Page 1

Modulation of pertussis and adenylate cyclase toxins by sigma factor RpoE in Bordetella 1 pertussis 2

3

Short title: Sigma factor RpoE of B. pertussis 4

5

Mariette Barbiera, Dylan Boehma, Emel Sen-Kiliça, Claire Bonnina, Theo Pinheiroa, Casey 6 Hoffmanb, Mary Grayb, Erik Hewlettb, F. Heath Damrona 7

a Department of Microbiology, Immunology, and Cell Biology, West Virginia University, 8 Morgantown, WV, USA 9

b Department of Medicine, Division of Infectious Diseases and International Health, University of 10 Virginia, Charlottesville, Virginia, USA 11

12

*Corresponding author 13

Email: fdamron@hsc.wvu.edu 14

15

16

IAI Accepted Manuscript Posted Online 14 November 2016Infect. Immun. doi:10.1128/IAI.00565-16Copyright © 2016, American Society for Microbiology. All Rights Reserved.

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 2

Abstract 17

Bordetella pertussis is a human pathogen that can infect the respiratory tract and cause the 18

disease known as whooping cough. B. pertussis uses pertussis toxin (PT) and adenylate 19

cyclase toxin (ACT) to kill and modulate host cells to allow the pathogen to survive and persist. 20

B. pertussis encodes many uncharacterized transcription factors and very little is known about 21

their functions. RpoE is a sigma factor which in other bacteria, responds to oxidative, heat, and 22

other environmental stresses. RseA is a negative regulator of RpoE that sequesters the sigma 23

factor to regulate gene expression based on conditions. In B. pertussis, deletion of the rseA 24

gene results in high transcriptional activity of RpoE and high amounts of secretion of ACT. By 25

comparing parental B. pertussis to a rseA gene deletion mutant (PM18), we sought to 26

characterize the roles of RpoE in virulence and determine the regulon of genes controlled by 27

RpoE. Despite high expression of ACT, the rseA mutant strain did not infect the murine airway 28

as efficiently as the parental strain and PM18 was killed more readily when inside phagocytes. 29

RNAseq analysis was performed and 263 genes were differentially regulated by RpoE and 30

surprisingly the rseA mutant strain where RpoE activity was elevated, expressed very little 31

pertussis toxin. Western blots and proteomic analysis corroborated the inverse relationship of 32

PT to ACT expression in the high RpoE activity rseA deletion strain. Our data suggest that 33

RpoE can modulate PT and ACT expression indirectly through, unidentified mechanisms, in 34

response to conditions. 35

36

37

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 3

Introduction 38

Whooping cough is a disease caused by an upper respiratory tract infection with the 39

obligate human pathogen Bordetella pertussis (1). In the past, incidence of whooping cough 40

was low due to an effective whole cell vaccine that was subsequently replaced by an acellular 41

vaccine in the 1990’s. Acellular pertussis vaccines are typically composed of 3-5 B. pertussis 42

protein antigens and are always provided as a combination vaccine with diphtheria and tetanus 43

toxoids (DTaP). Since the switch to the acellular vaccines, the incidence of whooping cough 44

has significantly increased to a 50 year high in 2012 of 48,277 cases in the US. Studies have 45

shown that protection against pertussis wanes during the five years after the fifth and final dose 46

of DTaP in children (2). There are multiple hypotheses regarding the resurgence of pertussis. 47

Recent studies in the baboon infection model indicate that while the acellular vaccine can 48

prevent disease manifestation (whooping cough), the acellular vaccine does not prevent 49

transmission and colonization with B. pertussis (3, 4) and as a result may be affording 50

asymptomatic transmission of the pathogen around the population. Overall, there are many 51

longstanding questions about pertussis that remain and need to be answered in order to 52

understand how this pathogen infects to develop more effective vaccines (5). 53

B. pertussis is transmitted by respiratory droplets and once inside the host, the bacteria 54

will adhere to the respiratory epithelia cells of the airway (6). After attachment, the pathogen 55

expresses multiple toxins (1) including: the pertussis toxin (PT) and adenylate cyclase toxin 56

(ACT). While PT has long range effects on the host, the ACT is thought to act more locally. 57

Both of these toxins are necessary for establishment of infection in murine models (7), but it 58

seems pertussis toxin plays an earlier role in the establishment of infection. The combo of 59

pertussis toxin and ACT provide a “one-two punch” that affects neutrophil recruitment and 60

survival of B. pertussis in the respiratory tract (7). The overall virulence of the pathogen is 61

linked to expression of these toxins through what is known as the Bordetella virulence gene 62

system, which is composed of the BvgS sensor kinase and the BvgA response regulator (1). 63

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 4

Several in vitro conditions have been shown to induce BvgS phosphorylation of BvgA, which 64

subsequently cause gene expression at specific genes throughout the genome. BvgAS controls 65

expression of hundreds of genes (8) that are necessary and sufficient for virulence of Bordetella 66

(9) in murine models of infection (10). These genes can be classified into three phenotypic 67

phases: Bvg- (ex vivo), Bvgi (transmission) (11), and Bvg+ phase (in vivo) (1). 68

It is apparent that Bvg-dependent genes play key roles during the infectious process; 69

however, there are likely other transcription factors that modulate gene expression and 70

influence B. pertussis virulence. Encoded in the reference Tohama I genome, there are at least 71

10 putative sigma factors, in addition to the housekeeping RpoD (data not shown). Little is 72

known about most of the sigma factors, except for the extracytoplasmic sigma factor RpoE (σE / 73

SigE). In B. brontiseptica SigE has been shown to facilitate survival of membrane perturbations 74

and is required for virulence in immunocompromised mice (12). It has also been shown that in 75

B. pertussis RpoE influences alleviation of membrane stress (13). RpoE is controlled by an 76

anti-sigma factor known as RseA. RseA sequesters RpoE to the inner-membrane which 77

renders it transcriptionally inactive; however, proteolytic cleavage of RseA or genetic mutation to 78

rseA results in transcriptional activation of RpoE. Deletion of rseA from B. pertussis strain UT25 79

resulted in an increased amount and diversity of periplasmic proteins (13). Surprisingly, it was 80

shown that the rseA gene deletion mutant (PM18) of the B. pertussis strain UT25 also released 81

a high amount of adenylate cyclase toxin (ACT) into the culture media (13). Not only was more 82

ACT released, but it was of greater stability(13). Furthermore, cell membrane stress caused by 83

hydrogen peroxide, activated increased release of ACT (13). 84

Here in this study, we observed that in the absence of the RseA anti-sigma factor (PM18 85

strain), B. pertussis was less able to infect the lungs of outbred mice or survive in J774A.1 cells 86

or neutrophils. Since RpoE is a sigma factor and could have global effects on gene and protein 87

expression.We hypothesized that there could be alterations in the pertussis toxin expression 88

profile of this strain. We used RNA sequencing (RNAseq) and shotgun proteomics to identify 89

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 5

the genes and proteins that are controlled by RpoE to expand upon what we know about B. 90

pertussis gene expression. We observed that in the absence of the anti-sigma factor RseA, 91

there was an inverse relationship between expression of genes encoding ACT and PT. While 92

we observed increased ACT, decreased PT was measured both at the level of mRNA and 93

protein. Our data indicate the RpoE system can influence expression of virulence factors and 94

likely plays a role in pathogenesis. 95

96

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 6

Materials and Methods 97

Bacterial strains and growth conditions for murine challenge. 98

B. pertussis strains BP338 (wild type; Tohama I), UT25 (UT25Sm1) (14) and PM18 99

(UT25Sm1ΔrseA) (13) were cultured on Bordet Gengou (BG) agar (15) (remel) supplemented 100

with 15% de-fibrinated sheep blood (Cocalico Biologicals) for 48 h at 36°C. Bacteria were 101

transferred from BG plates to three flasks of 12 ml of modified Stainer-Scholte liquid medium 102

(SSM) (16). SSM cultures were not supplemented with cyclodextrin unless indicated. SSM 103

cultures were grown for ~22 h at 36°C with shaking at 180 rpm until the OD600 reached 0.8. The 104

cultures were then diluted to provide a challenge dose of 2 x 107 CFU in 20 μl. Four week old 105

