The human innate immune protein calprotectin elicits a multi-metal starvation response in Pseudomonas aeruginosa

To combat infections, the mammalian host limits availability of essential transition metals such as iron (Fe), zinc (Zn), and manganese (Mn) in a strategy termed “nutritional immunity”. The innate immune protein calprotectin (CP) contributes to nutritional immunity by sequestering these metals to exert antimicrobial activity against a broad range of microbial pathogens. One such pathogen is Pseudomonas aeruginosa, which causes opportunistic infections in vulnerable populations including individuals with cystic fibrosis. CP was previously shown to withhold Fe(II) and Zn(II) from P. aeruginosa and induce Fe- and Zn-starvation responses in this pathogen. In this work, we performed quantitative, label-free proteomics to further elucidate how CP impacts metal homeostasis pathways in P. aeruginosa. We report that CP induces an incomplete Fe-starvation response, as many Fe-containing proteins that are repressed by Fe limitation are not affected by CP treatment. The Zn-starvation response elicited by CP seems to be more complete than the Fe-starvation response and includes increases in Zn transporters and Zn-independent proteins. CP also induces the expression of membrane-modifying proteins, and metal-depletion studies indicate this response results from the sequestration of multiple metals. Moreover, the increased expression of membrane-modifying enzymes upon CP treatment correlates with increased resistance to polymyxin B. Thus, response of P. aeruginosa to CP treatment includes both single and multi-metal starvation responses and includes many factors related to virulence potential, broadening our understanding of this pathogen’s interaction with the host. Importance Transition metals are critical for growth and infection by all pathogens, and the innate immune system withholds these metals from pathogens to limit their growth in a strategy termed “nutritional immunity”. While multi-metal depletion by the host is appreciated, the majority of metal depletion studies have focused on individual metald. Here we use the innate immune protein calprotectin (CP), which complexes with several metals including iron (Fe), zinc (Zn), and manganese (Mn), and the opportunistic pathogen Pseudomonas aeruginosa to investigate multi-metal starvation. Using an unbiased label-free proteomics response, we demonstrate that multi-metal withholding by CP induces a regulatory response that is not merely additive of individual metal starvation responses, including the induction of Lipid A modification enzymes.


49
Transition metals are essential for all life, and invading microbial pathogens must acquire these 50 nutrients to grow in the host and cause infection. The host innate immune system limits growth 51 of microbial pathogens by withholding essential transition metals through a strategy termed 52 "nutritional immunity" (1, 2). Metal-sequestering innate immune proteins are important 53 components of this host response. Nutritional immunity originally focused on the competition for 54 ferric iron [Fe(III)] wherein host proteins such as lactoferrin sequester Fe(III) and bacterial 55 siderophores scavenge Fe(III) from these host proteins and deliver it to bacteria via siderophore 56 receptors (1). Later, with the discoveries of additional metal-sequestering host proteins and 57 microbial metal-uptake systems, the model for nutritional immunity expanded to include other 58 nutrient metals such as Mn(II) and Zn(II) (2, 3). The S100 protein calprotectin (CP, 59 S100A8/S100A9 oligomer, MRP8/MRP14 oligomer) plays a central role in nutritional immunity

110
CP is found in high concentrations in the lungs of CF patients, which are commonly 111 infected by P. aeruginosa (35-37). Moreover, CP has been shown to reduce the antimicrobial 112 activity of P. aeruginosa toward another CF pathogen, Staphylococcus aureus, which was 113 attributed to its ability to inhibit the production of toxic secondary metabolites by P. aeruginosa, 114 including phenazines and 2-alkyl-4(1H)-quinolones (AQs) (38). We previously reported that CP 115 withholds Fe from P. aeruginosa and thereby causes an Fe-starvation response through 116 targeted analyses of known Fe-acquisition and -regulatory systems (10). Our study also 117 demonstrated a change in virulence factor production in response to CP with the 118 downregulation of the phenazines, which we determined to be a consequence of Fe  155 was grown for 8 hours, which afforded growth to early stationary phase (10). Cells were 156 harvested, and quantitative label-free proteomics was performed using nano ultra-performance 157 liquid chromatography coupled to high-resolution tandem mass spectrometry as previously   Table S1) or downregulated (Fig 1C, Table S1) upon CP treatment showed significantly 165 more interactions than if they were a random collection of genes from the genome, indicating a 166 biological relationship between the proteins in each analysis (Table S2)

