All experiments were performed using a local clinical isolate of Providencia stuartii (Ps, M21250) and Providencia rettgeri (Pr, M17517), and Escherichia coli ATCC^® 25922™ (Eco) as a positive control. Other clinical isolates of Pr (n=4) and Ps (n=4) were included for studying some of the parameters, and were obtained from patients of different hospitals from Buenos Aires between April 2013 to September 2017 (See table in Supplementary Material 1 for details of the isolates used). All Pr or Ps strains were different from each other according to their pulse field gel electrophoresis (PFGE) pattern. Bacterial identification was performed by typical biochemical methods and by Maldi TOF (Bruker, London, UK). Susceptibility testing was determined by disc diffusion following protocols and interpretation criteria established by CLSI.

The strains were cryopreserved at -80 °C. For use, they were seeded in tryptic soy agar (TSA) and incubated at 35 ± 2 °C for 18 h to obtain isolated colonies. Then, a single colony was inoculated in tryptic soy broth (TSB) and incubated at 35 ± 2 °C for 24 h in static conditions. After that, 100 µL of the culture was added into 10 mL of fresh TSB, and grown for another 4 h with agitation, until reaching log phase. Bacteria were pelleted by centrifugation at 9600 x g for 15 min, washed twice in phosphate buffered saline (PBS) 1 x and resuspended at the desired concentration in PBS 1 x. Bacteria concentration was determined by measuring O.D. at 600 nm, and adjusting to 0.09 absorbance units, that is equivalent to 1 x 10⁸ colony forming units (CFU)/mL. CFU concentration was confirmed by counting on TSA plates. Bacteria were prepared at the desire working concentration in RPMI 1640 supplemented with 2% of heat-inactivated fetal calf serum (FSC). All experiments were performed using a ratio PMN:bacteria of 1:10 unless otherwise specified.

Blood samples were obtained from healthy volunteer donors who had not taken any medication for at least 10 days before the day of sampling. Blood was obtained by venipuncture of the forearm vein and was drawn directly into citrated plastic tubes. This study was performed according to institutional guidelines (National Academy of Medicine, Buenos Aires, Argentina) and received the approval of the institutional ethics committee (N° 19/22/CEIANM), and written informed consent was provided by all the subjects.

PMN were isolated by Ficoll-Hypaque gradient centrifugation (Ficoll Pharmacia, Uppsala; Hypaque, Wintthrop Products, Buenos Aires, Argentina) and dextran sedimentation, as previously described (17). Contaminating erythrocytes were removed by hypotonic lysis. After washing, the cells (96% PMN on May Grünwald/Giemsa-stained Cyto-preps) were suspended in RPMI 1640 supplemented with 2% FCS and used immediately after. All experiments where PMN were used were performed in RPMI 1640 supplemented with 2% FCS.

Bacteria alone or bacteria in the presence of PMN in a 1:1 ratio were left for 1 h or 3 h at 37 °C, 5% CO_2. After incubation, serial dilutions in H₂O were performed for each sample, (bacteria grown alone for 1 or 3 h or bacteria incubated with PMN for 1 or 3 h) and the resultant solutions were plated in duplicates on TSA. CFU were counted the following day, and CFU were normalized using the following equation: CFU (normalized) = CFU  +  PMN CFU  −  PMN , where CFU + PMN was the number of CFU obtained for bacteria incubated for 1 or 3 h in the presence of PMN, and CFU - PMN was the number of CFU obtained for each bacterium grown for 1 or 3 h in the absence of PMN.

Chemotaxis was quantified using a modification of the Boyden chamber technique (18). PMN (1 x 10⁵) in 50 µL were placed in triplicate in the top wells of a 48-well micro-chemotaxis chamber. A PVP-free polycarbonate membrane (3 µm pore size; Neuro Probe Inc. Gaithersburg MD, USA) separated the cells from lower wells containing either RPMI or the different bacteria (1x10⁶). The chamber was incubated for 30 min at 37 °C in a 5% CO₂ humidified atmosphere. After incubation, the filter was stained with TINCION-15 (Biopur SRL, Rosario, Argentina), and the number of PMN on the undersurface of the filter was counted in a five random high-power fields (HPF) x 400 for each of triplicate.