CD1 Mice were obtained from Charles River. Mice were anesthetized by intraperitoneal 106

injection of ketamine and xyalzine in saline. Two 10 μl doses of the B. pertussis strain were 107

pipetted directly into each nostril of the mouse. Five to seven mice were infected with strains 108

UT25 or PM18 and at 24, 48, and 144 h post challenge, mice were euthanized for determination 109

of bacterial burden in the nasal wash, trachea, and lungs. To determine the number of B. 110

pertussis in the nares, 1 ml of PBS was flushed up through the nares and collected. Trachea 111

and lungs were extracted and homogenized. Serial dilutions in PBS were plated on BG 112

containing streptomycin (100 μg/ml) to ensure only UT25 or PM18 B. pertussis were cultured. 113

All murine infection experiments were performed according to protocols approved by the 114

University of Virginia Animal Care and Use Committee (protocol number 4004), conforming to 115

AAALAC International accreditation guidelines. 116

117

J774A.1 macrophage intracellular killing assay. 118

J774A.1 macrophages (ATCC) were cultured at 37°C in 5% CO2 in Dulbecco’s modified Eagle 119

medium (DMEM) containing 10% heat inactivated fetal bovine serum (FBS), and 1% 120

penicillin/streptomycin. J774A4.1 cells were counted with a hemocytometer and diluted, 106 121

cells well were added to a 24-well tissue culture plate, and incubated for 2 h. After which, 122

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 7

DMEM was removed and wells washed 3 times with PBS. B. pertussis strains were grown as 123

described above and diluted to 108 CFU/ml. Cells were infected with either UT25 or PM18 by a 124

multiplicity of infection (MOI) of 50. Dose of infection was quantified by plating of serial dilutions 125

on BG plates. To facilitate bacterial interaction with J774A.1, cells tissue culture plates were 126

centrifuged at 550 x g for 5 mins, then incubated for 1 h at 37°C in 5% CO2. Non-adherent 127

bacteria were removed with medium and wells were washed 3 times with PBS. DMEM with 100 128

µg/ml polymyxin B sulfate (TOKU-E) was added to each well to kill extracellular bacteria, and 129

incubated for 1 h. After incubation, antibiotic concentration was reduced to 5 µg/ml to stop 130

bacterial replication. Intracellular viable bacteria levels were determined at 2 and 24 h post 131

infection. At each time point, cells were washed 3 times with PBS, then lysed by 5 min 132

incubation in sterile water. Wells were scraped and serial dilutions of lysates were plated on BG 133

plates to determine colony forming unit counts (CFU). Student’s t-tests at a 95% confidence 134

level were used to analyze 3 biological replicates of each experimental condition. 135

136

Bone marrow derived neutrophil (BMDN) isolation. 137

Bone marrow was isolated from femurs and tibias of 2 wild type 129S2/SvHsd (Envigo) mice. 138

Bone marrow was flushed using 15 ml of DMEM. Cell solution was centrifuged at 1000 x g for 5 139

mins, then re-suspended in 5 ml of DMEM. Neutrophils from cell suspension were isolated by 140

addition of 5 ml of Lympholyte-M (Cedarlane) to cell suspension, then centrifuged at room 141

temperature for 20 mins at 1,200 x g. The gradient layer corresponding to the neutrophils was 142

re-suspended in red blood cell (RBC) lysis buffer (BD Pharm Lyse, BD Biosciences) and 143

incubated for 2 min at 37°C to lyse the red blood cells, then centrifuged at 200 x g for 5 min. 144

Supernatant was removed and neutrophils re-suspended in DMEM before counting by 145

hemocytometer. Neutrophil population was visually inspected after Kwik diff staining 146

(ThermoFisher) and light microscopy. 147

148

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 8

149

150

BMDN intracellular and extracellular killing assay. 151

Neutrophil cellular killing assays were performed similarly to the macrophage assays described 152

above. Neutrophil suspensions were diluted and 106 cells were added to 1.5 ml tubes. Cells 153

were centrifuged at 1,000 g for 5 mins, and supernatant removed. B. pertussis strains were 154

grown as described above and diluted to 107 CFU/ml for the extracellular killing assay and 108 155

CFU/ml for the intracellular killing assay. Cells were infected with a MOI of 200, and incubated 156

at 37°C in 5% CO2 for 1 h for the neutrophil uptake assay, and 2 h for the extracellular killing 157

assay. After incubation the cell suspensions were centrifuged at experimental conditions. At this 158

point the supernatant from the cell suspensions used for the extracellular killing assay were 159

serially diluted and plated on BG plates. Cell suspensions used for the neutrophil uptake assay 160

were washed once with PBS. Then DMEM with 100 µg/ml polymyxin B sulfate (TOKU-E) was 161

added to each well to kill extracellular bacteria, and incubated for 1 h. After incubation, cells 162

were centrifuged, lysed with sterile water, and then lysate serial dilutions plated on BG plates, 163

as previously described in the macrophage assay. Student’s t-tests at a 95% confidence level 164

were used to analyze 3 biological replicates of each experimental condition. 165

166

Growth conditions for RNA and protein sample preparation for RNAseq and proteomics. 167

B. pertussis strains UT25 and PM18 were cultured on BG agar at 36°C for 48 hr. B. pertussis 168

cells were collected from the plate using a swab and transferred to three flasks of 12 ml of 169

modified Stainer-Scholte liquid medium (SSM) (16). SSM cultures were grown for ~22 h at 170

36°C with shaking at 180 rpm at which time the OD 600 was 0.8. The SSM cultures were then 171

transferred to 200 ml of SSM and OD600 matched to 0.1. After 22 h at 36°C with shaking of 180 172

rpm, the OD600 of each culture was 0.8. One ml of cells from each culture was pelleted (20,000 173

x g 1 min) and re-suspended in RNAprotect Bacteria Reagent (Qiagen) to stabilize the RNA in 174

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 9

the cells. The suspension was then pelleted (20,000 x g 1 min) and 900 µl of RNAprotect 175

supernatants were removed by pippetting. The RNAprotect-cell samples were then stored at -176

80°C until RNA was isolated. The remaining ~200 ml of SSM culture was pelleted by 177

centrifugation (12,000 x g; 4°C; 1 h). 178

179

Isolation of RNA, library construction, and Illumina sequencing 180

RNA was isolated using RNeasy Mini Kit (Qiagen) as specified by the instructions of the 181

manufacturer. The resulting RNA was pooled from the three technical samples for each strain 182

and treated with RNase-free DNase (Qiagen). To remove the DNase, the samples were then 183

cleaned up on another RNeasy Mini column. The resulting RNA was quantified on a Nanodrop 184

ND-1000 (Nanodrop). Next, the RNA integrity was assessed by running the samples on an 185

Agilent BioAnalyzer RNA Pico chip. All samples were observed to have RNA integrity scores of 186

10 and then submitted to two rounds of Ribo-zero rRNA depletion (illumina) and reassessed for 187

RNA integrity. rRNA depleted mRNA samples were then fragmented and prepared into libraries 188

that were then sequenced on an Illumina MiSeq by the University of Virginia Department of 189

Biology Genomics Core. One biological sample of UT25 and PM18, were sequenced on one 190

Miseq lane, thus using three lanes in total for 81 million 2 x 76 bp reads. 191

192

RNAseq and bioinformatics analyses 193

The reads were aligned to the B. pertussis Tohama I genome (17) using CLC Genomics 194

Workbench version 8. Reads per kilobase per million (RPKM) and fold change for each gene 195

were calculated. Empirical analysis of digital gene expression (EDGE) was performed to 196

determine differentially expressed genes (18). Genes with EDGE test p value of <0.05 were 197

considered differentially regulated. Gene set enrichment (GSEA) (19), hypergeometric tests on 198

annotations (20), and STRING database analysis (21) were performed to determine the 199

systems of related genes and pathways that were differentially expressed. Subcellular 200

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 10

localization for the B. pertussis reference strain Tohama I gene products was obtained from 201

pSORTb V 3.0 (22). Raw data reads were submitted to the Sequence Read Archive (SRA) and 202

can be access under reference number SRP076127. 203

204

RNA isolation and reverse transcriptase PCR (qRT-PCR). 205

B. pertussis strains UT25 and PM18 were cultured in the same conditions described above for 206

RNAseq analysis. Briefly, cells were stored at -80°C in RNAprotect (Qiagen), cells were lysed 207

with lysozyme, and RNA was isolated with RNeasy Mini kits (Qiagen). DNA was digested using 208

off-column digestion with RNase-free DNase and re-isolated with another RNeasy column. 209

