Flagella‐mediated cytosolic motility of Salmonella enterica Paratyphi A aids in evasion of xenophagy but does not impact egress from host cells

Salmonella enterica is a common foodborne, facultative intracellular enteropathogen. Typhoidal serovars like Paratyphi A (SPA) are human restricted and cause severe systemic diseases, while many serovars like Typhimurium (STM) have a broad host range, and usually lead to self‐limiting gastroenteritis. There are key differences between typhoidal and non‐typhoidal Salmonella in pathogenesis, but underlying mechanisms remain largely unknown. Transcriptomes and phenotypes in epithelial cells revealed induction of motility, flagella and chemotaxis genes for SPA but not STM. SPA exhibited cytosolic motility mediated by flagella. In this study, we applied single‐cell microscopy to analyze triggers and cellular consequences of cytosolic motility. Live‐cell imaging (LCI) revealed that SPA invades host cells in a highly cooperative manner. Extensive membrane ruffling at invasion sites led to increased membrane damage in nascent Salmonella‐containing vacuole, and subsequent cytosolic release. After release into the cytosol, motile bacteria showed the same velocity as under culture conditions in media. Reduced capture of SPA by autophagosomal membranes was observed by LCI and electron microscopy. Prior work showed that SPA does not use flagella‐mediated motility for cell exit via the intercellular spread. However, cytosolic motile SPA was invasion‐primed if released from host cells. Our results reveal flagella‐mediated cytosolic motility as a possible xenophagy evasion mechanism that could drive disease progression and contributes to the dissemination of systemic infection.

(STM) often have a broad host range, typhoidal Salmonella (TS) serovars such as S. Typhi (STY) or S. Paratyphi A (SPA) are characterized by adaptation to primate hosts in which typhoid or paratyphoid fever are important systemic diseases.The strict host adaptation limits studies of the virulence mechanism of TS, and frequently infection models of STM in susceptible mouse strains are used as surrogate to investigate systemic Salmonella infections.However, SPA, STY and other TS are distinct from NTS by the presence of increased accumulation of pseudogenes (Holt et al., 2009;McClelland et al., 2004), additional virulence factors as Vi capsule (STY) or typhoid toxin (STY, SPA) and distinct regulation of expression of virulence functions (Cohen et al., 2022;Reuter et al., 2021).
Both NTS and TS invade non-phagocytic mammalian cells, such as epithelial cells, by trigger invasion mediated by translocation of effector proteins by the Salmonella pathogenicity island 1 (SPI1)encoded type III secretion system (T3SS).Host cell invasion is considered to initiate the intracellular lifestyle of Salmonella, and allows to breach epithelial barriers.Translocation of SPI1-T3SS effector proteins also evokes strong proinflammatory responses of epithelial cells, leading to intestinal inflammation, a hallmark of gastroenteritis by NTS.While SPI1-T3SS also mediates invasion by SPA and STY, intestinal inflammation usually is absent and other routes of entry appear to be used to breach epithelial barriers of the intestines in order to reach systemic sites (reviewed in Dougan & Baker, 2014).
Compiling data on the intracellular lifestyle in mammalian host cells reveal that STM is well-adapted to life inside a specific PCV, referred to as Salmonella-containing vacuole, or SCV.The SCV possesses canonical markers of late endosomal compartments, yet allows STM survival and proliferation.The manipulation of the host cell endosomal system mainly mediated by effector proteins of the SPI2-encoded T3SS is central to SCV formation and maintenance (Jennings et al., 2017).
In addition to survival and proliferation in the SCV, further intracellular fates of STM are observed.If the integrity of the SCV is not maintained, STM is exposed to host cell cytosol.This may evoke ubiquitination and autophagy of cytosolic STM (Birmingham et al., 2006), or induce pyroptotic cell death (Fink & Cookson, 2007), or may lead to cytosolic hyper-replication resulting in the release of highly infected enterocytes as observed for STM (Brumell et al., 2002;Knodler et al., 2010).
In an approach to understand specific virulence mechanisms of TS, a comprehensive comparative transcriptional analyses of intracellular STM and SPA was performed (Cohen et al., 2022).The data revealed various differences in the expression of metabolic functions and distinct patterns of expression of flagella genes.We followed potential phenotypic consequences of the distinct expression patterns and observed that a subpopulation of intracellular SPA expresses flagella and is motile in the host cell cytosol.Such flagellamediated cytosolic motility was not observed for STM present in host cell cytosol, and clearly is distinct from intracellular motility evolved by other pathogens, where host cell actin polymerization is hijacked by various surface proteins.Actin-based motility (ABM) serves as the exit strategy of many intracellular pathogens, and among others, Shigella flexneri and Listeria monocytogenes use ABM for intercellular spread, mediated by plasma membrane protrusions harbouring bacteria that exit a host cell and enter a naïve neighboring host cell (reviewed in Dowd et al., 2021).While our prior work did not evidence intercellular spread by flagella-mediated motility of SPA, we found that cytosolic SPA released from host cells were invasion-primed to efficiently infect new host cells (Cohen et al., 2022).Yet, the mechanism of host cell exit of SPA remains to be resolved.
Here, we set out to analyze why SPA is released into cytosol and how host cells respond to intracellular motile SPA.We observed that flagella-mediated motility enables SPA to avoid xenophagic recognition, and serves as pre-requisite for host cell exit of invasion-primed motile SPA.