PMN (5 x 10⁵) were incubated with the different bacterial strains in a PMN:bacteria ratio of 1:10 for 30 min at 37 °C in a 5% CO₂ humidified atmosphere. Then, a specific mouse anti-human CD11b antibody conjugated with phycoerythrin (PE) (Biolegend, San Diego, CA, USA Cat. 301306) was added, and cells were incubated in the dark at 4 °C for 30 min. The FSC and the mean fluorescence intensity of CD11b were determined on 50,000 events by flow cytometry after excluding debris by FSC-SSC using a FACSCanto (Becton Dickinson). The gating strategy used to exclude doublets is depicted in Supplementary Material 2 . The increase in the percentage of FSC and the CD11b mean fluorescence intensity (MFI) was analyze on single cells. For CD11b results were normalized according to the following formula: MFI CD11b (normalized) = MFI CD 11 b of the different treatments  MFI CD 11 b of untreated PMN  ( − ) .

PMN was challenged with bacteria in a 1:10 ratio for 1 h at 37 °C in 5% CO₂. Then cells were fixed, and processed for observation by transmission electron microscopy (TEM). The number of PMN with bacteria inside was quantified in 100 cells, and expressed as the percentage of phagocytosis. The number of bacteria inside individual PMN was also registered and expressed as the frequency of PMN with different number of bacteria (# bacteria/PMN).

ROS production was measured by flow cytometry (FACSCanto, Becton Dickinson) using Dihydrorhodamine 123 (DHR, Sigma Aldrich). PMN (5 x 10⁵) were incubated with 1 μM of DHR for 15 min at 37 °C. Subsequently, cells were incubated with or without the different bacterial strains for 30 min at 37 °C 5% CO₂ in a humidified atmosphere. Immediately after, the H₂O₂ production was determined in the FL-1 channel.

NETs by confocal microscopy were determined as previously reported (19). Briefly, PMN (5 x 10⁵) were seeded gently onto glass coverslips coated with 0.001% poly-L-lysine in a 24-well plate in triplicate, allowed to settle, and incubated in the presence of bacteria. Cells were incubated for 3 h at 37 °C, 5% CO₂. After the incubation period, samples were gently fixed with 4% paraformaldehyde (PFA) for 15 min at room temperature. After fixation, samples were washed three times with PBS 1x and incubated in a blocking buffer (PBS 1 x, 3% BSA) for 1 h at room temperature. Then, a primary rabbit anti-neutrophil elastase antibody (Merk KGaA, Darmstadt, Germany; Cat. 481001) was added and incubated for 1 h at room temperature. After washing with PBS 1 x three times, samples were incubated with a donkey anti-rabbit IgG secondary antibody conjugated with Dy-Light 649 (Biolegend, San Diego, CA, USA; Cat. 406406) for 1 hour at room temperature in the dark. After washing, samples were treated with triton x100 (0.25%) for 15 min, and were mounted with Vectashield mounting medium containing propidium iodide (Vector Laboratories, Inc. Burlingame, CA, USA; Cat H-1300). Images for NETs evaluation were acquired using a FluoView FV1000 confocal microscope (Olympus, Tokyo, Japan) equipped with a Plapon 60 x/1·42 objective lens and processed using Olympus. At least 10 different fields were observed in each triplicate (200 x). NETs areas were determined as previously reported in at least 5 microphotographs obtained in 200 x for each sample (20), using the wand tool of the FIJI software (21). The scale for the measurement was obtained from the data given in the confocal microscope image. To measure nuclear decondensation, PMN were incubated for 90 min with the different bacteria and samples were fixed, permeabilized with triton x 100 (0.25%) for 15 min and stained using the same protocol detailed above. Microphotographs were obtained in 600 x. The number of PMN with decondensed nucleus were quantified in at least 100 cells and expressed as a percentage.

