Salmonella Typhimurium effector SseI inhibits chemotaxis and increases host cell survival by deamidation of heterotrimeric Gi proteins

Salmonella enterica serotype Typhimurium (S. Typhimurium) is one of the most frequent causes of food-borne illness in humans and usually associated with acute self-limiting gastroenteritis. However, in immunocompromised patients, the pathogen can disseminate and lead to severe systemic diseases. S. Typhimurium are facultative intracellular bacteria. For uptake and intracellular life, Salmonella translocate numerous effector proteins into host cells using two type-III secretion systems (T3SS), which are encoded within Salmonella pathogenicity islands 1 (SPI-1) and 2 (SPI-2). While SPI-1 effectors mainly promote initial invasion, SPI-2 effectors control intracellular survival and proliferation. Here, we elucidate the mode of action of Salmonella SPI-2 effector SseI, which is involved in control of systemic dissemination of S. Typhimurium. SseI deamidates a specific glutamine residue of heterotrimeric G proteins of the Gαi family, resulting in persistent activation of the G protein. Gi activation inhibits cAMP production and stimulates PI3-kinase γ by Gαi-released Gβγ subunits, resulting in activation of survival pathways by phosphorylation of Akt and mTOR. Moreover, SseI-induced deamidation leads to non-polarized activation of Gαi and, thereby, to loss of directed migration of dendritic cells.

Introduction plays a crucial role in hydrolysis of GTP and in inactivation of heterotrimeric G proteins [24]. Thus, deamidation of the glutamine residue by PMT freezes the G protein in its active state.
Here, we elucidated the molecular mode of action of SseI. We report that the Salmonella SPI-2 T3SS effector deamidates heterotrimeric G proteins of the G i family in vitro and in vivo. Thereby, SseI is responsible for the activation of Akt kinase after target cell invasion and increases cell survival during Salmonella infection. Moreover, our studies reveal a pivotal role of SseI-induced deamidation of Gα i in the inhibition of dendritic cell migration.

Characterization of the deamidation of G proteins by SseI in vitro
The comparison of the amino acid sequence of the deamidase domain of PMT (residues 1144-1240) with SseI revealed high sequence similarity (S1A Fig). More importantly, amino acids known to be crucial for deamidase activity of PMT are conserved in SseI (e.g., C178, H216, and D231) [19]. Therefore, we utilized a previously described method to detect deamidation of G proteins by bacterial effectors [23,24]. To this end, SseI was coexpressed in E. coli with the α-subunit of G i2 . Deamidation of Gα i2 was determined by immunoblot analysis, utilizing a monoclonal antibody (GαQE) that specifically recognizes Gα after deamidation of a specific glutamine residue in the switch II region [25]. The GαQE antibody detected Gα i2 coexpressed with SseI in E. coli but not the solely expressed Gα i2 , indicating a deamidation activity of SseI ( Fig 1A). Purified Gα i2 was subjected to mass spectrometric (MS) analysis. MS analysis identified a tryptic peptide of Gα i2 (m/z 455.2170 2+ ) corresponding to amino acid residues 199-206. This peptide includes the glutamine residue (Gln-205) essential for the hydrolysis of GTP. Additionally, a second peptide was identified with a mass shift of 1 Da (m/z 455.7102 2+ ). Tandem MS analysis revealed that Gln-205 was deamidated, resulting in a glutamic acid residue at this position. No relevant deamidation occurred when Gα i2 was coexpressed with mutant SseI (SseI-C178A), which lacks deamidation activity (Fig 1B and S1B and S1C Fig). Similarly, recombinantly expressed wild type (wt) SseI (amino acids 137-322, SseI C ), but not the C178A, H216A, or D231A mutants, caused deamidation of purified Gα i2 (Fig 1C).