201
In addition to known Fe-and Zn-responsive proteins, proteins for the type VI secretion

204
PA1217, PA1218) were downregulated upon CP treatment (Fig 1, Table S1). CP treatment also 205 led to the upregulation of proteins involved in membrane remodeling and tolerance to CAMPs    Fe, Mn, or Zn or lacking all three metals to afford "Fe-depleted", "Zn-depleted", "Mn-depleted" 234 and "metal-depleted" CDM, respectively, and protein abundance was compared to that of PA14 235 grown in metal-replete CDM. Network analysis was used as described above to determine 236 biological connections within the upregulated and downregulated proteomes in response to Fe,

237
Zn, and Mn limitation (Table S1).  limitation were also downregulated in response to CP (Fig. 2, Table S3). These proteins 283 included phenazine biosynthesis and secretion proteins and the NRPS operon noted above.

284
Further analysis of the 99 downregulated proteins that appeared specific to Fe limitation 285 revealed that 9 of these proteins were significantly (p<0.05) downregulated upon CP treatment, 286 but the change did not meet the 1-LFC threshold, possibly indicating a weaker Fe-starvation 287 response to CP than induced by Fe-depleted CDM (Fig S1). Notably, several PrrF regulon

291
Previous studies using an AntR transcriptional and translational reporter strain showed that 292 AntR, which is downregulated by PrrF during growth in DTSB (57), is also downregulated in 293 response to CP treatment during growth in Tris:TSB (10). However, AntR and its regulatory 294 targets AntABC were not detected in this proteomics experiment. To determine whether the Fe-295 starvation response to CP during growth in metal-replete CDM included a decrease in AntR, we 296 used the same reporter strain and found that antR expression is similarly repressed by CP 297 treatment during growth in CDM (Fig S2), indicating that PrrF functions under these conditions. Combined, these data indicate that that CP elicits only a portion of the Fe-starvation response 299 that is observed upon Fe limitation in Fe-depleted CDM.

308
The upregulated proteins that were specific to the Zn-limitation proteome included Zn(II)-309 independent proteins (CynT2, FolE2 ) that were not significantly changed in response to CP.

310
Moreover, two ferredoxins (Fdx2, FdxA) were upregulated by Zn limitation and were not 311 significantly affected by CP treatment (Table S3) (Fig S4B). We performed real-time PCR (RT-PCR) 317 analysis to investigate this finding further and observed that expression of both the znuA and 318 zur mRNAs were induced by CP treatment and by Zn limitation (Fig S4C-D), suggesting that 319 Zur is post-transcriptionally downregulated upon Zn limitation and CP treatment. Together, 320 these observations suggest that the response of P. aeruginosa to Zn starvation is more complex 321 than currently appreciated; this notion warrants further investigation.

322
The comparisons between the responses of PA14 to CP, Fe limitation, and Zn limitation 323 also identified responses that were unique to CP treatment. Several T6SS proteins encoded by the HSI-II T6SS locus (Fig 2A, Fig 3A) were downregulated only in response to CP. This result 325 was surprising given that our recent proteomic studies showed that Fe starvation upregulates 326 the expression of the same T6SS proteins (47). However, the mRNAs encoding three of the 327 HSI-II T6SS proteins -lip2, clipV2, and hsiB2 -were not significantly changed in a subsequent 328 RT-PCR experiment (Fig 3B-D), contrasting with the recent Fe-regulation study (47) and the 329 current proteomics results (Fig 3A). The difference in gene expression between this work and 330 the previous study may result from differences in experimental conditions and post-331 transcriptional effects may be responsible for the distinct effects of CP on protein expression.

332
Also notable was the Feo Fe(II)-import system, which was upregulated only in response to CP 333 but not in response to Fe limitation (Fig 2A, Fig S1). Additionally, proteins involved in 334 membrane modification and CAMP resistance (PmrA, PmrB, ArnA, ArnB, SpeE2) were 335 upregulated only in response to CP (Fig 2B, Table S3). The Feo Fe(II)-transport system has 336 been shown to be regulated by PmrAB (58), possibly explaining the increase in its expression in 337 response to CP. Together, these data indicate that CP treatment elicits a physiological response 338 that overlaps with but is distinct from the P. aeruginosa Fe-and Zn-starvation responses.   (Fig 4B). A previous study with PAO1 360 reported the restoration of protease activity in the ∆znuA and ∆cntO Zn(II)-uptake mutants with 361 the addition of Zn(II) to the assay buffer but not in ∆znuA∆cntO (26). Similar to the ∆znuA∆cntO 362 mutant, we were unable to recover protease activity to Zn-depleted or CP-treated culture 363 supernatants by supplementing the assay buffer with Zn(II) (Fig 4C). The ∆znuA∆cntO double 364 mutant had a marked growth defect during growth in minimal media and significantly lower cell-365 associated Zn when compared to wild type and the single mutants, indicating it was more Zn starved than the single mutants (26). Our current results indicate that Zn-depleted and CP-367 treated cultures were similarly too Zn-starved to recover protease activity with the addition of 368 Zn(II) to the assay buffer.