PMN (1 x 10⁶) were incubated in the presence of the different bacteria for time indicated in each experiment at 37 °C, 5% CO₂. After the incubation period, a commercial micrococcal Nuclease S7 (4 U, Roche, Basilea, Suiza) was added for 15 min at 37 °C in order to cut and release the DNA associated with NETs from the cell body of PMN. After inactivation of the enzyme by the addition of 5 mM of EDTA, supernatants were collected and centrifuged twice, first at 900 x g and then at 9600 x g to eliminate bacteria. The presence of double stranded (d.s.) DNA was quantified in duplicates using the commercial kit Quant-iT™ PicoGreen^® (Invitrogen, Thermo Fisher, Waltham. MA. USA) that contains a DNA intercalator, which was read in a fluorimeter (DeNovix DS-11 spectrophotometer).

PMN (1 x 10⁶) were incubated with Pr or Ps in PMN:bacteria ratio of 1:10 for 60 min at 37 °C, 5% CO₂ for vital NETs induction. After the incubation period, NETs were cut from the PMN body as detailed above and a total of 250 µL of this supernatant was used for DNA extraction with the Bosphore Nucleic Acid Extraction Versatile Spin Kit (Anatolia Geneworks, Istanbul, Turkey), following manufacturer’s instructions. Subsequently, 30 ng of the extracted DNA was used in consecutive quantitative PCR (qPCR) reactions, each with a final volume of 20 µL. Reactions included 10 µM of specific primers targeting the actin for nuclear DNA determination (actin-FW 5´- atgtttgagaccttcaacacccc-3´, actin-RV 5´- gccatctcttgctcgaagtccag-3´) or ND1 for mitochondrial DNA detection (ND1-FW 5´-aacatacccatggccaacct-3´, ND1-RV 5´agcgaaqgggttgtagtagccc-3´). All qPCR assays were performed using the SsoAdvanced SYBR Green Supermix (2×) and run in the CFX96 Dx Real-Time PCR Detection System from Bio-Rad (San Francisco, CA, USA). Results were expressed as the fold increase using the following formula: Fold increase = 1 Ct normalized , where Ct normalized = Ct forPr or Ps  Ct for untreated PMN  ( − ) .

PMN (5 x 10⁵) were challenged with bacteria in a 1:10 ratio for 3 h at 37 °C in 5% CO₂. After the incubation period, propidium iodide (PI, 1 µg/mL) was added for 15 min to assess the loss of plasmatic membrane integrity (cell death) and the percentage of PI positive cells were determined by flow cytometry. In another set of experiments, the release of lactate dehydrogenase (LDH) was determined in PMN incubated with the different bacteria in a 1:10 ratio for 3 h at 37 °C in 5% CO₂. After, the incubation period, samples were centrifuged, and the supernatants were used for LDH determination in duplicates using the LDH-L kit (Wiener Lab., Argentina) following manufacturer’s instructions.

PMN were co-incubated with Pr o Ps and PMA (40 ng/mL) or Eco (ratio 1:1, Eco: PMN) at 37 °C, 5% CO₂. After 3 h, NETs area or d.s. DNA was determined as stated above. In some experiments, Ps and Pr were fixed with PFA (4%, 30 min), or bacterial supernatants (sn) were obtained from bacterial cultures grown in RPMI 1640 for 4 h, filtered using a 0.22 µm filter and then supplemented with 2% FCS. PMN in RPMI 2% FSC were then incubated with fixed bacteria or sn (1:1, vol/vol), and NETs formation assays using PMA were performed.

To assess the DNase activity of Pr and Ps or of bacterial sn, purified eukaryotic DNA (600 ng/mL) was incubated with bacteria or sn for 1 h at 37 °C adding Ca2₊ (50 mM) and Mg2₊ (50 mM), and then d.s. DNA was evaluated using the commercial kit Quant-iT™ PicoGreen^® (Invitrogen, Waltham, MA, USA) or an agarose gel was performed to visualized DNA. A commercial nuclease S7 (Nuc S7) (Roche, Basilea, Suiza) was used as a positive control. To reverse DNase activity, samples were treated with EDTA (50 mM).