Construction of a PMT-SseI chimeric toxin
The effector SseI is secreted by a type III secretion system of S. Typhimurium. Therefore, SseI is not taken up by eukaryotic cells as compared to AB-type bacterial exotoxins like PMT. To enable cellular uptake of SseI, we utilized the receptor binding and translocation domain of PMT [26]. The deamidation domain of SseI (SseI C ) was fused C-terminally to the N-terminal part of PMT (amino acids 1-505) (Fig 2A). This chimera, PMT-SseI C , was recombinantly expressed, purified and tested for cellular activity. Treatment of HEK-293 cells with increasing concentrations of PMT-SseI C led to a deamidation of Gα proteins as determined by immunoblot analysis using the GαQE antibody. Treatment of cells with the C178A mutant of PMT-SseI C exhibited no deamidation (Fig 2B). In a next step, we compared the activity of PMT-SseI C with that of PMT. HEK-293 cells were intoxicated either with PMT or with PMT-SseI C and cell lysates were tested for deamidated Gα proteins. PMT led to two deamidation signals migrating at the molecular mass of the Gα q and Gα i proteins ( Fig 2C). However, PMT-SseI C only induced one deamidation signal migrating at the molecular mass of Gα i .

Activation of G i proteins by the chimeric toxin PMT-SseI C
Deamidation of the essential glutamine residue in the switch II region impairs the GTP hydrolysis by the α-subunit resulting in a permanent active phenotype of the G protein [23]. Activated Gα i family members inhibit adenylyl cyclase leading to decreased cAMP levels. We utilized a FRET-based approach to determine how cAMP levels respond to SseI treatment. The FRET sensor Epac2-camps [27] exhibits a decreased FRET ratio, when cAMP is increased. When HeLa cells, expressing the FRET sensor, were treated with the adenylyl cyclase activator forskolin, the FRET ratio declined, indicating increased cAMP levels. In line with an activation of Gα i , pretreatment with PMT-SseI C or PMT attenuated the effect of forskolin, indicating an inhibitory effect on adenylyl cyclase (Fig 2D). Similar results were obtained by direct measurements of cAMP levels (see ELISA assay S2A Fig).
Activation of heterotrimeric G proteins leads to dissociation into Gα and Gβγ-subunits [28]. Both subunits can interact with their specific effectors and stimulate specific signaling cascades. Therefore, we studied the impact of SseI on phosphoinositol-3-kinase (PI3K)γ, a prototypical Gβγ effector [29][30][31]. Activation of Gβγ was monitored by translocation of the phosphatidylinositol-3, 4, 5-triphosphate (PIP 3 ) sensor protein GFP-Grp1 PH from the cytosol to the membrane after additional ectopic expression of p110γ without and with p101 ( Fig 2E and  S2B Fig). As a control, we used N-formylmethionine-leucyl-phenylalanine (fMLP), which stimulates Gi-coupled GPCRs (G protein-coupled receptors). fMLP induced translocation of Grp1 PH to the membrane in the presence of complete PI3Kγ (p101 with p110γ) (S2Bb

Activation of Akt signaling by SseI
Various recent studies suggested that Gα i/o proteins and PI3-kinases play crucial roles in regulation of immune cell signaling and cell survival in infection [32][33][34][35][36]. Therefore, we studied the downstream signaling of Gβγ subunits and of PI3Kγ in more detail. To this end, we transfected bone marrow-derived macrophages (BMDMs) or RAW264.7 macrophages with a GFP expression vector either harboring SseI or the catalytically inactive mutant (SseI-C178A). After starvation of cells for 4 h, phosphorylation of Akt, mTOR and the mTOR-effector S6 ribosomal protein was analyzed by immunoblotting. Transfection with SseI strongly increased phosphorylation of the PI3K downstream effectors Akt, mTOR, and S6 ribosomal protein in BMDMs ( Fig 2F and S2C Fig). Similar results were obtained in RAW264.7 macrophages (S2D and S2E  Fig).