RNA concentration and quality was assessed on a Molecular Devices i3 Spectramax Spectra 210

drop plate. To ensure RNA was DNA-free, 25 ng of RNA was checked by PCR amplification 211

and was only used for cDNA if no amplicon was observed and a CT of >32. cDNA was 212

synthesized using M-MLV reverse transcriptase (Promega) per the manufacturer’s instructions 213

using 250 ng of RNA and gene specific reverse primers for targets. Twenty five microliter qPCR 214

mixtures were setup with SYBR Green PCR master mix (Applied Biosystems), per 215

manufacturer’s instructions using 1 μl of cDNA. A minimum of three technical replicate 216

reactions were ran per gene target per sample on a Step One Plus qPCR thermocycler (Applied 217

Biosystems). Primers were designed on Primer3 (Primer-Blast; NCBI) and checked for 218

specificity by PCR. Melt curve analysis as well as subsequent agarose gel electrophoresis were 219

performed on all reactions. Gene expression was normalized to the rpoB reference using the 2-220

∆∆CT method (23). For statistical analysis, the ∆CT for the three biological replicate experiments 221

was calculated and a Student’s t-test was performed in Microsoft Excel 2013. Standard error of 222

mean was calculated based on the interval of the ∆CT of the three biological replicates. The 223

following primers sequences used in this study were described in a study by Bibova et al. (24): 224

cyaF(CGAGGCGGTCAAGGTGAT), cyaR(GCGGAAGTTGGACAGATGC), 225

ptxAF(CCAGAACGGATTCACGGC), ptxAR (CTGCTGCTGGTGGAGACGA), bvgAF 226

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 11

(AGGTCATCAATGCCGCCA), bvgAR (GCAGGACGGTCAGTTCGC), fhaBF 227

(CAAGGGCGGCAAGGTGA), fhaBR (ACAGGATGGCGAACAGGCT), rpoBF 228

(GCTGGGACCCGAGGAAAT), rpoBR (CGCCAATGTAGACGATGCC). Primers for BP2497 229

were designed for this study: BP2497F (TCGGATCGCACCAATTACTTC) and BP2497R 230

(CCTTGGCGATCAGCGAGTT). 231

232

Western blot analysis. 233

B. pertussis strains were grown on BG for 48 h and then transferred to SSM and grown at 36°C 234

until reaching an OD 600 of 0.4 PTX or 0.8 for ACT blots. A minimum of four biological replicates 235

were grown for each strain. Total culture samples were directly taken from the culture. A 236

replicate set of total cultures were pelleted at 6,000 x g 3 min at 4 °C. 15 μl of total culture, 237

pellets, and supernatant samples were mixed with 15 μl Laemmli buffer (25) and boiled for 5 238

min. Samples were then cooled on ice and loaded onto 12% polyacrylamide gels for 239

electrophoresis (Mini-PROTEAN II Cell; Bio-rad). Gels were then electroblotted onto PVDF 240

membranes (Micro Separations). To ensure proper loading and transfer, Memcode (Pierce) 241

was used to reversibly stain the PVDF membranes. The membrane was blocked with 5% nonfat 242

dry skim milk TBS-Tween (0.01%). The primary antibodies were diluted 1:2000 in blocking 243

buffer. The membrane was probed 2 hr with either mouse monoclonal anti-pertussis toxin 244

(ThermoFisher; clone number 1280/204) or mouse monoclonal antibody 3D1 (26). The 245

membranes were then washed with PBS-Tween (0.1%) and probed with goat anti-mouse IgG 246

HRP (Affinpure). ECL chemiluminescence (Amersham) was used to detect HRP-labeled 247

secondary antibodies. Signal was detected on a Bio-rad Chemidoc Touch imaging system. 248

249

Measurement of AC enzymatic activity. AC enzymatic activity was measured by conversion 250

of [α-32P]ATP to [32P]cAMP in a cell-free assay as described previously (27). 251

252

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 12

Biofilm formation assay. To measure the biofilm formed by B. pertussis strains, the microtiter 253

crystal violet plate assay was used. Bacteria were grown in 96-well polyvinylchloride (PVC), 254

non-tissue- culture treated microtiter plates (Nunc). 100 µL cultures were started at OD 600 of 255

0.1 from BG grown strains. Biofilm was measured at 96 h. Briefly, planktonic and loosely 256

adherent bacteria were removed and wells were washed. Bacteria remaining adherent to wells 257

were stained with 0.1% crystal violet, washed, and dye was solubilized with 95% ethanol. 258

Crystal Violet absorbance was measured at OD595 (28). 259

260

Proteomic mass spectrometry 261

Using replicate samples of UT25 and PM18 at OD600 0.8, protein lysates were prepared by 262

lysing cells in Mass Spec grade ProteaPrep Anionic Cell lysis buffer (protea). DNA was 263

fragmented by three 30 s ultrasonic bursts to decrease viscosity. DC protein assay (Bio-rad) 264

was used to measure amount of protein. Ten μg of UT25 and PM18 protein lysates were 265

loaded on 12% polyacrylamide gels for electrophoresis (Mini-PROTEAN II Cell; Bio-rad). The 266

gel was then stained with Coomassie brilliant blue to visualize protein bands. The University of 267

Virginia Proteomic Core performed the sample preparation, separation, and mass spectrometry 268

analysis. For each sample gel, pieces (10 slices per sample) were transferred to a siliconized 269

tube and washed and de-stained in 200 µL 50% methanol for 3 hours. The gel pieces were 270

dehydrated in acetonitrile, rehydrated in 30 µL of 10 mM dithiolthreitol in 0.1 M ammonium 271

bicarbonate and reduced at room temperature for 0.5 h. The DTT solution was removed and 272

the sample alkylated in 30 µL 50 mM iodoacetamide in 0.1 M ammonium bicarbonate at room 273

temperature for 0.5 h. The reagent was removed and the gel pieces dehydrated in 100 µL 274

acetonitrile. The acetonitrile was removed and the gel pieces rehydrated in 100 µL 0.1 M 275

ammonium bicarbonate. The pieces were dehydrated in 100 µL acetonitrile, the acetonitrile 276

removed and the pieces completely dried by vacuum centrifugation. The gel pieces were 277

rehydrated in 20 ng/µL trypsin in 50 mM ammonium bicarbonate on ice for 10 min. Any excess 278

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 13

enzyme solution was removed and 20 µL 50 mM ammonium bicarbonate added. The sample 279

was digested overnight at 37 °C and the peptides formed extracted from the polyacrylamide in 280

two 30 µL aliquots of 50% acetonitrile/5% formic acid. These extracts were combined and 281

evaporated to 15 µL for MS analysis. The LC-MS system consisted of a Thermo Electron 282

Orbitrap Velos ETD mass spectrometer system with a Protana nanospray ion source interfaced 283

to a self-packed 8 cm x 75 um id Phenomenex Jupiter 10 µm C18 reversed-phase capillary 284

column. A 7 µL aliquot of the extract was injected and the peptides eluted from the column by 285

an acetonitrile/0.1 M acetic acid gradient at a flow rate of 0.5 µL/min over 1.2 hours. The 286

nanospray ion source was operated at 2.5 kV. The digest was analyzed using the data 287

dependent capability of the instrument acquiring full scan mass spectra to determine peptide 288

molecular weights followed by product ion spectra to determine amino acid sequence in 289

sequential scans. This mode of analysis produces approximately 50,000 MS/MS spectra of ions 290

ranging in abundance over several orders of magnitude. Not all MS/MS spectra are derived 291

from peptides. The data were analyzed by database searching using the Sequest search 292

algorithm against Uniprot Bordetella pertussis Tohama I proteome. The peptide data was then 293

loaded into Scaffold software version 4. To normalize for protein length, peptide counts were 294

submitted to normalized spectral abundance factor (NSAF) (29). NSAF-normalized abundance 295

was then ranked as percentile within the sample. Percentile difference was calculated to 296

demonstrate the relative abundance differences between proteins in the UT25 and PM18 297

proteomes. 298

299

Results and Discussion 300

It was previously shown that the B. pertussis PM18 strain (UT25 with an in-frame rseA 301

gene deletion) produced and secreted more ACT than the parental UT25 strain (13). Based on 302

this, we hypothesized that the PM18 strain, would be highly fit in vivo and efficiently infect the 303

airways of mice, due to the roles of ACT in facilitating infection. To test this hypothesis, we 304

