IAI Accepts, published online ahead of print on 11 October 2010 Infect. Immun. doi:10.1128/IAI.00736-10 Copyright © 2010, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved.

1

Characterization of a Staphylococcus aureus surface virulence factor promoting resistance

2

to oxidative killing and infectious endocarditis

3

Natalia Malachowa1,2, Petra L. Kohler3, Patrick M. Schlievert3, Olivia N. Chuang3, Gary M.

4

Dunny3, Scott D. Kobayashi2, Jacek Miedzobrodzki1, Gregory A. Bohach4†‡, and Keun Seok

5

Seo4‡*

6

1

7

Jagiellonian University, 30-387 Krakow, Poland

8

2

9

of Allergy and Infectious Diseases, National Institute of Health, Hamilton, MT, 59840, USA

Department of Microbiology, Faculty of Biochemistry, Biophysics and Biotechnology,

Laboratory of Human Bacterial Pathogenesis, Rocky Mountain Laboratories, National Institute

10

3

11

55455, USA

12

4

13

Moscow, ID, 83844, USA

Department of Microbiology, University of Minnesota Medical School, Minneapolis, MN,

Departmement of Microbiology, Molecular Biology, and Biochemistry, University of Idaho,

14 15

Running title: SOK protein in staphylococcal virulence

16 17

Present addresses:

18

†

19

State, MS, 39762, USA

20

‡

21

Mississippi State, MS, 39762, USA

Department of Biochemistry and Molecular Biology, Mississippi State University, Mississippi

Department of Basic Sciences, College of Veterinary Medicine, Mississippi State University,

1

22

*

23

Mississippi State University, P. O. Box 6100, Mississippi State, MS, 39762, USA. Phone: 1-662-

24

325-1419; Fax: 1-662-325-1031, E-mail: [email protected]

25

Key words: endocarditis, neutrophils, oxidative stress, Staphylococcus

Corresponding author, Mailing address: Keun Seok Seo, Department of Basic Sciences,

2

26

ABSTRACT

27

Staphylococcus aureus is a prominent human pathogen and a leading cause of

28

community- and hospital-acquired bacterial infections worldwide. Herein we describe the

29

identification and characterization of the S. aureus 67.6 kDa hypothetical protein, named for

30

Surface factor promoting resistance to Oxidative Killing (SOK) in this study. Sequence analysis

31

showed that SOK gene is conserved in all sequenced S. aureus and homologous to myosin cross-

32

reactive antigen of Streptococcus pyogenes. Immunoblot and immunofluorescence analysis

33

showed that SOK was co-purified with membrane fractions and was exposed on the surface of S.

34

aureus Newman and RN4220. Comparative analysis of wild-type S. aureus and an isogenic

35

deletion strain indicated that SOK contributes to both resistance to killing by human neutrophils

36

and to oxidative stress. In addition, the S. aureus sok deletion strain showed dramatically reduced

37

aortic valve vegetation and bacterial cell number in a rabbit endocarditis model. These results,

38

plus the suspected role of the streptococcal homologue in certain diseases such as acute

39

rheumatic fever, suggest that SOK plays an important role in cardiovascular and other

40

staphylococcal infection.

3

41

INTRODUCTION

42

Staphylococcus aureus is a commensal that often colonizes skin and mucosal membranes

43

(11, 28). This species is usually benign in healthy individuals, but it is a high-risk pathogen for

44

immunocompromised individuals. As a consequence of its numerous virulence factors and

45

adaptability, S. aureus is one of the most significant human pathogens for both nosocomial- and

46

community-associated infections (20). Moreover, an increasing resistance to antibacterial agents,

47

and the adaptation and emergence of methicillin- and vancomycin-resistant S. aureus (MRSA and

48

VRSA, respectively) is alarming (2, 13).