Role of SseI in infection of macrophages by Salmonella Typhimurium
To study the role of SseI in infection of RAW264.7 macrophages and BMDMs, we employed wt-or ΔsseI-Salmonella Typhimurium strains and used immunofluorescence microscopy. As depicted in Fig 3A, S. Typhimurium were identified by Salmonella O-antigen antiserum inside SseI deamidates an essential glutamine residue in the switch II region of Gα i2 . (A) Western blot analysis of Gα i2 . GST-Gα i2 was coexpressed without (control) and with SseI in E. coli. Purified GST-Gα i2 was immunoblotted and detected by the Gα i2 -specific antibody (Gα i2 ) and by the deamidation-specific antibody (GαQE). (B) Gα i2 coexpressed with SseI was analyzed by HPLC-MS/MS spectrometry. Combined extracted ion chromatograms for m/z 455.2 (2+) and 455.7 (2+) , corresponding to the tryptic peptides MFDVGGQR and MFDVGGER (amino acids 199-206) of Gα i2 , are shown (see also S1B and S1C Fig). (Upper panel) Gα i2 coexpressed with inactive SseI-C178A. (Lower panel) Gα i2 coexpressed with active wt SseI. (C) Immunoblot analysis of recombinantly expressed Gα i2 incubated with wild type C-terminal part of SseI C (wt) or with 3 different mutant SseI C (C178A, H216A and D231A). https://doi.org/10.1371/journal.ppat.1007248.g001 Salmonella Typhimurium effector SseI deamidates G i proteins and outside of macrophages. Notably, we observed deamidated G proteins, determined by the GαQE antibody, only in macrophages infected with wt-Salmonella. Moreover, deamidation occurred strictly after internalization of bacteria (Fig 3A, S3A and S3B Fig). Salmonella Typhimurium effector SseI deamidates G i proteins Next, we studied the activation of the PI3K/Akt pathway by SseI during infection of RAW264.7 macrophages with wt-or ΔsseI-Salmonella strains. Immunoblot analysis revealed G protein deamidation as early as~2 h post infection (p.i.) with subsequent increase over time ( Fig 3B). Infection with wt-and ΔsseI-Salmonella strains led to phosphorylation of Akt at 1 h p. i.. Hereafter, phosphorylation of Akt decreased over time in macrophages infected with ΔsseI-Salmonella. In wt-Salmonella-infected macrophages, phosphorylation of Akt remained longer than in the ΔsseI-Salmonella-infected cells (Fig 3B). Immunofluorescence microscopic studies revealed that 5 h p.i. an increase in Akt-phosphorylation was seen in wt-Salmonella-infected cells, but not in non-infected or ΔsseI-infected cells (Fig 3C and S3C Fig). In line with an essential role of PI3K, LY294002 (LY29), a specific PI3K inhibitor, decreased phosphorylation of Akt in a concentration-dependent manner (Fig 3D and 3E).
Several publications have shown that activation of Akt leads to anti-apoptotic and pro-survival outcomes in various mammalian cell types [33,36]. Therefore, we were prompted to investigate, whether the infection with wt-or ΔsseI-Salmonella has an effect on the survival of host cells. To this end, we measured cell death of infected cells by determination of LDH release from infected RAW264.7 macrophages. This study revealed that infection with ΔsseI-Salmonella caused higher amounts of LDH release than with wt-Salmonella ( Fig 4A). Accordingly, cell viability measured by the metabolic capacity of the cells, was increased after infection with wt-Salmonella as compared to ΔsseI-infected cells ( Fig 4B). Inhibition of PI3K with LY29 diminished the effects of wt-Salmonella (Fig 4A and 4B).

