Dual perspective proteomics infectome profiling discovers Salmonella type III secretion system effector functions in macrophages

Intracellular bacterial pathogens have evolved sophisticated infection strategies, including the release and secretion of virulence factors to interfere with host cell functions and to perturb immune responses. For Salmonella enterica serovar Typhimurium (S. Typhimurium), the type III secretion systems encoded on Salmonella pathogenicity islands (SPI) 1 and 2 mediates invasion of the bacterium into innate immune cells and regulates bacterial replication and survival within the hostile environment of the host, respectively. Here, we explore the temporal and strain-specific dual perspective response of both the host and pathogen during cellular infection via quantitative proteomics. We report time- and pathogenicity island-specific expression and secretion of infection-associated proteins (i.e., SL1344_1263, SL1344_3112, SL1344_1563, and YnhG) and regulated immune response proteins in macrophage, including Cd86, Cd40, Casp4, C3, IL-1α, and Cd69). Through intracellular macrophage and in vivo murine models of infection, we reveal a role in virulence for three of the bacterial proteins (SL1344_1263, SL1344_1563, and YnhG), defining their importance as novel T3SS effectors. We characterize the temporal intra- and extracellular production of the effectors and identify their interaction networks in host cells, representing inhibitory and stimulatory pathways mounted by invading bacterial pathogens. Author Summary The relationship between a host and pathogen is intricate, and regulation of the host immune response correlates with the progressive timing of infection and tailored responses to the pathogen. Relying on detection and quantification of protein-level changes using mass spectrometry-based proteomics, we explore the production of known and novel effectors by Salmonella during intracellular survival within macrophage. Our results portray a role for these effectors in bacterial virulence using an in vivo murine model of infection, and we define a dynamic network of interaction between the effectors and host proteins. These interactions reveal opportunity for drug design to disrupt modulation of the host by the invading bacterium as a new strategy for combatting infection. Our approach is dynamic and universal, with the power to alter therapeutic discovery against infectious diseases.


Introduction
The enteric, pathogenic bacterium Salmonella causes a wide variety of diseases, ranging from gastroenteritis to systemic infections [1]. For this Gram-negative pathogen, two distinct type III secretion systems (T3SS) represent specialized virulence factors, which play an important role in delivering effector proteins to host cells. For Salmonella enterica serovar Typhimurium (S. Typhimurium), invasion of the host cell depends on bacterial proteins encoded in the chromosomal locus, Salmonella pathogenicity island 1 (SPI-1). The genes of SPI-1 encode for components of a T3SS, regulatory proteins, along with secreted effector proteins and their chaperones, which collectively establish the initial phase of infection [2]. In addition to inducing host cell death, SPI-1 mediates host cell invasion, autophagy, and inflammation [3][4][5]. Conversely, the Salmonella pathogenicity island 2 (SPI-2)-encoded T3SS permits replication of bacterial cells once inside macrophages by formation of a protective compartment, the Salmonella containing vacuole (SCV) [6][7][8]. Once inside the SCV, Salmonella is able to replicate and disseminate to systemic sites, including the spleen and liver, causing severe diseases (e.g., bacteremia) [9].
Currently, the increasing prevalence of global infection and multi-drug resistant Salmonella strains demands the discovery of new treatment options [10]. Understanding the mechanisms associated with Salmonella infection suggests innovative strategies for identifying novel therapeutics.
Previous work provides critical insight into the mechanisms of, and effectors associated with, the T3SS in Salmonella [6,9,[11][12][13][14]. In addition, recent studies of the interaction between Salmonella and the host have defined effector-specific cellular reprogramming and options for drug discovery [15,16]. The development of network resources (e.g., SalmoNet) for biochemical modeling, host-pathogen interaction studies, drug-discovery, experimental validation of novel interactions, and information to reveal new pathological mechanisms improves our understanding 5 of Salmonella pathogenesis [17]. To uncover such information, mass spectrometry-based proteomics provides a robust, sensitive, and unbiased platform for profiling protein-level differences in complex biological samples [18]. Information about abundance, interactions, regulation, and post-translational modifications are defined, providing novel biological insight into bacterial pathogens and their relationship with the host [19,20]. The majority of experiments focus on a single perspective (i.e., identification of pathogen or host proteins) given limitations of detection of pathogenic proteins in the presence of highly abundant host cells. For example, cellular proteome profiling of Salmonella from infected host epithelial cells focuses on the bacterial adaptive response to infection [21,22]. Other studies use quantitative proteomics combined with high-throughput techniques for the characterization of host gene responses and host phosphoproteome dynamics following Salmonella infection by focusing on the impact of specific effectors [11,23]. In addition, secretome profiling of Salmonella identifies a novel virulence factor and defined host epithelial cell response to infection, and proteomic profiling of Salmonellainduced filaments reveals critical information about bacterial survival mechanisms [24][25][26][27].
Moreover, interactions between the host and pathogen are explored using affinity purification and modern proximity labeling techniques (e.g., BioID), which provide valuable information into the complex network of interactions between a single effector and the host during infection [28,29].
Building upon this previous knowledge and providing a universal platform to investigate the proteome of host-pathogen interactions from dual perspectives, we unbiasedly define the interaction between Salmonella and primary macrophages for both pathogenicity islands in a single experiment. In addition, our in-depth infectome profiling defines known and novel Salmonella effectors regulated by the pathogenicity islands and we uncover new temporal and strain-specific responses of the host to infection. Moreover, we characterize the novel effectors 6 with in vitro and in vivo infection assays to reveal roles in virulence and visualize their production in host cells. Interactome profiling of the bacterial effectors reports unique points of biochemical contact within host cells, suggesting new opportunities to perturb such interactions as novel antivirulence strategies for therapeutic intervention. Overall, this study is the first to report extensive protein-level dynamics during Salmonella infection simultaneously from global, host, and pathogen perspectives in consideration of both the SPI-1 and SPI-2 modes of action. Finally, our approach is universal for the discovery of novel infection-associated proteins and has applications with medical and agricultural relevance to combat microbial infections on a global scale.

