CRISPR/Cas9 screen reveals host cell determinants of macrophage infection

To begin investigating phagocytosis of non-opsonized Lm, we first assessed uptake of various Listeria species using a gentamicin protection assay. Primary murine bone marrow-derived macrophages (BMDMs) were infected, followed immediately by short centrifugation to control for any motility differences between species. Gentamicin was added 30 minutes post-infection to kill extracellular bacteria and intracellular bacteria were enumerated 1 hour post-infection. Lm was internalized more efficiently than either L. ivanovii or L. innocua, with L. innocua being taken up only ~15% as efficiently as Lm (Fig 1A). These results suggest distinct mechanisms of phagocytosis between Listeria species under non-opsonizing conditions, although the bacterial and host factors involved in this differential uptake are unknown.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Genome-wide CRISPR/Cas9 screen identifies novel determinants of Lm macrophage infection.

(A) Gentamicin protection assay measuring bacterial uptake by BMDMs 1 hour post-infection. Bacteria (MOI = 1) were centrifuged onto BMDMs at 300 x g for 3 minutes prior to incubation. Data are normalized to uptake of Lm. (B) Schematic of the CRISPR/Cas9 screen in iBMMs. Three independent iBMM knockout libraries were generated using the genome-wide Brie sgRNA library. Cells were infected with mock or GFP-Lm at MOI = 10 for 4 hours. 6 hours post-infection, uninfected cells (GFP^–) were isolated by FACS. sgRNA abundances in sorted cells versus mock-infected controls were measured by Illumina sequencing and MAGeCK analysis. (C) MAGeCK score of each gene in the Brie library from three independent screen replicates. Gene names are depicted for the top 14 hits (lowest MAGeCK scores), several related genes, and specific genes identified by Metascape analysis. The size of each colored circle represents the number of sgRNAs targeting that gene which were enriched in the GFP⁻ population. (D) Validation of CRISPR/Cas9 screen. iBMMs were transduced with empty vector or sgRNAs targeting selected hits. For each gene, two independent sgRNAs were used to generate two knockout cell lines. Cell lines were infected with GFP-Lm as in (B) and analyzed by flow cytometry. Data are normalized to iBMMs transduced with empty vector. Non-significant results are depicted in gray. (E) Gentamicin protection assay measuring bacterial uptake by iBMMs treated with 10 μg/mL actinomycin D, 1 μg/mL cycloheximide, 5 μM bpV(pic), 10 μM CHIR99021, or 10 μM parthenolide. iBMMs were infected at MOI = 1 for 30 minutes and CFU were quantified 1 hour post-infection. All data displayed are the means and standard error of the mean (SEM) of at least three biological replicates. *p<0.05, ***p<0.001 as determined by unpaired t tests. p values for significant results in (D) ranged from <0.0001 to 0.025.

https://doi.org/10.1371/journal.ppat.1011058.g001

To identify macrophage factors required for phagocytosis of Lm, we performed a genome-wide CRISPR/Cas9 knockout screen in immortalized bone marrow-derived macrophages expressing Cas9 (iBMM) [22]. iBMMs exhibited similar phagocytic capacity as BMDMs (S1A Fig) and uptake of Lm was enhanced in these cells as compared to other Listeria species (S1B and S2A Figs). Three independent iBMM knockout libraries were generated using the murine Brie sgRNA library [23]. To select for iBMM mutants defective for Lm uptake, we constructed a reporter strain of Lm constitutively expressing GFP (GFP-Lm) in an actA-deficient background to eliminate intercellular spread between different mutant cells. Using standard infection conditions of a multiplicity of infection (MOI) of 1 for 1 hour resulted in only ~5% of macrophages becoming GFP⁺ (S2B Fig). To overcome this bottleneck and maintain coverage of the knockout library, each iBMM knockout library was infected with GFP-Lm using an MOI of 10 for 4 hours, which resulted in approximately 80% GFP⁺ iBMMs (S2B Fig). 6 hours post-infection, phagocytosis-deficient (i.e. GFP^–) cells were isolated using fluorescence-activated cell sorting (FACS). sgRNAs were amplified from sorted cells and mock-infected control libraries and abundance in each population was measured by Illumina sequencing (Fig 1B). Importantly, 1000-fold coverage of the library was maintained at each step, including transduction, passaging, and sorting.