49

S. aureus is the causative agent of diverse human and animal maladies including, but not

50

limited to, abscesses, food poisoning, toxic shock syndrome, septicemia, and endocarditis (3, 46,

51

49). This cadre of diseases results from S. aureus strain heterogeneity. Although numerous, most

52

S. aureus virulence factors are categorized into one of the following groups according to their

53

function: (i) surface proteins that promote adhesion, internalization, and colonization; (ii) toxins

54

and enzymes that promote tissue damage, inflammation, and invasion and dissemination; (iii)

55

surface factors that affect phagocytosis by leukocytes; (iv) factors that enhance survival in

56

phagocytes; or (v) superantigens and other molecules that modulate the immune system by

57

altering the function of lymphocytes and antigen presenting cells (1, 12, 44).

58

Our bioinformatics analysis of thirteen S. aureus genomic sequences in search of potential

59

virulence factors for staphylococcal-induced cardiovascular diseases revealed a conserved open

60

reading frame (ORF, 96-100 % identity among all S. aureus sequences). The predicted translation

61

products from these ORFs share 59 % identity with the 67-kDa myosin cross-reactive antigen

62

(MCRA) of Streptococcus pyogenes (19). The S. aureus homologue (ORF SA0102) was reported

63

initially by Kuroda, et al, (25) in reference to the N315 strain genome sequence. They described

4

64

SA0102 as one of two MHC class II β-chain homologues in the N315 genome. The 67-kDa

65

Streptococcus pyogenes protein and the SA0102 predicted translation product share 62% and 34%

66

similarity (19 % and 21.2 % identity), respectively, to murine β1 domain of the mouse I-Au chain

67

(19; our unpublished results).

68

The 67-kDa streptococcal homologue is a putative virulence factor, and hybridization

69

studies suggested that related proteins exist in streptococcal groups A, C, and G (19). This protein

70

is a member of extensive MCRA protein family. It reacts with sera from patients with acute

71

rheumatic fever (ARF), acute glomerulonephritis, and active streptococcal infections (19). It also

72

reacts with anti-myosin antibody in sera of patients with ARF. Recently, this streptococcal protein

73

was described as fatty acid double bond hydratase (47). Although members of MCRA protein

74

family are widely distributed among bacteria, only three proteins from this family were

75

biochemically characterized (6, 47) and the exact role of the vast majority of proteins and

76

homologues belong to this family remains unknown. The objective of this study is to characterize

77

the 67-kDa myosin-cross-reactive homologue of S. aureus. To address this goal, we constructed

78

deletion mutants in two well-characterized S. aureus strains, RN4220 and Newman, compared the

79

properties of the parental and mutant strains, and investigated the molecular relatedness of the

80

structural gene in a number of clinical isolates. The data suggest that this protein is ubiquitous

81

among S. aureus clinical isolates and could contribute to infectious diseases such as endocarditis

82

by promoting survival in phagocytes and resistance to oxidative killing.

83

property, we tentatively designated the S. aureus protein SOK, a Surface factor promoting

84

resistance to Oxidative Killing.

5

Due to this latter

85

MATERIALS AND METHODS

86

Bacterial strains, plasmids, and growth conditions. Strains and plasmids used are

87

described in Tables 1. Escherichia coli was grown using Luria–Bertani (Difco Laboratories,

88

Detroit, MI, USA) medium that was supplemented with ampicillin (Am; 100 µg/ml) or

89

chloramphenicol (Cm; 34 µg/ml) when necessary. Except where indicated, S. aureus was

90

propagated using tryptic soy (TS) media (Difco Laboratories, Detroit, MI, USA). Plasmid

91

selection for S. aureus was conducted with erythromycin (Em), Cm, or tetracycline (Tc) (5, 10, or

92

10 µg/ml, respectively).

93

Bioinformatics analysis. The search for homologues was performed using BLAST on the

94

NCBI web server (http://www.ncbi.nlm.nih.gov/), and sequences were aligned by the ClustalW

95

program (http://www.ebi.ac.uk/Tools/clustalw/index.html). The following tools in the ExPASy

96

Proteomic Server were used to analyze various properties of the SOK protein including isoelectric

97

point and molecular weight (Compute pI/Mw tool), signal peptide cleavage site was calculated by

98

artificial neural networks (NN) and hidden Markov models (HMM) that are used by SignalP 3.0

99

server, hydrophobicity (ProtScale), and transmembrane topology (MEMSAT, The PSIPRED

100

Protein Structure Prediction Server). BPROM program (Softberry, Inc. Mount Kisco, NY, USA)

101

was used to predict bacterial promoter.

