Zebrafish (Danio rerio) embryos were obtained by natural spawning of Tab5 and Tg(BACmpo:mCherry) lines (Renshaw et al., 2006). Fertilized eggs were raised in petri dishes containing E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl₂, 0.3 mM MgSO₄) and 0.1% methylene blue until 72 hour post-fertilization (HPF). All procedures complied with national guidelines and were approved by the Animal Use Ethics Committee of the University of Chile and the Bioethics Advisory Committee of Fondecyt-Conicyt (the funding agency for this work).

P. aeruginosa PAO1 was used for the injection and immersion assays. Avirulent Escherichia coli (strain DH5α) was used as a control. The bacteria were grown in LB medium overnight at 37°C and with shaking at 180 RPM. The cultures were washed and subsequently suspended in PGS medium (PGS (↓Pi); 2.5 g/L peptone, 3 g/L NaCl, 1 mM MgSO₄, 1 mM CaCl₂, 2% glycerol, pH 6.0); these suspensions were used to inoculate a liquid culture in PGS medium or PGS medium supplemented with inorganic phosphate (PGS (↑Pi); 25 mM potassium phosphate buffer pH 6.0) in a 1:100 ratio. These cultures were grown at 37°C with shaking at 180 RPM for 18 h.

For the immersion assays (Varas et al., 2017), cultures were washed and re-suspended in E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl₂, 0.33 mM MgS0₄, pH 7.0), then adjusted to an optical density at 600 nm (OD₆₀₀) of 1.4 equivalent to (1.1 ± 0.23)E9 CFU/mL for PGS (↓Pi) and (1.1 ± 0.22)E9 CFU/mL for PGS (↑Pi) for P. aeruginosa and equivalent to (1.2 ± 0.25)E9 CFU/mL for PGS (↓Pi) and (1.0 ± 0.15)E9 CFU/mL for PGS (↑Pi) for E. coli. To determine CFU/mL, the bacterial suspensions were plated on LB-agar. Zebrafishes of 3 DPF were washed with sterile E3, and 10 larvae were placed per well in a 6 well plate. The wells were filled with the bacterial solution to obtain an adjusted OD₆₀₀ of 0.7, 0.525 and 0.35 within a final volume of 8 ml, corresponding to (5.5 ± 0.22)E8 CFU/mL, (4.1 ± 0.20)E8 CFU/mL, (2.8 ± 0.23)E8 CFU/mL, respectively. The zebrafish larvae were incubated at 20°C for 30 h.

For the injection assays, the bacterial cultures were washed and subsequently suspended in phosphate-buffered saline (PBS) and adjusted to an OD₆₀₀ of 5.0. Zebrafish of 3 DPF were anesthetized with 0.01% tricaine and mounted in low melting point agarose. The zebrafish were kept under anesthesia for the duration of bacterial injections. Between 2,000 and 6,000 CFU were injected into the caudal artery. To determine the CFU injected, the droplets were also injected in sterile PBS and plated on LB-agar. The injected zebrafish were kept at 20°C for 72 h.

Survival curves were analyzed using the Kaplan–Meier method (GraphPad Prism version 6.00 for Mac OS X, GraphPad Software, La Jolla California USA, www.graphpad.com). All experiments were done in triplicate for a total of 30 larvae per condition.

Recruitment of neutrophils was observed using the Tg(BACmpo:mCherry) zebrafish line. The larvae were observed using an Olympus MVX10 (Japan) fluorescence microscope at 2, 4, 7, and 22 h post-injection (HPI). To determine the number of neutrophils in circulation, injected Tg(BACmpo:mCherry) zebrafish larvae were anesthetized with 0.01% tricaine for 2 min and the number of neutrophils that passed through the site of injection by the caudal artery in 1 min were counted. Statistical differences were assessed using a two-way ANOVA followed by Turkey's multiple comparisons test using GraphPad Prism version 6.00 for Mac OS X, GraphPad Software, La Jolla California USA, (www.graphpad.com).

For the proteomic analysis we selected and froze samples in a dry methanol bath at −80°C, including: 10 larvae injected with 2,000–6,000 CFU of P. aeruginosa [grown in medium PGS (↓Pi)] at 22 HPI; 10 larvae injected with sterile PBS at 22 HPI; 10 larvae at 22 h post-exposure (HPE) to ~2.5 × 10⁸ CFU/mL of P. aeruginosa [grown in medium PGS (↑Pi)] by static immersion; and 10 larvae incubated for 22 h in sterile E3 medium.