Using Model-based Analysis of Genome-wide CRISPR/Cas9 Knockout (MAGeCK) [24], we identified 557 sgRNAs enriched in the GFP⁻ populations compared to the mock-infected libraries (p < 0.01), corresponding to 235 genes (Fig 1C and S1 Table). Initial examination of these candidates found several sets of genes belonging to the same protein family or protein complex: TFIID complex proteins (Taf1d, Taf2, Taf3, Taf4a, Taf5, Taf8, Taf10, and Taf13); SAGA complex proteins (Usp22, Supt7l, Supt20, Taf5l, Taf6l, Atxn7l3, Tada2b, and Tada1); negative elongation factor (Nelfa, Nelfb, Nelfcd, and Nelfe); Ras-related GTP-binding proteins (Rraga and Rragc); glycogen synthase kinase 3 (Gsk3a and Gsk3b); and translation initiation factors (Eif3h, Eif4g2, and Eif5). Identification of multiple independent genes within the same complex suggests that our genome-wide screen was robust. In addition, we performed gene ontology analysis using Metascape to identify biological pathways enriched in our data set [25]. Several enriched pathways were involved in chromatin remodeling, mRNA transcription, and mRNA processing (S2C Fig), suggesting that proper host cell transcription may be required for phagocytosis of Lm. Interestingly, we also identified signaling pathways involving major regulators of cell proliferation and survival, including PTEN (Pten), p53 (Trp53), and NF-κB (Rela) (Figs 1C and S2C). Overall, analysis of the screen provides confidence in our approach and indicates that we have identified several pathways that have not been previously implicated in Lm uptake.

To validate the screen, individual knockout cell lines of a subset of genes were generated using 2 sgRNAs per gene and genome editing was confirmed using Inference of CRISPR Edits (ICE) analysis [26]. Knockout iBMMs were infected with GFP-Lm in the same manner as the original screen, and GFP⁺ cells were quantified using flow cytometry. Of the eleven candidate genes tested, eight exhibited significantly reduced intracellular Lm with both sgRNAs and one with a single sgRNA (Fig 1D). The two candidate genes which had no change in Lm uptake with either sgRNA (Nelfcd and Taf13) had <30% CRISPR-InDel frequency, as measured by ICE analysis, or were unable to be confirmed by sequencing. Together, these data validate the power of our screen to identify novel genes involved in Lm macrophage infection.

Based on the length of infection used in the screen, we anticipated the possibility of identifying genes involved not only in phagocytosis but also in Lm vacuolar escape and cytosolic growth. In addition, we expected to identify genes required for the general process of phagocytosis along with those specific for internalization of Lm. To test uptake more directly and assess the specificity of the identified pathways, gentamicin protection assays were used to measure internalization of Lm, L. innocua, and the related Gram-positive bacterium Bacillus subtilis. As an orthologous approach to genetic knockouts, we used chemical inhibitors targeting transcription (actinomycin D), translation (cycloheximide), PTEN [bpV(pic)], GSK3 (CHIR99021), and NF-κB (parthenolide). All inhibitors tested blocked internalization of Lm during the 1-hour infection, suggesting the involvement of these pathways in uptake of Lm rather than later stages of infection (Fig 1E). Interestingly, inhibitors targeting transcription, PTEN, and GSK3 had no effect on uptake of L. innocua or B. subtilis, indicating their specificity for internalization of Lm. We also found that host cell translation and NF-κB were required for uptake of all three species (Fig 1E). Collectively, these data demonstrate the breadth of our screen in identifying regulators of phagocytosis both specific to Lm and involved more broadly in phagocytosis of bacteria.

