Petrobactin Protects against Oxidative Stress and Enhances Sporulation Efficiency in Bacillus anthracis Sterne

Bacillus anthracis causes the disease anthrax, which is transmitted via its dormant, spore phase. However, conversion from bacillus to spore is a complex, energetically costly process that requires many nutrients, including iron. B. anthracis requires the siderophore petrobactin to scavenge iron from host environments. We show that, in the Sterne strain, petrobactin is required for efficient sporulation, even when ample iron is available. The petrobactin biosynthesis operon is expressed during sporulation, and petrobactin is biosynthesized during growth in high-iron sporulation medium, but instead of being exported, the petrobactin remains intracellular to protect against oxidative stress and improve sporulation. It is also required for full growth and sporulation in blood (bovine), an essential step for anthrax transmission between mammalian hosts.

Bacillus anthracis is a gram-positive bacillus that under conditions of environmental 29 stress, such as low nutrients, can convert from a vegetative bacillus to a highly durable spore that 30 enables long-term survival. The sporulation process is regulated by a sequential cascade of 31 dedicated transcription factors but requires key nutrients to complete, one of which is iron. Iron 32 acquisition by the iron-scavenging siderophore petrobactin is the only such system known to be 33 required for vegetative growth of B. anthracis in iron-depleted conditions, e.g., in the host. 34 However, the extent to which petrobactin is involved in spore formation is unknown. This work 35 shows that efficient in vitro sporulation of B. anthracis requires petrobactin, that the petrobactin 36 biosynthesis operon (asbA-F) is induced prior to sporulation, and that petrobactin itself is 37 associated with spores. Petrobactin is also required for both oxidative stress protection during 38 late stage growth and wild-type levels of sporulation in sporulation medium. When considered 39 with the petrobactin-dependent sporulation in bovine blood also described in this work, these 40 effects on in vitro growth and sporulation suggest that petrobactin is required for B. anthracis 41 transmission via the spore during natural infections in addition to its key functions during active 42 anthrax infections. 43

Importance: 44 45
Bacillus anthracis causes the disease anthrax, which is transmitted via its dormant, spore 46 phase. However, converting from bacilli to spore is a complex, energetically costly process that 47 requires many nutrients including iron. B. anthracis requires the siderophore petrobactin to 3 scavenge iron from host environments. We show that in the Sterne strain, petrobactin is required 49 also for efficient sporulation, even when ample iron is available. The petrobactin biosynthesis 50 operon is expressed during sporulation, and petrobactin is biosynthesized during growth in high 51 iron sporulation medium but instead of being exported, the petrobactin remains intracellular to 52 protect against oxidative stress and improve sporulation. It is also required for full growth and 53 sporulation in blood (bovine), an essential step for anthrax transmission between mammalian 54 hosts. 55 9 After observing that the asb operon is expressed during late stage growth in sporulation 163 medium, we next sought to determine if any of the predicted sporulation sigma factors were 164 required for asbA-F expression. Here, we used plasmid-based transcriptional reporter constructs 165 where the 260bp upstream of asbA (Figure 2A, vertical line) were fused to gfpmut3α, cloned into 166 the pAD123 expression vector, and expressed in a wild-type B. anthracis Sterne background. 167 This construct lacks the predicted binding sites for sporulation-specific sigma factors σ G and σ K 168 but retains predicted binding sites for Fur, σ B , and PerR ( Figure 2A). 169 To measure expression of asbA-F by this construct, strains of wild-type, the 170 transcriptional reporter, and a promoter-less gfpmut3α were grown in sporulation medium as 171 described and the RFUs were similarly calculated. Overall growth kinetics were similar and the 172 260bp asb promoter was sufficient for asb transcription during late stage growth ( Figure 2C, 173 black arrow). The observed increase in RFU for Figure 2B versus 2C is likely an artifact from 174 increased copy numbers of plasmid-based reporters. Together these data suggest that the high 175 iron levels in the sporulation medium do fully repress the asb operon by Fur and that sporulation-176 specific sigma factors are not required for expression of asbA-F during these conditions. 