Effect of SseI on migration of dendritic cells
Recent studies from the Monack laboratory showed that SseI inhibits directed cell migration [10,13]. Therefore, we were prompted to study the role of G proteins in the action of SseI in more detail. To this end, we investigated the migration of dendritic cells in a 3D collagen model. We infected bone-marrow derived mature dendritic cells (DCs) with wt-or ΔsseI-Salmonella, seeded the cells into collagen gels and applied a CCL19 chemokine gradient. ΔsseIinfected cells migrated with comparable speed and directness towards the chemokine compared to non-infected cells. In contrast, wt-infected DCs completely lost their ability to migrate in a directed manner, confirming previous results [10,13] (Fig 5A-5C; S1 Video). As found for macrophages, time-resolved Western blot analysis revealed deamidation of Gα-subunits by wt-Salmonella in DCs ( Fig 5D). Moreover, very similar as observed in macrophages, downstream signaling of the PI3K pathway was elevated in wt-Salmonella-infected DCs, exhibiting increase in Akt phosphorylation ( Fig 5D).
To investigate whether the inhibition of migration observed with wt-Salmonella was due to the enzymatic activity of SseI, we ectopically expressed wt SseI or the inactive C178A-SseI mutant in DCs. As shown in Fig 5E and 5F (see also S2 Video), transfection with wt sseI led to a similar loss of chemotaxis as compared to infection with wt-Salmonella, while cells transfected with SseI-C178A mutant migrated normally. Interestingly, we did not detect a difference in the speed of migration between wt-and inactive sseI-transfected cells (Fig 5G).