PTEN is required for uptake of Lm by macrophages

Given that Lm is internalized by macrophages using distinct mechanisms (Fig 1A), we aimed to further characterize pathways which uniquely regulate Lm uptake. Metascape analysis identified the PTEN regulation pathway as enriched in our dataset (Figs 2A and S2C), and two members of this pathway, GSK3 and PTEN itself, were both required specifically for phagocytosis of Lm by macrophages (Fig 1E). Therefore, we investigated the involvement of PTEN in Lm uptake. PTEN (Phosphatase and Tensin homolog) is a dual-specificity phosphatase primarily known for its tumor suppressor activity [27]. PTEN blunts the PI3K signaling cascade, halting the promotion of cell survival and proliferation by this pathway [28,29]. While the role of PTEN in cancer biology has been very well-studied, relatively little is known about the role PTEN might play in the regulation of bacterial phagocytosis by macrophages.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. PTEN is a novel host factor which enhances phagocytosis of specific Lm strains.

(A) Hits from CRISPR/Cas9 screen belonging to the PTEN regulation pathway (n = 16) as identified by Metascape analysis. Dashed line indicates p-value cutoff (p<0.01). (B) Immunoblot of PTEN in WT and Pten^–/– iBMM clones. β-actin is used as a loading control. (C) Gentamicin protection assay measuring uptake of Lm by Pten^–/– iBMMs. iBMMs were infected at MOI = 1 for 30 minutes and CFU were quantified 1 hour post-infection. Two independent single-cell clones were analyzed. Data are normalized to WT iBMMs. (D) Uptake of IgG-opsonized FITC-labeled latex beads by WT and Pten^–/– iBMMs. Beads were added to cells for 30 minutes and uptake was quantified by flow cytometry. (E) Gentamicin protection assay measuring uptake of Lm by iBMMs, BMDMs, and PMA-differentiated THP-1 cells. Cells were treated with vehicle or 5 μM bpV(pic). Data are normalized to vehicle-treated cells. (F) Gentamicin protection assay measuring uptake of Lm by TIB73 and Caco-2 cells. Cells were infected at MOI = 10 (Caco-2) or MOI = 50 (TIB73) for 1 hour and CFU were quantified 90 min post-infection. Cells were treated with vehicle or 5 μM bpV(pic). Data are normalized to vehicle-treated cells. (G) Gentamicin protection assay measuring uptake of bacteria by WT and Pten^–/– iBMMs. Data are normalized to WT iBMMs for each species. (H) Gentamicin protection assay measuring uptake of Lm strains by WT and Pten^–/– iBMMs. Data are normalized to WT iBMMs for each strain. (I) Gentamicin protection assay measuring uptake of un-opsonized and antibody-opsonized Lm 10403S by WT and Pten^–/– iBMMs. Data are normalized to WT iBMMs. All data are means and SEM of at least three biological replicates. *p<0.05, **p<0.01, ***p<0.001, n.s. = non-significant, as determined by unpaired t tests.

https://doi.org/10.1371/journal.ppat.1011058.g002

To elucidate the role of PTEN in the context of Lm infection, we first derived single-cell clones from the iBMM Pten CRISPR knockout population (Fig 1D) and confirmed the absence of PTEN protein by immunoblot for two independent clones (Fig 2B). A small amount of PTEN protein at a lower molecular weight was detected for Clone 2, however sequencing analysis revealed a 45 base pair deletion in the phosphatase domain of this clone. Each cell line was infected with Lm and uptake was assessed by gentamicin protection assay. Compared to wildtype (WT) iBMMs, Pten^–/– iBMMs exhibited a dramatic reduction in Lm uptake (Fig 2C), consistent with chemical inhibition of PTEN. As both clones had similar phenotypes, iBMM Pten^–/– Clone 1 was used for the remainder of our experiments. To assess the overall phagocytic ability of Pten^–/– iBMMs, we measured phagocytosis of fluorescently-labeled latex beads, an established method for studying basal phagocytic activity [30]. IgG-opsonized FITC-labeled beads were added to WT and Pten^–/– iBMMs for 30 minutes, after which excess beads were removed. Fluorescence of any remaining extracellular beads was quenched with trypan blue and FITC fluorescence was quantified by flow cytometry. Phagocytosis of fluorescent beads was not significantly different between WT and Pten^–/– iBMMs (Fig 2D), indicating that basal phagocytic activity was not altered by loss of Pten. In addition, these results suggest that PTEN is not required for Fcγ receptor-mediated phagocytosis in iBMMs.