Genomic DNA was extracted using the QIACube DNAMini Kit (Qiagen, Germantown, MD, USA). WGS was performed on Ps M21250 and Pr M17517 using the Nextera XT DNA library preparation kit and the Illumina MiSeq Platform (Illumina, San Diego, CA, USA). FastQC and Trim Galore were used for quality assessment and trimming. Reads were de novo assembled using Spades and confirmed for species with Kraken. Automated annotation was performed with Prokka. Average nucleotide identity (ANI) values were determined using the ANI calculator and ANI clustermap tool. Genome analysis was performed using the Comprehensive Genome Analysis tool offered by Patric 3.6.8 and with the ANLIS-Malbrán analysis pipeline, including ARIBA (MLST and resfinder) and AMRfinderPlus. Other bioinformatics tools were used to search for or compare specific genes or amino acid sequences of capsule and nucleases, such as BLAST and UniProt, using the appropriate reference for each species (search words: nuclease, endonuclease, and exonuclease).

BALB/c mice were bred in the animal facility of IMEX-CONICET. Male mice aged 9–16 weeks and weighing 20–25 g were used. They were maintained under a 12-hlight–dark cycle at 22 ± 2 °C and fed with standard diet and water ad libitum. The experiments performed were conducted according to the principles set forth in the Guide for the Care and Use of Laboratory Animals (National Research Council, 1996). Bacterial strains were prepared as stated above, and 1 x 10⁷ CFU/mice were inoculated by intraperitoneal (i.p.) injection. After 24 h, mice were sacrifice and the peritoneal liquid (pl) was collected. PMN were counted with turk´s solution using an optical microscope. MPO was determined incubating 40 µL of the pl with 50 µL of the specific commercial substrate 3, 3’, 5, 5’-tetramethyl-benzidine (1 x TMB) (Invitrogen, Thermo Fisher, Ealtham, MA, USA) for 15 min at room temperature in the dark. After the incubation period, the reaction was stopped using 30 µL of H₂SO₄ (1 M), measuring the resultant absorbance at 450 nm, subtracting the absorbance at 570 nm using a microplate reader (Varioskan™ LUX, Thermo Scientific). Released d.s. DNA was measured as detailed above. The spleen and lungs were also obtained and processed to obtain single cell suspensions. To estimate dissemination from the inoculation site, CFU in the lung, spleen or pl recovered at 24 h were determined by serial dilution in TSA plates, and expressed as the relative bacterial distribution, calculated as follows: Relative bacterial distribution (%) = CFU recovered from the lung or spleen or pl total CFU recovered  ( lung + spleen + pl )    x100.

Statistical differences were determined using the Prism 5.0 software (Graph Pad Software, La Jolla, CA) after applying the Shapiro-Wilk normality test. When normality was determined, the one-way analysis of variance (ANOVA) followed by the Tukey´s post-test for multiple comparisons was performed. Otherwise, the non-parametric Kruskal-Wallis test followed by the Dunn´s post-test for multiple comparisons was performed. P values less than 0.05 were considered significant, and the exact p values are shown in the graphs. The eta-squared (η²) size effect of the ANOVAs performed were calculated as the sum of squares for the effect divided by the total sum of squares. This statistical parameter together with the confidence intervals of the comparisons made for each figure are included in the Supplementary Material 3 .

In this study, we selected one clinical isolate of Providencia rettgeri (from now on termed as Pr) and one of Providencia stuartii (from now on termed as Ps) to study how PMN respond to challenges with these isolates. As a positive control, we used an Escherichia coli ATCC (Eco) strain that has been used in previous studies and fully activates PMN (22, 23). We determined whether PMN could efficiently eliminate Pr or Ps performing killing assays. For this purpose, Pr, Ps, or Eco were left alone or were confronted with PMN for 1 h or 3 h, and total CFU were determined and expressed as described in Material and Methods. As depicted in Figure 1 , 1 h after challenge, Eco and Ps were partially eliminated, whereas Pr was not. However, at 3 h post-challenge, Ps increased its survival, reaching values similar to those obtained for Pr. These results indicate that even though Ps is partially contained by PMN at early time points, both Pr and Ps finally subvert PMN-mediated killing, evidencing immune evasion mechanism that allow them to survive.