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 14

infected outbred CD1 female mice with 2 x 107 CFU of either UT25 (parental strain) or PM18 305

(UT25∆rseA) by intranasal infection and determined the bacterial burden of the nares (by nasal 306

wash), trachea, and lungs at 24, 48, and 144 h post challenge. In the nasal wash, we observed 307

statistically fewer viable PM18 at 144 hr post challenge, but otherwise no difference at 24 or 48 308

h (Fig. 1A). It is important to point out that a nasal wash only contains the bacteria that can be 309

flushed out of the nares and likely does not equal the total number of viable bacteria in the 310

nares. In the trachea, we observed no statistical differences between the bacterial burdens and 311

any of the three time-points (Fig. 1B). Unlike the trachea, the bacterial burdens in the lungs 312

were strikingly different. At 24 h post challenge, the amount of viable PM18 bacteria in the lung 313

was statistically increased over UT25, which would be consistent with our hypothesis. At 48 h 314

there was a statistically significant increase in viable UT25 bacteria compared to PM18. These 315

data, while puzzling to rationalize, suggest that in the absence of rseA, highly activated RpoE 316

affected gene expression to an extent that altered virulence, even though both strains were in 317

Bvg+ phase, upon visual inspection on BG agar, both before and after the murine challenges. 318

Based on the murine challenge data, we next wondered if PM18 was more susceptible to 319

cellular killing by macrophages or neutrophils. To test this, J774A.1 cells were infected with 320

UT25 or PM18 (Fig. 1D). At both 2 and 24 h post infection, intracellular PM18 was more 321

susceptible to killing by the J774A.1 cells. To expand on these data, bone marrow derived 322

neutrophils were isolated and infected in a similar manner to the J774A.1 assay. Intracellular 323

and extracellular survival of UT25 and PM18 were determined. Statistically less PM18 were 324

viable after 2 hr inside the neutrophils (Fig. 1E). However, extracellular UT25 and PM18 both 325

survived 2 hr incubation with neutrophils. The decreased survival in mice and inside cells were 326

the exact opposite results that would be predicted by our hypothesis. For this reason, we 327

conducted additional analyses to identify the genes controlled by the RpoE sigma factor to 328

understand how this transcription factor may impact the virulence of B. pertussis. 329

330

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 15

Mapping the RpoE regulon of B. pertussis with Illumina RNAseq. In many bacteria, RpoE 331

controls expression of factors which can ameliorate membrane stress, heat shock, and 332

desiccation. We sought to profile the effect of RpoE on B. pertussis gene expression by 333

comparing strain UT25 to its isogenic rseA deletion mutant, PM18 by RNAseq. To perform 334

RNAseq, RNA was prepared from B. pertussis strain UT25 and PM18, growing in SSM liquid 335

media. Three pools of RNA were collected from cultures grown up to OD600 0.8. Three 336

separate cDNA libraries were generated and each sequenced on the Illumina MiSeq platform 337

with one lane. Each library was sequenced with ~27 million reads (2 x 76 bp). For 338

standardization and clarity, the reads were mapped to the reference Tohama I genome. On 339

average, 83% of all reads mapped to Tohama I. Likely some genes are present or absent in 340

UT25 which could account for these statistics. 341

RNAseq analysis of UT25 and PM18 resulted in read data for 90.8% of the genome 342

(3502 B. pertussis genes). If the IS481 transposase and ribosomal genes are excluded, then 343

only 2.4% of genes encoded were not detected. The B. pertussis Biocyc annotation indicates 344

that 10% of Tohama I genes are Pseudo-genes. However, based on our RNAseq analysis, it is 345

possible that as many as 7% of these genes are expressed at the RNA level. ssrA encodes a 346

regulatory RNA which tags proteins for degradation and ssrA was the highest expressed 347

transcript in all libraries sequenced with an average RPKM of 609,274 (Table S1). 348

349

Effect of rseA deletion and subsequent RpoE activation on the transcriptome of B. 350

pertussis. Comparing the gene expression profiles of UT25 to PM18 we identified 263 genes 351

(Fig. 2AB) that were differentially expressed between UT25 and PM18 (p>0.05) (Table S1). Of 352

this set of genes, 138 were repressed and 125 were activated (Fig. 2A). The predicted 353

localizations of the products encoded by these differentially expressed genes, were used to 354

determine which cellular compartments were most affected by RpoE activation in PM18. Most 355

activated genes correspond to the cytoplasmic membrane whereas most repressed genes 356

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 16

encode cytoplasmic proteins (Fig. 2B). For extracellular, outer-membrane, periplasmic, and 357

unknown products, equal numbers of genes activated and repressed were observed (Fig. 2B). 358

To better understand the classes of genes dysregulated, hypergeometric tests on annotations 359

(19) were performed. In this analysis, we compared the activated and repressed gene sets to 360

the B. pertussis whole annotation to determine if annotation classes were over-represented. 361

Within the repressed gene set, 28 annotations of 525 were found to be enriched. The term with 362

the smallest p value was pathogenesis (p value 6.5e-10). There are 20 genes in the B. 363

pertussis genome that were annotated with the term pathogenesis. Within the set of 138 364

repressed genes, 10 are annotated as being involved in pathogenesis (Table S1). Of these 365

genes, almost all genes associated with pertussis toxin were repressed (Fig. 3A and Table S1). 366

These data suggest that RpoE repressed factors, such as pertussis toxin, that facilitate 367

virulence. Surprisingly, 11 annotations corresponding to ATP synthesis/electron transport were 368

statistically overrepresented in the suppressed gene set. This suggests that RpoE expression 369

may lead to decrease in expression of genes involved in ATP synthesis in B. pertussis, thereby 370

affecting metabolism (Fig 3A). The activated gene set (125 genes) had 11 annotations that 371

were overrepresented with statistically significant p values (<0.05) with 18 genes in this set are 372

annotated as being involved in transport. Additionally, proline catabolism and siderophore 373

biosynthesis/transport genes were also overrepresented (Fig 3B). In the B. pertussis genome, 3 374

genes are annotated as being involved in biofilm formation and within our activated gene set, 2 375

were observed. This suggested RpoE may have a role in expression of biofilm synthesis genes. 376

377

Role of RpoE on secretion of adenylate cyclase toxin and implications on biofilm 378

formation. The expression data and bioinformatics analysis suggested that biofilm synthesis 379

genes were upregulated in PM18. We therefore wondered if PM18 would form a more robust 380

biofilm than UT25. Interestingly, strain PM18 formed less biofilm than UT25 as determined by 381

the standard crystal violet assay (Fig. 4A). Studies have indicated that adenylate cyclase toxin 382

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 17

(ACT) decreased Bordetella biofilms (30). When recombinant ACT was supplemented to 383

laboratory strain BP338, less biofilm was detected (Fig. 4A). Strain UT25 forms a robust biofilm 384

compared to BP338 (Fig. 4A) and was also susceptible to addition of exogenous ACT. A 385

previous study showed that ACT secretion was increased in PM18 although the gene encoding 386

ACT, cyaA, was not up-regulated (13). Here in our study, western blot analysis revealed that 387

PM18 expressed more ACT as compared to UT25 and a large amount was secreted into the 388

supernatant (Fig 4B). These data were corroborated by measuring the cyclase activities of the 389

UT25 and PM18 samples (Fig. 4B). Here using RNAseq, we corroborated the previous study 390

and observed that cyaA expression was not increased when comparing UT25 and PM18 391

transcriptomes (Fig 4C). Interestingly, although cyaA expression was not statistically 392

upregulated in PM18 (Fig. 4C), two of the three genes that encode the type I secretion system 393

that are responsible for secreting ACT were upregulated in PM18 with significant p values (Fig. 394

4C). The increased expression of type I secretion genes could be responsible for the increased 395