Bacterial strains and growth conditions.
S. Typhimurium SL1344 wild-type (WT) (streptomycin resistant), SL1344 spi1 (kanamycin resistant) by deletion of the pathogenicity island, and SL1344 spi2-defective (kanamycin resistant) by transposon insertion in ssrA by signature-tagged mutant (STM) construction, and LT2 were used for this study (strains generously gifted from A. Zychlinsky's lab) [30,31]. Salmonella strains were maintained on Luria broth (LB) plates with the respective antibiotics, followed by inoculation of LB broth (with antibiotics) with shaking (200 rpm) overnight (18 h) at 37°C. For Salmonella in vitro cultures, WT, spi1, and spi2-defective were grown in LB (without antibiotic) to early stationary phase and collected for proteome analysis. The in vitro growth assay was performed in biological quadruplicate and the experiment was performed in triplicate.
For infection, BMDMs were seeded in 6-well plates at 1×10 6 cells/well in BMDM-DMEM media without pen/strep overnight at 37°C, 5% CO 2 . Bacterial cells were grown in LB broth to early stationary phase (same cultures used for in vitro growth described above), spun down, and resuspended in phosphate-buffered saline (PBS) and diluted in BMDM-DMEM media. Cells were infected at a multiplicity of infection (MOI) of 50:1 (bacteria:macrophage) for 30 min at 37°C, 5% CO 2 . Subsequently, cells were washed with PBS and incubated at 37°C, 5% CO 2 in growth medium containing 100 µg/mL gentamicin for 1 h. Medium was replaced to decrease the gentamicin concentration to 10 µg/mL for later time points. Cultures were collected at 0, 0.5, 1, 2, and 6 hours post infection (hpi) (i.e., hours post infection calculated following: 30 min co-culture, 1 h gentamicin treatment, add fresh DMEM) for proteome analysis. The BMDM infection assay was performed using biological quadruplicates and the experiment was performed in duplicate.

Proteomic sample preparation
Total proteome, secretome, and infectome profiling was performed as previously described with minor modifications [32,33]. Cell pellets (bacterial and macrophage) were collected and washed with PBS, resuspended in 100 mM Tris-HCl, pH 8.5 with proteinase inhibitor capsule (PIC) and mechanically disrupted by probe sonication in an ice bath. Next, 2% (final) sodium dodecyl sulphate (SDS) was added, followed by reduction with 10 mM dithiothreitol (DTT) at 8 95°C for 10 min at 800 rpm, alkylation with 55 mM iodoacetamide (IAA) at room temperature in the dark, and acetone precipitation (80% final) at -20°C overnight. Precipitated proteins were pelleted, washed twice with 80% acetone and air-dried, prior to solubilization in 8 M Urea and 40 mM HEPES. LysC was added at 1:100 µg protein for 3 h at room temperature, followed by addition of 300 µl 50 mM ammonium bicarbonate (ABC) and 1:100 µg protein of trypsin overnight at room temperature. Protein concentration was determined by BSA tryptophan assay [34]. For secretome samples, the supernatant from bacterial cell pellets above was filtered (0.22 µm syringe filter tip), reduced with 10 mM DTT for 30 min at room temperature, alkylated for 20 min with 55 mM IAA in the dark, and then subjected to LysC and trypsin digestion overnight as above.
Digestion was stopped by addition of 10% v/v trifluoroacetic acid (TFA) per sample and the acidified peptides were loaded onto Stop-And-Go extraction tips (StageTips) (containing three layers of C 18 ) to desalt and purify according to the standard protocol [35]. Approx. 50 µg of sample was loaded onto StageTips and stored at 4°C until LC-MS/MS measurement.

Mass spectrometry
Samples were eluted from StageTips with Buffer B (80% acetonitrile (ACN) and 0.5% acetic acid), dried with a SpeedVac concentrator, and peptides were resuspended in Buffer A (2% ACN and 0.1% TFA). Approx. 3-5 µg of protein was subjected to nanoflow liquid chromatography on an EASY-nLC system (Thermo Fisher Scientific, Bremen, Germany) on-line coupled to an Q Exactive HF quadrupole orbitrap mass spectrometer (Thermo Fisher Scientific). A 50 cm column with 75 μm inner diameter was used for the chromatography, in-house packed with 3 μm reversed-phase silica beads (ReproSil-Pur C 18 -AQ, Dr. Maisch GmbH, Germany). Peptides were separated and directly electrosprayed into the mass spectrometer using a linear gradient from 5% to 60% ACN in 0.5% acetic acid over 120 min (bacterial cellular proteome and 9 secretome) or 180 min (infectome) at a constant flow of 300 nl/min. The linear gradient was followed by a washout with up to 95% ACN to clean the column. The QExactive HF was operated in a data-dependent mode, switching automatically between one full-scan and subsequent MS/MS scans of the fifteen most abundant peaks (Top15 method), with full-scans (m/z 300-1650) acquired in the Orbitrap analyzer with a resolution of 60,000 at 100 m/z.