177 Given these conclusions, we next wanted to better understand the population dynamics 178 and kinetics for asb expression relative to sporulation. The chromosome-based translational 179 reporter and wild-type strains grown in sporulation medium were imaged with phase-contrast 180 and fluorescence microscopy at six, eight, ten, and 12 hours post-inoculation. Individual bacilli 181 were scored for Gfpmut3α expression (positive is at least 1.4x above background fluorescence) 182 and sporulation (if they contained phase bright spores) (representative images in Figure 3B and the upregulation of asbA-F during this period suggest that petrobactin is synthesized and 204 may be present in the culture medium. However, the petrobactin-specific catechol moiety 3,4-205 dihydroxybenzoate, was not detected in sporulation medium at 12 hours post-inoculation by the 206 colorimetric catechol assay (data not shown). This could be due to either assay interference by 207 the medium, petrobactin levels below the limit of detection, or suggest an intracellular role for 208 petrobactin. To confirm petrobactin biosynthesis and address these possibilities, we used laser 209 ablation electron spray ionization mass spectroscopy (LAESI-MS) to detect petrobactin both in 210 the spent culture medium and the cell pellets of B. anthracis wild-type and asb mutant strains 211 grown in sporulation medium for 12 hours (33). When compared against our negative control, 212 the petrobactin-null asb mutant strain, LAESI-MS confirmed the catechol assay results as it did 213 not detect petrobactin in the spent culture medium from the wild-type strain ( Figure 4A), 214 indicating no discernable export of this siderophore took place. However, petrobactin was 215 detected in cells of the wild-type strain thus confirming synthesis ( Figure 4A). 216 The use of petrobactin intracellularly might result in association of petrobactin with the 217 B. anthracis Sterne spore so we also subjected wild-type and asb mutant strain spores to LAESI-218 MS (n=3) analysis. This experiment detected petrobactin in wild-type, but not asb mutant strain 219 spores ( Figure 4B). This phenotype could be restored by supplementing growth and sporulation 220 of the asb mutant strain with 25µM of purified petrobactin (n=1). Complete ablation of the 221 spores was confirmed by an abundance of the spore-core-component calcium dipicolinic acid in 222 the chromatograph (data not shown). These data indicate that while petrobactin is not exported 223 into the medium at detectable levels, it is biosynthesized but remains associated with the spore. 224 To this point in our studies, the mechanism of asb expression and role for petrobactin 225 biosynthesis in sporulation remains unclear. Binding sites for asbA-F regulators in the plasmid-226 based asb transcriptional reporter include PerR, an oxidative stress response regulator, and σ B , a 227 general stress response regulator ( Figure 2A). In B. subtilis, σ B is active during early sporulation, 228 but is not required for either sporulation or an oxidative stress response, likely since most σ B -229 regulated genes can be activated by other transcription factors (40, 41). However, Lee et al., 230 found that oxidative stress can induce petrobactin expression and synthesis, even in high iron 231 conditions (35). While sporulation is not known to be preceded by oxidative stress, B. subtilis 232 cells become resistant to oxidative stress upon entry to the stationary phase (41-43). 233 Additionally, oxidative stress protective enzymes are induced during late stage growth of B. 234 anthracis, maintained during sporulation, and two superoxide dismutases become incorporated in 235 the exosporium (27, 28). Taken together with evidence of intracellular petrobactin, we predicted 236 that petrobactin is protective against oxidative stress. 237 To test this hypothesis, wild-type, asb mutant ± 25µM petrobactin, and dhb mutant 238 strains were tested for resistance to the oxidative stressor hydrogen peroxide (H 2 O 2 ) at eight 239 hours of growth (i.e., before sporulation) in sporulation medium. Percent survival was calculated 240 by comparing the treated CFU/mL (those exposed to 10mM H 2 O 2 ) to untreated CFU/mL (water). 241 While about 50% of the wild-type and the dhb mutant strain populations survived oxidative 242 stress exposure, less than 1% of the asb mutant strain population survived ( Figure 4D). This was 243 due to a four-log decrease in the CFU/mL of the asb mutant strain following treatment with 244 10mM H 2 O 2 ( Figure 4C). The defect in survival was rescued by supplementation of the asb 245 mutant strain with 25µM of purified petrobactin to the medium at the time of inoculation ( Figure  246 4C, 4D). These data confirm our hypothesis that petrobactin is protective against oxidative stress 247 during stationary phase-but prior to sporulation-in sporulation medium, which likely supports 248 efficient sporulation and thus transmission between mammalian hosts. 249

Petrobactin is preferred for rapid growth and sporulation in bovine blood. 250