ACT secretion. Additionally, further studies would be required to define if RpoE plays a role in 396

post-transcriptional regulation of ACT expression by increasing translation efficiency, altering 397

small RNA binding and activity, or altering transcript stability. 398

399

rpoE expression is not auto-regulated in B. pertussis. In Gram negative pathogens, such 400

as P. aeruginosa, expression of the rpoE gene is auto-regulated and can drive expression of its 401

own gene through a specific promoter (31). This allows for amplification of rpoE expression. If 402

RpoE were auto-regulated in B. pertussis, deletion of rseA would result in extremely high rpoE 403

gene expression; however, rpoE expression remained stable in the absence of rseA (Table 1). 404

This suggests that B. pertussis rpoE expression is not auto-regulated. If the genomic 405

organization of B. pertussis rpoE is compared to either Escherichia coli or Pseudomonas 406

aeruginosa it is apparent that the genes immediately upstream of rpoE are different (data not 407

shown). This could suggest that some genetic rearrangements due to the organism’s evolution 408

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 18

could have influenced how RpoE is regulated. Since rseA is deleted from PM18, no rseA RNA 409

was detected, as would be expected (Table 1). The mucB gene encodes a negative regulator 410

of RseA and in the absence of rseA, mucB expression was slightly statistically increased (Table 411

1; 1.57 fold ). Since it is clear rpoE is not auto-regulated, then an rpoE dependent promoter 412

must reside somewhere in the coding sequence of rpoE to drive expression of mucB. In silico 413

promoter prediction suggests there are several promoters with strong prediction scores in the 414

rpoE coding region (data not shown). Downstream of mucB, another negative regulator of 415

RseA is encoded in the mucD gene. MucD (DegP) is a chaperone protease that resides in the 416

periplasm. In E. coli, DegP has been extensively characterized and can either refold or digest 417

aberrant proteins (32). In P. aeruginosa, inactivation of mucD results in high rpoE sigma factor 418

directed transcription (33). In general, MucD is classified as an indirect negative regulator of 419

RpoE activity because the main role of MucD is to control aberrant proteins that activate 420

proteolysis of the RseA anti-sigma factor. As expected, deletion of rseA resulted in increased 421

mucD expression (Table 1). It appears there were several promoters with strong prediction 422

scores in the mucB coding region (data not shown) which could be responsible for mucD 423

expression. Collectively, these results suggest that B. pertussis rpoE is not auto-regulated, but 424

RpoE does drive expression of the mucB and mucD genes that encode negative regulators of 425

RpoE. 426

427

Comparing the proteomes of UT25 and PM18. RNAseq was useful for identifying which 428

genes of B. pertussis are controlled, either indirectly or directly, by RpoE. In order to increase 429

the robustness of these analyses and better understand the role of RpoE, mass spectrometry 430

was performed on total protein lysates of the UT25 and PM18 strains. SDS-PAGE was 431

performed and gel fractions for each strain were submitted for total mass analysis. In total, 432

1658 B. pertussis proteins were detected with 63,006 total peptides. To characterize the 433

subcellular localizations of the proteins identified, Uniprot predictions were supplemented with 434

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 19

additional manual predictions based on the most reasonable assumption of where the protein 435

would be in the B. pertussis cell. To analyze the total proteomic dataset, we set a cut-off value 436

of proteins at a minimum of 10 peptides. Only differences in peptide numbers between UT25 437

and PM18 higher than 5 peptides were considered. Using these stipulations, 186 activated 438

proteins and 243 repressed proteins was used to compare the subcellular localizations (Fig. 439

5A). 11% of the total activated subset were outer membrane localized compared to 4% in the 440

repressed subset. These data suggest that RpoE controls remodeling of the outer membrane in 441

B. pertussis. Also of note, extracellular proteins were enriched in the repressed subset (Fig. 442

5A). In particular, the PM18 strain had decreased pertussis toxin which is an extracellular 443

protein, and was identified in the RNAseq as repressed in PM18. 444

To compare the findings of the RNAseq and proteomics, fold changes were calculated 445

based on the normalized spectral abundance factor (NSAF) of the UT25 and PM18 datasets 446

(Table S2). In order to overlay the activated genes (RNAseq) and proteins, a minimum average 447

RPKM value of 10 (54 of 125 statistically significant activated genes) was used as a cutoff and 448

compared against proteins with a minimum 10 peptides and fold change (PM18/UT25 NSAF) of 449

1.5 (115 of 1658 identified proteins). For the repressed gene to protein comparisons, a 450

minimum average RPKM value of 10 (137 of 138 statistically significant activated genes) was 451

used as a cutoff and compared against proteins with a minimum 10 peptides and fold change 452

(PM18/UT25 NSAF) of -1.5 (185 of 1658 identified proteins). Based on these cut-off values, 453

7.1% and 8.4% of the RNAseq and proteomics, respectively, correlated between the datasets 454

(Fig. 5B). These correlations suggest that RpoE might drive direct expression and also 455

influence expression indirectly. The NSAF value of each protein was used calculate a fold 456

change between PM18 and UT25. These data were then plotted based on the corresponding 457

gene start site of the B. pertussis chromosome (Fig. 5C). Proteins with expression values that 458

correlated to the RNAseq dataset are highlighted in black. In both the RNAseq and proteomic 459

analysis, the BP2497 putative zinc protease was highly expressed in PM18 (Table 2). In terms 460

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 20

of peptides identified in PM18, BP2497 was the third most abundant protein detected behind 461

GroEL and EF-Tu. Due to the high expression in PM18, it is possible that BP2497 expression is 462

directly controlled by RpoE or alternatively it is possible that an unknown intermediate 463

transcription factor is activating BP2497 expression. BP2497 is a part of the M16 peptidase 464

family (MEROPS Accession MER001222) which consists of mostly uncharacterized proteases. 465

At this stage, based on our data, we cannot propose a function for BP2497; however, due to the 466

increased release of ACT in PM18, we speculate that BP2497 could have a role in this process. 467

Current efforts are underway to investigate this possible role. In a recent study, BP2497 was 468

identified as a potential vaccine candidate because it was both increased in mRNA and protein 469

expression in low sulfate conditions suggesting that it may be Bvg regulated (34). Using locked 470

phase mutants, BP2497 has also been shown to be Bvg regulated (35). In the PM18 strain, 471

most Bvg regulated genes/proteins were decreased in expression yet, BP2497 was highly 472

expressed. Sulfate can decrease Bvg expression and here we observed Sulfate-binding protein 473

(Sbp) was more highly expressed in UT25 than PM18. These results suggest that Sbp could 474

alter intracellular sulfate and Bvg status; however, the BvgA regulator protein expression 475

increased in PM18. These seemingly confounding data indicate that molecular underpinnings 476

of RpoE modulation of virulence factors in B. pertussis may be independent of Bvg regulation, 477

therefore increasing the virulence versatility of the pathogen. 478

The colonization defects observed with PM18 in the CD1 murine infection studies lead 479

us to hypothesize that PM18 would have decreased virulence factor expression; however, there 480

are several factors that were expressed higher in PM18 than UT25 (Table 2). In regards to ACT 481

and pertussis toxin, the proteomic analysis corroborated the RNAseq analysis. When all of the 482

pertussis toxin and secretion system proteins are counted, PM18 had substantially less 483

pertussis toxin protein components. 484

485

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 21

qRT-PCR and Western blot analyses corroborate RNAseq and proteomics of PM18 486

and UT25. In order to support the RNAseq and proteomic findings, mRNA expression levels 487

were profiled for cyaA, ptxA, BP2497, mucD, bvgA, and fhaB in UT25 and PM18 at OD600 0.4 488

and 0.8 (Fig 6A). At 0.4 cyaA expression was repressed in PM18 but increased at 0.8. In 489

accordance with the RNAseq and proteomics, ptxA was 20-fold lower in PM18 at OD600 0.4 (Fig 490

6A). Both BP2497 and mucD mRNA expression correlated with RNAseq and proteomics. To 491

confirm the proteomics, PtxA expression was determined by Western blot in UT25 and PM18 492

from both pellet and supernatant, supernatant only, and pellet only protein lysates (Fig. 6B). As 493

observed with the qRT-PCR data, pertussis toxin expression was decreased in strain PM18 in 494

total, supernatant, and pelleted cells (Fig. 6B). 495

Summary and conclusion. Here in this study, we used a murine challenge model to 496

determine if the sigma factor RpoE was involved in regulating the virulence of B. pertussis. To 497

our surprise, we observed that the PM18 strain with high RpoE activity and high amounts of 498