Mass spectrometry data analysis and statistics
Raw mass spectrometry files were analyzed using MaxQuant software (version 1.6.10.43) [36]. The derived peak list was searched with the built-in Andromeda search engine [37] against the reference S. Typhimurium SL1344 proteome (downloaded from Uniprot, June 30, 2020; 4,658 sequences) and Mus musculus proteome (downloaded from Uniprot, June 30, 2020; 55,462 sequences). The parameters were as follows: strict trypsin specificity was required with cleavage at the C-terminal after K or R, allowing for up to two missed cleavages. The minimum required peptide length was set to seven amino acids. Carbamidomethylation of cysteine was set as a fixed modification (57.021464 Da) and N-acetylation of proteins N termini (42.010565 Da) and oxidation of methionine (15.994915 Da) were set as variable modifications. Peptide spectral matching (PSM) and protein identifications were filtered using a target-decoy approach at a false discovery rate (FDR) of 1%. 'Match between runs' was enabled with a match time window of 0.7 min and an alignment time window of 20 min. 'Split by taxonomy ID' was enabled when two proteome FASTA files were loaded in the same experiment (e.g., S. Typhimurium and M. musculus). Relative label-free quantification (LFQ) of proteins was performed using the MaxLFQ algorithm integrated into MaxQuant using a minimum ratio count of 1, enabled FastLFQ option, LFQ minimum number of neighbors at 3, and the LFQ average number of neighbors at 6 [38].
Statistical analysis of the MaxQuant-processed data was performed using the Perseus software environment (version 1.6.2.2) [39]. Hits to the reverse database, contaminants, and proteins only identified by site were removed. LFQ intensities were converted to a log scale (log 2 ), and only those proteins present in triplicate within at least one sample set were used for further statistical analysis (valid-value filter of 3 in at least one group). Missing values were imputed from a normal distribution (downshift of 1.8 standard deviations and a width of 0.3 standard deviations). A Student's t-test was performed to identify proteins with significant differential abundance (p-value ≤ 0.05) (S0 = 1) between samples employing a 5% permutationbased FDR filter [40]. A Principal Component Analysis (PCA) was performed, as well as Pearson correlation with hierarchical clustering by Euclidean distance to determine replicate reproducibility within Perseus. STRING database was reference for interaction mapping (https://STRING-db.org) [41].

Mutant construction
Mutants were first constructed in the S. Typhimurium LT2 background using the lambda Red recombinase method [42]. LT2 was used to improve transformation efficiency and ensure correct subsequent gene knockout in the SL1344 strains. Briefly, oligos were designed for homologous recombination of each gene (i.e., SL1344_3112, SL1344_1263, SL1344_1563, and ynhG) and used to amplify the kanamycin resistant cassette from pKD4 as previously described [15]. The WT LT2 was transformed with pKD46 (i.e., arabinose inducible lambda Red recombinase) at 30°C, followed by transformation of WT pKD46 with PCR product and plating on LB+Kan 25 plates. Colonies were cured (37°C, overnight) of the pKD46 plasmid and confirmed for resistance cassette insertion into the gene of interest by colony PCR. Next, to transform into S.
Typhimurium SL1344, the kanamycin resistance cassette was amplified from the LT2 background 11 and transformed into SL1344 using the methods described above. At least two independent mutants were generated for each gene and mutants were confirmed by colony PCR and sequencing of the insertion sites. All primer sequences used in this study are provided (Supp. Table 1).

Cytotoxicity assays
Murine primary macrophages were seeded onto a 12-well plate at a density of 0.1 × 10 6 cells per well and incubated overnight at 37°C. Before infection with the bacteria, the medium was replaced with fresh BMDM-DMEM media (without antibiotics). Infection of cells with Salmonella was done as described [43], by adding bacteria 50:1 MOI onto the cells followed by incubation at 37°C. After 30 min of incubation, gentamicin (100 μg/ml) was added to kill extracellular bacteria, followed by incubation in fresh BMDM-DMEM media with gentamicin (10 μg/ml) . At the indicated time points (1, 3, 6, and 18 hpi), culture supernatants were collected for analysis as the experimental release samples. Cytotoxicity was quantified colorimetrically with the CytoTox96 lactate dehydrogenase (LDH)-release kit (Promega). The percentage of cytotoxicity was calculated with the formula: 100 × (absorption of supernatant/absorption whole cell lysate).
The cytotoxicity assay was performed in biological triplicate and the experiment was performed in duplicate.