| Early escape to host cell cytosol depends on invasion mechanism of SPA
Although STM and SPA belong to the same species, the disease they cause in humans is very different ranging from a self-limiting gastroenteritis in case of STM infection, to severe systemic disease with potentially lethal outcome caused by SPA.We recently compared the intracellular gene expression profiles of STM and SPA during infection and identified differences in the expression of flagella/chemotaxis, SPI1 and carbon utilization pathways, and demonstrated flagellamediated movement of SPA in cytosol of host cell (Cohen et al., 2022).
Flagella-mediated motility was not observed for intracellular STM, despite a small proportion of STM that also enters host cell cytosol.
To further investigate the phenotypic heterogeneity of intracellular SPA, we performed live-cell imaging (LCI) of infected LAMP1-GFP expressing HeLa cells, and compared SPA WT to isogenic mutant strains deficient in flagella synthesis (ΔfliC), torque generation (ΔmotAB), or certain effector proteins of SPI1-T3SS (ΔsopE, ΔsopE2) or SPI2-T3SS (ΔsifA) crucial for early and late SCV integrity.
Furthermore, an SPA strain with a synthetic zipper invasion mechanism was investigated, that is, SPA ΔinvA [P invF ::Y.p. inv] with defective SPI1-T3SS and Yersinia pseudotuberculosis (Y.p.) Invasin protein Inv synthesized under control of SPI1 promoter P invF .
We quantified the intracellular phenotypes in categories of 'cytosolic motile' and 'cytosolic non-motile', while 'other' includes SPA residing in an SCV or showing cytosolic hyper-replication (Figure 1).In line with previous findings (Cohen et al., 2022), about 38% of SPA WT-infected host cells harbored cytosolic SPA exhibiting intracellular motility, while in 5% of infected cells cytosolic SPA WT were non-motile (Figure 1a,b).
SPA mutant strains with defects in maintaining SCV integrity (ΔsifA), or lacking SPI1-T3SS effectors (ΔsopE, ΔsopE2) showed similar results of phenotypes with cytosolic motile SPA ΔsopE, ΔsopE2, or ΔsifA in 30%, 39% or 40% of infected cells, respectively.This population was absent in cells infected with non-motile mutant strains lacking the flagella filament subunit FliC (Figure 1a DNA replication during infection.Because ΔmotAB and ΔfliC strains are non-motile, we also tested if motility during invasion affects intracellular phenotypes.For this, expression of motAB was placed under control of promoter P tetA , and induced by addition of anhydrotetracycline (AHT) (Schulte et al., 2019).We induced expression of motAB with AHT in bacterial culture and omitted AHT during infection and the following incubation time to avert motility in infected HeLa cells.
SPA ΔmotAB [P tetA ::motAB] showed increased invasion compared with ΔmotAB (Figure S1), but also completely lacked the cytosolic motile subpopulation.Further, we set out to evaluate the contribution of the mode of invasion to cytosolic release and subsequent intracellular motility using the mutant strain ΔinvA incapable of translocation of trigger invasion-mediating SPI1-T3SS effectors.We introduced a plasmid for expression of Y.p. inv under control of SPI1 promoter P invF , allowing expression under similar conditions as the SPI1-T3SS, and thereby conferring zipper invasion to SPA.Interestingly, for this strain the subpopulation of cytosolic motile bacteria was highly reduced to 1% of infected cells, the cytosolic subpopulation was almost absent (5%), and SPA predominantly resided in SCV (Figure 1a,d).These data confirm previous observations that identified trigger invasion as driving force for cytosolic release (Röder & Hensel, 2020).

| Co-operative trigger invasion by SPA
It was previously described that the mode of trigger invasion and also strain-specific equipment with SPI1-T3SS effectors affect cytosolic release following host cell entry, and that trigger invasion is often cooperative between Salmonella (Lorkowski et al., 2014;Röder & Hensel, 2020).The STM strain SL1344 harboring both sopE and sopE2 induced more pronounced actin rearrangements and membrane ruffles (Clark et al., 2011), and more frequently is released to host cell cytosol compared with STM NCTC12023 harboring only sopE2 (Röder & Hensel, 2020).We set out to investigate membrane ruffling during invasion by LCI of different STM serovars and SPA WT, as well as SPA invading through zipper mechanism (Figure 2, Figure S2, Movies S1-S4).We observed a rather small size of membrane ruffles for STM NCTC 12023 (Figure 2a respectively.Statistical analysis compared with SPA WT was performed with unpaired two-tailed t test and is indicated as *** for p < 0.001.at invasion sites (mean = 1.6 STM cells per site), whereas during the invasion of STM SL1344 multiple bacteria were involved in the process (mean = 3.2 STM cells per site).SPA WT showed highest accumulation of bacteria at sites of membrane rearrangement (mean = 7.3 SPA cells per site).If SPA invasion was mediated by Y. pseudotuberculosis Inv, that is by zipper invasion, bacterial counts per invasion site were much lower (mean = 2.1 SPA cells per site).

| Trigger invasion enhances membrane damage at nascent SCV
Previous work revealed that the SCV is prone to rupture, and damaged compartments are targeted by membrane damage sensors and repair mechanisms such as galectins, sphingomyelinases, or the ESCRT machinery (Ellison et al., 2020;Göser et al., 2020;Paz et al., 2010;Thurston et al., 2012).We used an engineered version of equinatoxin II (EqtSM) as a rapid reporter of membrane damage.EqtSM is binding cytosol-exposed sphingomyelin in damaged endosomal membranes (Deng et al., 2016;Niekamp et al., 2022).HeLa cells expressing Halotagged EqtSM were infected by SPA WT or SPA ΔinvA [P invF ::Y.p. inv] to assess potential contribution of the invasion mechanism to membrane damage of the nascent SCV (Figure 3).LCI revealed that trigger invasion by SPA WT led to 45% of SCV positive for EqtSM, indicating transient membrane damage, while only 14% of SCV were EqtSM-positive if SPA invaded through Y.p. inv-mediated zipper mechanism (Figure 3d).

| Motility of cytosolic SPA interferes with xenophagic capture
A subset of intracellular pathogens such as Listeria monocytogenes (L.m.) or Shigella flexneri actively escape the early PCV.As the host cell cytosol is a hostile environment for cytosolic bacteria, for example, due to recognition and clearance by cell-autonomous defense mechanism such as the autophagosomal machinery, these pathogens evolved sophisticated defense mechanisms to avoid decoration and degradation by expression of proteins interfering and inhibiting xenophagy.Such mechanisms have not been reported for cytosolic Salmonella.An event preceding autophagic recognition of cytosolic bacteria is ubiquitylation of their surface structures (Perrin et al., 2004), such as LPS (Otten et al., 2021).We also observed ubiquitylation of SPA located in host cell cytosol (Figure S3), and analyzed if subsequent xenophagic recognition occurs.Alternatively, we hypothesized that cytosolic motility of SPA may possibly interfere with xenophagy.
We analyzed the effect of inhibition of xenophagy on intracellular proliferation of SPA and STM.Infected cells were treated with 3-methyladenine or Wortmanin (WTM), commonly used inhibitors of autophagy (Klionsky et al., 2021).Because presence of 3-methyladenine appeared to increase loss of infected host cells, we focused on inhibition by WTM with more consistent results.
Imaging intracellular proliferation of SPA or STM in HeLa cells with or without WTM treatment indicated increased proliferation of SPA WT in WTM-treated cells (Figure S4).The effect of WTM on STM was less pronounced, while SPA invaded via Yersinia Inv showed reduced intracellular proliferation in presence of WTM (Figure S4).