Global proteomic profiles were obtained by services from Bioproximity, LLC (USA). Protein denaturation, digestion and desalting of samples were prepared using the filter-assisted sample preparation (FASP) method (Wisniewski et al., 2009). Briefly, the samples were digested using trypsin, and each digestion mixture was analyzed by UHPLC-MS/MS and a quadrupole-Orbitrap mass spectrometer (Q-Exactive, Thermo Fisher). Mass spectrometer RAW data files were compared with the most recent protein sequence libraries available from UniProtKB. Proteins were required to have 1 or more unique peptides across the analyzed samples with E-value scores of 0.0001 or less.

The differences between the proteome of the larvae injected with P. aeruginosa and PBS, and between the larvae exposed to P. aeruginosa by static immersion were determined. Significance cut-offs for the ratios were set at 1.5-fold change. The proteins up-regulated and down-regulated exclusively in one of the conditions were determined. Gene ontology (GO) analysis were performed using Panther web-based software (Mi et al., 2013a,b). For enrichment analysis the cut of was set to P < 0.05.

To analyze the P. aeruginosa proteome, proteins in the infected larvae but not in the respective control were analyzed using Panther web-based software and a cut-off set to P < 0.05. For visualizing changes in the GO categories in both methods, Treemap (Version 3.7.2 for Mac) were used for the hierarchical representation of enriched biological processes and cellular components. Items were grouped by category and their size was proportional to the P-value.

P. aeruginosa becomes more virulent when grown with limited inorganic phosphate (Pi) in the medium (Long et al., 2008; Zaborin et al., 2012). Therefore, in order to obtain a wide range of outcomes to compare the injection and static immersion methods, the bacterial cells were grown in PGS medium lacking Pi or in the same medium supplemented with Pi, PGS (↓Pi) or PGS (↑Pi), respectively.

For the immersion assay, zebrafish larvae at 72 HPF were immersed in high concentrations of P. aeruginosa or E. coli (in the order of 10⁸ CFU/ml) for 30 h at 20°C. Mortality of zebrafish was observed when they were immersed with P. aeruginosa, but not with E. coli (Figures 1A,B). As expected, the larvae survival was dependent on the concentration of bacteria and the virulence of P. aeruginosa was increased when they were grown in Pi limitation. Before the death of the larvae, their tail was curved and there was damage and necrosis in the fins, tail and head (Figure 1C). Using the transgenic line Tg(BACmpo:mCherry) we did not find significant migration of neutrophils or changes in the distribution of these cells when the zebrafish was immersed with P. aeruginosa (Figure 1D).

The region of the caudal artery was chosen as the site of injection due to the reduced pigmentation in comparison to the caudal vein. Between 2,000 and 6,000 CFU of P. aeruginosa or E. coli cells were injected in zebrafish larvae at 72 HPF, and sterile PBS was injected as control. Under the tested conditions, death of the larvae was observed only when P. aeruginosa grown in PGS (↓Pi) was injected (Figure 2A). Several larvae injected with P. aeruginosa presented damage and necrosis in the tail before their death (Figure 2B). In the first HPI, a circulatory blockage at the site of infection was observed when injected with P. aeruginosa, E. coli or sterile PBS. However, this blockage remained only for a few hours in the larvae injected with PBS sterile, but it was observable until 28 HPI in the larvae injected with P. aeruginosa.

In order to compare both methods of exposure and considering the absence of zebrafish mortality before 40 HPI, we address if in previous HPI there was some response in the larvae, specifically an inflammatory response. To do so, a transgenic line that has the neutrophils marked with the fluorescent protein mCherry was used and injected with P. aeruginosa, E. coli or sterile PBS. In the larvae injected with P. aeruginosa there was a clear recruitment of neutrophils to the site on injection that remained until 28 HPI. This also occurred after an injection with sterile PBS, however under these conditions it was barely observable at 6 HPI and it was totally absent at 28 HPI (Figure 3B). These results suggest that an injection with P. aeruginosa generates an important inflammatory response at the site of injection. To quantify this, the number of neutrophils that passed trough the site of injection in 1 min was determined (Figure 3A). At 22 HPI and 28 HPI there was a significant difference between the number of neutrophils circulating in the larvae injected with P. aeruginosa grown in PGS (↓Pi) or injected with E. coli with respect to the control. In contrast, there was no difference between the larvae injected with P. aeruginosa grown in PGS (↑Pi) with those injected with E. coli.

