a strain containing a lacZ fusion to the promoter region of eut (Peut-lacZ) and consistently measurements of β-galactosidase activity revealed substantially higher transcription levels of Peut-lacZ in ΔarcA cells than in the WT (Fig. 2D).

Bioinformatics analysis led us to identify a putative ArcA-binding site located upstream of the predicted −35 element of the eut promoter region (Fig. 3A). To examine whether ArcA directly regulates the eut operon by binding to its promoter, we then performed the in vitro electrophoretic mobility shift assay (EMSA). Upon incubation of the DNA fragment with recombinantly purified and phosphorylated ArcA (ArcA-P), a retarded band was clearly observed corresponding to the DNA-protein complexes of ArcA-P with Peut (Fig. 3B). Collectively, these findings suggest that S. Typhimurium AcrA directly binds to the eut promoter and represses its transcription under anaerobic conditions.

Fig. 3.

Repression of the eut operon by direct binding of ArcA to its promoter region and physiological relevance of such regulation in S. Typhimurium. A, Bioinformatics analysis of the promoter region (Peut). The sequence of the eut promoter region of 500 bp upstream of the start codon is shown and the putative ArcA-binding site is underlined. Putative - 10 and - 35 elements are circled. B, EMSA of phosphorylated ArcA (ArcA-P) with the promoter region (Peut). The arrow denotes free DNA and the asterisk indicates DNA-protein complexes. NC: negative controls in which a DNA fragment with random sequence was utilized. C, A schematic diagram of the eut* construct. D, RT-qPCR analyses of eutA mRNA levels in the WT, eut*, and ΔarcA strains under anaerobic conditions. E, Competitive growth index of the WT versus eut* strains at various time points in vitro. F, Competitive growth index of the WT versus eut* strains at 24 and 48 h upon infection of C. elegans.

To test the physiological significance of this genetic repression, we constructed a mutant strain (designated as eut*) in which the original eut promoter was replaced by a constitutive one (Fig. 3C). RT-qPCR validated that the expression of the eut genes (using eutA as an example) in this strain was de-repressed in the LB-MOPS-X medium under anaerobic conditions, slightly higher than that in the ΔarcA strain (Fig. 3D). Next, we performed competition assays of the eut* strain with its parental strain both in vitro and in an in vivo C. elegans infection model. Indeed, the eut* mutant exhibited competitive disadvantages both in vitro (Fig. 3E) and in the C. elegans intestine in vivo (Fig. 3F). Taken together, these results suggest that the regulation of the eut operon by ArcA is important to S. Typhimurium physiology and fitness in the host.

ArcA Regulates the TCA Cycle In Both Direct and Indirect Manners

The largest group of ArcA-repressed proteins encompasses enzymes in the central carbon metabolism (Table I), in particular those in the TCA cycle. For instance, both GltA and IcdA were of higher levels in the ΔarcA cells than in the WT (Fig. 4A). Consistently, RT-qPCR analysis showed the repression of these genes by ArcA on the transcript level (Fig. 4B). Intriguingly, a transcriptional regulator YdcI was also elevated in the ΔarcA cells (Fig. 4C). Previously in E. coli, YdcI was shown to regulate the expression of gltA and control carbon flux to the TCA cycle (25). S. Typhimurium YdcI shares 83% identity with its E. coli homologue. To examine its contribution to the ArcA regulatory network, we constructed strains expressing chromosomally 3×FLAG-tagged YdcI in both the WT and ΔarcA backgrounds. Immunoblotting analysis confirmed the increased expression of YdcI in the ΔarcA cells (Fig. 4D). Consistently, RT-qPCR analyses revealed substantially higher levels of the ydcI transcript in the ΔarcA cells than in the WT (Fig. 4E). Together these data suggest that S. Typhimurium YdcI may also play a role in regulating the TCA cycle.

Fig. 4.

Repression of the TCA cycle enzyme by ArcA. A, Representative mass spectra of dimethyl labeled peptides from GltA and IcdA. Peptides in the WT and ΔarcA samples with light and heavy labels are indicated by open and filled triangles respectively. B, RT-qPCR analyses of icdA, icdA, and sdhC mRNA levels in the WT and ΔarcA bacteria grown under anaerobic conditions to an OD₆₀₀ of ∼0.3. C, Representative mass spectra of dimethyl labeled peptides from YdcI. Peptides in the WT and ΔarcA samples with light and heavy labels are indicated by open and filled triangles respectively. D, Immunoblotting analyses of YdcI-3×FLAG in the WT and ΔarcA strains. GroEL was used as a loading control. E, RT-qPCR analyses of ydcI mRNA levels in the WT and ΔarcA strains. F, RT-qPCR analyses of the mRNA levels of those genes associated with the TCA cycle in the WT and ΔydcI strains. G, The incoherent feed-forward regulation loop in the regulation of TCA enzymes by AcrA and YdcI.