Quantification of intracellular proliferation by CFU counts in host
cell lysates indicated that presence of WTM led to a small, but not significant, increase of intracellular proliferation of SPA WT, STM WT and SPA ΔmotAB [P tetA ::motAB] (Figure S5).The intracellular proliferation of STM ΔsifA was not increased, but we observed that SPA ΔinvA and STM ΔinvC invading via Yersinia Inv showed highly reduced proliferation in the presence of WTM.We anticipated increased proliferation of STM ΔsifA in presence of WTM, since this mutant strain is more frequently cytosolic and exposed to autophagy.In contrast, SPA and STM invading via Yersinia Inv are rarely located in host cell cytosol.These data indicate that population-wide analyses of effects of xenophagy on the fate of cytosolic SPA is not suitable.
To investigate the capture of SPA by autophagosomal membranes on single cell level, we deployed the late-autophagosomal membrane protein LC3B as a marker, generated a HeLa cell line stably expressing LC3B-GFP, and used this cell line for infection by SPA and LCI.
These analyses followed prior work on the fate of intracellular STM (Birmingham et al., 2006;Perrin et al., 2004).Individual intracellular SPA were scored for motility and association with LC3B-positive membranes (Figure 4, Movies S5 and S6).We compared SPA WT with SPA ΔmotAB [P tetA ::motAB] (motile during infection, but no intracellular motility), and SPA WT treated with cefotaxime and ciprofloxacin (non-motile and non-replicative) regarding levels of LC3B decoration.SPA-infected cells showed heterogenous populations as illustrated by representative images (Figure 4d-f).The number of bacteria and also the degree of LC3B decoration varied from cell to cell (Figure 4d-f, Movies S5 and S6).
Image analysis of more than 170 infected cells with over 1200 bacteria per condition (Figure 4a,b) revealed that residing in an SCV efficiently avoids LC3B decoration for all strains (4.5% LC3Bpositive SPA WT, 5.7% LC3B-positive SPA ΔmotAB [P tetA ::motAB]), with the highest rate of LC3B decoration observed in cefo/ciprotreated SPA WT (9% LC3B-positive, Figure 4c), possibly due to failure of maintaining SCV integrity in presence of antibiotics.
However, the cytosolic populations of the respective strains showed significant differences regarding LC3B decoration.The entire cytosolic population of SPA WT showed about 41.5% LC3Bpositive bacteria, whereas cytosolic populations of SPA ΔmotAB [P tetA ::motAB], and SPA WT + cefo/cipro both showed high frequencies of LC3B decoration with 70.6% and 79.7%, respectively (Figure 4c).We next dissected the cytosolic population into groups of non-motile SPA and motile SPA to analyze effects of motility on xenophagic escape.Surprisingly, SPA WT showed a high frequency of LC3B decoration for the non-motile subpopulation (87.7%), but low frequency for the motile subpopulation (18.6%).
In contrast, the motile fraction was almost absent for SPA ΔmotAB [P tetA ::motAB] and SPA WT + cefo/cipro strains with only 31 and 8 bacteria scored, respectively (Figure 4b).Thus, the frequency of LC3B decoration of the cytosolic population of these strains reflects that of the non-motile fraction with 71.1% and 80.2%, respectively.We conclude that cytosolic motility leads to reduced xenophagic capture of intracellular SPA.
To further investigate the heterogenous fate of individual intracellular SPA subpopulations, correlative light and electron microscopy (CLEM) analyses were performed to reveal ultrastructure and membrane association (Figure 5).Light microscopy allowed identification of cytosolic motile and non-motile, as well as SCV-bound SPA for subsequent analyses by transmission electron microscopy (TEM).
Whereas cytosolic motile SPA showed no contact to membranous structures, we observed SCV harboring SPA with distinct single membrane compartment in close contact to the bacterial envelope (Figure 5d,e).Cytosolic non-motile SPA showed association with LC3B-positive membranes (Figure 5f).The LC3B signal correlated to double membrane structures characteristic for autophagosomes.
We also detected SPA that were partially enclosed by LC3B-positive membranes, indicating either ongoing autophagosomal capture, or futile xenophagy.gradients.We set out to investigate the maximum velocity of flagellamediated movement of cytosolic SPA.As previously demonstrated for multiple other bacteria, swimming speed is dependent on the type of flagellation and rotation speed of the flagellum (reviewed in Wadhwa & Berg, 2022).First, we compared maximum velocity during different growth conditions in LB medium (Figure 6a).SPA cultures grown under aerobic conditions showed a median maximum velocity of 11 μm × s −1 when grown overnight, and 42 μm × s −1 when subcultured for 3.5 h, while SPI1-inducing microaerophilic conditions led to a median maximum velocity of 26 μm × s −1 .Other intracellular pathogens such as L.m. utilize actin polymerization to energize locomotion in host cell cytosol.We observed median maximum velocity of cytosolic L.m. of 0.5 μm × s −1 (Figure 6b) which was within the range of previously published data (Lacayo & Theriot, 2004).In contrast, cytosolic SPA were able to move almost as fast as SPA grown in media under microaerophilic conditions (25 μm × s −1 ; Figure 6c).

| Increasing LC3B decoration correlates with reduced motility of cytosolic SPA
As described above, a subpopulation of LC3B-decorated SPA are still motile (Figure 4, Movies S5 and S6).We hypothesized that increased autophagosomal capture delimits intracellular motility of SPA, or in turn, intracellular motility prevents autophagosomal capture.To test these possible correlations, we determined the signal intensity of SPA-associated LC3B-GFP as proxy of degree of autophagosomal capture of individual SPA, and determined by tracking of the same cytosolic SPA cells the maximum velocity (Figure 6d, Figure S6).Plotting the LC3B-GFP signal intensity versus maximal velocity of individual SPA indicated that velocity decreases with increasing association with autophagosomal membranes.Calculation of linear regression resulted in a negative slope (y = −19.5x+ 21.9) with maximum velocity dropping to <5 μm × s −1 at a relative LC3B-GFP fluorescence of 0.86, which reflects an almost complete enclosure of SPA.Pearson correlation coefficient analysis was performed to test correlation between LC3B decoration and maximum velocity of cytosolic STM.This resulted in a moderately negative correlation between the two variables (r = −0.412,n = 63) with significant relationship (p < 0.001).We conclude that flagella-mediated motility is a factor that reduces decoration of cytosolic SPA by autophagosomal membranes or alternatively, recruitment of autophagosomal membranes of cytosolic SPA hinders flagella-mediated motility.