First, the zebrafish proteomes were analyzed. The proteome profile of the larvae injected with P. aeruginosa was compared with larvae injected with sterile PBS (Figure 4A). A total of 12,677 proteins were up-regulated (fold-change>1.5) and 13,292 proteins were significantly down-regulated (fold change<-1.5). From these, 8,575 proteins were up-regulated and 9,363 proteins were down-regulated exclusively in this condition, and therefore not in the larvae exposed by immersion. On the other hand, when the proteome of the larvae immersed with P. aeruginosa was compared to the proteome of the larvae immersed in sterile E3 medium (control), a total of 13,607 proteins were up-regulated and 13,136 proteins were down-regulated. From these 9,505 proteins were up-regulated and 9,208 proteins were down-regulated exclusively in the larvae exposed by immersion.

The proteins up-regulated and down-regulated in each condition were categorized according to GO-terms of biological process (Supplementary Figures 1–3). No appreciable differences between the proteins up-regulated and down-regulated were found in each condition and neither between conditions in regard of biological process. The majority of the proteins were categorized under “metabolic process (GO:0008152)” and cellular process (GO:0009987)” with ~25 and 20%, respectively in all the analysis. The GO-term “immune system process (GO:0002376)” represented only around 5% of the proteins and the GO-term “response to stimulus” only around 6% of the proteins.

Afterwards, an overrepresentation test was carried out and the proteins were analyzed by Pathway (Table 1, Supplementary Table 1). In the up-regulated proteins of the larvae exposed by static immersion the “hypoxia response via HIF activation pathway (P00030)” was enriched. In contrast, in the down-regulated proteins of the larvae exposed by injection the categories “B cell activation (P00010)” and “endothelin signaling pathway (P00019)” were enriched. There were no significantly enriched groups in the up-regulated proteins by injection.

To address the differences in the proteins that changed their levels in one method of exposure but not in the other, the proteins up-regulated only in larvae exposed by injection or in larvae exposed by immersion were analyzed by Pathway. In the proteins up-regulated exclusively in injected larvae the pathways “axon guidance mediated by netrin (P00009),” “inflammation mediated by chemokine and cytokine signaling pathway (P00031)” and “Cytoskeletal regulation by Rho GTPase (P00016)” were enriched. On the other hand, in the proteins up-regulated exclusively in the larvae exposed by immersion, “integrin signaling pathway” and “angiogenesis (P00005)” were enriched. No pathways were significantly enriched when proteins that were down-regulated exclusively in one of the conditions or the proteins that were up-regulated in both conditions were analyzed.

Several genes or gene products that have been reported as genetic markers for exposure to bacteria by injection were found in the zebrafish proteome (Figure 4B). Particularly, the levels of the transcripts of the mmp9 and il8l1genes were reported as up-regulated in zebrafish after an exposure to S. typhimurium by injection (Stockhammer et al., 2009). However, neither of the proteins encoded by these genes were significantly up-regulated in larvae injected with P. aeruginosa at 22 HPI. IL-8L1 (N0GW19) was only detected in the larvae immersed with P. aeruginosa and in the larvae injected with PBS, but not in the control immersed in E3 medium or injected with P. aeruginosa. MMP9 (P14780) was slightly down-regulated (-2.04-fold change) in the larvae exposed by static immersion. The proteins IL-1ß (E0WCW4), TNF-ß (Q1JQ40), interferon-induced GTP-binding protein MxC (Q6DKF0), STAT1A, and STAT1B were reported as molecular markers in zebrafish injected either with S. typhimurium (Stockhammer et al., 2009) or E. tarda (van Soest et al., 2011). We found all these proteins up-regulated in the larvae injected with P. aeruginosa. Interestingly, they were all down-regulated in the larvae immersed in P. aeruginosa, except for STAT1B (-1.24-fold change). In addition, IL-1ß and TNF-α were reported as molecular markers for the response of the zebrafish injected at 50 HPF with P. aeruginosa PA14 into the yolk circulation valley (Clatworthy et al., 2009). Both proteins were up-regulated in our injection assay, but not in the larvae exposed by static immersion. These results suggest that the expression of known components in the response toward a bacterial injection is dependent on the types of pathogens infected. To our knowledge, the expression of cyp1a is the only proposed marker for an exposure by immersion to a E. tarda (van Soest et al., 2011). However, in this study CYP1A was only detected in the larvae incubated in sterile E3 medium.

