Fluorescent protein-based reporters reveal stress response of intracellular Salmonella enterica on single cell level

Intracellular bacteria such as Salmonella enterica are confronted with a broad array of defense mechanisms of their mammalian host cells. The ability to sense host cell-imposed damages, and to mount efficient stress responses are crucial for survival and proliferation of intracellular pathogens. The various combinations of host defense mechanisms acting on intracellular bacteria and their individual response also explain the occurrence of distinct subpopulations of intracellular S. enterica such as dormant or persisting, slowly or rapidly replicating cells. Here we describe a set of fluorescence protein (FP)-based reporter strains that were used to monitor the expression of cytoplasmic or periplasmic stress response systems on a single cell level. This is mediated by a fast maturing FP as reporter for induction of stress response genes. We evaluated slower maturing FPs for a second function, i.e. the analyses of the status of intracellular proliferation of pathogens. The combination of two FPs allows, on a single cell level, the interrogation of stress response and intracellular proliferation. Application of these reporters to S. enterica allowed us to detect and quantify distinct intracellular subpopulations with different levels of stress response and proliferation. Importance Sensing of, and responding to host-mediated damages are important defensive virulence traits of bacterial pathogens. Intracellular pathogens such as Salmonella enterica are exposed to various types of antimicrobial host cell defenses that impose, among other, periplasmic and cytosolic stresses. Intracellular S. enterica form distinct subpopulations that differ in proliferation rate, metabolic activity and persister formation. Here we deploy fluorescence protein-based reporter strains to monitor, on a single cell level, the response of intracellular S. enterica to periplasmic or cytoplasmic stress. A second fluorescent protein reports the biosynthetic capacity of individual intracellular S. enterica. The dual fluorescence reporters can be deployed to characterize by flow cytometry phenotypically diverse subpopulations and stress responses in intracellular bacteria.

The ability to sense environmental cues and to respond to potentially detrimental factors is of 51 vital importance for bacterial cells. In addition to various common stress factors, pathogenic and 52 facultative intracellular bacteria such as Salmonella enterica serovar Typhimurium (STM) face 53 various defense mechanisms of the host that normally result in clearance of bacteria that breach 54 borders within a host. The proper response to environmental or host-imposed damages thus is an 55 important part of the defensive virulence factors. As attack of host cells against intracellular 56 pathogens is multifactorial and acting on various bacterial structures, stress responses have to 57 perform efficiently in order to mount protection and repair that is coordinated in space and time. 58 Gram-negative bacteria possess an array of stress response systems (SRS) that can sense harmful 59 conditions and damages to the cell envelope, or to periplasmic or cytosolic components. 60 Periplasmic stress often results in the accumulation of misfolded proteins that cannot be properly 61 assembled into polymeric surface structures such as fimbriae, or inserted into the outer 62 membrane. A common response to the accumulation of misfolded or otherwise damaged proteins 63 in the periplasm is activation of proteolytic degradation (reviewed in 1). Here, periplasmic 64 protease HtrA, a.k.a. DegP has a key role as multifunctional protein quality control factor. HtrA 65 functions both as proteolytic enzyme and chaperone with broad substrate specificity, but only 66 targeting misfolded and/or mislocalized proteins. Such proteins are recognized by the specific 67 PDZ domain 1 found in proteins important for signal transduction and proteolytic activity. HtrA 68 has a unique molecular architecture, i.e. a serine protease domain and two PDZ domains per 69 monomer. Their activity is regulated by the conserved mechanism by oligomerization. To 70 support refolding of denatured proteins, peptidyl-prolyl isomerases such as FkpA are present in 71 the periplasm. Their main mode of action is the cis-trans isomerization of peptidyl-prolyl bonds. 72 It is a crucial step in protein folding because both cis and trans peptide bonds are found in 73 prolines of native proteins. However, the isomeric state of prolines in each protein is 74 characteristic and important for proper protein function. FkpA belongs to the FKBP family 75 because of its binding to FK506. In addition, FkpA acts as chaperone with activity independent 76 of its peptidyl-prolyl isomerase activity (2). Expression of factors responding to periplasmic 77 stress is coordinately regulated by DegS and the sigma factor E pathway, the two-component 78 system BaeRS, and two-component system CpxAR with its periplasmic accessory component 79 Another form of damage is the loss of turgor resulting from hyperosmotic environments, or 81 damage of the cytoplasmic membrane barrier. As one form of counteraction, the periplasmic 82 trehalase TreA is expressed that catalyzes hydrolysis of osmo-protectant trehalose into two 83 glucose molecules (reviewed in 4). 84 The cytosolic stress response is required, among others, for repair of oxidized cysteine residues 85 in cytosolic proteins that is achieved by functionally partially redundant thioredoxins (Trx) and 86 glutaredoxins (Grx) (reviewed in 5, 6). Effector proteins translocated by the SPI2-T3SS remodel the host cell endosomal system, 111 resulting in the formation of an extensive network of interconnected membrane vesicles termed 112 Salmonella-induced filaments (SIF). This action of the SPI2-T3SS has been linked to the 113 nutrition of intracellular STM, the avoidance of host cell defense mechanism, and reduction of 114 exposure to stress factors (11-13). 115 Although STM is equipped with this range of defensive and offensive virulence factors and able 116 to deploy these, the individual fate of intracellular STM is surprisingly heterogenous (14). While 117 a subset of STM cells initiate rapid intracellular proliferation, another subset remains non-118 replicating or dormant, yet another subset is killed by antimicrobial defenses of the host cell. If 119 the divergent intracellular fate of STM is a consequence of the ability to respond to stress factors 120 remains to be determined. 121 Analyses of mutant strains deficient in SRS have been performed, but did not reveal the role of 122 stress response on single cell level. We aimed to generate a reporter system for single cell 123 analyses of response of STM to stress within host cells, and as well allows quantification of 124 bacterial proliferation. Here we introduce a dual fluorescence reporter system that enables, for 125 single intracellular STM cells, analysis of activation of specific stress responses and intracellular 126 proliferation. 127  (Fig. S 1A). Despite the reported differences in toxicity of some FP, all FP were sufficiently 145 well tolerated by STM at the expression levels used for the dual color reporters. 146