102

DNA isolation. Genomic DNA was isolated by using Genomic DNA Prep Plus kits (A&A

103

Biotechnology, Gdynia, Poland) facilitated by a method to promote S. aureus lysis (41). Plasmids

104

were isolated with a MiniPrep isolation kit (Qiagen GmbH, Hilden, USA) using the

105

manufacturer’s protocol. S. aureus cells were treated with lysostaphin in the kit resuspension

106

buffer. DNA was quantified using a NanoDrop ND-1000 (Nanodrop Tech., Wilmington, DE).

107

Restriction fragment length polymorphism (RFLP) analysis. A 1,789 bp fragment 6

108

containing predicted sok coding region plus flanking promoter and ribosome binding sites was

109

amplified using primers 12631 and 12632 (Table 2). PCR products were purified, quantified (see

110

above), and digested with Csp6I, Hin6I or TaqI. These enzymes were selected based on two

111

criteria; (i) at least three cleavage sites exist within the PCR-amplified sok gene, and (ii) predicted

112

digestion patterns are readily discernable by agarose gel electrophoresis. Fragments were

113

separated by electrophoresis in 2 % agarose gels to compare sok genotypes according to digestion

114

profiles. At least one band difference was used as the criterion for designating unique RFLP

115

profiles. Concordance between spa typing results and PCR-RFLP was calculated as described

116

previously (30).

117

RNA isolation and real-time PCR (RT-PCR). RNA was isolated with the RNeasy Mini

118

Kit (Qiagen Science, Germantown, MD, USA) according to the manufacturer’s protocol. Bacteria

119

were disrupted using a FastPrep FP120 (ThermoSavant, Holbrook, NY, USA) (45 sec, 6.0 m/sec).

120

DNA was removed using RNase free DNaseI (Ambion, Austin, TX, USA), and RNA was further

121

purified by treatment again with the same RNA isolation protocol. RNA samples with

122

OD260/OD280 ≥ 2.0 were used. First-strand cDNA was synthesized from 1 µg of RNA using

123

Superscript Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA), and 5 µl of a 100-fold diluted

124

aliquot was used as template (final sample volume 25 µl; MicroAmp Optical 96-Well Reaction

125

Plate, Applied Biosystems). RT-PCR was performed in an ABI Prism 7500 system using SYBR

126

green mix as recommended (Applied Biosystems, Foster City, CA).

127

Cloning and purification of SOK. Recombinant SOK (rSOK) was expressed in E. coli

128

BL21(DE3) pLysS, using pGEX- 5T which expresses proteins with an N-terminal His-tag and a

129

glutathione S-transferase label (5).

130

polymerase (Fermentas, Lithuania); the 1,827 bp product was digested with BamHI and XhoI.

The predicted SOK ORF was amplified with Pfu DNA

7

131

The resulting 1,776 bp fragment was cloned into pGEX-5T and transformed to E. coli DH5α cells

132

(Life Technologies, Inc., Gaithersburg, MD). After confirming the appropriate construct by PCR

133

and sequencing, the plasmid was purified and transformed into E. coli BL21 (DE3) pLysS

134

(Stratagene, USA). Expression was induced with 1 mM IPTG when the OD600 of the culture

135

reached 0.9.

136

centrifugation, washed with phosphate-buffered saline (140 mM NaCl, 2.7 mM KCl, 10 mM

137

Na2HPO4, 1.8 mM KH2PO4, pH 7.3), and disrupted using a French Press (Thermo Fisher

138

Scientific, Inc., Waltham, MA). rSOK was purified on the Glutathione Sepharose 4B resin,

139

according to the manufacturer's recommendation (Amersham, Piscataway, NJ) and GST label was

140

cleaved with Thrombin (Sigma-Aldrich, St Louis, MO). The recombinant protein was dialyzed

141

against 10 mM MOPS buffer (pH 7.0) containing 10 mM NaCl and 1 mM EDTA. The apparently

142

pure rSOK, assessed by SDS-PAGE (26), was stored in dialysis buffer containing 10% glycerol.