ACT had impaired survival in the CD1 mouse and in vitro against J774A.1 macrophages and 499

neutrophils (Fig.1). Next we used transcriptomics and proteomics to characterize the genes and 500

proteins controlled by RpoE. While it is clear that BvgAS controls many key toxins and 501

virulence factors, our understanding of B. pertussis virulence is lacking due to the fact that in the 502

past vaccines were highly effective at preventing infections. RNAseq, proteomics, qRT-PCR, 503

and Western blot analyses confirmed that the PM18 strain expressed significantly less pertussis 504

toxin than the parental strain, yet more ACT protein (Fig. 4A, 5C, and 6AB). In addition strain 505

PM18 also was less able to form in vitro biofilms (Fig. 4). It is possible that the decreased 506

pertussis toxin production and lowered ability to form biofilms contributed to the decreased 507

bacterial survival of PM18 in the CD1 mouse. Overall, these data suggest that RpoE can affect 508

both a wide array of genes and cause differential expression of the B. pertussis toxins. 509

Our data present anidea that contradicts the current dogma that pertussis toxin and ACT 510

are controlled primarily by the Bvg system. We observed that cyaA gene expression was not 511

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 22

significantly increased in PM18, (Fig. 4C). However, increased ACT protein can be detected 512

(Fig. 4B). How or why is more ACT translated and secreted in the rseA mutant? Bibova et al. 513

showed that in the absence of the RNA chaperone Hfq, that less ACT was produced suggesting 514

that a small RNA may post-transcriptionally control ACT translation (24). However, we did not 515

observe differential expression of hfq in our RNAseq analysis (Table S1). In the same study, it 516

was observed that loss of hfq resulted in decreased pertussis toxin gene and protein expression 517

(24). Due to these observations, it does not seem likely that hfq is responsible for the inverted 518

pertussis toxin to ACT expression that we observe in the PM18 strain. Our expression data 519

suggest that the increase in ACT is post-transcriptional and additional studies are required to 520

determine if this increase is associated with changes in posttranslational regulation. Next, we 521

observed a slight decrease in bvgA expression (p value 0.04) that could account for decreased 522

pertussis toxin gene expression. However, we do not understand how an increase in RpoE 523

could lead to a decrease in bvgA. We now hypothesize that local stress conditions such as 524

oxidative stress, may lead to RseA degradation resulting in ACT release. Another logical next 525

step would be to directly determine the target promoters that RpoE binds. It is clear to us that 526

more work needs to be done to understand the transcriptional and metabolic networks of B. 527

pertussis. 528

If RseA acts upon RpoE as it does in other Gram-negative organisms, then loss of 529

RseA, should significantly increase RpoE activity. This ultra-high RpoE activity would not likely 530

be a constitutive physiological condition. Normally, RseA inhibition of RpoE in the wild type 531

situation would be dynamic and when conditions warranted RseA would impose its inhibition 532

thereby decreasing RpoE activity. Such regulation allows for efficient yet dynamic modulation of 533

factors needed during stress conditions. In the case of P. aeruginosa, the enzymatic machinery 534

to synthesize and secrete the exopolysaccharide alginate is controlled by the RpoE ortholog, 535

AlgU(T) (36). In conditions where alginate is needed for protection, the MucA anti-sigma factor 536

is proteolytically cleaved to release AlgU (37). We hypothesize that stress conditions may 537

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 23

induce the proteolytic cleavage of B. pertussis RseA. In Figure 7, a schematic of the putative 538

regulators of RpoE in B. pertussis are indicated. RseA is the main negative regulator of RpoE 539

and its function, based on other characterized systems, would be to sequester RpoE to the 540

inner membrane preventing its association with RNA polymerase. To activate RpoE, MucB is 541

released from RseA by some yet unknown mechanism, and the AlgW (DegS) protease would 542

then cleave RseA to activate the proteolytic process. Subsequent RseA cleavage by the inner 543

membrane protease, MucP, and the cytoplasmic ClpXP would then result in an uninhibited 544

RpoE that would drive expression of its target genes. Hanawa et al. demonstrated that 545

hydrogen peroxide can induce RpoE activity in B. pertussis (13) which suggests that RseA 546

inhibition can be removed, likely by proteolytic cleavage; however, direct experiments will be 547

necessary to establish this mechanism. It is possible that during the infectious process, B. 548

pertussis encounters stress conditions that induce RpoE activity. Based on this, it is likely that 549

proteins expressed by RpoE may be expressed at key steps during infection and therefore 550

these may turn out to be useful vaccine antigens. Here in this study, we have used 551

transcriptomics, proteomics, and validation experiments to establish a regulon of genes and 552

proteins under the control of RpoE. Future studies will be focused on dissecting the regulatory 553

networks of B. pertussis in the context of RpoE and other transcription factors so that a more 554

complete landscape of how B. pertussis is able to infect can be determined. 555

556

Funding Information 557

This work was supported by funding from the NIH/NIAID grant 5 RO1 AI1018000 [E.L.H.] and 558

laboratory startup funds from West Virginia University to F.H.D. 559

560

Acknowledgements 561

B. pertussis strains UT25 and PM18 were kindly provided by Dr. Sandra Armstrong (University 562

of Minnesota). We would like to thank Dr. AnhThu Nguyen (University of Virginia) for next 563

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 24

generation library preparation / illumina sequencing and Dr. Nicholas Sherman (University of 564

Virginia) for mass spectrometry proteomic analysis. Author contributions: M.B. designed / 565

performed murine experiments, analyzed data, and composed the manuscript. E.S. and D.T.B 566

performed pertussis toxin Western blots and analyzed data. C.B. and T.P. performed qRT-PCR 567

analysis and analyzed data, C.H. performed biofilm assays and analyzed data. M.G. performed 568

cyclase assays and ACT Western blots, E.L.H designed experiments and composed the 569

manuscript. F.H.D designed / performed experiments (murine infection model, RNAseq, 570

proteomics, and etc), analyzed data, and composed the manuscript. 571

References 572

1. Melvin JA, Scheller EV, Miller JF, Cotter PA. 2014. Bordetella pertussis pathogenesis: current 573

and future challenges. Nat Rev Microbiol 12:274-288. 574

2. Klein NP, Bartlett J, Rowhani-Rahbar A, Fireman B, Baxter R. 2012. Waning protection after 575

fifth dose of acellular pertussis vaccine in children. N Engl J Med 367:1012-1019. 576

3. Warfel JM, Beren J, Kelly VK, Lee G, Merkel TJ. 2012. Nonhuman primate model of pertussis. 577

Infect Immun 80:1530-1536. 578

4. Warfel JM, Beren J, Merkel TJ. 2012. Airborne transmission of Bordetella pertussis. J Infect Dis 579

206:902-906. 580

5. Hewlett EL, Burns DL, Cotter PA, Harvill ET, Merkel TJ, Quinn CP, Stibitz ES. 2014. Pertussis 581

pathogenesis--what we know and what we don't know. J Infect Dis 209:982-985. 582

6. Paddock CD, Sanden GN, Cherry JD, Gal AA, Langston C, Tatti KM, Wu KH, Goldsmith CS, Greer 583

PW, Montague JL, Eliason MT, Holman RC, Guarner J, Shieh WJ, Zaki SR. 2008. Pathology and 584

pathogenesis of fatal Bordetella pertussis infection in infants. Clin Infect Dis 47:328-338. 585

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 25

7. Carbonetti NH, Artamonova GV, Andreasen C, Bushar N. 2005. Pertussis toxin and adenylate 586

cyclase toxin provide a one-two punch for establishment of Bordetella pertussis infection of the 587

respiratory tract. Infect Immun 73:2698-2703. 588

8. Nicholson TL. 2007. Construction and validation of a first-generation Bordetella bronchiseptica 589

long-oligonucleotide microarray by transcriptional profiling the Bvg regulon. BMC Genomics 590