Design of fluorescent protein-based reporters for stress response of intracellular
We analyzed the effect of the bacterial growth phase on dilution of DsRed (Fig. S 1B) Fig. 1E) demonstrated that mCherry and sfGFP showed rather constant FP intensities, with 155 decrease to 54.5 % and 28.9 % of initial x-median RFI, respectively. The red FP tagRFP-T and 156 mCherry2 revealed more pronounced decrease to 11.8 % and 12.9 % of initial intensity at 4 h of 157 subculture. While the decrease of x-median RFI in log phase culture was comparable for both 158 FP, tagRFP-T showed more rapid increase in RFI than mCherry2 in STM entering the stationary 159 phase. Of the FP analyzed, DsRed showed the strongest growth phase-dependent decrease of x-160 median RFI to a minimum of 3.0 % of initial fluorescence (Fig. 1E). 161 For further correlation of growth kinetics to DsRed dilution, we cultured STM in media that lead 162 to different growth rates (Fig. S 1CD). Growth in synthetic PCN (Phosphate, Carbon, Nitrogen) 163 medium resulted in culture yield comparable to LB medium, but altered growth kinetics. Lowest 164 growth rate was observed in PCN with maltose as C-source (0.66), moderate growth rate in PCN 165 with glucose (0.98), and highest growth rate in LB (1.33) (Fig. S 1C). Quantification of DsRed 166 fluorescence again showed a sharp decrease in RFI during culture in LB with a minimum at 4 h 167 of subculture 3.6 % of the initial DsRed intensity of the inoculum (Fig. S 1D DTT or polymyxin B (Fig. S 2B). Furthermore, the growth phase-dependent expression of the 205 various stress reporters was analyzed (Fig. S 2C). The analyses revealed that P treA was highly 206 induced in stationary cultures (5.5-fold), while P htrA and P msrA showed about 2.4-fold increased 207 induction in stationary phase cells. The other reporters showed no increased induction in 208 stationary phase (Fig. S 2D). 209 Based on these data, STM was exposed to 1 mM H 2 O 2 or 0.3 mM methyl viologen in the late log 210 phase of subculture in chemically defined media to induce cytosolic stress. For periplasmic stress 211 induction, STM was exposed to 3 mM DTT, or 10 µg x ml -1 polymyxin B (PM). Furthermore, 212 osmotic stress was imposed by 1 M sucrose (suc), or heat shock by shift of cultures to 42 °C for 213 3 h. For additional induction of treA, culture growth to stationary phase was used (stat). The 214 levels of sfGFP expression in non-stressed and stress-exposed STM were determined by FC. As 215 indicated in Fig. 2A and B, each of the stress reporters showed a specific induction by stressor 216 addition, or stress exposure. However, the levels of induction varied between the reporters. For 217 CSR, we found that the P msrA fusion resulted in the highest induction (3.9-fold) of sfGFP 218 fluorescence, while 5.5-fold induction was observed for PSR sensor P treA in stationary cultures. 219 In addition, stress reporters were checked for cross-induction to other stressors than what they 220 should be reacting to (Fig. 2C). For that, PSR reporters were tested for response to CS, and vice 221 versa. Concentrations and application of stressors was performed as described above and the 222 levels of sfGFP expression were determined by FC. Reporters for PSR (P htrA and P fkpA ) only 223 showed induction upon exposure to periplasmic stressors DTT and PM, whereas no induction 224 could be observed when H 2 O 2 or MV was added, indicating specificity of the reporters. The 225 P treA ::sfgfp fusion was induced when H 2 O 2 , MV or DTT were added, but only low induction was 226 observed after PM treatment. The reporters for the CSR (P msrA , P trxA , P grxA , and P bisC ) only 227 responded to CS triggered by H 2 O 2 or MV. Increased induction after addition of the PS inductor 228 PM was not observed. However, after stressing with DTT, slight induction of P msrA and P trxA , and 229 higher induction of P grxA and P bisC was detected. Thus, also the reporters P msrA and P trxA for CRS 230 showed specificity to CS. A low degree of cross-induction was observed for CSR reporters P grxA 231 and P bisC after addition of the periplasmic stressor PM. 232