To test this hypothesis, we constructed a bacterial strain lacking ydcI (ΔydcI) and measured the mRNA levels of all nine genes in the TCA cycle (gltA, acnB, icdA, sucA, sucC, sdhC, fumA, fumB, and mdh) (Fig. 4F). Notably, we observed significantly higher transcript levels of these genes in the ΔydcI cells compared with the WT, indicating their transcriptional repression by YdcI. The transcription of acnB, icdA, sdhC, and fumA genes was most repressed by YdcI (about 50–100 fold higher in the ΔydcI strain). Nevertheless, the overall magnitude of the transcriptional changes of these TCA genes in the ΔydcI strain was markedly less than that in the strain lacking arcA. For instance, the gltA transcript was 25-fold more abundant in the ΔydcI strain whereas its level was 479-fold higher in the ΔarcA mutant compared with the WT. Collectively, these data suggest that the repression of the TCA cycle by ArcA is, at least in part, mediated by its regulation on YdcI. In other words, ArcA may both directly and indirectly repress the expression of the TCA genes (Fig. 4G). Also observed in other regulatory systems (26), such regulation patterns were termed as an incoherent feed-forward regulation loop (Fig. 4G), providing a means of fine-tuning the gene expression during stress adaptations.

Involvement of Post-transcriptional Regulation of SPI-1-encoded Virulence Factors and Other Pathways In the ArcA Regulon

Network analyses of ArcA-activated proteins revealed several distinct clusters including ribosomal proteins, hydrogenases, proteins involved in maltose utilization and sulfate assimilation, and SPI-1-encoded virulence factors (Fig. 1B). To determine whether similar regulation occurs on the transcript level, we performed RT-qPCR analyses of selective genes in these pathways. Consistent with their proteomic changes, the mRNA levels of the ribosome genes rplC and rplX were substantially lower in the ΔarcA cells (Fig. 5A). In comparison, the mRNA level of hybA dropped by ∼ 60% in the ΔarcA cells compared with the WT, and that of hybE was slightly decreased (Fig. 5B). Furthermore, the transcripts of the maltose utilization genes malE, lamB, and the sulfate assimilation gene cysN were more abundant in the ΔarcA cells compared with the WT (Fig. 5C–5D), opposite to the down-regulation of their protein products. These discrepancies between mRNA and proteomic measurements may indicate the presence of post-transcriptional regulation.

Fig. 5.

Proteins or pathways positively regulated by ArcA. RT-qPCR analyses of the mRNA levels of selected genes in anaerobically grown WT and ΔarcA strains: A, Ribosomal proteins, B, Hydrogenase-2 proteins, C, Maltose utilization proteins, D, Sulfate assimilation proteins, and E, SPI-1-associated virulence factors. *, p < 0.05. **, p < 0.01. ***, p < 0.01. p values were calculated by using the Student's t test. F, Immunoblotting analyses of SipA-3×FLAG in the culture supernatants and cell lysates of the WT and ΔarcA strains. GroEL was used as a loading control.

Furthermore, SPI-1-encoded proteins such as PrgH, HilA and SipA were less abundant in the ΔarcA cells. RT-qPCR analysis found that the prgH transcript was largely unchanged whereas the mRNA levels of hilA and sipA were higher in the ΔarcA mutant than in the WT cells (Fig. 5E). Next, we constructed bacterial strains expressing chromosomal SipA-3×FLAG in both the WT and ΔarcA backgrounds. Immunoblotting analyses of total cell lysates showed markedly lower levels of SipA in the ΔarcA mutant cells relative to the WT (Fig. 5F). Furthermore, we probed SipA in bacterial culture supernatants and found significantly less secretion of this effector in the ΔarcA cells as well (Fig. 5F). Collectively, these data suggest the repression of S. Typhimurium SPI-1 T3SS in the ΔarcA cells is likely because of post-transcriptional regulatory mechanisms.

Salmonella Lacking arcA Exhibits Competitive Disadvantage and Lower Cellular ATP Levels

Given the global regulatory roles of ArcA in carbon and energy metabolism as well as protein synthesis, next we examined the physiological consequences of arcA deletion in S. Typhimurium. We first measured the competition index of the WT versus ΔarcA cells when both strains were co-cultured under anaerobic conditions. As shown in Fig. 6A, ΔarcA cells were significantly outcompeted by the WT bacteria. Furthermore, we measured the cellular ATP production and found that the ATP level in cells lacking arcA was only half of that in the WT bacteria (Fig. 6B), supporting the role of ArcA in regulating central metabolism of S. Typhimurium.

Fig. 6.

Impacts of arcA-deficiency on S. Typhimurium physiology. A, Competitive growth index of the WT and ΔarcA strains at various time points. B, Cellular ATP levels of the WT and ΔarcA strains grown anaerobically to an OD₆₀₀ of ∼0.3.