8:220. 591

9. Weiss AA, Hewlett EL, Myers GA, Falkow S. 1984. Pertussis toxin and extracytoplasmic 592

adenylate cyclase as virulence factors of Bordetella pertussis. J Infect Dis 150:219-222. 593

10. Merkel TJ, Stibitz S, Keith JM, Leef M, Shahin R. 1998. Contribution of regulation by the bvg 594

locus to respiratory infection of mice by Bordetella pertussis. Infect Immun 66:4367-4373. 595

11. Cotter PA, Miller JF. 1997. A mutation in the Bordetella bronchiseptica bvgS gene results in 596

reduced virulence and increased resistance to starvation, and identifies a new class of Bvg-597

regulated antigens. Mol Microbiol 24:671-685. 598

12. Barchinger SE, Zhang X, Hester SE, Rodriguez ME, Harvill ET, Ades SE. 2012. sigE facilitates the 599

adaptation of Bordetella bronchiseptica to stress conditions and lethal infection in 600

immunocompromised mice. BMC Microbiol 12:179. 601

13. Hanawa T, Yonezawa H, Kawakami H, Kamiya S, Armstrong SK. 2013. Role of Bordetella 602

pertussis RseA in the cell envelope stress response and adenylate cyclase toxin release. Pathog 603

Dis doi:10.1111/2049-632X.12061. 604

14. Brickman TJ, Armstrong SK. 1996. The ornithine decarboxylase gene odc is required for alcaligin 605

siderophore biosynthesis in Bordetella spp.: putrescine is a precursor of alcaligin. J Bacteriol 606

178:54-60. 607

15. Bordet J, O G. 1906. Le microbe de la coqueluche. Ann Inst Pasteur (Paris) 20:731-741. 608

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 26

16. Stainer DW, Scholte MJ. 1970. A simple chemically defined medium for the production of phase 609

I Bordetella pertussis. J Gen Microbiol 63:211-220. 610

17. Parkhill J, Sebaihia M, Preston A, Murphy LD, Thomson N, Harris DE, Holden MT, Churcher CM, 611

Bentley SD, Mungall KL, Cerdeno-Tarraga AM, Temple L, James K, Harris B, Quail MA, Achtman 612

M, Atkin R, Baker S, Basham D, Bason N, Cherevach I, Chillingworth T, Collins M, Cronin A, 613

Davis P, Doggett J, Feltwell T, Goble A, Hamlin N, Hauser H, Holroyd S, Jagels K, Leather S, 614

Moule S, Norberczak H, O'Neil S, Ormond D, Price C, Rabbinowitsch E, Rutter S, Sanders M, 615

Saunders D, Seeger K, Sharp S, Simmonds M, Skelton J, Squares R, Squares S, Stevens K, Unwin 616

L, Whitehead S, Barrell BG, Maskell DJ. 2003. Comparative analysis of the genome sequences 617

of Bordetella pertussis, Bordetella parapertussis and Bordetella bronchiseptica. Nat Genet 35:32-618

40. 619

18. Robinson MD, Smyth GK. 2008. Small-sample estimation of negative binomial dispersion, with 620

applications to SAGE data. Biostatistics 9:321-332. 621

19. Tian L, Greenberg SA, Kong SW, Altschuler J, Kohane IS, Park PJ. 2005. Discovering statistically 622

significant pathways in expression profiling studies. Proc Natl Acad Sci U S A 102:13544-13549. 623

20. Falcon S, Gentleman R. 2007. Using GOstats to test gene lists for GO term association. 624

Bioinformatics 23:257-258. 625

21. Szklarczyk D, Franceschini A, Kuhn M, Simonovic M, Roth A, Minguez P, Doerks T, Stark M, 626

Muller J, Bork P, Jensen LJ, von Mering C. 2011. The STRING database in 2011: functional 627

interaction networks of proteins, globally integrated and scored. Nucleic Acids Research 628

39:D561-D568. 629

22. Yu NY, Wagner JR, Laird MR, Melli G, Rey S, Lo R, Dao P, Sahinalp SC, Ester M, Foster LJ, 630

Brinkman FS. 2010. PSORTb 3.0: improved protein subcellular localization prediction with 631

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 27

refined localization subcategories and predictive capabilities for all prokaryotes. Bioinformatics 632

26:1608-1615. 633

23. Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expression data using real-time 634

quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25:402-408. 635

24. Bibova I, Skopova K, Masin J, Cerny O, Hot D, Sebo P, Vecerek B. 2013. The RNA chaperone Hfq 636

is required for virulence of Bordetella pertussis. Infect Immun 81:4081-4090. 637

25. Laemmli UK. 1970. Cleavage of structural proteins during the assembly of the head of 638

bacteriophage T4. Nature 227:680-685. 639

26. Lee SJ, Gray MC, Guo L, Sebo P, Hewlett EL. 1999. Epitope mapping of monoclonal antibodies 640

against Bordetella pertussis adenylate cyclase toxin. Infect Immun 67:2090-2095. 641

27. Goldhammer AR, Wolff J, Hope Cook G, Berkowitz SA, Klee CB, Manclark CR, Hewlett EL. 1981. 642

Spurious protein activators of Bordetella pertussis adenylate cyclase. Eur J Biochem 115:605-643

609. 644

28. O'Toole GA. 2011. Microtiter dish biofilm formation assay. J Vis Exp doi:10.3791/2437. 645

29. Zybailov B, Mosley AL, Sardiu ME, Coleman MK, Florens L, Washburn MP. 2006. Statistical 646

analysis of membrane proteome expression changes in Saccharomyces cerevisiae. J Proteome 647

Res 5:2339-2347. 648

30. Irie Y, Mattoo S, Yuk MH. 2004. The Bvg virulence control system regulates biofilm formation in 649

Bordetella bronchiseptica. J Bacteriol 186:5692-5698. 650

31. Schurr MJ, Yu H, Boucher JC, Hibler NS, Deretic V. 1995. Multiple promoters and induction by 651

heat shock of the gene encoding the alternative sigma factor AlgU (sigma E) which controls 652

mucoidy in cystic fibrosis isolates of Pseudomonas aeruginosa. J Bacteriol 177:5670-5679. 653

32. Huber D, Bukau B. 2008. DegP: a Protein "Death Star". Structure 16:989-990. 654

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 28

33. Damron FH, Yu HD. 2011. Pseudomonas aeruginosa MucD regulates alginate pathway through 655

activation of MucA degradation via MucP proteolytic activity. J Bacteriol 193:286-291. 656

34. de Gouw D, de Jonge MI, Hermans PW, Wessels HJ, Zomer A, Berends A, Pratt C, Berbers GA, 657

Mooi FR, Diavatopoulos DA. 2014. Proteomics-identified Bvg-activated autotransporters 658

protect against Bordetella pertussis in a mouse model. PLoS One 9:e105011. 659

35. Cummings CA, Bootsma HJ, Relman DA, Miller JF. 2006. Species- and strain-specific control of a 660

complex, flexible regulon by Bordetella BvgAS. J Bacteriol 188:1775-1785. 661

36. Martin DW, Holloway BW, Deretic V. 1993. Characterization of a locus determining the mucoid 662

status of Pseudomonas aeruginosa: AlgU shows sequence similarities with a Bacillus sigma 663

factor. J Bacteriol 175:1153-1164. 664

37. Damron FH, Goldberg JB. 2012. Proteolytic regulation of alginate overproduction in 665

Pseudomonas aeruginosa. Mol Microbiol 84:595-607. 666

667

Figure Legends 668

Figure 1. Colonization of the murine airway or survival in J774A.1 and neutrophils by B. 669

pertussis strain UT25 and the isogenic PM18 (∆rseA) strains. CD1 mice were infected with 670

2 x107 CFU of the B. pertussis strains UT25 and PM18. At 24, 48 and 144 h post infection the 671

bacterial burdens of the nasal wash (A), trachea (B), and lungs (C) were determined. Four to 672

five mice were used for each time-point and all of the values obtained are indicated. At 144 hr 673

PM18 measured by the nasal wash (1ml) was significantly lower than the parental UT25. In the 674

lungs, PM18 was present in a higher amount at 24 hr but then decreased as UT25 increased. (* 675

p<0.05 and ** p<0.01). J774A.1 cells were allowed to phagocytose UT25 or PM18. Less viable 676