Following growth at 20° C for an additional 10 h, cells were collected by

143

Antibody production. Spraque-Dawley rats (6-8 week old; Simonsen Laboratories, Inc.,

144

San Diego, CA) were given biweekly subcutaneous injections of rSOK (100 µg) in Freund's

145

incomplete adjuvant (GIBCO, Grand Island, NY, USA). One week after the fourth boost, sera

146

were harvested and pooled. The anti-rSOK titer, 15,000, was determined by immunoblots (9), and

147

diluted (1:10,000) antisera were typically used for experiments.

148

Immunofluorescence analysis of S. aureus. Cell samples were prepared by the method

149

of Hiraga, et al. (17). Slides were visualized using a Zeiss LCSM 5 PASCAL version 4.0 SP2

150

(Carl Zeiss MicroImaging GmbH 1997-2006, Heilderberg, Germany) following treatment with

151

anti-rSOK rat sera (above) and AlexaFluor488-conjugated goat anti-rat IgG (H+L, 2 mg/ml)

152

secondary antibodies (Molecular Probes Inc., Eugene, OR).

153

SOK subcellular localization. Cells and supernates from overnight (O.N.) cultures (25

8

154

ml) were separated by centrifugation (8,000 × g, 10 min). Culture supernatant proteins were

155

precipitated (2 h; -20° C) with 9 volumes of trichloroacetic acid: acetone (1:8; v/v), pelleted, and

156

resuspended in distilled water (1 ml). The cell pellet was washed with deionized water and treated

157

(18,000 psi) in a French Press (7), followed by adding Protease Inhibitor Cocktail and Nuclease

158

Mix (Amersham Bioscience Corp, Piscataway, NJ) and centrifugation (75,000 × g; 30 min). The

159

obtained supernate containing soluble cytoplasmic proteins was stored at -80° C until needed. The

160

resulting pellet, containing crude membrane and wall fractions, was resuspended in rehydration

161

buffer (7 M urea; 2 M thiourea; 2 % ASB-14; 0.5 % triton X-100; 2 mM tributylphosphine; 1 %

162

Bromophenol Blue), and incubated at room temperature for 2 h. The sample was clarified by

163

centrifugation (75,000 × g, 20 min) and supernatant fluids were recovered for analysis of integral

164

membrane proteins. Cell wall proteins in the pellet were released by incubating for 3 h in TE

165

buffer (50 mM Tris-HCl and 10 mM EDTA, pH 8.0) with lysostaphin (40 µg/ml). Proteins were

166

analyzed by SDS-PAGE and immunoblotting (see above).

167

SOK deletion mutagenesis. DNA fragments, 918 (5’) and 698 (3’) bp in length, were

168

amplified by PCR from within sok through external flanking regions. The 5’ fragment included

169

nts -154 through 764; the 3’ fragment encompassed nts 1,138 through an additional 59 nt after the

170

predicted stop codon (nt numbering is relative to the predicted ATG initiation codon, see Fig. S1A

171

in the supplemental material). PCR products were digested with appropriate enzymes (Table 2)

172

and ligated on opposite ends of the Tcr cassette in pDG1515 (50) resulting in pSOK1 vector. This

173

construct was propagated in E. coli DH5α and digested with BamHI and KpnI. The Tcr cassette

174

plus flanking sok fragments were cloned into the pCL10 shuttle vector resulting in pSOK2. The

175

plasmid was then electroporated into S. aureus RN4220 and Newman as recommended (ECM

176

600; BTX Molecular Delivery Systems). Transformed cells were grown in TS broth (TSB) with

9

177

Tc (24 h; 43° C) and plated on TS agar (TSA) containing Tc to select colonies with the first

178

recombination. One colony was transferred to TSB and grown (30° C; 5 days) with daily