To study what PMN responses might be related to the ability of Pr and Ps to survive in the presence of PMN, we first analyzed the capacity of PMN to migrate towards the bacterial strains using a chemotaxis chamber. Medium alone or the different bacteria were used as the chemotactic stimulus. PMN were allowed to migrate across a membrane, and migrated PMN were quantified. As observed in Figure 2A , PMN migrated towards Pr, Ps, and Eco compared to medium alone, but chemotaxis towards Ps and Pr was less effective than towards Eco. Moreover, Ps was less chemotactic than Pr.

Activation of PMN causes changes in cell size, reflected as an increase in the forward scatter (FSC), and the up-regulation of surface molecules, such as CD11b, that mediate endothelial binding necessary to transmigrate to the site of infection. To determine if Pr or Ps activated PMN, we measured the percentage of PMN with a high FSC after bacterial challenge. We found that Pr, Ps, and Eco were similarly able to cause an increase in the % of PMN with high FSC ( Figure 2B ). As can be observed in the representative dot-plots, Eco also affected the side scatter (SSC) profile, whereas Pr and Ps did not. Moreover, when we measured the changes in CD11b expression, we found that compared with untreated cells, all bacteria caused a statistically significant increase in the expression of CD11b. However, both Pr and Ps caused a lower up-regulation compared to Eco ( Figure 2C ).

These results indicate that both Pr and Ps can be initially recognized by PMN, migrating towards bacteria and becoming activated by them. However, the induced response is less robust than the one observed with Eco, which fully activates PMN.

To determine whether the PMN responses induced by Pr and Ps chosen for this study were representative of other clinical isolates, we determined the effect of Pr and Ps on CD11b expression using 4 other isolates of Pr and Ps (see Supplementary Material 1 , Supplementary Material 4A ). Although in some cases particular isolate presented slight differences compared to the others, the effect of Pr and Ps on CD11b expression with the pooled isolates was similar to that seen for the chosen isolates.

After initial contact with bacteria, PMN must ingest microorganisms to kill them intracellularly by triggering the production of reactive oxygen species (ROS). In this sense, we first evaluated whether Pr and Ps were phagocytosed by PMN. For this, we measured the percentage of PMN with ingested bacteria by transmission electron microscopy (TEM) after 1 h of incubation. As Figure 3A shows, the percentage of PMN with intracellular bacteria (% Phagocytosis) was similar for Pr, Ps, and Eco. However, when the number of bacteria inside PMN was counted, we found that the percentage of PMN with<2 bacteria per PMN predominates for Pr, but for Ps and Eco PMN with 6 or 7 bacteria were also observed ( Figure 3B ), indicating that even if the % of PMN that ingested bacteria was similar for all groups, less Pr was internalized compared to Ps or Eco.

When the percentage of ROS-producing PMN was evaluated, all bacterial strains induced a statistically significant increase compared to unstimulated PMN. However, the % induced by Pr was lower compared to Ps and Eco, and the % induced by Ps was lower compared to Eco ( Figure 3C ). When the amount of ROS per PMN (MFI ROS) was determined ( Figure 3D ), Pr showed no differences compared to unstimulated PMN. Although the amount of ROS induced by Ps was higher compared to untreated PMN, it was lower than that observed for Eco. These results show that Pr and Ps are poor ROS inducers, being Pr even less stimulatory than Ps, and these differences could be associated with the differences in the number of bacteria phagocytosed by PMN.

When we determined the % of ROS producing PMN with 4 other isolates of Pr and Ps ( Supplementary Material 4B ), we found that the pooled samples showed the same response pattern as the one found with the isolates used throughout this study.