We next investigated the role of PTEN in Lm internalization in various cell types. iBMMs, BMDMs, and THP-1 human monocyte-derived macrophages were treated with the well-characterized potent competitive PTEN inhibitor bpV(pic) [31]. Uptake of Lm was reduced to similar levels in all three macrophage cell types tested (Fig 2E), indicating that regulation of Lm uptake by PTEN is conserved in primary murine macrophages as well as human macrophages. Lm also induces phagocytosis-like internalization into non-phagocytic cells (S1A Fig) [32]. In contrast to macrophages, TIB73 murine hepatocytes and Caco-2 human colonic epithelial cells exhibited increased Lm internalization following treatment with bpV(pic) (Fig 2F). These data demonstrate that PTEN is required for phagocytosis of Lm specifically by macrophages.

Chemical inhibition of PTEN specifically blocked uptake of Lm but not L. innocua or B. subtilis (Fig 1E). To confirm the specificity of PTEN for regulation of Lm uptake, we assessed phagocytosis of various Listeria species and B. subtilis in WT and Pten^–/– iBMMs. Consistent with our previous results, Pten^–/– iBMMs were deficient for uptake of the pathogenic species Lm and L. ivanovii, but not the non-pathogens L. innocua, L. riparia, L. seeligeri, L. newyorkensis, or B. subtilis (Fig 2G). Given that Lm and L. ivanovii are taken up more efficiently by macrophages than L. innocua, we compared uptake efficiency in WT and Pten^–/– iBMMs between these species by determining the percentage of the initial inoculum that was internalized under each condition. Uptake efficiency of Lm and L. ivanovii by Pten^–/– iBMMs was reduced to that of L. innocua in WT cells (S3 Fig). These data suggest that the enhanced macrophage uptake of Lm and L. ivanovii relative to L. innocua observed in Fig 1A is PTEN-dependent. In addition to assessing uptake of different species, we also measured phagocytosis of various Lm strains in Pten^–/– iBMMs. Similar to the lab strain 10403S used in all prior experiments, four other Lm strains exhibited reduced uptake in Pten^–/– iBMMs (Fig 2H). In contrast, uptake of three Lm strains (HER1082, HER1083, and Li2) remained unchanged in Pten^–/– iBMMs. Similar to our species-level analysis, Lm strains which required PTEN for uptake were internalized more efficiently by macrophages, and this advantage was lost in Pten^–/– iBMMs (S3 Fig). Overall, these data demonstrate that PTEN enhances uptake of specific species and Lm strains by macrophages.

We hypothesized that specific bacterial factors present in these species and strains confer the ability to be taken up by PTEN-dependent phagocytosis. Bacterial factors required for PTEN-dependent phagocytosis could be secreted factors or surface molecules that directly interact with macrophages. To test this, we opsonized Lm 10403S with Listeria antisera to block direct binding to the Lm cell surface and force uptake through Fc receptors. Opsonized Lm no longer required PTEN for uptake (Fig 2I), indicating that direct interaction between the macrophage surface and Lm cell wall is required for uptake via PTEN-dependent phagocytosis.

PTEN promotes adherence of Lm to macrophages

Given that Pten^–/– iBMMs have reduced intracellular CFU as early as 1 hour post-infection, we hypothesized that PTEN is not involved in escape from the vacuole. To directly test this, we compared uptake of wildtype Lm and a listeriolysin O-deficient strain (Δhly), which is unable to escape the phagocytic vacuole. If the defect in Pten^–/– iBMMs was due to impaired vacuolar escape, we would expect to see no difference in intracellular CFU between WT and Pten^–/– iBMMs during infection with the Δhly mutant that is trapped in the vacuole. However, the Δhly strain displayed a similar reduction in CFU in Pten^–/– iBMMs compared to wildtype Lm (Fig 3A), indicating that PTEN does not play a role in vacuolar escape of Lm. In addition, there was no difference in the intracellular growth rate of wildtype Lm between WT and Pten^–/– iBMMs (Fig 3B). Taken together, these data confirm that while PTEN promotes internalization of Lm, PTEN is not important for Lm escape from the vacuole or replication in the host cytosol.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. PTEN promotes adherence of Lm to macrophages.