179

transferring to fresh TSB. After 5 days, an aliquot was grown on TSA with Tc. Colonies were

180

screened for a Tcr, Cms phenotype indicating the second recombination resulted in deletion of a

181

404nt sok sequence and insertion of Tc cassette. To complement the mutation, a fragment

182

representing the sok ORF with 260 nts upstream was amplified using MExpF and MExpR primers

183

and ligated into the BamHI and SalI sites of pMIN164 (18) resulting in pSOK3. This plasmid was

184

electroporated into S. aureus RN4220 and Newman, respectively (selection was accomplished on

185

TSA with Tc and Em).

186

Polymorphonuclear leukocyte (PMN) killing assay. Human PMNs were isolated from

187

heparinized venous blood of four different donors in accordance with a human subject protocol

188

approved by the University of Idaho Institutional Review Board for Human Subjects (approval

189

number 05-056). Donors were informed the procedure risks and provided a written consent prior

190

to participation. Killing of bacteria by human PMNs was determined as described (22) with the

191

following modifications. PMNs (106) were combined with opsonized bacteria (107) in 96-well

192

plates which were centrifuged at 400 × g for 5 min, and incubated at 37° C for up to 205 min.

193

PMNs were treated with 400 µM gentamicin (Sigma-Aldrich Co.) 15 min following addition of S.

194

aureus to remove any remaining extracellular bacteria. Cells were incubated for an additional 10

195

min at 37o C with gentamicin prior to commencement of the assay (T=0) and further incubation

196

for time points up to 180 min. At indicated times, gentamicin was removed by aspiration, and

197

cells were gently washed with PBS, lysed in sterile water, and bacteria plated on TSA. Colony-

198

forming units (CFU) were enumerated following overnight incubation, and percentage of bacteria

199

killed was calculated by using the equation (CFUPMN+/CFUT=0) × 100, where CFUT=0 indicates

10

200

number of enumerated colonies for time point zero and CFUPMN+ is number of CFU for each

201

analyzed time point. The assay measures percentage of total number of viable ingested bacteria

202

compared to time point zero.

203

Reactive oxygen species (ROS) analysis. Human PMNs (above) were mixed with 10

204

mM 2,7-dichlorodihydrofluorescein diacetate (DCF) (Molecular Probes Inc., Eugene, OR) and

205

incubated for 30 min at room temperature in the dark (8). Opsonized bacteria (above) were mixed

206

with the PMNs at10:1 ratio, and transferred to pre-coated wells of 96-well plate. The plate was

207

centrifuged (5 min; 700 × g; 4o C) to synchronize phagocytosis. ROS production was monitored

208

during incubation (37° C) using a SpectraMax M2 plate reader (Molecular Devices, Sunnyvale,

209

CA) with 485 nm excitation and 538 nm emission wavelengths. Data were analyzed by SoftMax

210

Pro software, version 5.0.1 (Molecular Devices) and presented as the rate of change (Vmax) in

211

fluorescence over time.

212

H2O2 and 1O2 susceptibility assays. S. aureus cells (mid-exponential growth phase) were

213

harvested by centrifugation, washed once with PBS, and adjusted to 1 × 108 or 2 × 109 cells/ml,

214

respectively for H2O2 and 1O2 assays (27). To assess the effect of H2O2 on S. aureus viability, the

215

bacteria were incubated with various concentrations of H2O2 (Sigma-Aldrich Co., St Louis, MO,

216

USA) in glass tubes for 1 h at 37° C, followed by addition of Micrococcus lysodeikticus catalase

217

(1000 U/ml) (Sigma-Aldrich Co.) to quench the remaining H2O2. The percentage of surviving

218

bacteria was calculated by using the equation (CFUH2O2+/CFUT=0) × 100, where CFUT=0 indicates

219

number of enumerated colonies for time point zero and CFUH2O2+ is number of CFU after

220

incubation with H2O2 for 1 h. For 1O2 susceptibility, bacteria were incubated in 24-well plates (for

221

30 or 60 min at 37° C) with various concentrations (0.25 – 6.0 µg/ml) of methylene blue (Sigma-

222

Aldrich Co.). The plates were placed 10 cm from a 100 watt incandescent light bulb followed by

11

223

plate counting to measure survival.