Macrophage association, invasion, and replication
For phenotyping S. Typhimurium association, invasion, and replication (AIR) in primary macrophages, we followed the protocol previously described [44]. Briefly, primary macrophages were seeded at 0.1 x 10 6 in BMDM-DMEN media overnight at 37°C, 5% CO 2 and co-cultured with S. Typhimurium strains WT, spi-1, spi-2, SL13441563, SL13441263, SL13443112, and ynhG as described above. 'Association' plates were collected after 30 min co-culture, macrophage cells were lysed with 1% Triton X-100 for 10 min and released associated-Salmonella were serially diluted in PBS and plated on LB (WT) or LB+Kan 25 (mutants). After 1 h of incubation in gentamicin, the 'invasion' plate was collected and treated as above, and after an additional 22.5 h, the 'replication' plate was collected, and serial dilutions were plated. Agar plates were incubated overnight at 37°C and CFUs were counted. Bacterial counts were normalized to macrophage cell counts. Comparison of association, invasion, and replication of normalized CFUs by Student's t-test compared each strain to the WT and reported statistical significance (p-value <0.05). The experiment was performed in biological triplicate and repeated twice.

Murine competitive index assays
All animal experiments were performed in accordance with the guidelines of the Canadian Council on Animal Care and the University of British Columbia (UBC) Animal Care Committee (certificate A17-0228). Mice were ordered from Jackson Laboratory (Bar Harbor, ME) and maintained in a specific pathogen-free facility at UBC. Ten six-week-old female C57BL/6 mice were orally gavaged with 100 µl PBS containing a total of 3×10 6 or 3×10 7 CFU of S. Typhimurium

FLAG-tag proteins
Plasmids encoding C-terminally 3×FLAG-tagged versions of YnhG, SL1344_1263, and SL1344_1563 were generated by PCR amplification of the target gene from WT SL1344 genomic DNA. Each gene was expressed under the control of its native promoter; sequence encoding the 3×FLAG antigen was engineered into the reverse primer. PCR products were restriction digested and ligated into pACYC184 using standard techniques. The sequence of all inserts was verified by Sanger sequencing (UBC Sequencing + Bioinformatics Consortium).

Western blots
Whole cell extracts and culture supernatants from S. Typhimurium WT and FLAG-tagged strains YnhG, SL1344_1263, and SL1344_1563 were fractionated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane using a transfer apparatus according to the manufacturer's protocols (Bio-Rad). After incubation with 3% non-fat milk in 1X TBS (50 mM Tris, 150 mM NaCl, pH 7.5) overnight at 4°C, the membrane was washed five times with TBST (1X TBS, 0.05% Tween-20), and incubated with ANTI-FLAG M2 antibody (Sigma-Aldrich) for 1 h. Membranes were washed three times for 5 min and incubated with 1:3000 dilution of horseradish peroxidase-conjugated Goat anti-mouse IgG Fc secondary antibody (Invitrogen) for 1 h. Blots were washed with TBST three times and developed using an alkaline phosphate (AP) substrate buffer (1 M Tris, pH 9.5, 5 M NaCl, 1 M MgCl 2 , 50 mg/ml 5-bromo-4-chloro-3-indolyl phosphate, and 5% nitrotetrazolium blue chloride). Experiment was performed in biological and technical triplicates.

Immunofluorescence and imaging
For bacteria-only immunostaining, cultures were set overnight in LB broth prior to subculturing. At select times, bacterial cultures were collected and fixed in 80% methanol. Protocol for immunostaining was adapted from a previously described method [45]. Briefly, fixed cells were plated onto a cell slide coated with 0.1% poly-L-lysine (Sigma-Aldrich). Plated cells were incubated with buffer (25 mM Tris-HCl pH 7.6, 10 mM EDTA and 2 mg/mL lysozyme). Slides were submerged in 99% methanol, followed by 100% acetone prior to blocking with 10% goat serum in PBS. Antibodies were diluted in 1% BSA in PBS and incubated for 1 h. Cells were incubated with monoclonal ANTI-FLAG M2 antibody (Sigma-Aldrich) diluted to 1:100 followed For infection immunostaining, C57BL/6 macrophages were maintained as described above. To minimize autofluorescence, cells were grown in FluoroBrite DMEM media (ThermoFisher Scientific) supplemented with 10% FBS 1 d prior to infection and during the infection protocol as described above. Samples were collected at 1 and 6 hpi and fixed overnight in 4% paraformaldehyde in PBS. Protocol for immunostaining was adapted from previously described methods and as outlined above [45,46]. Cells were incubated with monoclonal Anti-FLAG antibody (Sigma-Aldrich) diluted to 1:100 followed by incubation with Alexa Fluor 488 diluted to 1:200. Cells were counterstained with DAPI (10 µg/ml) for 1 min. Between incubation steps, slides were washed three times with PBS. A drop of SlowFade TM Gold antifade (Life Technologies) was added prior to mounting coverslips.
Slides were imaged using a Leica DM5500B microscope, equipped with a Hamamatsu 3CCD digital camera operated through Volocity software ver. 6.3 (Quorum Technologies). For imaging macrophage samples, a fixed exposure of 500 ms was used to detect Alexa Fluor 488 bound to bacterial cells to account for the autofluorescence.