Discussion
Our studies elucidate the molecular mechanism of the S. Typhimurium SPI-2 T3SS effector SseI. This effector protein plays a crucial role in S. Typhimurium infection after invasion of the pathogen and modulates the immune responses of the host [9]. Here, we show that SseI deamidates a specific glutamine residue (Gln205 in Gα i2 ) in the α-subunits of heterotrimeric G proteins, which is involved in GTP hydrolysis, thereby the G protein is persistently activated. Thus, T3SS effector SseI exhibits the same mode of action as exotoxin PMT from Pasteurella multocida. This finding is in line with the structural similarity of both toxins, although the primary sequence identity of the catalytic domains is only~20% [16]. Exchange of the conserved catalytic amino acid residues C178, H216 and D231 inhibited the deamidase activity of SseI. Because SseI is per se not able to enter target cells, we constructed a chimeric toxin (PMT-SseI C ), consisting of the N-terminal binding and translocation domain of PMT and the deamidating domain of SseI. Additionally, we transfected mammalian cells with a GFP-SseI construct. These experiments confirmed our in vitro results in intact cells, showing that SseI acts as a deamidase on heterotrimeric G proteins.
While PMT deamidates several G proteins, including G i/o , G q/11 and G 12/13 [23,24], our data indicate that in intact cells, SseI is specific for G i proteins. In line with activation of G i , SseI decreased cAMP levels in cells. Besides inhibition of adenylyl cyclase, G i is crucially involved in activation of PI3 kinase via release of Gβγ subunits [29,40]. Accordingly, we observed the translocation of a PIP 3 sensor protein (GFP-Grp PH ) to the cell membrane after treatment with PMT-SseI C . Many reports describe increase in Akt/mTOR signaling in Salmonella-host interactions [33,36,41]. We found that SseI increased phosphorylation and activation of Akt, its target mTOR and S6 ribosomal protein. These effects depended on the active SseI protein. The inactive mutant SseI-C178A had no effect.
To study the role of SseI in infection, we employed S. Typhimurium, in which SseI was deleted. These studies confirmed that after infection of macrophages with wt-S. Typhimurium, G i proteins were deamidated, while this was not observed with ΔsseI-Salmonella. Importantly, deamidation of G proteins was detected only after uptake of Salmonella into host cells, indicating the importance of invasion for activation of SPI-2 T3SS and translocation of the effector SseI. wt-Salmonella strongly activated the pro-survival kinase Akt in macrophages and DCs. Notably, also infection with ΔsseI-Salmonella resulted in Akt phosphorylation. However, the phosphorylation of Akt, observed in the absence of SseI, diminished with time, whereas wt-Salmonella caused longer lasting Akt phosphorylation. Short-term Akt activation by ΔsseI-Salmonella is probably caused by Salmonella outer protein B (SopB), which is an effector of SPI-1 T3SS and a strong activator of Akt [42,43]. Activation of Akt by SopB differs from SseI and might be independent of class I PI3 kinase [44,45]. In contrast, we observed that activation of with indicated antibodies (GαQE, p-Akt, Akt and actin). (E-G) Migration of DCs after ectopical expression of wild type (wt) SseI or SseI-C178A (C178A). Cells were transferred into a collagen matrix, a CCL19 gradient was applied, and directed migration (E) was determined by time-lapse microscopy (wt SseI, n = 43; SseI-C178A, n = 49). Cells were randomly selected from 3 independent experiments. (F, G) Quantification of the chemotactic index (F) and speed (G) of DCs. Statistical significance was assessed using unpaired, two-tailed Students t-test. https://doi.org/10.1371/journal.ppat.1007248.g005 Salmonella Typhimurium effector SseI deamidates G i proteins Akt by SseI is inhibited by the class I PI3K inhibitor LY29. In line with the activation of the PI3K-Akt signaling pathway, we observed increased survival of host cells after infection with wt-Salmonella, as compared to ΔsseI-infection.
Previously, it was reported that SseI inhibits the directed motility of macrophages and DCs towards a chemokine gradient [13]. Using an infection model, we confirmed that wt-Salmonella but not ΔsseI-Salmonella inhibited directed migration of DCs. Until recently, the molecular mechanism by which SseI affects directed migration was largely enigmatic. Our data indicate that the SseI-catalyzed deamidation of G i proteins is responsible for this effect. Concomitantly with inhibition of directed migration, infection with wt-Salmonella but not with ΔsseI-Salmonella caused deamidation of G i in DCs.
It is widely accepted that G i proteins play a pivotal role in chemotaxis and most, if not all, chemokine receptors are coupled to G i proteins [37,46]. Three G i proteins (G i1-3 ) exist, which share 85-95% amino acid sequence identity in their Gα subunits. While G i1 is only poorly expressed in leukocytes, G i2 and G i3 are highly expressed [37] (http://www.immgen.org/ databrowser/index.html). Especially, G i2 appears to be essential for immune cell signaling. Several studies showed that G i2 plays an important role in directed migration and homing of B and T-cells [47][48][49][50] and in macrophage migration [39]. Similarly, we show that deletion of Gα i2 inhibits the migration of DCs towards a CCL19 gradient. In line with recent publications, deletion of Gα i3 had minor effects on directed DC migration [39,51]. However, it is notable that wt-Salmonella infection was still able to inhibit directed migration even in Gnai2 -/-DCs, possibly indicating that Gα i3 is able to substitute in part for Gα i2 . When wild type Gα i2 was reexpressed in Gnai2 -/-DCs, directed migration was partially rescued. Importantly, expression of the mutant Gα i2 -Q205E that mimics the action of SseI, inhibited DC migration. Moreover, our studies reveal that SseI-induced deamidation of G i proteins inhibits directed migration but does not affect the migration speed. Also pertussis toxin was shown to inhibit directed migration of DCs without affecting migration speed [52]. However, in contrast to SseI, pertussis toxin inactivates Gi proteins by blocking the interaction with GPCRs.
Thus, how might SseI, which activates Gi proteins, inhibit directed cell migration? McLaughlin and coworkers demonstrated that SseI binds to the scaffold protein IQGAP1 (Ras GTPase-activating-like protein) [13], which interacts with numerous proteins (e.g. Rho proteins) involved in regulation of actin dynamics [53]. Because IQGAP1 also binds to the inactive SseI mutant, this cannot explain the action of SseI [13]. Earlier studies, which were mainly based on two-hybrid screen analyses, found interactions of SseI with the actin-binding protein filamin and the LIM domain fragment of TRIP6 [54,55]. However, these studies did not consider an enzyme activity of SseI. We propose that inhibition of directed migration by SseI is caused by blockade of the GTPase cycle and the non-polarized persistent activation of G i . SseI blocks rapid deactivation and activation of G i , which is probably essential to follow a chemoattractant gradient. In line with this hypothesis are recent findings about RGS (regulators of G protein signaling) proteins in immune cells. RGS proteins decrease the activity of G proteins by accelerating GTP hydrolysis and, thereby, accelerate G protein inactivation but also its activation [56][57][58], which appears to be essential for chemotaxis [8,37]. It was shown that immune Salmonella Typhimurium effector SseI deamidates G i proteins cells, harboring the Gα i2 G184S mutant, which causes resistance to RGS proteins, exhibit major alteration in chemotaxis [8]. The active state of the G184S Gα i2 mutant is prolonged, a condition mimicking the effects of SseI that persistently activates G i proteins. Thus, a similar scenario as seen with the G184S Gα i2 mutant but even stronger, might be induced by SseI, resulting in insufficient downstream signaling from GPCRs-involved in chemotaxis and loss of polarization necessary for directed migration. In line with this hypothesis are also previous findings obtained with constitutively activated PI3Kγ, which was fused with the CAAX box of K-Ras resulting in persistent cell membrane localization and non-polarized enhancement of PIP 3 production [59]. Leukocytes with constitutively activated PI3Kγ lost cell polarity and failed to follow a chemotactic gradient but they did not show differences in mean migration velocity. Surve and coworkers reported that selective activation of Gβγ by compound 12155 causes directed migration of neutrophils [60]. This effect was inhibited by Wortmannin, underlining the essential involvement of PI3K. PI3K is able to activate Rac, which is crucially involved in polarity control during neutrophil chemotaxis [61]. A recent study performed with optogenetic activation of PI3K sheds some light on the complexity of signaling pathways involved in directed migration downstream of PIP 3 [62]. The authors showed that PI3K activates Rac mainly via the guanine nucleotide exchange factor P-Rex-1. However, Rac activation by chemoattractants (involving GPCRs and not only PI3K) depended on multiple Rac regulators indicating involvement of a signaling network.
Taken together our data elucidate the molecular mode of action of the SPI-2 effector SseI by deamidation of a crucial glutamine residue in the α-subunits of G i proteins of the host, which results in inhibition of the turn-off mechanism of the G protein. The finding that G i proteins are persistently activated by SseI will offer a new perspective in studies on the biological effects of this SPI-2 T3SS effector and may help to identify further signal factors involved in chemotaxis, acting downstream of G i proteins.