224

Endocarditis model. New Zealand white rabbits (2-3 kg) were anesthetized with xylazine

225

and ketamine (20 mg/kg each) and subjected to transaortic catheterization for 2 h to damage the

226

aortic valves (43). After removal of the catheters and closing the animals, a washed suspension of

227

S. aureus [2 ml; 1.0 × 109/ml (RN4220 strains) or 5.0 × 108/ml (Newman strains) in PBS] was

228

administered to each rabbit in the marginal ear vein. Animals were observed daily. Except for

229

rabbits infected with the parental Newman strain, which succumbed on day 2-3, animals were

230

sacrificed on day 4-5 to assess vegetation formation. Aseptically harvested vegetations from all

231

animals were weighed, and bacteria were enumerated by plate counts.

232 233

Statistical analysis. Student t-test was carried out using GraphPad Prism Software version 4.02 (GraphPad Software, Inc., San Diego, CA).

234

12

235

RESULTS

236

Bioinformatics analysis results. The initial analysis of published S. aureus genome

237

sequences revealed that the sok gene is ubiquitous and the sequence is highly conserved in these

238

strains (96 to 100%). In all thirteen strains, a 591-residue protein is predicted to be translated from

239

a 1,776 bp gene (Fig. S1A). The theoretical isoelectric point (pI) and molecular mass of the S.

240

aureus Newman sok translation product are 5.0 and 67,648 Da, respectively. A potential signal

241

peptide cleavage site predicted using the Neural Networks (NN) algorithm is located between

242

positions 36 and 37 (SLA-AA). Subsequent to the removal of potential signal peptide, the

243

deduced molecular size of putative mature SOK protein in Newman strain is 63,663 Da. One

244

potential transmembrane spanning region contains an internal helix cap (residues 25-28), a central

245

transmembrane helix (residues 29-38), and an external helix cap (residues 39-42). This is

246

consistent with the hydrophobicity profile of SOK, which predicts a highly hydrophobic region

247

between residues 25 and 42 (Results not shown). In the S. aureus Newman sequence (accession

248

number AP009351) (4), sok is located between ORFs for two hypothetical proteins (NWMN0049

249

and NWMN0051), which by BLAST sequence comparison resemble a Na/P co-transporter and

250

metabolic/drug transporter, respectively. sok homologues are found in a variety of genome

251

sequences including those of Gram-positive and Gram-negative bacteria (Fig. S1B). S. aureus

252

SOK sequences share 89.9-91.2 % homology with homologues in other staphylococcal species

253

and >33.8 % homology with those of other bacterial species (Fig. S1B).

254

RFLP analysis. The molecular genetic variability of sok was evaluated using an RFLP

255

technique to analyze a 59-member culture collection, from the National Institute of Public Health,

256

Warsaw, Poland, comprised predominantly of methicillin-resistant and methicillin-sensitive S.

257

aureus human clinical isolates. The isolates are from a variety of human infections, predominantly

13

258

from Poland, but also from other European countries collected from 1992 through 2001. Initial

259

characterization of these isolates by pulsed-field gel electrophoresis (PFGE), PCR typing, multi-

260

locus sequence typing (MLST), and spa typing was previously reported (30). RFLP analysis of

261

the sok locus reveals that the 59 isolates segregate into eight unique RFLP types, and this

262

segregation correlates closely with the clustering of spa types (Table 3). RFLP analysis of sok

263

gene reveals 97 % concordance with spa clusters deduced previously by Malachowa et al. (30).

264

The largest RFLP group correlates with the S2 spa cluster that includes 44% of all isolates. This

265

particular spa cluster contains spa types comprised of the identical or similar repeats profile (23,

266

30). Nearly all sok RFLP patterns correlate with a specific spa cluster with two exceptions:

267

isolates 794 and 3502 belong to the S3 cluster, yet they fell into different groups by RFLP

268

analysis of the sok locus. Only the BN4 isolate was not classified to any cluster type, either by spa

269

typing or sok PCR-RFLP.