Immunoprecipitation
FLAG-tagged S. Typhimurium strains and WT (as a negative control) were used to infect C57BL/6 macrophages as outlined above. At 1 and 6 hpi, cells were washed twice with wash buffer 1 (150 mM NaCl, 50 mM Tris, pH 7.5, 0.25% NP-40) and lysed with lysis buffer (150 mM NaCl, 50 mM Tris, pH 7.5, 5% glycerol, 1% NP-40, 1 mM MgCl 2 , and a PIC (Roche)). Lysates were diluted 1:1 with wash buffer 1 and incubated with Anti-DYKDDDDK Magnetic Agarose beads (Pierce) for 80 min with end-over-end rotation at 4°C. Beads were washed twice with wash buffer 1 and four times with wash buffer 2 (150 mM NaCl, 50 mM Tris, pH 7.5) followed by on-bead digestion as previously described [47]. Digested samples were purified by STAGE-tip and subjected to the LC-MS/MS protocol outlined above.

Results
We devised a mass spectrometry-based proteomics approach to profile the abundance of bacterial proteins during laboratory growth (i.e., cell culture medium) combined with the infectome from dual perspectives (i.e., host and pathogen) in a single mass spectrometry measurement over the time course of infection (Fig. 1A). Our approach enables detection and tracking of both SPI-1-and SPI-2-associated bacterial proteins during infection of primary macrophages from invasion to replication, as well as the identification and characterization of novel infection-associated proteins.

Defining pathogenicity island-associated proteins in vitro
First, we analysed proteins produced in the cellular proteome and secretome across the three bacterial strains (Fig. 1B). From the cell pellet, we observed 2,499 bacterial proteins after filtering for valid values (representing 54% of the gene-encoding regions) and 11 uniquely secreted proteins, including proteins associated with the T3SS (prgI, orgC, invJ, prgJ, spaO), bacteriophage (SL1344_1940, SL1344_1941, SL1344_1953, SL1344_1958), and flagella (fliK). Among the Salmonella strains, we identified a core cellular proteome consisting of 2,813 proteins with 23, 49, and 66 proteins unique to WT, spi1, and spi2-defective strains, respectively (Fig. 1C). Second, for the secretome, we observed consistent identification of 232 proteins with 14, 5, and 47 proteins unique to the WT, spi1, and spi2-defective strains, respectively (Fig. 1C). Importantly, many of the unique proteins identified have roles as effectors or encoded genes for the respective pathogenicity islands, as evident in profiling differences in abundance of SPI-1-and SPI-2associated proteins relative to WT in the cellular proteome (Fig. 1D) and secretome (Fig. 1E). As anticipated, these data show increased production of SPI-1-and SPI-2-associated proteins in the WT strain relative to the mutant strains, along with potential cross-regulation between the pathogenicity islands in the absence of spi1, possibly attributed to transcriptional regulation by hilD, as previously reported [48]. Notably, fewer SPI-2-associated effectors were detected in the secretome likely due to the selected growth condition (i.e., rich medium).

Global profiling demonstrates depth and dynamic range of the infectome
To assess the depth of proteome coverage for both the host and pathogen in a single-run mass spectrometry experiment, we determined 78% were host proteins (6,852 host/8,833 total proteins) and 22% were bacterial proteins (1,981 bacterial/8,833 total proteins) (Fig. 1F). Notably, this represents detection of almost 80% of the total measurable bacterial proteome in the presence of the host. We observed a dynamic range spanning >5 orders of magnitude with distribution of both host and bacterial proteins (Fig. 1G). Finally, a principal component analysis (PCA) of the infectome revealed a clear distinction between infected vs. non-infected macrophages (component 1, 53.4%), as well as temporal and strain-specific responses (component 2, 4.2%) (Fig. 1H).