Following death of an infected mammal, blood laden with B. anthracis is exposed to the 251 atmosphere by either hemorrhagic draining or the activity of scavengers on the carcass (11, 44, 252 45). Since vegetative bacilli are not easily infectious, B. anthracis transmission requires 253 sporulation in aerated blood, a process triggered when the blood-borne CO 2 reported to suppress 254 sporulation decreases following death, thus triggering the sporulation cascade in a race against 255 decomposition (44). Experiments to test in vivo sporulation are ethically and technically 256 challenging, so to determine the relevance of each iron acquisition system-petrobactin, hemin, 257 and bacillibactin-to disease we measured sporulation in bovine blood. Cultures of wild-type B. 258 anthracis Sterne, the asb mutant, the dhb mutant, and the isd mutant strain (a mutant in hemin 259 utilization) were grown in defibrinated bovine blood with shaking for three days. Every 24 hours, 260 the total and sporulated CFU/mL were enumerated. 261 Compared to wild-type at 24 hours, growth of the asb mutant strain was reduced by one 262 log ( Figure 5A) with two log fewer spores ( Figure 5B) whereas all other strains-the isd and the 263 dhb mutant strains-had equivalent CFU/mL. While percent sporulation at 24 hours is low, 264 generally < 25%, most sporulation in the wild-type strain appears to occur during the first 24 265 hours of incubation, after which non-sporulated cells begin to die thus reducing the total 266 CFU/mL and increasing percent spores. Conversely, the asb mutant strain demonstrated delayed 267 sporulation, gaining an additional log of spores between the 24 and 48-hour timepoints, but the 268 percent sporulation at 48 hours and 72 hours was < 25% compared to the wild-type strain at 80% 269 ( Figure 5C). Percent sporulation for both the isd mutant strain and the dhb mutant strain were 270 about 50%, though these were not statistically significant from the wild-type strain ( Figure 5C). 271 Additionally, both total and spores CFU/mL for both the isd and the dhb mutant strains were like 272 wild-type, suggesting that petrobactin is a preferred iron gathering system during growth in 273 bovine blood ( Figure 5A, 5B). 274 The growth defect and delayed sporulation of the asb mutant strain could be due to 275 oxidative stress, a lack of available iron, or a combination of the two stresses. To separate the 276 effects of petrobactin supplementation on iron acquisition and protection from oxidative stress, 277 the asb mutant strain was supplemented with 25µM of either petrobactin or hemin (n=2). Hemin 278 is the oxidized form of heme, which is released into blood by the lysis of red blood cells and can 279 be bound by B. anthracis hemophores, making it biologically relevant (46, 47). However, hemin 280 isn't known to protect against intracellular oxidative stress, so we predicted that if petrobactin 281 were only required for iron acquisition, then hemin supplementation should complement the asb 282 mutant strain phenotype. 283 Supplementation of the asb mutant with hemin did not affect overall growth but appeared 284 to enhance early sporulation whereas supplementation with petrobactin rescued both growth and 285 sporulation ( Figure 5A-C). These data suggest that the iron provided via hemin may allow for 286 efficient sporulation while the dual benefits of petrobactin iron acquisition plus protection from 287 oxidative stress enable continued growth prior to the onset of sporulation. 288

Discussion: 289
In this work, we show that petrobactin is not required for B. anthracis Sterne germination 290 ( Figure 1A) but is required for efficient sporulation in sporulation medium ( Figure 1B). Using 291 fluorescent asbA:gfpmut3α reporter fusions, we also show that asb is both transcribed and 292 translated during late stage growth of B. anthracis Sterne prior to sporulation in a sporulation-293 sigma-factor independent manner ( Figure 2B-C, 3). Unlike during vegetative growth, petrobactin 294 is not exported during sporulation but remains intracellular ( Figure 4A) where it has a significant 295 role in protecting against oxidative stress ( Figure 4C-D) and eventually associates with the spore 296 ( Figure 4B). These findings may have relevance to transmission since petrobactin is also 297 required for efficient sporulation in bovine blood ( Figure 5), a pre-requisite for survival and 298 transmission of the pathogen (11, 44). We believe this to be the first demonstration that a 299 siderophore is induced in preparation for sporulation and present in the mature spore. 300 The iron-gathering capacity of siderophores has long been appreciated for their role in 301 pathogenicity and since their discovery, evidence for alternate functions has accumulated. 