Response of reporters to stress exposure in intracellular Salmonella 233
We next analyzed the response of stress reporters in intracellular STM. STM WT harboring the 234 various reporters for PSR or CSR were used to infect the human epithelial cell line HeLa, or the 235 murine macrophages-like cell line RAW264.7 (Fig. 3). The intracellular bacteria were released 236 from host cell 8 h p.i. and subjected to FC. For comparison, the bacterial inoculum was analyzed 237 using the same settings for FC. STM WT harboring a plasmid for constitutive expression of 238 DsRed, but lacking expression of sfGFP served as negative control for adjustment of FC gating. 239 For all stress reporters investigated, we observed increased induction of sfGFP levels in 240 intracellular STM (P fkpA ::sfgfp and P mrsA ::sfgfp were induced 7.32-fold and 7.34-fold, 241 respectively, in HeLa cells, and 6.21-fold and 7.02-fold for P trxA ::sfgfp and P grxA ::sfgfp, 242 respectively, in RAW264.7 cells. For most sensors, the overall sfGFP expression in intracellular 243 bacteria was higher in RAW264.7 cell than in HeLa. For P bisC and P fkpA , induction was 244 comparable for intracellular bacteria in both cell lines. Since STM grown to stationary phase 245 were used for RAW264.7 infection, the expression levels in the inoculum were higher than in 246 late log phase cultures used for invasion of HeLa cells. 247 We conclude that the various promoter fusions are suitable to report the stress exposure of 248 intracellular STM. 249