270

SOK is surface-exposed and co-purifies with membrane fractions. sok in S. aureus

271

strain Newman were disrupted with a Tcr cassette introduced by allelic replacement, as described

272

previously, to generate strain NM10 (∆sok). Strain NM10 was transformed with a plasmid

273

harboring promoter region plus coding region of sok to produce the complemented strain

274

designated NM20. RT–PCR confirmed the lack and restoration of detectable sok expression in the

275

mutant and complemented strain, respectively, plus the absence of any detectable effect on

276

expression of the adjacent downstream gene (Results not shown).

277

Gene expression studies were consistent with immunofluorescence microscopy using rat

278

antisera prepared against rSOK.

Newman, NM10, NM20, and a strain lacking protein A

279

(DU5875) were examined for SOK by immunofluorescence microscopy (Fig. 1). The S. aureus

280

parental and complemented strains expressed detectable levels of SOK on the cell surface. SOK

14

281

was also detectable on the surface of the strain lacking protein A. In contrast, SOK was not

282

detected on the surface of the sok - strain (Fig. 1).

283

Proteins in culture supernates and subcellular fractions were resolved by SDS-PAGE (Fig.

284

2A) and analyzed by immunoblotting (Fig. 2B). The SOK protein was detected in integral

285

membrane fractions but not in fractions containing cytoplasmic proteins or proteins covalently

286

bound to the cell wall (Fig. 2). The membrane fraction contained an immunoreactive band with a

287

migration consistent with a ~55 kDa protein, which is smaller than predicted (67.6 kDa) based on

288

bioinformatic analysis (see above). Internal controls for an extracellular protein (ScpA, 44 kDa),

289

an integral membrane protein (AgrC, 42 kDa), and a covalently bound membrane protein (SrtA,

290

24 kDa) were appropriately detected from the corresponding protein fractions (Fig. 2C).

291

SOK affects S. aureus survival and ROS production in PMNs. Human PMNs were

292

cultured under conditions to promote synchronized phagocytosis of opsonized S. aureus. After 15

293

min, extracellular bacteria were killed with gentamicin, and the percentages of viable intracellular

294

bacteria were quantified after incubation for additional time periods. The parental strain displayed

295

nearly linear killing kinetics throughout the 3 h incubation period following addition of

296

gentamicin, after which 43.7 % of intracellular bacteria remained viable (Fig. 3A). The isogenic

297

sok - strain (NM10) was more sensitive to PMN killing; 49.5 % were killed after 65 min and after

298

3 h, only 12.7 % remained viable. The complemented strain (NM20) showed similar killing

299

kinetics with the parental strain.

300

Disruption of the sok gene affected the PMN ROS production following phagocytosis of S.

301

aureus Newman. Phagocytosis of the parental Newman isolate rapidly induced PMN ROS

302

production compared to resting PMNs (Fig. 3B). Interestingly, ROS levels in cultures harboring

303

phagocytosed S. aureus lacking SOK, were dramatically different compared to cultures containing

15

304

the parental strain (Fig. 3B). Despite nearly identical levels of ROS for the first 40 min, levels

305

rose dramatically for the next 30 min until eventually declining to or below those of PMN cultures

306

harboring the phagocytosed Newman parental strain. The pattern of ROS production was restored

307

to the parental strain by SOK complementation.

308

SOK affects S. aureus sensitivity to 1O2 killing, but not H2O2. Since SOK expression

309

influenced ROS levels and survival in PMNs harboring intracellular staphylococci, we assessed

310

whether SOK expression affects susceptibility to the ROS molecules H2O2 or 1O2. Analysis of the

311

bacterial cells viability of Newman and NM10 strain after 1 h incubation with various

312

concentrations of H2O2 indicated that SOK expression did not affect the viability of S. aureus cells

313

when exposed to H2O2. All strains showed dose-dependent killing effects with respect to

314

increasing concentrations of H2O2. Following exposure to 10 mM H2O2, all strains showed

315

approximately 50 % survival (49.5 ± 10.3, 53.8 ± 4.3, and 53.1 ± 6.1 for Newman, NM10, and

316

NM20, respectively). Following exposure to 100 mM H2O2, all strains showed approximately 5 -6

317

% survival (5.6 ± 0.68, 6.6 ± 1.2, and 5.5 ± 0.95 for Newman, NM10, and NM20, respectively)

318

(Fig. 4A).