DISCUSSION

S. Typhimurium has evolved several signal transduction systems to sense the transition from ambient conditions in vitro to anaerobiosis in vivo, leading to the reprogramming of bacterial respiratory pathways (27). One of these is the ArcAB two-component system, which controls the expression of a wide spectrum of important bacterial pathways. We present a proteomic landscape of S. Typhimurium ArcA regulon by quantitative profiling of differentially expressed proteins in an arcA deletion strain. In addition to those findings overlapping with previous transcriptomic studies, importantly we found several distinct pathways under the regulation of ArcA. Furthermore, some pathways exhibit different/opposing regulation patterns between proteins and corresponding transcripts.

Ethanolamine catabolism is an important physiological process employed by many bacteria residing in the intestines of mammals, including both commensal and pathogenic species (28–31). Ethanolamine is a compound that can be readily derived from the breaking down of phosphatidylethanolamine in the membrane lipids of both host and microbiome (32). Genes encoded in the eut operon are responsible for the catabolism of this carbon and nitrogen source to yield ammonia and acetyl-CoA (33). Nonetheless, in the intestinal tract, whether S. Typhimurium utilizes/catabolizes ethanolamine depends on the diet of the host (i.e. whether other carbon sources are available such as glucose, lactose etc.) and the physio-immunological status of the intestines (i.e. whether the intestine is undergoing infection and inflammation etc.). The genetic content and the local regulator of the eut operon differs in different species (34). For instance, the eut operon in the commensal bacterium Enterococcus faecalis is primarily regulated by the two-component system EutVW (35). In S. Typhimurium it is regulated by an AraC family transcription regulator EutR, which is in an adjacent, separate operon from that of the functional eut genes (36) (Fig. 2A). The entire eut operon in S. Typhimurium includes 17 genes and the enzymes are organized in a multiprotein complex termed as carboxysome or metabolosome (37). Given the large number of genes in this operon, it is not surprising that in addition to the local regulator EutR, its regulation is also coordinated with the global carbon and energy metabolism of S. Typhimurium (i.e. its expression is regulated by the global regulator ArcA).

Our finding of ArcA repression of the eut operon is, however, in direct contrast to the previous transcriptomic study (14). One reason that may account for this discrepancy is the different strains used (i.e. 14028s versus SL1344). We performed RT-qPCR analysis following the description of Evans et al., though we were unable to achieve the amplicon of the reference gene rpoD used in their study. Nonetheless, our in vivo transcription assay using P_eut-lacZ confirmed the repression of the eut genes by ArcA, and further EMSA experiments showed the direct binding of AcrA to the promoter of the eut operon. These findings, combined with the absence of ethanolamine in the culture medium used in our assay (and in the assay by Evans et al.) and the large number of genes in the eut operon (i.e. their expression would be energy-consuming), strongly support the notion that ArcA represses the expression of eut genes in the rich medium (LB-MOPS-X) under anaerobic conditions. Our competition assays further proved that dysregulation of the eut operon leads to competitive disadvantages both in vitro and in C. elegans intestine during in vivo infection, underlying the importance of ArcA-mediated repression of this operon in the physiology and fitness of S. Typhimurium.

In addition, our proteomic dataset suggests ArcA activates the expression of S. Typhimurium SPI-1 T3SS, which is required for bacterial invasion of intestinal epithelial cells (38) and controlled by the transcriptional regulator HilA (39). Western-blot analysis further confirmed the decreased production and secretion of SipA. In contrast, RT-qPCR analysis revealed higher transcript levels of hilA and sipA in the ΔarcA cells, suggesting the involvement of post-transcriptional or post-translational regulation of these proteins. Known components that could contribute to the post-transcriptional regulation of SPI-1 genes include the regulators HilD, HilC, and RtsA which individually exert their regulatory effects through HilA (40). However, we did not detect the expression changes of these proteins. Another protein InvB (a chaperone) was shown to affect the abundance of SipA at the post-transcriptional level (40). It was reported that InvB does not alter the transcription of sipA but its absence results in reduced SipA levels in S. Typhimurium (41). However, InvB was not detected in our study probably owing to its low abundance. Given the important roles of SPI-1 T3SS in S. Typhimurium virulence, further work would be warranted to identify the regulatory circuits that mediate the production and secretion of SipA post-transcriptionally and/or post-translationally. Nonetheless, it is not unexpected that global anaerobic regulators participate in the regulation of bacterial virulence. For example, another global anaerobic regulator FNR has been shown to mediate the virulence of Shigella flexneri in the low oxygen environment of animal intestinal tract through regulating T3SS components and Ipa effector proteins (42). Together, these findings highlight the highly orchestrated regulation of bacterial metabolism, physiology, and virulence at the host-pathogen interface, which is increasingly recognized in the field.

DATA AVAILABILITY

The proteomics data reported in this paper have been deposited to the iProx database (URL:http://www.iprox.org/page/HMV006.html) and are available under the accession number IPX0001126001.