Host response to S. Typhimurium infection reveals temporal and strain-specific regulation
During infection, dynamics proteome changes are measured from both the host and pathogen perspectives to define how the host cells protect themself from invasion and how the pathogen evades or adapts to the host response. Here, we focus on infection from the host's perspective by profiling changes in protein abundance over time and in the presence of the different S. Typhimurium strains. By PCA, we observed three distinct clusters of samples based on time and strain ( Fig. 2A) Next, clustering by protein IDs via hierarchical clustering highlighted protein groups associated with the temporal host response (Fig. 2B). A closer look at host proteins with increased abundance over the time course of infection revealed a subset of proteins with significant increases in abundance at 6 hpi regardless of the bacterial strain applied to the macrophage (Fig. 2C). These proteins included induction of host immunity (e.g., IL-1⍺, Nos2, Cd40) and defense response (e.g., Sp140) (Supp. Table 2). Next, we noticed a trend for proteins to significantly increase in abundance over the time course of infection, including innate immune effector (e.g., Il1B) ( Fig. 2D; Supp. Table 3) and immune response-associated proteins (i.e., Sp140, Pfkfb3). Hierarchical clustering of immune-associated proteins demonstrated clustering by Euclidean distance (Fig. 2E).
To explore significant changes in abundance of proteins unique to each of the collected time points, we generated stacked volcano plots to highlight time-specific host responses (Fig. 2F).
This analysis identified significantly different proteins unique to each time point, including 1 hpi (five proteins), 2 hpi (four proteins), and 6 hpi (153 proteins) (Supp. Table 4). Notably, at 0.5 hpi, no significant changes in abundance were observed. As anticipated, at 6 hpi, we observed the largest change in unique host protein response with involvement of many immune associated proteins (e.g., Cd86, Cd40, Casp4, C3, IL-1⍺), as well as proteins involved in signaling cascades (e.g., Cks1b) and regulatory networks (e.g., Rin2), demonstrating a specific and complex response to bacterial infection over time. Taken together, these data demonstrate time-specific responses induced within the host during infection, regardless of the bacterial strain, and represents key points for further exploration as potential inhibitory or stimulatory host targets in defense against invading bacterial pathogens.
Next, to assess unique host responses dependent on the S. Typhimurium strain, we generated stacked volcano plots to highlight strain-specific significantly different proteins (Fig.   2G). For each strain, we observed unique host cell expression pattern (Supp. Table 5). For example, in the presence of WT, macrophages uniquely produced Cd69, an early activation antigen, suggesting activation of this host protein is related to T3SS function. For spi1 infection, we observed four uniquely produced proteins with diverse roles in cellular processes, including a signaling inhibitor (Cdkn1a). Notably, given the distinction of WT to invade macrophages vs. entry only by phagocytosis for spi1 due to a lack of invasion properties, these protein differences may be linked to unique intracellular locations of the bacteria within the macrophage. For spi2-defective infection, we observed the unique production of 49 proteins (28 with increased abundance; 21 with decreased abundance), indicating specific host defense programs to S.
Typhimurium in the absence of bacterial cell replication, including specific reduction in proinflammatory signals (e.g., S100a9, S100a8) and induction of inflammasome components (e.g., Casp4). Overall, these bacterial strain-specific responses demonstrate remodeling of the host proteome in response to the presence or absence of pathogenicity islands and provide new insight into the complementary roles SPI-1 and SPI-2 play in Salmonella infection.

Co-regulation of novel infection-associated proteins and Salmonella pathogenicity islands
Here, we aimed to identify novel infection-associated proteins influenced by the presence or absence of the pathogenicity islands during a time course of infection of primary macrophages.
To identify novel infection-associated proteins, we compared the in vitro cellular proteomes and secretomes of spi1 and spi2-defective to the respective infectomes, relative to WT. This approach defines proteins with increased abundance during infection and identifies a subset of previously uncharacterized proteins with posssible roles in virulence. We selected candidates for functional follow-up based on abundance changes, predicted secretion or membrane location for interaction with the host cell, potential roles in virulence, and regulation influenced by SPI-1 and/or SPI-2.
As expected, we also identified known virulence factors, serving as positive controls for our approach, and allowed for further characterization of their roles in Salmonella virulence.
A comparison of WT-spi1 cellular proteome vs. infectome identified 28 proteins with higher abundance during infection in the absence of spi1 (Fig. 3A). Of these proteins, 17 are associated with SPI-1 and SPI-2. We prioritized two proteins (SL1344_1263, SL1344_3112) which are produced in vitro, displaying higher abundance during infection, and potentially interact with the host via localization (i.e., membrane) and secretion (i.e., signal peptide) (Supp. Table 6).
Of the remaining nine proteins, roles in motility, biosynthesis, and metabolism were described based on Gene Ontology Biological Process (GOBP). A comparison of WT-spi2-defective cellular proteome vs. infectome identified 316 proteins with higher abundance during infection in the disruption of spi2, supporting a larger discrepancy between the impact of infection from WT vs. SPI-2, compared to WT vs. SPI-1 (Fig. 3B). Of these proteins, 32 are associated with SPI-1 and SPI-2, two proteins were also detected in the SPI-1 comparisons (SL1344_1263, SL1344_3112) and two were hypothetical proteins with signal peptides (YnhG, SL1344_1563) (Supp. Table 7).
Of the remaining 280 proteins, roles in biosynthesis, metabolism, motility, transcription, translation, and enzymatic activity are described by GOBP, suggesting opportunity for curated future exploration. Secretome analyses supported these findings (Fig. 3C & 3D; Supp. Time-resolved SPI-1 and SPI-2 dependent protein profiles revealed abundance changes for SL1344_1563 (i.e., increased abundance at 6 hpi), strain-and time-associated response for SL1344_1263 (i.e., influenced by absence of SPI-1 and SPI-2 over the time course of infection), as well as strain-specific responses (i.e., SL1344_3112 regulated by the absence of SPI-1 and SPI-2, and ynhG with a delayed response when SPI-2 is absent) (Fig. 3E). Profiling of novel infectionassociated bacterial protein abundance in the cellular proteome and secretome, align with these infectome observations in WT, spi1, and spi2-defective strains, supporting potential coregulatory roles (Supp. Fig. 1). Taken together, our comparison of proteins produced during in vitro bacterial growth vs. co-infection with primary macrophages revealed a sub-set of novel infection-associated proteins with regulatory patterns influenced over the time course of infection and dependent on the pathogenicity islands in Salmonella.