| DISCUSS ION
Our study investigated the causes of the recently described cytosolic motility of SPA (Cohen et al., 2022) and its contributions to intracellular lifestyle.We demonstrated that mode of invasion and equipment with SPI1-T3SS effector proteins are substantial for membrane damage at the nascent SCV membrane and cytosolic release in early stage of infection.Furthermore, we showed that non-motile SPA are targeted by autophagosomal membranes.However, the population of cytosolic The main findings of this study are summarized in Figure 7.Our data suggest that release of SPA into cytosol occurs early after invasion, and is likely due to the inability to generate a stable nascent SCV (Figure 7e).We frequently observed clusters of invading SPA that induced high levels of local actin recruitment, and extensive membrane ruffles.Such events may lead to incomplete SCV formation, and may allow clusters of bacteria to directly enter the cytosol.This early exit from a labile nascent SCV occurs so rapidly that the transcriptional profile cannot be adjusted to the SCV-specific repression of SPI1 and flagella regulons.We propose that a large proportion of intracellular SPA maintains the transcriptional profile of invading bacteria.This hypothesis may be tested by labelling of flagella, infection, and LCI to follow the fate of flagella on SPA single cells after invasion.
SPA possesses both SopE and SopE2.We and others previously demonstrated that STM harboring SopE2 and SopE induce more severe membrane ruffles compared with the majority of STM strains only harboring SopE2 (Clark et al., 2011;Röder & Hensel, 2020).The more overt trigger invasion by SopE-positive STM was associated with increased cytosolic release, also occurring early after invasion (Röder & Hensel, 2020).This may be similar to damage of the nascent SCV induced by SPA.However, in contrast to other reports (Knodler et al., 2010(Knodler et al., , 2014)), we did not observe intracellular FBM of SopE/ SopE2-positive STM SL1344 (Cohen et al., 2022), indicating further differences between STM and SPA.SPA possesses functional sopE and sopE2, while STY strains possess functional sopE and pseudogenic sopE2 (Valenzuela et al., 2015), thus damage of the nascent SCV after invasion of STY may be less pronounced compared with SPA.
SopF is a recently identified SPI1-T3SS effector protein involved in stabilization of the nascent SCV and controlling autophagosomal recognition of intracellular STM (Lau et al., 2019;Xu et al., 2019).While sopF is present in the genome of the SPA isolate investigated here, expression of sopF and translation of SopF may be altered in SPA.Future work has to reveal a potential role of SopF in the intracellular lifestyle of SPA.A further factor possibly contributing to decreased stability of the nascent SCV is the altered function of SptP, an SPI1-T3SS effector protein with activity antagonistic to SopE.Prior work showed that host cells fail to restore normal actin cytoskeleton organization when SptP is defective and SopE activity overrules (Kubori & Galan, 2003).Johnson et al. (2017) demonstrated that SptP in STY and SPA have several AA exchanges that affect the binding of chaperone SicP, and by this the efficiency of translocation.

Recent work by
Virtually, all intracellular bacterial or parasitic pathogens adapted to cytosolic lifestyle have convergently evolved mechanisms to recruit host cell G-actin for polar polymerization to F-actin, resulting in actin-based motility (ABM) (reviewed in Dowd et al., 2021).ABM enables cell-to-cell spread, and thus enables infection of new host cells without demanding exit and exposure to antimicrobial mechanisms acting on extracellular bacteria.FBM is rarely observed for intracellular bacteria.In the late phase of the intracellular replication of Legionella pneumophila, FBM supports the host cell exit and released motile bacteria initiate infection of new host cells (Molmeret et al., 2010).FBM was reported for STM also in the late phase of the intracellular lifestyle, for example within epithelial cells extruded from the epithelial layer (Knodler et al., 2010).Interestingly, Burkholderia thailandensis use ABM, but can also activate a cryptic flagella synthesis system that mediates FBM (French et al., 2011).
We did not observe intercellular spread of SPA.While ABM is slower compared with the cytosolic FBM of SPA, the higher force generated by actin polymerization leads to generation of membrane protrusions and ultimately to intercellular spread.Furthermore, switching between swimming and tumbling motility restricts generation of membrane protrusions that are productive in infection of neighboring cells.Yet, we have to consider that the applied model observations were made for STM in gallbladder epithelium (Knodler et al., 2010).In contrast to STM, SPA showed lower intracellular replication (Reuter et al., 2021).Defects in metabolic utilization of nutrients in host cell cytosol such as glucose-6-phosphate and amino acids, and low SPI2-mediated replication inside SCV may contribute to reduced replication numbers of intracellular SPA (Cohen et al., 2022;Forest et al., 2010;Holt et al., 2009).
The cytosolic presence of pathogens such as L. monocytogenes or S. flexneri induce host cell-intrinsic defense mechanisms, resulting in ubiquitination and xenophagy.Xenophagy is partially avoided by ABM (Yoshikawa et al., 2009), but additional mechanisms appear required to efficiently avoid autophagosomal degradation.These include the T3SS effector protein IcsB of S. flexneri (Ogawa et al., 2005) with an acyltransferase activity affecting small GTPases and membrane fusion machinery (Liu et al., 2018), or secreted Listeriolysin O and phospholipases of L. monocytogenes targeting autophagosomal membranes (Birmingham et al., 2007).For SPA, the high speed of FBM is sufficient to enable protection of a subpopulation of cytosolic SPA against xenophagy (Figure 7f).Xenophagic recognition and potential subsequent clearance of cytosolic SPA is likely an arms race between the host cell and the pathogen.If the autophagosomal machinery is rapidly delivered to SPA and its flagella filament, motility is stopped and autophagy can be completed before SPA motility and proliferation initiates (Figure 7g in SPA-infected cells may be more complex and inhibition of autophagy could lead to massive intracellular proliferation and subsequent killing of host cells, thus biasing intracellular proliferation assays.For several Yersinia strains, it was reported that degradation by fusion with lysosomes is inhibited and Yersinia is able to replicate in a nonacidic autophagosome (Connor et al., 2018;Lemarignier & Pizarro-Cerda, 2020;Moreau et al., 2010;Straley & Harmon, 1984).Further investigations assessing acidification of LC3B-positive compartments and interfering with autophagosome formation using Lysotracker, siRNA approaches, and more selective autophagy inhibitors, in combination with image-based analyses could help to understand the hostpathogen interactions in SPA in more detail.
In addition to xenophagy, other host cell defense mechanisms restrict proliferation of cytosolic pathogens.One example are guanylate-binding proteins (GBP) that belong to the large family of interferon-inducible dynamin-related proteins.GBP1 can polymerize to a protein coat on envelopes of Gram-negative cells, and was shown to interfere with ABM of S. flexneri (reviewed in Kutsch & Coers, 2021).It will be of interest for future research to analyze the effect of GBP1 on FBM of SPA.
During systemic infections by TS, induction of massive inflammatory response due to bacteria-induced extrusion of infected cells was not observed, and this is considered to contribute to stealth strategy of TS (reviewed in Dougan & Baker, 2014;Hiyoshi, Tiffany, et al., 2018).Certain features of immune evasion, such as avoidance of neutrophil oxidative burst have convergently evolved, that is, by expression of Vi capsule (STY) or the very long O-antigen of LPS (SPA) (Hiyoshi, Wangdi, et al., 2018).Intracellular FBM has not been reported for STY, and we did not observe increased numbers of cytosolic STY, nor did we detect flagella expressed by intracellular STY.
Functional flagella are expressed by intracellular Legionella pneumophila during transition from the replicative to non-replicative, transmissive forms, and a recent study showed that flagellated subpopulations emerge in the Legionella-containing vacuole (Schell et al., 2016;Striednig et al., 2021).This transition occurs at the end of the intracellular replication cycle, and stimulates host cell pyroptotic death, and release of L. pneumophila for new rounds of host cell infection (Schell et al., 2016).We did not observe induction of host cell death by cytosolic motile SPA, but this host cell response is highly cell type dependent.Further understanding of physiological consequences of intracellular motility of SPA demand analyses in improved infection models such as organoids of human origin (Sepe et al., 2020) that are capable to simulate tissues closely related to in vivo conditions.