(A) Gentamicin protection assay measuring uptake of Lm wildtype (wt) and Δhly strains by WT and Pten^–/– iBMMs. iBMMs were infected at MOI = 1 for 30 minutes and CFU were quantified 1 hour post-infection. Data are normalized to WT iBMMs for each strain. (B) Intracellular growth of wildtype Lm in WT and Pten^–/– iBMMs. Data are normalized to CFU at 30 minutes for each genotype. (C) Immunofluorescence microscopy of infected WT or Pten^–/– iBMMs. Cells were infected by centrifugation with mCherry-Lm at MOI = 5 and fixed 15 min post-infection. Extracellular bacteria were labeled with Listeria antisera and a fluorescent secondary antibody. Macrophage nuclei were stained with DAPI. Scale bars indicate 25 μm. (D) Quantification of adherence in (C). The total number of bacteria per 100 macrophages was counted for at least three fields of view (approximately 600 macrophages) per genotype and averaged for each experiment. (E) Quantification of Lm internalization 1 hour post-infection. The percentage of intracellular bacteria was calculated for at least three fields of view per genotype and averaged for each experiment. All data are means and SEM of at least three biological replicates. **p<0.01, ***p<0.001, n.s. = non-significant, as determined by unpaired t tests (A) or ratio paired t tests (D, E).

https://doi.org/10.1371/journal.ppat.1011058.g003

To further investigate the role of PTEN in Lm internalization, we employed an immunofluorescence strategy to quantify intracellular and extracellular bacteria over time during macrophage infection. WT and Pten^–/– iBMMs were cultured on glass coverslips and infected with mCherry-expressing Lm to visualize all adherent and intracellular bacteria. At various timepoints, coverslips were washed to remove non-adhered bacteria, fixed, and immunostained under non-permeabilizing conditions to label only extracellular adherent bacteria. Importantly, gentamicin was not used in these experiments, allowing simultaneous measurement of both adherence and internalization. At 15 minutes post-infection, Pten^–/– iBMMs had far fewer bacteria present overall (Fig 3C and 3D), indicating a defect in Lm adherence to Pten^–/– macrophages relative to WT iBMMs. To measure efficiency of internalization, we calculated the percentage of total bacteria that were intracellular for each genotype. There was no difference in the percentage of intracellular bacteria between WT and Pten^–/– macrophages (Fig 3E), suggesting that PTEN does not affect the internalization rate of Lm. We observed similar results over an entire 1 hour time course for both adherence and internalization, with the adherence defect in Pten^–/– iBMMs apparent as early as 5 minutes post-infection (S4 Fig). Collectively, our data indicate that PTEN enhances phagocytosis by promoting adherence of Lm to macrophages.

The lipid phosphatase activity of PTEN promotes Lm uptake

PTEN is a protein and lipid phosphatase which participates in major phospho-signaling cascades in mammalian cells [27]. To begin dissecting the PTEN-dependent signaling pathways involved in regulation of Lm adherence, we first investigated the enzymatic activity of PTEN required for Lm uptake. The following PTEN variants were constitutively expressed in Pten^–/– iBMMs: a CRISPR-resistant PTEN allele (PTEN^CR); PTEN^CR G129E, which retains protein phosphatase activity but is unable to dephosphorylate lipids; or PTEN^CR G129R, which lacks both protein and lipid phosphatase activity [33]. Expression of each PTEN variant was confirmed at the protein level by immunoblot (Fig 4A). Each cell line was infected with Lm and uptake was measured by gentamicin protection assay. Ectopic expression of PTEN^CR in Pten^–/– iBMMs restored uptake of Lm to that of WT iBMMs (Fig 4B). In contrast, neither the lipid phosphatase mutant (G129E) nor the enzymatically dead mutant (G129R) variants of PTEN rescued Lm uptake in Pten^–/– iBMMs, demonstrating that the lipid phosphatase activity of PTEN is required for enhancement of Lm phagocytosis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. The role of PIP₃ in regulating phagocytosis of Lm by macrophages.