Suicidal NETs release is one of the main mechanisms of extracellular PMN-mediated killing. Together with other granular proteins, elastase co-localizes with DNA fibers after NETs release. To determine suicidal NETs formation, we measured NETs area of released DNA positive for propidium iodide (PI) and elastase by confocal microscopy, and also quantified double-strand (d.s.) DNA released from NETs after 3 h of bacterial challenge. As observed in Figures 4A, B , no NETs were found when PMN were challenged with Pr and Ps compared to Eco, which was a strong inducer of NETs release. Reinforcing this result and taking into account that the release of suicidal NETs leads to cell death, we found that after 3 h of incubation with the different bacterial strains, Pr and Ps did not cause PMN death as measured by the lack of propidium iodide (PI) internalization by flow cytometry, whereas Eco induced a high percentage of PI+ PMN ( Figure 4C ). Additionally, cell death was also assessed by the release of LDH in PMN after 3 h of incubation with the different bacteria, and no LDH release was observed by Pr and Ps (LDH (AU) (–): = 0.285 ± 0.005; Pr = 0.300 ± 0.010; Ps = 0.295 ± 0.005; Eco = 0.705 ± 0.025*, n=3, *p<0.0001 vs (–).). The absence of NETs was not associated with a lack of stimulus, as increasing the bacteria:PMN ratio did not lead to an increase in NETs formation ( Supplementary Material 5A ). Moreover, to investigate if factors present in human serum (HS) could reverse the absence of NETs observed with Pr or Ps, these assays were performed in 50% of HS but no increase in d.s. DNA was observed in these conditions either ( Supplementary Material 5B ).

In line with the results shown in previous sections for other parameters, when PMN challenged for 3 h with 4 other isolates of Pr or Ps, the same results were found as with the chosen isolates and no d.s. DNA release was observed ( Supplementary Material 4C ).

These results indicate that suicidal NETs are not observed and PMN remain viable when challenged with Pr or Ps.

Up to this point, Pr and Ps can induce some degree of activation in PMN, but NETs formation seems to be the most affected PMN function. Therefore, we next asked whether, in addition to not being induced, Pr and Ps could affect the formation of NETs induced by other stimuli. For this purpose, Pr or Ps were confronted with PMN stimulated with Eco or the well-known NET-inducer PMA. As shown in Figure 5 , both Pr and Ps decreased NETs area and d.s. DNA triggered by Eco ( Figures 5A, B ) or PMA ( Figures 5C, D ). These results indicate that Pr and Ps affect NETs triggered by different stimuli, evidencing the presence of NETs evasion strategies that interferes with this type of extracellular-killing mechanism.

To assess if the decrease of NETs in response to PMA caused by Pr and Ps was related to bacterial viability, Pr and Ps were fixed, and dead bacteria were used together with PMA to determine NETs formation. As observed in Figure 6A , dead Pr or Ps could not decrease NETs release triggered by PMA, indicating that bacterial viability is necessary to reduce NETs. Moreover, when cell-free supernatants (sn) obtained from Pr or Ps cultures were used in the presence of PMA, we found that these sn decreased d.s. DNA induced by PMA ( Figure 6B ). This decrease was not cause by the release of toxic factors that may cause PMN death, as PMN viability was not affected in PMN incubated for 3 h with Pr- or Ps-derived sn (% of PI⁺ PMN: Untreated PMN (–) = 3.18 ± 0.34; sn Pr = 2.38 ± 0.20; sn Ps = 2.520 ± 0.03; n=3). These results indicate that live Pr and Ps actively release a factor into the supernatant that can decrease NETs in response to PMA.