| Bacterial strains and growth conditions
In this study, Salmonella enterica serovar Typhimurium (STM) NCTC 12023, STM SL1344 and S. enterica serovar Paratyphi A (SPA) 45,157 were used as wild-type (WT) strains.All mutant strains are isogenic to the respective WT and Table 1 shows the characteristics.STM and SPA strains were routinely cultured at 37°C on Luria-Bertani (LB) agar or in LB broth using a roller drum at 60 rpm.Listeria monocytogenes was grown on brain heart infusion (BHI) agar or in BHI broth using a roller drum at 60 rpm.Antibiotics for maintenance of plasmids listed in Table 2 were added to LB in concentrations of 50 μg × mL −1 carbenicillin, 50 μg × mL −1 kanamycin, or 12 μg × mL −1 chloramphenicol.

| Generation of bacterial strains
Isogenic mutant strains were generated by λ Red recombineering for insertion of kanamycin resistance (aph) cassettes amplified from template plasmids pKD4 or pKD13 as described before (Chakravortty et al., 2002;Datsenko & Wanner, 2000) using oligonucleotides listed in Table 3. Insertion of aph cassettes was confirmed by colony PCR.
If required, the aph cassette was removed by introduction of pE-FLP for FLP-mediated recombination.

| Construction of plasmids
Plasmids were generated by Gibson assembly as described before (Röder et al., 2021;Röder & Hensel, 2020;Schulte et al., 2019) using oligonucleotides listed in Table 3.Plasmids used in this study are listed in Table 2 and were introduced into the respective strains by electroporation.

| Host cell culture and infection
Host cells were cultured and infected as described previously (Cohen et al., 2021).In short, HeLa cells (ATCC no.CCL-2) were seeded in surface-treated 8-well chambered coverslips (ibidi) 24 or 48 h prior to infection to reach ~80% confluency (~80,000 cells) and were then used for infection.Cells were continuously maintained in high glucose (4.5 g × L −1 ) DMEM (Merck) supplemented with 10% inactivated fetal calf serum (iFCS, Sigma) at 37°C in a humidified atmosphere with 5% CO 2 .If indicated, cells were pulse-chased with 100 μg × mL −1 fluid phase marker Dextran AlexaFluor 647 (Thermo Fisher Scientific) 16 h prior infection.STM strains for infection experiments were subcultured from an overnight culture (1:31) in fresh LB medium and grown for 3.5 h at 37°C under aerobic conditions using a roller drum at 60 rpm.SPA strains for infection were grown for 8 h under aerobic conditions as described above and subcultured (1:100) in fresh LB medium and stationary phase subcultures were grown for 16 h under microaerophilic conditions as described in Elhadad, Desai, et al. (2015).For infection with Listeria monocytogenes, overnight cultures grown under aerobic conditions were used.Bacteria were adjusted to an optical density of 0.2 at 600 nm in PBS and used for infection with the respective MOI (between 5 and 90, dependent on strain and experiment).Bacteria were centrifuged onto the cells for 5 min at 500× g to synchronize infection and incubated for 25 min.
After washing thrice with PBS, cells were incubated in medium containing 100 μg × mL −1 of gentamicin to kill extracellular bacteria.
Afterwards, cells were maintained in medium containing 10 μg × mL −1 gentamicin until fixation or LCI.To image invasion of HeLa cells by Salmonella, cells were infected directly on the microscope stage and imaged for ~1 h.subsequent fluorescence-activated cell sorting (FACS).Supernatant containing lentivirus was generated by seeding 3.8 × 10 5 cells × mL −1 HEK 293FT cells (Invitrogen R700-07) in 10 cm tissue culture plates (10 mL per plate, TPP) in antibiotic-free growth media (high glucose DMEM + 10% iFCS).After 24 h, cells should be 70%-80% confluent and were then transfected with lentiviral packaging plasmid (9 μg psPAX2), envelope plasmid (0.9 μg pMD2.G) and pLX304 carrying eGFP-LC3B (9 μg) using Opti-MEM (Fisher Scientific) and FuGENE HD transfection reagent (Promega).The transfection mix was incubated for 30 min at room temperature before being added to the cells dropwise.Cells were incubated for at least 18 h and medium was changed to 15 mL growth medium containing 30% iFCS.