319

Although sok did not affect survival of bacteria during incubation with H2O2, additional

320

experiments indicated that the sok - strain was more sensitive to 1O2 than the parental strain (Fig.

321

4B and C). Methylene blue releases singlet oxygen (1O2) species when exposed to light and was

322

therefore used to measure the susceptibility of S. aureus strains to 1O2. In the presence of 0.5 and 1

323

µg/ml methylene blue at 60 min incubation time, the survival rate of both strains dropped

324

significantly, when compared to the same concentration of methylene blue for 30 min incubation.

325

The strain lacking SOK (NM10) was more sensitive to 1O2 than the parental strain after incubation

326

for 30 min with 0.5 and 1 µg of methylene blue by 4 and 24 times, respectively. The highest level

16

327

(498 times) of difference in susceptibility to 1O2 between mutant and wild type strains was

328

observed when bacteria were incubated with 1 µg methylene blue for 60 min. Longer incubation

329

times (2 and 3 h) as well as higher concentrations of methylene blue (3 and 6 µg/ml) were tested

330

for Newman and NM10, but no bacteria were able to survive under such conditions (Results not

331

shown).

332

SOK enhances virulence in a model of staphylococcal endocarditis. To investigate the

333

effect of sok, and thus SOK, on virulence of S. aureus, an infective endocarditis model was used.

334

New Zealand white rabbits, with aortic valve leaflets previously damaged by cardiac

335

catheterization, were challenged with the parental S. aureus strains RN4220 or Newman, or sok-

336

derivatives of the parent strains (NM1 or NM10, respectively). Animals were also challenged with

337

sok- strains complemented with a plasmid encoding sok (NM2 or NM20). Rabbits were

338

challenged intravenously with 1 × 109 CFU of Newman or Newman derivatives. RN4220 and its

339

derivatives were used to infect rabbits at a dose of 2 × 109 CFU per animal. Hearts were harvested

340

immediately from animals that died and from survivors that were euthanized after infection. Heart

341

tissues were examined, and vegetations on aortic valves were removed, weighed, and

342

homogenized to enumerate the bacteria contained in the vegetations. If vegetations were not

343

observed the aortic valves were removed from hearts and homogenized to enumerate bacteria

344

adhering to host tissue.

345

Infection with parent S. aureus RN4220 caused vegetations in all animals (45.2 ± 5.1 mg).

346

In contrast, vegetations were observed in only one of six rabbits infected with NM1 (sok -) (6.0

347

mg) (Fig. 5A). Animals infected with the parental RN4220 showed greater weight loss than

348

animals infected with NM1, and had diarrhea, and mottled faces (Fig. 5B). Vegetations were also

349

observed in animals infected with the complemented strain (NM2) (40.0 ± 7.9 mg). Statistical

17

350

analysis of the vegetation sizes by unpaired t-test showed that parental RN4220 strain and NM2

351

strain produced vegetation sizes that were significantly larger (P< 0.0001 and P=0.0003,

352

respectively) than the vegetations produced by the sok deletion mutant, but that were not

353

significantly different (P=0.58) from each other. Consistent with vegetation sizes, S. aureus

354

RN4220 and NM2 also produced vegetations with larger bacterial loads (CFU log10 of 4.90 ± 0.35

355

and 3.71 ± 0.71/heart, respectively) than the sok mutant (log10 CFU of 1.49 ± 0.18/heart). As

356

measured by unpaired t-test, the bacterial loads of vegetations produced by RN4220 and NM2

357

were significantly different than the bacterial loads of vegetations produced by the sok deletion

358

mutant NM1 (P