To suggest potential novel molecular markers that are exclusive either for the injection or static immersion methods for exposure we look for proteins that fulfill the following conditions: (1) the protein should participate in pathways that are solely enriched in the larvae infected by one of the exposure methods, (2) the protein or its human homolog should have a known biological function related with an infective process and (3) the levels of protein expression be distinct in each method of exposure.

Considering these conditions, we suggest that proteins MYLKA, CHUK, CXCL11.6, CISH, and IFIT5 can be used as noel molecular markers for the injection method (Figure 4B). MYLKA had a 16.7-fold change were it was compared to the control. CHUK was up-regulated in injected larvae (1.93-fold change). CXCL11.6 had a 5.30-fold change in injected larvae. IFIT5 had a 121-fold change. The proteins COL8A2, COL141A, COL5A2B, VWA1, FYNA, FYNB, RAF1A, and MAP2KA are potential novel markers for an infection by static immersion (Figure 4B). These proteins were significantly up-regulated in the larvae exposed by static immersion and down-regulated in the injected larvae, except for MAP2KA that was not down-regulated in the injected larvae (1.14-fold change).

We next analyzed P. aeruginosa proteomes to address the differences in the proteins expressed in the P. aeruginosa injected and the P. aeruginosa in the immersion assay. A total of 1,159 P. aeruginosa PAO1 proteins were exclusively found in the injected larvae, and 1,276 in the larvae exposed by immersion (Supplementary Table 2). All of these proteins were analyzed by the GO-terms of component and biological process (Figure 5). Several GO-term categories associated with the bacterial-type flagellum were enriched in the P. aeruginosa injected, but not in the immersion assay. In regard of the GO-term by biological process, the categories “protein secretion by the type III secretion system” (GO:0030254), “bacterial-type flagellum organization” (GO:0044781), “putrescine metabolic process” (GO:0009445) and “response to drug” (GO:0042493) were enriched in the injected P. aeruginosa, but not in the immersion assay. On the other hand, the categories “single-species biofilm formation” (GO:0044010), “pathogenesis” (GO:0009405), “cellular response to antibiotic” (GO:0009267), and “cellular response to starvation” (GO:0009267) were enriched in the P. aeruginosa co-incubated with the zebrafish, but not in the injected bacteria. The categories “biological adhesion” (GO:0022610), “cellular response to stress” (GO:0033554), “pilus assembly” (GO:0009297), “biofilm formation” (GO:0042710), “cell communication” (GO:0007154), “type IV pilus-dependent motility” (GO:0043107), “cilium or flagellum-dependent cell motility” (GO:0001539), “protein secretion by the type II secretion system” (GO:0015628), “secondary metabolite biosynthetic process” (GO:0044550), “siderophore metabolic process” (GO:0009237), “pyoverdine metabolic process” (GO:0002048), “response to antibiotic” (GO:0046677), and “peptidoglycan metabolic process” (GO:0000270) were enriched in both groups of proteins.

To explain how these processes were regulated after an infection by injection or static immersion, the bacterial transcription regulators detected in the infected larvae but not in the respective controls were analyzed (Figure 5B). A total of 15 bacterial transcription regulator families were detected in the larvae injected with P. aeruginosa and in the larvae exposed by static immersion (LysR, LuxR, AraC, GntR, TetR, AsnC, XRE, Fis, DeoR, IclR, Cro/CI, HTH-type, ArsR, PvdS, and AlpA), which suggest that several processes are regulated in a similar way in both methods of exposure. A total of 7 bacterial transcription factor families were detected only in the in the larvae injected with P. aeruginosa (RpiR, MvfR, Sigma-54 dependent family, PfeR, BvgA, and RhlR). Twelve bacterial transcription regulator families were detected only in the larvae in the larvae exposed by static immersion (PhoB, FleQ, AlgB, PtxS, MerR, ExsA, CopG, MarR, LacI, SoxR, ArsC, and DctD).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.