| Generation of HeLa cells stably expressing LC3B-GFP
Medium containing lentiviruses was harvested after 24 and 48 h and centrifuged at 350× g for 5 min to pellet any residual packaging cells.Supernatant was stored in sterile polypropylene storage tubes and stored at −80°C.Half of the harvested supernatant was filtered using fast flow & low binding filters (Merck Millipore).
HeLa cells were seeded in 24-well plates (1 × 10 5 cells × mL −1 ) and incubated for 24 h before 8 μg × mL −1 Polybrene (Sigma) was added to the cells.Filtered supernatant containing lentiviruses was added to cells 48 h after seeding.48 h after lentiviruses were added to the cells, medium was exchanged to DMEM containing 10 μg × mL −1 Blasticidin (Invitrogen) and incubated for 72 h.Medium was changed to DMEM without antibiotics to allow growth for 24 h before cells were detached for culture in tissue culture flasks (TPP) to obtain sufficient cells numbers for FACS.
For FACS of lentivirus-transfected HeLa cells, cells were de-

| Correlative light and electron microscopy (CLEM)
CLEM of HeLa LC3B-GFP infected by SPA was performed as previously described (Krieger et al., 2014).Briefly, HeLa LC3B-GFP cells were grown on MatTek dishes with gridded coverslips.Leica, Wetzlar, Germany) and collected on formvar-coated copper slot grids (Plano).Grids were post-stained for 30 min with 2% uranyl acetate (Roth) and 20 min with 3% lead citrate (Leica).TEM images were acquired using a Zeiss Leo 912 Omega (Zeiss, Oberkochen, Germany) equipped with a CCD camera (TRS, Moorenwies, Germany).Overlays of the light and electron microscopic images were generated using LAS AF (Leica) and Photoshop CS6 (Adobe).et al., 2017) with minimal displacement threshold of 1 μm.For tracking of intracellular L.m., infected cells were imaged with setup as described for intracellular SPA at ~0.1 fps over 10-30 min.Tracking analyses were performed in FIJI with manual tracking plugin.

| Analyses of LC3B decoration
For analyses of LC3B-decoration of intracellular SPA, randomly selected infected cells were imaged 3-4 h p.i. as described above and were analyzed manually regarding LC3B-GFP signal at SPA cell bodies.100 μg × mL −1 cefotaxime and ciprofloxacin were added 2 h p.i. if indicated.
For correlation of maximum velocity and LC3B decoration of cytosolic SPA, individual bacteria were analyzed regarding LC3B-GFP fluorescence intensity using ZEISS Efficient Navigation (ZEN) software.Manual tracking with the same bacteria was performed as described above.

| Transfection of HeLa cells with Halo-tagged EqtSM
HeLa cells were maintained and infected as described above.In short, cells were seeded in 8-wells 2 days prior infection.Cells were transfected with Halo-tagged EqtSM expression construct 1 day prior infection using FuGENE HD in 1:2 ratio.Infected cells were labeled with 100 nM Janelia Fluor 646 HaloTag ligand (Promega, GA1120) for 30 min, washed five times with PBS and subsequently imaged using the Zeiss Cell Observer SD microscope set-up.
Cultivation of bacteria and infection of HeLa cells was carried out as described above.Infected cells were fixed with 3% PFA at the desired time point.After washing thrice with PBS, cells were incubated in blocking solution (2% goat serum, 2% BSA, 0.1% saponin in PBS) for 30 min.Next, cells were incubated for 1 h at RT with mouse monoclonal antibodies against Ubiquitin (Biomol PW 8810).After washing cells thrice with PBS, cells were incubated with primary antibodies against Salmonella O-Antigen of SPA (BD Difco 229471) for 1 h.After washing thrice with PBS, cells were stained with the appropriate secondary antibodies for 1 h (Invitrogen A11031, Dianova 111-607-003).After three washing steps with PBS, coverslips were mounted with Fluoroshield ,c), or defective in energizing flagella rotation (ΔmotAB), similar to cells exposed to the antibiotics cefotaxime and ciprofloxacin affecting bacterial cell wall synthesis and F I G U R E 1 Quantification of intracellular SPA phenotypes.(a-d) HeLa cells stably expressing LAMP1-GFP (green) were infected with Salmonella Paratyphi A (SPA) WT or isogenic mutant strains expressing mCherry (red) at MOI 30, 60, or 90 for WT, ΔinvA [P invF ::Y.p. inv], or ΔfliC, ΔmotAB, respectively.(a) Quantification of intracellular phenotypes was performed during LCI.From 4 to 6 h p.i., at least 100 infected cells per strain were examined for intracellular phenotypes.Data are means and standard deviations (SD) from three independent experiments.Statistical analysis was performed with one-way ANOVA and is indicated as n.s., not significant; *p < 0.05; **p < 0.01; ***p < 0.001.(b, c) Time-lapse images of infected HeLa-LAMP1-GFP cells showing movement of three cytosolic SPA WT (b, turquoise, magenta and yellow arrowhead), no movement of cytosolic SPA ΔfliC (c, red and blue arrowhead) and SPA ΔfliC residing in SCV (c, green arrowheads).(d) Maximum intensity projection of HeLa-LAMP1-GFP cells infected with SPA ΔinvA [P invF ::Y.p. inv] residing in SCV (green arrowheads).Scale bars: 10 μm.
, Figure S2, Movie S1), but more extensive membrane ruffling for STM SL1344 (Figure 2b, Figure S2, Movie S2).Prolonged and extensive membrane ruffling of a larger cell area was observed for SPA WT (Figure 2c, Figure S2, Movie S3).SPA ΔinvA [P invF Y.p. inv] showed lowest actin rearrangements at invasion sites (Figure 2d, Figure S2, Movie S4).Image analysis of foci of invasion indicated that STM NCTC 12023 entry occurs mostly by single bacteria F I G U R E 2 SPA induces massive membrane ruffles during trigger invasion, and invades in clusters of bacteria.(a-d) HeLa cells stably expressing Lifeact-eGFP (green) were infected with STM NCTC 12023 (a), STM SL1344 (b), SPA WT (c), or SPA ΔinvA [P invF ::Y.p. inv] (d) expressing mCherry (red) at MOI 75.Cells in chambered coverslips were infected on the microscope stage and imaged for 1 h by spinning disc confocal microscopy with intervals of 3-5 min between images.Still images from time-lapse are shown.Arrowheads indicate sites of ongoing invasion.Scale bars: 10 μm.(e) Image series were assessed for numbers of invading bacteria per ruffle.Shown are Tukey's box plots of invading bacteria per ruffle during infection with error bars including data of 1.5 × interquartile range (IQR).Outliers beyond the 1.5 × IQR are shown as dots.Middle lines and "+" denote median and mean value, respectively.Compiled from three independent experiments, 200, 169, 170 and 130 individual invasion events were analyzed for STM NCTC 12023, STM SL1344, SPA WT and SPA ΔinvA [P invF ::Y.p. inv], The results underline the critical role of entry mechanism for SCV integrity, and for subsequent intracellular lifestyle of Salmonella.F I G U R E 3 Trigger invasion by SPA destabilizes nascent SCVs.(a) To localize sites of SPA-induced membrane damage and to quantify the frequency of SCV damage, HeLa cells stably expressing LAMP1-GFP (b) or Lifeact-eGFP (c) (green) were transfected for expression of EqtSM-HaloTag and labeled with Janelia Fluor 646 HaloTag ligand (blue).Subsequently, cells were infected with SPA WT or SPA ΔinvA [P invF ::Y.p. inv] expressing mCherry (red).Infected cells were imaged 1-2 h p.i. and representative infected cells are shown as maximum intensity projection.Arrowheads indicate intact SCVs (green), or SCVs targeted by EqtSM (orange).Scale bars: 10 μm.(d) SCVs were scored for association with EqtSM-HaloTag.Data are means and SD from three independent experiments.Statistical analysis was performed with unpaired two-tailed t test and is indicated as * for p < 0.05.
For many bacteria of various cell shapes, flagellar rotation provides force for locomotion within their habitat.The rotation of one or multiple flagellar filaments mediates phases of straight swimming, alternating with phases of tumbling allowing orientation in chemical F I G U R E 4 Cytosolic non-motile SPA are targeted by autophagosomal membranes.(a-c) HeLa cells stably expressing LC3B-GFP (green) were pulse-chased with Dextran-Alexa647 (blue) and infected with SPA WT or SPA ΔmotAB [P tetA ::motAB] expressing mCherry (red) at MOI 30. 100 μg × mL −1 cefotaxime and ciprofloxacin were added to infected cells at 2 h p.i. if indicated.At 3-4 h p.i., randomly selected cells were imaged and later assessed for SPA associated with LC3B-positive membranes.(a, b) Quantification of SPA with (dashed bars) or without (bars) LC3B decoration, compiled from three independent experiments.(a) Quantification of SCV-bound versus cytosolic SPA without or with LC3B decoration.(b) Cytosolic SPA quantified in A were further scored as motile or non-motile, and assessed for LC3B decoration.(c) Percentage of SPA targeted by LC3B-containing membranes calculated from data shown in a and b.Data are means and SD of 195 infected cells with 2450 bacteria for WT, 171 infected cells with 3360 bacteria for ΔmotAB [P tetA ::motAB], and 215 infected cells with 1291 bacteria for WT + cefo/cipro from three independent experiments.Statistical analysis was performed with one-way ANOVA and is indicated as n.s., not significant; *p < 0.05; ***p < 0.001.(d-f) Representative images showing subpopulations in SPA-infected cells.Arrowheads indicate SPA residing in SCV (blue), cytosolic SPA (red), LC3B-decorated SPA (yellow).Images are shown as single Z-plane from stack (d) or as maximum intensity projection (e, f).Scale bars: 10 μm.