PM18 were observed at both 2 and 24 h post infection (D). Bone marrow derived neutrophils 677

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 29

were isolated and allowed to phagocytose UT25 or PM18 and similar to in J774A.1 cells, PM18 678

was more efficient killed intracellular (E) but not when extracellular (F). 679

Figure 2. RNAseq analysis of PM18 compared to UT25 reveals the genes controlled by 680

RpoE. In panel A, a volcano plot of genes statistically significant repressed (138; blue) or 681

activated (125; red) are indicated relative to the observed fold changes and their respective p 682

values. In panel B, the predicted or known localizations of each of the genes that were 683

differentially regulated are shown. 684

Figure 3. String analysis reveals the systems of genes that are repressed or activated by 685

RpoE. In panel A, the 138 repressed genes and in panel B the 125 activated genes are 686

grouped with the systems or families to which they correspond. Notably pertussis toxin and 687

metabolism/stress genes are repressed, and most of the activated genes do not belong to 688

known systems except for iron acquisition and branched chain amino acid uptake. 689

Figure 4. Biofilm formation, ACT and cyaABDEX expression in relation to RpoE in B. 690

pertussis. In panel A, the biofilm forming capabilities of each strain are shown. Strain UT25 691

biofilm formation can be suppressed by exogenous ACT. PM18 formed biofilms to the same 692

magnitude as UT25 with exogenous ACT which correlates with the high amount of ACT 693

secreted by PM18 (panel B). Panel C shows the relative read coverage of one of the RNAseq 694

replicates with respect to the cyaC-cyaABDEX operons. Only the T1SS system encoding genes 695

cyaB and cyaD were statistically activated in PM18. 696

Figure 5. Analysis of the UT25 and PM18 proteomes. Panel A shows the breakdown of 697

activated and repressed proteins with respect to their predicted subcellular localizations. The 698

RNAseq and proteomics were correlated and Venn diagrams are shown in B. The NSAF 699

normalize fold change of the protein found in both proteomes is shown in C and are plotted by 700

the location on the chromosome. Proteins that correlated with RNAseq analysis are highlight in 701

black halos. 702

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 30

Figure 6. qRT-PCR analysis of selected genes and Western blot measurement of 703

pertussis toxin. UT25 and PM18 were grown up to 0.4 and 0.8 and three biological replicates 704

were analyzed with three technical replicates each. (* p<0.05 , ** p<0.01, ***p<0.001). In panel 705

B, a representative Western blot of pertussis toxin is shown for strains UT25 and PM18 where 706

supernatants or cell pellets were used. UT25 has more pertussis toxin expression than PM18 707

which corroborates RNAseq, proteomics, and qRT-PCR analysis. 708

Figure 7. Hypothetical schematic of RpoE-RseA system proteins. The respective B. 709

pertussis RseA proteases: BP1776 (AlgW), BP1426 (MucP), and ClpXP (BP1776/BP1775) are 710

indicated. RpoE would be tethered to the inner membrane by RseA. AlgW, MucP and then 711

ClpXP would sequentially cleave RseA to release RpoE to drive expression at target promoters. 712

It is possible that this system can be activated during stress conditions to modulate pertussis 713

and adenylate cyclase toxin expression in B. pertussis. 714

715

716

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 31

Tables 717

Table 1. Differential gene expression due to absence of rseA in B. pertussis. 718

gene Database object name Fold change (PM18/UT25) P-value

RpoE system and gene regulation rpoE RNA polymerase sigma factor 1.02 3.97E-06 rseA RpoE anti-sigma factor -1,233.91 -1.03E-04 mucB Negative regulator of RseA 1.57 1.18E-04 mucD Serine protease 2.31 8.07E-04

Membrane associated gene products BP2497 Putative zinc protease 4.19 1.65E-08 BP0840 Outer membrane porin protein BP0840 -3.18 4.34E-05 BP0664 Putative exported protein -4.42 1.74E-09 BP1709 Putative permease of ABC transporter -5.15 6.40E-10 BP1708 Putative ABC transporter -6.38 6.68E-12 BP2830 Putative ABC transport protein, inner mebrane component 35.64 1.20E-04 BP2652 Putative exported protein 3.32 0.01

Iron acqusition alcA Alcaligin biosynthesis enzyme 6.39 0.01 alcB Alcaligin biosynthesis protein 11.34 3.84E-03 alcC Alcaligin biosynthesis protein 6.81 0.01 alcD Alcaligin biosynthesis protein 6.76 0.01 alcE Putative iron-sulfur protein 4.22 0.03 bfrF Putative ferric siderophore receptor 4.9 0.04

T1SS adenylate cyclase toxin cyaD Protein CyaD 2.38 1.23E-03 cyaB Cyclolysin secretion/processing ATP-binding protein CyaB 2.06 6.05E-03

Virulence associated genes bvgA Virulence factors putative positive transcription regulator BvgA -1.69 0.04 fhaB Filamentous hemagglutinin -1.76 0.02 vag8 Autotransporter -2.71 2.29E-04 ptxA Pertussis toxin subunit 1 -3.13 1.85E-05 ptxB Pertussis toxin subunit 2 -3.29 3.89E-06 ptxD Pertussis toxin subunit 4 -3.41 9.78E-07 ptxC Pertussis toxin subunit 3 -3.41 2.29E-06 ptxE Pertussis toxin subunit 5 -3.42 1.21E-06

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 32

Table 2. Proteomic analysis B. pertussis and the effect of loss of rseA. 719

Peptides

Protein names uniprot Gene Tohama locus PM18 UT25 Total FOLD

(PM18/UT25)Highly abundant proteins

Putative zinc protease Q7VVY4 BP2497 BP2497 892 171 1063 5.04 Elongation factor Tu (EF-Tu) Q7TT91 tuf1; tuf2 BP0007; BP3611 565 580 1145 -1.06 60 kDa chaperonin (GroEL protein) P48210 groL BP3495 1617 1504 3121 1.04 Outer membrane porin protein BP0840 Q04064 BP0840 BP0840 416 445 861 -1.11 ABC transporter amino acid-binding protein Q7VSU1 BP3831 BP3831 356 502 858 -1.46

RpoE system RNA polymerase sigma factor Q7VW35 rpoE BP2437 6 9 15 -1.55 Putative sigma factor regulatory protein Q7VW37 mucB BP2435 36 19 55 1.83 Serine protease Q7VW38 mucD BP2434 575 91 666 6.10

Adenylate cyclase toxin Bifunctional hemolysin/adenylate cyclase (ACT) P0DKX7 cyaA BP0760 257 248 505 1.00 Cyclolysin-activating lysine-acyltransferase CyaC P0A3I5 cyaC BP0758 2 1 3 1.93 Protein CyaD P0DKX9 cyaD BP0762 3 0 3 Protein CyaE P0DKY0 cyaE BP0763 2 0 2

Pertussis toxin All pertussis toxin or secretion system subunits 68 154 222 -2.02

Other virulence factors Sulfate-binding protein Q7VZE6 sbp BP0966 170 65 235 2.53 Filamentous hemagglutinin P12255 fhaB BP1879 949 1525 2474 -1.66 Adhesin Q7VVJ2 fhaS BP2667 35 87 122 -2.57 BrkA autotransporter Q45340 brkA BP3494 840 617 1457 1.32 Pertactin autotransporter P14283 prn BP1054 273 231 504 1.14 Probable TonB-dependent receptor BfrD P81549 bfrD BP0856 232 109 341 2.06 Protein TolB Q7VU03 tolB BP3343 193 144 337 1.29

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

Page 33

Autotransporter subtilisin-like protease Q7W0C9 sphB1 BP0216 144 135 279 1.03 Transcription regulator BvgA P0A4H2 bvgA BP1878 155 118 273 1.27 Tracheal colonization factor Q79GX8 tcfA BP1201 362 255 617 1.37 Autotransporter Q79GN7 vag8 BP2315 222 289 511 -1.35 Dermonecrotic toxin Q7VTS2 dnt BP3439 27 44 71 -1.69 Putative toxin Q7VYQ9 BP1251 BP1251 16 15 31 1.03

720

721

722

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from

on May 14, 2018 by guest

http://iai.asm.org/

Dow

nloaded from