(A) Immunoblot of PTEN in Pten^–/– iBMMs expressing CRISPR-resistant PTEN (PTEN^CR), PTEN^CR G129E, or PTEN^CR G129R. β-actin was used as a loading control. (B) Gentamicin protection assay measuring uptake of Lm by WT iBMMs or Pten^–/– iBMMs expressing PTEN variants. iBMMs were infected at MOI = 1 for 30 minutes and CFU were quantified 1 hour post-infection. Data are normalized to WT iBMMs. (C) Immunoblot of phosphorylated Akt (Ser473) in BMDMs in response to Lm infection. BMDMs were mock infected or infected with MOI = 100 and lysed at the indicated timepoints. Total Akt and β-actin were probed as loading controls. (D) Gentamicin protection assay measuring uptake of Lm by macrophages treated with 100 nM wortmannin. Data are normalized to vehicle-treated cells. (E) Gentamicin protection assay measuring uptake of bacteria by iBMMs in the presence of 100 nM wortmannin. Data are normalized to vehicle-treated cells. (F) Gentamicin protection assay measuring uptake of Lm by iBMMs in the presence of 1 μM INK1117, 10 μM TGX221, 10 μM IC87114, or 50 nM IPI549. Data are normalized to DMSO-treated controls. All data are means and SEM of at least three biological replicates. **p<0.01, ***p<0.001, as determined by unpaired t tests.

https://doi.org/10.1371/journal.ppat.1011058.g004

Class 1A PI3K activation is required for Lm phagocytosis by macrophages

The lipid phosphatase activity of PTEN specifically dephosphorylates the D3 position of 3-phosphorylated phosphatidylinositol species, mainly phosphatidylinositol (3,4,5)-triphosphate (PIP₃) [28]. PIP₃ is transiently produced upon phosphorylation of PI(4,5)P₂ (PIP₂) by activated PI3K and turnover of PIP₃ back to PIP₂ is mediated by PTEN via dephosphorylation [27]. We hypothesized that turnover of PIP₃ by PTEN is important for enhancement of Lm phagocytosis. To begin investigating the role of PIP₃ in Lm uptake, we first determined whether PIP₃ is produced during Lm infection of macrophages. Upon PI3K activation and subsequent PIP₃ production, Akt is recruited to the plasma membrane where it undergoes phosphorylation [34]; therefore, detection of phosphorylated Akt is a common method of measuring PI3K activation. Akt phosphorylation was measured by immunoblot analysis at 15, 30, and 60 minutes post-infection. Infection with Lm induced strong Akt phosphorylation 15 minutes post-infection, which waned at 30 minutes and returned to baseline levels by 60 minutes post-infection (Fig 4C). These data indicate that Lm rapidly and transiently activates PI3K and PIP₃ production during macrophage infection.

We next modulated PIP₃ levels during infection and measured the effect on Lm uptake. First, PIP₃ production was blocked by treating macrophages with the PI3K inhibitor wortmannin. iBMMs, BMDMs, and THP-1 cells all exhibited reduced Lm uptake when pre-treated with wortmannin (Fig 4D), indicating that PI3K activation is required for phagocytosis of Lm by macrophages. Given that PTEN was specifically required for uptake of Lm compared to other bacterial species, we also assessed the specificity of the requirement for PI3K. In contrast to PTEN, PI3K activation was required for uptake of L. ivanovii, L. innocua, and B. subtilis (Fig 4E), suggesting that PI3K activation is broadly important for phagocytosis of bacteria by macrophages.