Materials
PCR primers were from Apara (Denzlingen, Germany). All other reagents were of analytical grade and purchased from commercial sources. All Antibodies used in this study are listed in S1 Table. Bacterial strains S. enterica serovar Typhimurium NCTC12023 (ATCC 14028) was used as parent strain for the ΔsseI strain, which was created by deletion of the SseI coding sequence [63].
Dendritic cells: Cell concentration was adjusted to 2.5 x 10 5 cells/ml and 10 ml suspension was added into 10 cm bacterial petri dishes. GM-CSF was added for a final concentration of 20 ng/ml. At day 3, 10 ml medium with 20 ng/ml GM-CSF was added. At day 6, 10 ml medium were carefully removed and 10 ml fresh medium with 20 ng/ml GM-CSF were added. At day 8-9, non-adherent cells were collected and centrifuged (300 x g, 5 min). Pellets were resuspended in 10 ml fresh medium with 20 ng/ml GM-CSF without P/S and transferred into 6 cm cell culture dishes. LPS (200 ng/ml) were added for 24 h for further maturation of the DCs, when needed. Only non-adherent cells were used for experiments.
Bone marrow-derived macrophages: Cell concentration was adjusted to 5-7 x 10 5 cells/ml and 10 ml suspension was added into 10 cm tissue culture dishes. M-CSF was added for a final concentration of 20 ng/ml. At day 2, 10 ml fresh medium with 20 ng/ml M-CSF were added. At day 4, 10 ml medium were removed and exchanged by 10 ml fresh medium with 20 ng/ml M-CSF. At day 6-8, medium and non-adherent cells were discarded, adherent cells were detached with trypsin/EDTA and plated in medium with 20 ng/ml M-CSF without P/S at desired concentrations for experiments.