Characterization of novel infection-associated bacterial proteins reveals roles in intracellular lifestyle and virulence
To evaluate whether the novel infection-associated bacterial proteins impact host cell response, we generated bacterial deletion mutants and performed in vitro and in vivo loss-offunction infection experiments. Our initial analysis monitored induced cell death by lactate dehydrogenase (LDH) release over the time course of infection for the Salmonella strains. We report similar macrophage cytotoxicity for spi1 and uninfected cells, as anticipated given the inability of spi1 to invade, along with similar trends in macrophage cytotoxicity amongst WT, spi2-defective, and novel infection-associated proteins (Fig. 4A). Next, we performed an association, invasion, and replication (AIR) assay to monitor difference in functions among the tested strains with spi1 and spi2-defective serving as controls for invasion and replication, respectively (Fig. 4B). We observed similar rates of association among the strains compared to WT with significant increase in association for spi1. For invasion, we observed a significant reduction in spi1, SL1344_3112, and SL1344_1263, supporting connections with SPI-1 regulation, as confirmed by infectome profiling presented above, as well as a reduction in invasion for spi2-defective. Lastly, we showed a significant reduction in replication for spi2-defective, as anticipated, and SL1344_1563, further supporting our observation of SL1344_1563 and coregulation with SPI-2.
To examine the contribution of the candidate proteins in virulence, we used a typhoid model of murine infection by competitive index with WT, causing moderate colonization of the gut and systemic dissemination [49]. In tissues representing initial infection (i.e., ileum, cecum, colon), commonly influenced by bacterial strain invasion, we observed no significant change in abundance for any of the strains in the ileum (Fig. 4C) but a significant increase in abundance of the SL1344_1263 strain, relative to WT, in the cecum, further supporting a role of co-regulation between SL1344_1263 and SPI-1 (Fig. 4D). We also observed a significant decrease in abundance of SL1344_1563, compared to WT, in the colon (Fig. 4E). Next, we assessed bacterial load in tissues associated with systemic infection (i.e., spleen, liver) and survival fitness of the bacterium (i.e., replication). Similar to our observation in the colon, we showed a significant decrease in bacterial load counts of SL1344_1563 compared to WT in the spleen (Fig. 4F) and liver (Fig.   4G), as well as a significant increase in abundance of ynhG in the spleen (Fig. 4F) and liver (Fig.   4G), supporting roles in bacterial replication and co-regulation with SPI-2. Notably, SL1344_3112 did not show any significant differences in virulence within the tested models and was not used for further analyses. Importantly, of the bacterial strains evaluated, only SL1344_1563 demonstrated a reduction in virulence, whereas the other strains showed heightened virulence in the absence of the proteins, suggesting co-regulation with SPI-1 and SPI-2 in positive and negative manners. Overall, these data reveal novel T3SS effector roles for infection-associated proteins in bacterial survival and virulence in the host.

Time-dependent bacterial protein production confirms proteome profiling
Based on our in vitro and in vivo data, we assessed the production of the novel T3SS effector proteins, proteins in vitro (i.e., during bacterial growth). Here, FLAG-tagged effector proteins under regulation of the native promoter allowed for tracking and quantifying production over a time course of bacterial growth (Fig. 5A). We observed a time-dependent increase in protein production for SL1344_1563 over 6 h of bacterial growth, corroborating with the time-dependent infectome protein profile (Fig. 3) and roles in virulence during systemic infection (Fig. 4). For SL1344_1263, we observed an early induction of the protein at 1 h, which decreased over time, aligning with infectome abundance (Fig. 3) and co-regulation by SPI-1, and early colonization of 23 the gut (Fig. 4). Lastly, we profiled YnhG production over time and observed consistent abundance throughout the growth curve, suggesting on/off production influenced by the presence of the host.
We confirmed protein production by quantifying the proportion of cells with FLAG production (Fig. 5B) and Western blot (Fig. 5C) and showed increased secretion of SL1344_1563 in culture supernatant over time (Fig. 5D). Together, profiling of FLAG-tagged protein abundance within Salmonella validates infectome and virulence characterization and supports SPI-1 and SPI-2 patterns of co-regulation and provides evidence of protein secretion.

Bacterial protein secretion and interactions within host cells as molecular checkpoints for perturbation of infection-associated proteins
We next investigated production and secretion of the infection-associated proteins in the presence of the host. We showed that both SL1344_1563 and YnhG are produced in the presence of macrophages at 6 hpi (Fig. 6A). Whereas SL1344_1563 emerged as a bacterium affiliated within the macrophage, supporting its role as a transmembrane protein involved in amino acid transport [50], YnhG was secreted outside the macrophage. Production of SL1344_1263 was not detected.

Discussion
In this study, we profile differences of the bacterial proteome and infectome in consideration of both SPI-1 and SPI-2 over a time course of infection in primary macrophages.
We define time-dependent and strain-specific responses of the host to Salmonella infection and uncover unique patterns of cellular regulation dependent upon these parameters. In addition, we report novel infection-associated bacterial proteins by comparing abundance during bacterial growth in vitro (i.e., cell culture) to co-culture with primary macrophages and we demonstrate coregulation with SPI-1 and/or SPI-2 for these effectors. We dissect the role of the infectionassociated proteins in bacterial virulence using in vitro (i.e., LDH & AIR assays) and in vivo (i.e., murine infection) models by generation of bacterial knockout strains, and we observe temporal production differences associated with induction of SPI-1 and SPI-2 in vitro. Moreover, we characterize the role of two infection-associated proteins, SL1344_1563 and YnhG, within the macrophages by visualizing their production and secretion, as well as generating a network of novel interactions between the pathogen and host proteins. Overall, our approach identifies and characterizes novel T3SS effectors regulated by SPI-1 and SPI-2. The method is universal with potential in diverse biological systems to discover new infection-associated proteins and delinate their roles in cellular regulation, as well as interactions with the host that are critical for infection.
Such connectivity between the host and pathogen suggests opportunity for perturbation to disrupt the observed interactions as a method of regulating bacterial virulence and controlling progression of disease.