369
While these results were generally consistent with previous studies, they did not explain 370 the increase in abundance of the cell-associated proteases. We considered that the expression 371 of secretory systems responsible for export each of these proteases may be downregulated. AprF was downregulated below our threshold of 1 LFC (Fig S5A). In contrast, HxcV was 378 strongly repressed when PA14 was cultured in Zn-depleted medium. This protein was not 379 detected in the CP-treated samples; thus, we cannot exclude the possibility that CP similarly 380 repressed its expression. Further gene expression analysis showed that expression of xcpT did 381 not change upon CP treatment or Zn limitation, whereas xcpP expression increased a small but 382 statistically significant amount in response to both CP treatment and Fe limitation (Fig S5B, Fig   383   S5C). Similarly, expression of genes within the xcp operons were not significantly changed 384 under any condition tested (Fig S5D-G). Overall, these data suggest that decreased expression 385 of proteins that secrete Zn(II)-dependent proteases is unlikely to contribute to the decreased 386 protease activity in culture supernatants despite increased levels of these proteins within cells.   low-Mg CDM (Fig 5B, Table S4). PA14 pre-cultured in either metal-depleted CDM or pre-

475
We also observed a robust Zn-starvation response to CP, evidenced by the upregulation 476 of multiple Zn(II)-uptake systems by PA14. In contrast to the incomplete CP-induced Fe-477 starvation response described above, the Zn-starvation response elicited by CP was 478 comparable to that caused by Zn limitation, suggesting a robust response to Zn(II) withholding 479 by CP. Notably, the Zn-starvation response is induced in the absence of a decrease in cell-

499
A previous study demonstrated that CP chelated Zn(II) from proteases when supernatant was 500 treated with CP (39). Further work will be needed to determine the mechanism of decreased 501 protease during Zn limitation and to determine the mechanism of decreased proteased activity 502 during CP treatment of cultures.

514
We also identified what appears to be a multi-metal limitation response upon CP 515 treatment with the induction of the PmrAB two-component regulatory system and its targets 516 ArnA, ArnB, and SpeE2 (Fig 2, Fig 5).

536
In closing, the results from this proteomics study provide new insights into host-pathogen 537 interactions by systematically evaluating how multi-metal withholding by CP impacts P.  551 was grown as previously described in chemically-defined media (CDM). CDM was made as 552 previously described (10, 40) with some modifications (Table S4)

611
Azocasein protease activity assay. Five biological replicates of PA14 were grown as 612 described above. After 8 hours of growth, cultures were centrifuged at 16,000 x g for 10 613 minutes. The supernatants were sterile filtered using 0.2-µm-pore-size filters (Costar) and 614 stored at 4°C overnight. Total protease activity was quantified as previously described (82).

615
Briefly, 20 µL of supernatant was added to 0.5 mL of the 0.3% azocasein solution (50 mM Tris-616 HCl (pH 7.2) 0.5 mM CaCl2 buffer and supplemented with or without 10 µM ZnCl2) and 617 incubated for 30 minutes at 37°C. The reaction was stopped with 0.5 mL 10% tricholoracetic 618 acid. The quenched reaction mixture was centrifuged at 16,000 x g for 20 minutes and the 619 supernatant absorbance was read at 400 nm using a Biotek Synergy HT plate reader. One unit 620 of enzyme activity changes the A400 by 0.01. The enzyme activity was normalized to OD600.

622
Polymyxin killing assay. Three biological replicates of PA14 were grown as described 623 above in CDM, metal-depleted, and low-Mg CDM, and in CDM in the presence of 10 µM CP.

624
After 8 hours of growth, the cultures were centrifuged, and the cells were resuspended in PBS.

625
The OD600 was measured and PBS was inoculated to an OD600 of 0.05, which was subsequently 626 diluted 1:10 into CDM with no CaCl2 supplementation. The 1:10 dilution was used to inoculate