F
I G U R E 5 Various degrees of xenophagic capture of distinct subpopulations of intracellular SPA.HeLa cells expressing LC3B-GFP (green) were pulse-chased with Dextran-Alexa647 (blue) and infected with SPA WT expressing mCherry (red) at MOI 30.Cells were fixed 4 h p.i. during LCI (a), and subsequently processed for transmission electron microscopy (TEM) of ultrathin sections, which allowed relocation of the same cell by EM (b).LCI (a) and TEM (b, c) images were superimposed for correlation (d) and details are shown at higher magnification (e, f).Arrowheads indicate SPA residing in SCV (blue), cytosolic SPA (red) and SPA targeted by LC3B-positive autophagosomal membrane (yellow).Partial enclosure by autophagosomal membranes is indicated by green arrowheads.Scale bars: 10 μm (a, b), 5 μm (c, d), 1 μm (e), 500 nm (f).
motile SPA is able to delay or escape xenophagic recognition as infection progresses.Our work adds flagella-based motility (FBM) of intracellular bacteria to the mechanisms of evasion of host cell xenophagy.
of confluently cultured non-polarized epithelial cells may not reflect host cell types relevant during infection of human hosts as main host cells of SPA are phagocytic cells, leading to dissemination through the lymphatic system and further colonization of internal organs such as the liver, spleen, bone marrow and gall bladder.Motility of SPA inside the phagocytic cell line U937 has only been observed in spacious vacuoles rather than in the cytosol of host cells(Cohen et al., 2022), probably because presence of bacteria in the cytosol of phagocytic cells most likely results in inflammatory cell death, socalled pyroptosis(Castanheira & Garcia-Del Portillo, 2017).Similar F I G U R E 6 Xenophagic capture reduces velocity of intracellular motile SPA.Single bacterial cells after culture in broth media (a), or within infected host cells (b, c) were registered by LCI.Subsequently, bacterial motility was analyzed by automated (a, c) or manual (b) tracking analyses.(a-c) Maximum velocity of single bacterial cells from tracking SPA grown in culture (a), or intracellular SPA (b) or L.m. (c).For single cell analyses, tracks of 476 SPA WT from O/N culture, 7058 SPA WT from late log culture, 2998 SPA WT from microaerobic culture were collected from one experiment.For intracellular bacteria, tracks of 16,248 cytosolic SPA WT were collected from three independent experiments, and 42 tracks for L.m. from one experiment.Data are shown as Tukey's box plots with error bars including data of 1.5 × IQR.Outliers beyond the 1.5 × IQR are shown as dots.Middle line denotes median and "+" mean value.(d) Correlation between SPA-associated relative LC3B-GFP intensity to maximum velocity of individual cytosolic SPA.Representative images from measurement of fluorescence intensity are shown as single Z-plane from image stacks.Linear regression and 95% confidence bands were calculated for analyses of 63 cytosolic SPA from two independent experiments.
).The mode of trigger invasion by clusters of SPA, and resulting simultaneous cytosolic presence of several SPA cells likely contributes to rapid exhaustion of xenophagic capacity of the host cell, and inability to control SPA proliferation.We showed that the cytosolic population is escaping xenophagic recognition by FBM.Motile SPA showed the same maximum velocity in host cell cytosol as in culture media, interfering with decoration by autophagosomal membranes, although cytosolic SPA are readily ubiquitylated.If decoration by LC3B-positive membranes is only delayed to late time points of infection, and if this also leads to lysosomal degradation of engulfed bacteria deserves further investigation.Our experiments with inhibition of autophagy by Wortmanin did not reveal gross increase of STM or SPA intracellular proliferation upon inhibition.Prior work showed that autophagy mainly affects cytosolic STM in the early phase of infection, and a ΔsifA strain with increased release into host cell cytosol in the later phase of infection was rarely targeted by autophagy (Birmingham et al., 2006).The host pathogen-interplay F I G U R E 7 Model for SPA entry into host cell cytosol and xenophagic escape.Zipper or trigger invasion with limiting extend of membrane ruffling (a, b) primarily lead to an intracellular lifestyle with intact SCV, activity of SPA within the SCV such as SPI2-T3SS-mediated translocation, and maturation of the SCV (d).Trigger invasion with extensive membrane ruffling and increased clustered invasion (c) lead to membrane damage, rupture of the nascent SCV (e) and cytosolic lifestyle of SPA.Flagellated and motile cytosolic SPA delay or escape xenophagic capture (f).Non-motile cytosolic SPA are captured by LC3Bpositive membranes, while flagellamediated motility permits evasion of xenophagy (g).