Plasmids
SseI encoding DNA was amplified from S. enterica serovar Typhimurium NCTC12023 and cloned into the bacterial expression vector pCOLDII and mammalian expression vectors pEGFP-C1 and pcDNA3.1 by standard cloning techniques. Additionally, a C-terminal fragment of SseI harboring the homology domain to PMT-C3, starting with amino acid 137 was cloned into pCOLDII. Chimeric PMT-SseI C consists of the N-terminal portion (amino acid 1-505) of PMT and the PMT-C3 homology domain of SseI (amino acids 137-322). PMT-SseI C was cloned in accordance to the method described before for other PMT fusion proteins into the pGEX2T vector [26]. The biological inactive mutant of SseI (C178A) was constructed by site-directed mutagenesis. Gα i2 DNA (with optimized codon usage for prokaryotic expression) was ordered as gBlock (Integrated DNA Technology, IDT) and cloned into pCOLDII. All oligonucleotides used in this study are listed in S2 Table.

Transfection, infection
HeLa cells were transfected using PEI as described previously [67]. Transfection of RAW264.7 cells was performed utilizing lipofectamin2000 (Thermo Fisher Scientific) according to the manufacturer's instructions. Transfection of BMDMs and BMDCs was done using amaxa nucleofector kits (Lonza) according to manufacturer's instructions. For in vitro infection studies S. enterica serovar Typhimurium NCTC12023 and the isogenic variant MvP393ΔsseI were used. BMDMs or RAW 264.7 cells were infected with Salmonella with indicated MOIs for 30 min. Then cells were washed 3 times with warm PBS and incubated with DMEM (RAW264.7) or RPMI-1640 (BMDMs) supplemented with 100 μg/ml gentamicin for 60 min. For the remaining time of the experiment, the medium was changed to DMEM or RPMI supplemented with 10 μg/ml gentamicin. For immunofluorescence detection of deamidation in RAW264.7 macrophages, these cells were infected with Salmonella for 5 h (MOI = 1). Dendritic cells were infected with a MOI = 30 for 30 min. Due to non-adherence of stimulated DCs, cells were resuspended in infection-medium and centrifuged (300 x g / 5 min). Pellets were resuspended in RPMI-1640 containing 100 μg/ml gentamicin and further treatment of the cells was similar to treatment of BMDMs.

FRET measurements
cAMP was measured by fluorescence resonance energy transfer (FRET) using a single chain cAMP sensor as described before [27]. The construct EPAC2-camps was kindly provided by Dr. Viacheslav O. Nikolaev (University of Göttingen, Germany).

cAMP measurements
HEK-293 cells were incubated with wildtype PMT-SseI or inactive mutant (PMT-SseI-C178A) for 8 h. Then forskolin (10 μM) and 3-isobutyl-1-methylxanthin (IBMX 100 μM) were added to the medium to induce adenylyl-cyclase activity and cells were incubated for further 45 min. Medium was discarded and cells were lysed with lysis buffer from the cAMP Parameter Kit (Biotechne). Further treatment of the samples was performed according to the manufacturer's instructions (cAMP Parameter Kit/R&D Systems Biotechne).

Confocal Laser Scanning Microscopy of GFP-Grp1 PH translocation
The subcellular distribution of GFP-Grp1 PH reflects the activity of PI3Ks, i.e. enhancement of membrane-associated fluorescence was taken as an indicator of PI3K activation. Assay was performed as previously described [31].

Fluorescence microscopy
For fluorescence microscopy, cells were prepared as described before [26]. Fixed samples were analyzed with an inverted Axiovert 200 M microscope (Carl Zeiss) with a Yokogawa (Tokyo, Japan) CSU-X1 spinning disc equipped with an emission filter wheel and 405 nm / 488 nm / 561 nm laser lines.