Role of detoxification and repair mechanisms 250
As pathogen adapted to life within host cells, STM efficiently deploys stress responses, 251 detoxification and repair mechanisms to thrive in the vacuolar environment of the SCV. 252 Therefore, we investigated the stress response induction of STM WT within InterferonIFN-253 induced RAW264.7 macrophages at early time points post infection used P msrA as representative 254 for SR induction (Fig. 4A). We observed higher induction of P msrA ::sfgfp reporter when analyzed 255 in IFN-induced RAW264.7 macrophages. At 2, 4, 6, and 8 h p.i., the x-median RFI of sfGFP 256 was significantly higher for STM in activated compared to resting RAW264.7 macrophages. 257 Because the difference for P msrA ::sfgfp induction was most pronounced at 2 h p.i. we next 258 analyzed the induction of all stress reporter fusions at 2 h p.i. in resting and activated 259 macrophages (Fig. 4B). At this time point, all reporter fusions apart from P treA ::sfgfp showed an 260 increased x-median RFI of sfGFP in STM in IFN-induced RAW264.7 compared to resting 261 macrophages. 262 In addition, we anticipated that in absence of functional detoxification and/or repair mechanisms, 263 intracellular STM will be exposed to higher levels of stress, and in turn show For the stress reporters P msrA , P htrA , P trxA , and P treA , we observed higher induction in sodAB and 271 dksA strains (Fig. 4C). Strongest effects were determined in dksA background with 2.12-and 272 2.27-fold increased x-median RFI for P htrA ::sfgfp and P treA ::sfgfp reporters, respectively. These 273 data confirm that stress sensors report stress exposure of intracellular STM, and that 274 functionality of detoxification and repair mechanisms affects expression levels. 275

Role of the SPI2-T3SS in ablation of stress in intracellular Salmonella 276
In addition to SRS that are conserved in many bacteria, specific virulence factors also contribute 277 to avoid exposure of intracellular bacteria to host defense mechanisms, and by this exposure to 278 stress factors such as antimicrobial peptides, acidic pH, ROS and RNS, and nutritional deprivation. A role of the SPI2-T3SS and its effector proteins in converting the SCV into an 280 environment permissive for survival and proliferation been proposed (11). 281 We used P htrA and P msrA as reporters representative for PSR and CSR, respectively. The reporters 282 were introduced in STM WT, a ssaV-deficient strain unable to translocated SPI2-T3SS effector 283 proteins, the sifA-deficient strain unable to induce endosomal remodeling, and the sseF-deficient 284 strain that exhibit a reduced capacity in endosomal remodeling. While ssaV and sifA mutant 285 strains are highly attenuated in virulence in murine models of systemic salmonellosis, the sseF 286 mutant strain shows moderate attenuation (18, 21, 22). 287

Analyses of stress reporter induction for intracellular STM indicated strongest induction in the 288
ssaV background in RAW264.7 cells (Fig. 5, Fig. S 3, 1.6-fold and 1.4-fold for P msrA and P htrA , 289 respectively). These reporters showed also slightly higher induction in sifA (only P msrA, , 1.22-290 fold) and sseF background compared to STM WT (1.22-fold and 1.1-fold for P msrA and P htrA , 291 respectively). As shown in Fig. 3, induction of the reporters in HeLa cells was less pronounced. 292 Comparison of induction in STM WT, ssaV, sifA and sseF strains revealed only minute 293 strain-specific differences. Expression of the P htrA fusion in the sifA background was 294 significantly lower 0.75-fold) than for sseF and ssaV strains in RAW264.7 cells, and 295 significantly lower compared to all other strains in HeLa cells. This reflects the release from the 296 SCV for a part of the intracellular population of STM sifA and relief from stress imposed 297 within the SCV. 298 To test the coordination of stress response by intracellular STM, we analyzed the P msrA ::sfgfp 299 reporter in the background of a phoPQ strain (Fig. S 4). This strain is highly attenuated due to 300 the defect in a two-component system with global control of expression of stress response 301 functions and virulence genes (reviewed in 23). In the phoPQ background, the induction of 302 We conclude that the ability of intracellular STM to manipulate the endosomal system of host 304 cells by action of the SPI2-T3SS effector proteins contributes to reduction of stress exposure, 305 and is especially important in host cells such as phagocytes with a complex set of antimicrobial 306 defense mechanisms. 307 Analyses of the proliferation rate of intracellular Salmonella using dual color reporters 308 The We gated subpopulation 1 and subpopulation 2. Quantification of RFI determined 1.61-fold 321 higher sfGFP x-median RFI (Fig. 6DE), and 1.59-fold lower DsRed x-median RFI (Fig. 6DE) 322 for subpopulation 1 compared to subpopulation 2. 323 For further test of the reporter system, we investigated the reporter plasmid in STMsifA 324 recovered after infection of HeLa cells (Fig. 6FGHIJ). This mutant strain shows hyper-325 replication resulting from fast cytosolic proliferation (see Fig. 1H). We were able to distinguish a 326 smaller subpopulation 1 with high DsRed and sfGFP levels, and a larger subpopulation 2 with 327 about 2-fold reduced sfGFP levels and highly reduced x-median RFI for DsRed (Fig. 6IJ). in intensities can be observed in either subpopulation (Fig. 6). We conclude that subpopulation 1 334 of STM WT in Fig. 6AB consists of intracellular proliferating STM with dilution of DsRed. 335 Subpopulation 1 is responding to stress factors in the intracellular environment. In contrast, 336 subpopulation 2 shows reduced intracellular proliferation, resulting in less DsRed dilution and 337 higher levels per cells (Fig. 6C). The response to intracellular stress in subpopulation 2 is 338 reduced (Fig. 6E). 339 With the dual color fluorescence reporters, we have at hand a system that allows single cell 340 analyses of response of intracellular STM to environmental stimuli, as well as the interrogation 341 of the proliferation state of the pathogen. 342