tached from tissue culture flask and resuspended in appropriate amount of medium to reach concentration of 1 × 10 7 cells × mL −1 .A 100 μm cell strainer (BD Falcon) was used to prevent aggregation of cells prior to FACS.Cells were sorted for GFP fluorescence with BD FACS Aria.A mixed population of cells with different GFP expression levels was obtained.GFP-positive cells were cultured again in tissue culture flasks before a second round of FACS with the Sony SH800S was conducted.Single cells were sorted into 96-well plates to obtain single clones of LC3B-GFP expressing cells.Cell clones were analyzed regarding GFP intensity, cell division, and phenotypes of intracellular Salmonella prior to use in further experiments.TGT TGA CAA ATA AAG TCG TTA AAG ATT TTG TGT AGGCTGGAGCTGCTTC sopE-Del-Rev AGGAA GAG GCT CCG CAT ATT TTT TGG TTT TTC AGT GTT CAC ATA TGAATATCCTCCTTAG sopE2-Red-Del-For AAAGT GTA GCT ATG CAT AGT TAT CTA AAA GGA GAA CTA CCG TGT AGGCTGGAGCTGCTTC sopE2-Red-Del-Rev TTAAT TCA TAT GGT TAA TAG CAG TAT TGT ATT TAC TAC CAC ATA TGAATATCCTCCTTAG MotA-Del13-For CAACA GCG GAA GGA TGA TGT CGT GCT TAT CTT ATT AGG TTA TTC CGGGGATCCGTCGACC MotB-Del13-Rev TTCCG CTT TTG GCG ATG TGG GTA CGC TTG CCG GCG GGG CTT GTA GGCTGGAGCTGCTTCG uhpT Del13 For CCATT CGC AGG TAT AAA AAT TAG CTC AGG AGT AAT CCA TGA TTC CGGGGATCCGTCGACC uhpT Del13 Rev CGTTA CCA AAT GCA CAC ATT TAA GCG ATA TTG ACT TGC TGT GTA GGCTGGAGCTGCTTCG AAA AAG CCG GAT TAA TAATCTG 1f-SPA-PsopE GATTA TTA ATC CGG CTT TTT TAT TAT TTT CTC GGC CAG TGTACGTTCAA 1r-SPA-PsopE TCCCG GGT ACG TAG GAT CGG TAA TGA TCC TTT TAT ATGTACATAAC Vf-PtetA TCCGG CGA TTG ATT CAC CGAC Vr-PtetA TTCAC TTT TCT CTA TCA CTG ATA GGG AGTGGTA 1f-ptetA-SPAmotAB CCCTA TCA GTG ATA GAG AAA AGT GAA GTG CTT ATC TTA TTAGGTTACCTGGT 1r-SPAmotAB-ptetA GGTGA ATC AAT CGC CGG ATC ACC TCG GTT CCGCTTTTG 4.6 | Live-cell imaging Prior to LCI, medium of infected HeLa cells was changed to high glucose DMEM without phenol red supplemented with 30 mM HEPES.LCI was performed at 37°C and an atmosphere of 5% CO 2 with a Cell Observer microscope (Zeiss) equipped with a Yokogawa Spinning Disc Unit CSU X1a5000, an incubation chamber, 63× objective (α-Plan-Apochromat, NA 1.46) and 40× objective (Plan-Apochromat, NA 1.4), two ORCA Flash 4.0 V3 cameras (Hamamatsu) and appropriate filters for the respective fluorescence proteins or dyes.
Cells were fixed with 4% paraformaldehyde (Electron Microscopy Sciences) and 0.1% glutaraldehyde (Electron Microscopy Sciences) in 200 mM HEPES directly on the microscope stage during imaging of cytosolic motile SPA.Cells were incubated for 30 min and were then rinsed thrice with 200 mM HEPES buffer.Cells were subsequently fixed with 2.5% glutaraldehyde in 200 mM HEPES for 1 h, washed thrice with buffer and post-fixed with 2% osmium-tetroxide (Electron Microscopy Sciences), 0.1% ruthenium red (Applichem) and 1.5% potassium ferrocyanide (Sigma) in 200 mM HEPES on ice.After several washing steps, the cells were dehydrated in a cold graded series of ethanol and finally one rinse in anhydrous ethanol and two rinses in anhydrous acetone at room temperature.Infiltration was performed with increasing concentrations of EPON812 (Sigma) in anhydrous acetone.Utilizing the coordinate system, the ROI from light microscopy was relocated on the EPON block, trimmed and 70 nm serial sections were cut with an ultramicrotome (Leica EM UC7,

For
microscopic tracking analyses, bacterial cultures were imaged with a Cell Observer microscope (Zeiss) equipped with 100× objective (α-Plan-Apochromat, NA 1.46) and CoolSNAP camera at about 8 frames per second (fps) over 1 min.For tracking of intracellular SPA, infected cells were imaged with a Cell Observer microscope (Zeiss) equipped with a Yokogawa Spinning Disc Unit CSU X1a5000, an incubation chamber, 63× objective (α-Plan-Apochromat, NA 1.46), two ORCA Flash 4.0 V3 cameras (Hamamatsu) and appropriate filters for the respective fluorescence proteins at about 18 fps over 1 min.Tracking analyses were performed in Fiji with TrackMate v5.0.2 plugin (Tinevez HeLa cells stably expressing LC3B-GFP were generated using 3rd generation lentiviral vectors (The RNAi Consortium, 2015) and Bacterial strains used in this study.Plasmids used in this study.