Protein expression
SseI and Gα i2 (His) were expressed as N-terminal His 6 -tagged proteins and purified by affinity chromatography via a Ni-NTA column. Expressions of PMT or chimeric constructs of PMT and SseI were performed as glutathione-S-transferase fusion proteins in accordance with the method previously described [67]. Coexpression of SseI or the inactive mutant (SseI-C178A) with GST-Gα i2 was done as previously described [23]. G proteins used for the in vitro deamidation assay in S4A Fig were purified as GST-fusion proteins. The GST-tag was cleaved with Thrombin.

Immunoblot analysis
For immunoblotting, samples were subjected to SDS-PAGE and transferred onto polyvinylidene difluoride-membrane. Deamidation specific antibody anti-Gα q Q209E (3G3) was used as described before [25]. Binding of the appropriate horseradish peroxidase-coupled secondary antibody was detected by using enhanced chemiluminescent detection reagent (New England Biolabs) and the imaging system LAS-3000 (Fujifilm).

Mass spectrometric analysis
Gel bands were destained for 10 min using 30% acetonitrile / 70% 0.1 M NH 4 HCO 3 , followed by 10 min washing with 0.1 M NH 4 HCO 3 ; this procedure was repeated twice. Prior to drying in a Vacuum Concentrator (Concentrator 5301, Eppendorf), gel bands were shrunk with 100% acetonitrile. By applying 0.1 μg trypsin per gel band, proteolytic digest was performed overnight at 37˚C in 0.1 M NH 4 HCO 3 . Peptides were extracted from the gel matrix with 5% formic acid. Peptides were analyzed by LC/MS on a Q-TOF mass spectrometer (Agilent 6520, Agilent Technologies) coupled to a 1200 Agilent nanoflow system via a HPLC-Chip cube electrospray ionization interface carrying a ProtID-II-Chip (Agilent Technologies). For peptide separation, a linear gradient ranging from 3% acetonitrile (ACN) to 30% ACN running for 30 min at a flow rate of 300 nl/min was used. MS spectra were acquired from m/z 50 to m/z 3000 while operating the Q-TOF in the 2 GHz extended dynamic range mode. Internal mass calibration was enabled. Three multiple charged peptides were selected in a data-dependent manner from each survey MS-scan for acquisition of fragment mass spectra. For peak list generation, raw data files were processed using Mascot Daemon2.4.0 with default settings for the Agilent Q-TOF. Using Mascot Server 2.4.1 database searches were performed with tryptic specificity and two missed cleavages against a small database comprised of known contaminants and the target sequence. Mass tolerances of 50 ppm and 0.05 Da were allowed for peptide and fragment masses, respectively. As variable modifications deamidation of glutamine and phosphorylation of serine and threonine were selected in addition to Carbamidomethyl (C), Gln-> pyroGlu (N-term. Q) and oxidation (M). Extracted Ion Chromatograms were generated using the Qualitative Analysis module within the MassHunter B04.00 software.

3D dendritic cell migration
Chemotactic migration assays of dendritic cells were performed as described before [64]. Briefly, pre-treated dendritic cells were mixed with bovine collagen (1.5 mg/ml) and incubated in self-built migration chambers at 37˚C, 5% CO 2 until the collagen network was polymerized. The collagen-cell mixture was covered with medium, containing CCL19 (500 ng/ml), which diffuses into the gel matrix and thus forming the chemokine gradient. Cell migration was monitored with a Lionheart FX automated microscope. Cell tracks were analyzed with ImageJ and the Manual Tracking plugin, followed by evaluation of the data with Ibidi chemotaxis and migration tool. Calculation of the chemotactic indices:

Ethics statement
All animal experiments were performed in compliance with the German animal protection law (TierSchG). The animals were housed and handled in accordance with good animal practice as defined by FELASA (www.felasa.eu) and the national animal welfare body GV-SOLAS (www.gv-solas.de). The animal welfare committees of the University of Freiburg as well as the local authorities (Regierungspräsidium Freiburg, licenses X-13/03A and X-17/01F) approved all animal experiments.