Generation of reporter plasmids 446
For the generation of dual color fluorescence reporter plasmids, p4889 (P EM7 ::dsred P uhpT ::sfgfp) 447 with a constitutive expression of DsRed and GOI-regulated expression of sfGFP was used as 448 described before (Noster et al., 2019). Via Gibson Assembly (GA) of PCR fragments, the uhpT 449 promoter was replaced by promoter fragments of htrA, fkpA, bisC and grxA. Primers are listed in 450 Table S 1, and the resulting dual color reporter plasmids are listed in Table 2. 451

Generation of complementation plasmid 452
For the generation of a complementation plasmid, low-copy vector pWSK29 was used as a 453 vector backbone. By GA, the promoter, open reading frame and terminator of the dksA gene was 454 inserted as one fragment into pWSK29. Primers are listed in Table S 1, and the resulting  455 complementation plasmid is listed in Table 2. To release the intracellular bacterial population, infected cells were lysed at the indicated time 531 points using 0.5 % Triton X-100 in PBS for 10 min at RT with shaking. The lysate was 532 transferred to a test tube, and after pelleting of host cell debris by centrifugation for 5 min at 500 533 x g, bacteria were recovered from supernatant. Bacteria were further centrifuged for 5 min at 534 20,000 x g and fixed in 3 % PFA in PBS for 15 min at RT. After fixation and a further 535 centrifugation step, fixed bacteria were resuspended in 250 µl 100 mM NH 4 Cl in PBS for 536 quenching of residual free aldehydes. After that, samples were directly subjected to FC. To 537 measure the pre-induction of dual color reporters, bacteria from the inoculum were directly 538 harvested by centrifugation and fixed as described above. 539 To measure intensity decrease of red fluorescence during growth of a subculture of STM WT 540 within LB or PCN medium supplemented with glucose or maltose or to measure the impact of 541 the maturation time of various red fluorescent proteins on the intensity decrease during growth, 542 STM WT harboring the respective plasmids was grown o/n in the respective growth medium, 543 diluted to an OD 600 of 0.05 in fresh medium and subcultured over time. Every hour the OD 600 544 was determined and samples were taken, diluted in PBS and directly subjected to FC.    were taken and subjected to flow cytometry (FC) to determine the constitutive fluorescent 792 events per ml (bacteria x ml -1 ) B) STM WT harboring plasmid with constitutive expression of 793 dsred was cultured as described above in LB medium. FC measurements were performed as 794 described above to measure the constitutive DsRed intensity. In addition, the OD 600 was 795 determined. DsRed intensity course shows rapid decrease of DsRed intensity during 796 logarithmic growth phase and recovery of intensity after entering stationery phase. C) STM 797 WT harboring plasmid with constitutive expression of dsred was cultured as described above 798 in various growth media. OD 600 and FC measurements were performed as described above.