A common strategy to subvert NETs-mediated killing is to affect any of the components that comprise NETs. In this sense, one well-known mechanism used by pathogens is to degrade the NET (24). Considering the above finding, we studied whether Pr and Ps release DNases into the sn that degrade the DNA-fibers of NETs. For this purpose, we incubated isolated eukaryotic DNA with filtered sn obtained from bacterial cultures and used a Nuclease S7 (Nuc S7) for comparison. Besides, according to the requirement of DNases to use bivalent cations for optimal activity, we evaluated the impact of EDTA, a bivalent cation chelator. As depicted in Figures 6C, D , the sn of Pr and Ps were able to degrade eukaryotic DNA. Moreover, DNA degradation was reversed by the addition of EDTA. As a complementary analysis of these findings, genomic DNA was extracted and whole gene sequencing was performed on Pr and Ps and their sequences were submitted to GenBank under BioProject PRJNA1304938. The analysis of these sequences resulted in the identification of different genes encoding endo- or exo-nucleases in Pr and Ps, which were included in Supplementary Material 1 .

These results indicate that Pr and Ps can release DNases to the extracellular medium that may degrade the eukaryotic DNA released in NETs.

In light of the results presented above, revealing the presence of DNases in the sn of Pr and Ps, we wondered whether the lack of NETs at 3 h could be due not to a lack of induction but to an underdetection caused by their degradation. In this sense, the presence of decondensed DNA co-localizing with elastase in PMN nuclei would indicate that NETs release had occurred, but was not detected because of degradation by bacterial DNases. Conversely, if PMN nuclei possess intact non-decondensed DNA, this would confirm the lack of induction of suicidal NETs formation. To clarify this issue, and considering that the release of suicidal NETs is preceded by nuclear decondensation and migration of elastase to the nucleus, we determine the number of PMN with decondensed nuclei and nuclear elastase by confocal microscopy 90 min after bacterial challenge. As depicted in Figures 7A, B , neither Pr nor Ps increased the percentage of PMN with nuclear decondensation co-localizing with elastase, whereas Eco and PMA did. This result reinforces our previous observation indicating the lack of suicidal NETs induction by Pr and Ps.

In view of the potential role that DNases could have in the context of immune evasion, and as the early induction of another type of NETs, called vital NETs, has been reported for other bacteria, we next investigated the possibility that Pr or Ps were inducers of this mechanism. For this purpose, PMN were challenged with bacteria and extracellular d.s. DNA was measured between 5 min and 3 h. As shown in Figure 7C , we found that both Pr and Ps induced a statistically significant release of d.s. DNA, detected early at 30 min, compatible with vital NETs. This NETs remained high at 1 h, and then decreased over time. However, Eco did not induce an early release of d.s. DNA, but then extracellular DNA increased gradually, being maximal at 3 h. In line with this result, the release of extracellular DNA representing vital NETs in Pr or Ps-treated PMN was observed at 30 min in confocal microscopy experiments, where DNA was stained with PI ( Figure 7D ). To determine the origin of the DNA released in vital NETs, we performed qPCR using primers that amplify nuclear (actin) or mitochondrial (ND1) genes. The results shown in Supplementary Figure 6 indicate that the DNA released in response to Pr and Ps is of nuclear origin. Altogether, the results presented indicate that Pr or Ps induce the early release of vital NETs that could be degraded by secreted bacterial DNases.

To study the relevance of the evasion mechanisms described and validate our in vitro results, we inoculated Pr and Ps intraperitoneally (i.p.) in mice and used Eco for comparison. After 24 h, mice were sacrificed, and the peritoneal liquid (pl), the lungs, and spleen were obtained. As observed in Figure 8 , the number of PMN that migrate to the pl was similar for Pr, Ps, and Eco ( Figure 8A ). However, only Eco increased MPO and the release of d.s. DNA in the pl, evidencing a local PMN activation and NETs formation ( Figures 8B, C ). In addition, to reveal the ability of Pr and Ps to escape the local PMN response, we studied the dissemination of bacteria by measuring the CFU present in the collected organs. As shown in Figure 8D , Pr was almost completely absent in the peritoneum and was found mostly in the spleen and to a minor extent in the lungs. Ps was distributed more evenly between the organs, showing an intermediate pattern, and Eco was mainly present in the peritoneum, being absent in the lungs and spleen. Altogether, these results indicate that Pr and Ps induce a poor PMN response at the primary infectious site, favoring their dissemination to other organs, whereas Eco triggers PMN activation and is contained at the inoculation site.