New insights for a known virulence-associated bacterial protein, SL1344_1563
Previous structural and biochemical characterization of the periplasmic-binding protein component of a D-alanine ABC transporter, deemed DalS (also known as SL1344_1563 in strain SL1344), was co-regulated with the SPI-2 virulence locus through direct activation of response regulator, SsrB [50]. In the current study, we observe similar co-regulation with SPI-2, along with abundance of immune related proteins during infection of primary macrophages. We observe a fitness defect for dalS in vivo for overall survival and significant differences in competitive index in the liver and cecum [50]. We build upon these observations by reporting significant reduction in competitiveness of the SL1344_1563 strain in the spleen and liver, supporting a role in systemic infection, which aligns with the ability of bacteria to survive and replicate in the host, as associated with the SPI-2 virulence locus. An additional study explored the protective power of DalS for Salmonella from D-amino acid oxidase-dependent killing in neutrophils [54]. Here, we provide additional characterization of links between DalS and the host immune system by defining 26 a role in macrophage infection at the protein level, generating protein expression profiles within the bacteria as well as during co-culture with macrophages, and determining unique interactions between DalS and the host. Specifically, we report the interaction between SL1344_1563 and the bacterial protein, RplI, a 50 S ribosomal protein associated with translation, as well as two host proteins, including Rpp38, a ribonuclease P protein involved in rRNA processing, and Col6a1, a collagen alpha-1(VI) chain protein involved in cell adhesion and cellular response to amino acid stimulus [55]. Col6a1 was previously linked to an organ atlas of systemic bacterial infection with group A Streptococcus, as well as a recent connection of host infection with SARS-CoV-2 as a druggable target, implicating important roles in disease [56,57]. Further experimentation into the importance of the SL1344_1563-Col6a1 interaction is underway as a potential target to interrupt the interaction between pathogen and host during infection. For example, emphasis on the benefits of anti-virulence strategies to combat microbial infections have taken hold based on the potential to reduce virulence factor effectiveness without directly killing the pathogen, providing a selective therapeutic advantage without affecting commensal bacteria (e.g., microbiota) [58]. Such approaches lead to enhanced killing by the host's immune system and an overall reduction in selective pressure towards the development of antimicrobial resistance. Moreover, anti-virulence strategies may require rational drug design to specifically target the interaction between the host and pathogen proteins, a process fully supported through developments in mass spectrometrybased proteomics [59].

New insights for YnhG -an exported bacterial protein with implications in intercellular signaling and clearance of infection
YnhG is classified as a hypothetical exported protein involved in the peptidoglycan biosynthetic process as defined by GOBP. The presence of a signal peptide and LysM domain, a 27 highly-conserved carbohydrate binding module recognizing polysaccharides containing Nacetylglucosamine residues found in the bacterial cell wall, support its roles in building the bacterial cell surface, a critical point of interaction with the host [60]. Molecules exposed on bacterial surfaces elicit innate immune response through pattern recognition receptors, induction of cytokines and apoptosis, as well as activation of antimicrobial activity [61]. In addition, Lys-M domain-containing proteins of the host (e.g., LysMD1, LysMD2, LysMD3, and LysMD4) are assumed to be involved in immune responses against bacterial infection [62,63]. We also observed interaction of YnhG with several bacterial proteins at early (1 hpi) and late (6 hpi) time points of co-infection with connections to pathogenesis, including StpA and PrgI, suggesting a role of this protein in Salmonella virulence. In addition, YnhG interacts, either directly or indirectly with a plethora of host proteins, suggesting additional intercellular interaction with the host through secretion from the bacterial cell. Deletion of ynhG resulted in increased bacterial load counts in the spleen and liver following competitive index assays with a slight reduction in colonization of the ileum. These findings suggest a role for YnhG negative regulation of the host's immune response against the pathogen and therefore, strategies to enhance production of YnhG may alter disease outcome.

Conclusion
Overall, we report time-dependent and strain-specific proteomic responses of the host to Salmonella infection, and we discover and characterize new T3SS effector proteins regulated by SPI-1 and SPI-2 with implications in drug design. We provide a new strategy for defining the intricate relationship between a host and pathogen from a global and unbiased perspective to advance our understanding of infectious diseases and reveal new opportunities to combat infection.

Ethics Statement
For murine models of infection, experiments were performed in accordance with the guidelines of the Canadian Council on Animal Care and the University of British Columbia (UBC) Animal Care Committee (certificate A17-0228).

Availability of Data
The mass spectrometry proteomics data have been deposited in the PRIDE partner repository for the ProteomeXchange Consortium with the data set identifier: PXD027653 Reviewer login: reviewer_pxd027653@ebi.ac.uk Password: CegwyaGp