302 Multiple reports have demonstrated roles for siderophores in: cell signaling, sporulation 303 initiation, protection from copper and oxidative stress, the generation of oxidative stress against 304 competitors, and, most recently, in survival via spores (48, 49). 305 In early 2017, Grandchamp et. al showed with B. subtilis that siderophore 306 supplementation (including with the native bacillibactin) caused the onset of sporulation to occur 307 earlier (49). Since this enhancement required import of the siderophore into the bacterial cell and 308 iron removal by corresponding hydrolases, these authors hypothesized that the extra intracellular 309 iron acted as a signal for the onset of sporulation (49). However, their study did not address 310 bacillibactin regulation, export during sporulation, nor the cell stresses associated with 311 sporulation. So, to our knowledge, this is the first demonstration that a siderophore is induced to 312 protect against oxidative stress prior to sporulation in high iron conditions. 313 As noted in the introduction, siderophores are primarily regulated by the iron-dependent 314 repressor Fur. However, some siderophores, such as petrobactin, are biosynthesized in response 315 to oxidative stress conditions and other catecholate-containing siderophores (e.g., enterobactin 316 and salmochelin) are protective against reactive oxygen species (35, 50-55). This protection is 317 not due to iron sequestration that prevents additional Fenton reactions but is a function of the 318 antioxidant properties of catechols (50, 53). Supplementation with free catechols doesn't rescue 319 the protective function of enterobactin, which requires import and hydrolysis for effective 320 oxidative stress protection (53, 54). It is unclear whether petrobactin requires additional 321 processing to become active against oxidative stress, though the detection of petrobactin 322 associated with spores by mass spectrometry suggests it does not. 323 There are two non-exclusive hypotheses for siderophore upregulation during oxidative 324 stress: one, that superoxide radicals oxidize iron co-factors thus inactivating key enzymes and 325 two, that upregulation of the enzymes to mitigate oxidative stress require metallic (e.g., iron and 326 manganese) co-factors. Both of these would reduce the intracellular iron pool and thereby relieve 327 iron from Fur to enable iron acquisition system expression (35, 51). In the case of Bacillus spp., 328 the intracellular iron pool is further depleted during the onset of sporulation due to upregulation 329 of aconitase, an iron-rich citrate isomerase and stabilizer of σ K -dependent gene transcripts (56, 330

57). 331
While the demand for iron during oxidative stress and/or sporulation may relieve 332 negative regulation by Fur, it is likely that asbA-F expression is induced by an oxidative stress 333 regulator such as PerR. Enterobactin is positively regulated by the oxidative stress response and 334 there is compelling evidence linking Azotobacter vinelandii catecholate siderophores to similar 335 regulation (53-55). The observed phenotype for those siderophores is similar to that observed by 336 Lee et al. for petrobactin: high iron repression of the siderophore can be overcome by oxidative 337 stress (35, 53, 55). Petrobactin biosynthesized within the cell may then become randomly 338 associated with the prespore. More work is needed to better characterize regulation of the asb 339 operon and petrobactin biosynthesis. 340 As growth in blood marks an endpoint for an anthrax infection, the bacilli must not only 341 grow well, but also prepare for survival and transmission between hosts. Evidence in the 342 literature suggests that exposure of blood-borne bacilli to oxygen as a dying host bleeds out 343 begins the signaling cascade for sporulation creating a direct link between growth and 344 sporulation in blood and transmission (11, 44). It's known that petrobactin is required for growth 345 in macrophages and iron-depleted medium, but that requirement had not been demonstrated for 346 growth or sporulation in blood prior to these experiments. Our data suggest that petrobactin is the 347 preferred iron acquisition system for growth and sporulation in bovine blood, despite multiple 348 potential iron sources. While, petrobactin was required to achieve wild-type growth of 10 8 349 CFU/mL in blood, the asb mutant was still able to grow to 10 7 CFU/mL suggesting that another 350 iron acquisition source was functioning, likely either the isd system or bacillibactin. More work 351 is needed to fully understand the contributions of each iron system to growth and sporulation and 352 to verify these findings in other B. anthracis strains.  The asb translational reporter strain was grown in ModG sporulation medium with both phase 643 contrast and Gfpmut3α fluorescent micrographs taken at six, eight, ten, and 12 hours of growth 644 and the bacteria scored for fluorescence and sporulation. A) Pooled data from two replicates of