In mammals, PI3Ks encompass a large family of enzymes consisting of three classes that phosphorylate different phosphoinositides, all of which are inhibited by wortmannin. The majority of PI3K isoforms belong to Class I, which produces PIP₃ from PI(4,5)P₂, while Class II and Class III PI3Ks produce other phosphoinositide species such as PI3P and PI(3,4)P₂ [35]. The catalytic subunits of Class I PI3Ks are further divided into Class 1A isoforms (p110α, p110β, and p110δ) and a single Class 1B isoform (p110γ). To identify whether PIP₃-producing PI3K isoforms are involved in promoting phagocytosis of Lm, we pre-treated iBMMs with isoform-specific PI3K inhibitors targeting all four Class I PI3K isoforms and measured uptake of Lm. Inhibition of p110β blocked uptake of Lm in iBMMs, while inhibition of p110α, p110δ, or p110γ had no effect (Fig 4F). Collectively, these data show that PIP₃ production by the Class 1A PI3K p110β is required for phagocytosis of Lm by macrophages.

Class 1A PI3Ks are canonically activated by phosphorylated tyrosine residues on receptor tyrosine kinases and other cell surface receptors [35]. Additionally, Class 1A PI3Ks are activated downstream of toll-like receptors (TLRs) in response to bacterial infection [36,37]. Given that Lm is known to activate TLR2 and TLR5 on the surface of macrophages [38,39], we hypothesized that Lm-induced PI3K activation in macrophages is mediated by TLRs. To test this hypothesis, we measured uptake of Lm and PI3K activation in Tlr2^–/– macrophages. For these studies, BMDMs from Tlr2^–/– mice were isolated and absence of TLR2 was confirmed by immunoblot analysis (S5A Fig). Interestingly, uptake of Lm was unchanged in Tlr2^–/– BMDMs compared to WT BMDMs (S5B Fig). While we did not directly test Tlr5^–/– macrophages, studies from our lab have found that Lm lacking the TLR5 ligand flaA are taken up by macrophages with equal efficiency as wildtype Lm [40]. To account for the possibility of redundancy between TLR2 and TLR5, we infected wildtype and Tlr2^–/– BMDMs with ΔflaA and found that uptake remained unchanged in the absence of TLR2 and TLR5 recognition (S5C Fig). These results indicate that phagocytosis of Lm is TLR-independent.

To test the role of TLRs in mediating PI3K/PTEN signaling in macrophages during Lm infection, we pre-treated WT and Tlr2^–/– BMDMs with PI3K and PTEN inhibitors prior to infection with ΔflaA. Wortmannin and bpV(pic) inhibited uptake of ΔflaA to a similar extent in both WT and Tlr2^–/– BMDMs (S5D Fig), suggesting that PI3K signaling is intact in the absence of TLR stimulation. To test this, we monitored Akt phosphorylation in WT and Tlr2^–/– BMDMs infected with wildtype Lm and ΔflaA and observed Akt phosphorylation over background in the absence of TLR2 and TLR5 recognition (S5E Fig). Together, these data demonstrate that Lm-induced PI3K activation in macrophages occurs independently of TLR2 and TLR5. Given that specific Lm strains are phagocytosed using a distinct PTEN-dependent pathway compared to other Lm strains or L. innocua, we tested whether PTEN-independent uptake of the latter was mediated by TLRs. To that end, we compared uptake of Lm 10403S, Lm Li2, and L. innocua in WT and Tlr2^–/– BMDMs and found that uptake of all three bacterial strains was TLR2-independent (S5F Fig). While we did not directly test a role for TLR5 in PTEN-independent uptake, Lm Li2 and L. innocua do not produce flagella at 37°C [41,42], strongly suggesting that these strains do not engage TLR5 in our assays. Overall, our data show that uptake of Listeria species in murine macrophages is TLR-independent.

Myeloid PTEN promotes bacterial restriction during murine foodborne listeriosis

To investigate the role of PTEN-dependent phagocytosis of Lm in vivo, we generated myeloid-specific conditional Pten knockout mice. C57BL/6 mice expressing Cre recombinase from the myeloid-specific Lyz2 promoter (LysM-Cre^+/+) were crossed with C57BL/6 mice containing loxP sites flanking exon 5 of Pten (Pten^fl/fl) to produce LysM-Cre^+/+ Pten^fl/fl myeloid knockout mice (PTEN^M-KO). LysM-Cre^+/+ Pten^wt/wt mice which express Cre but do not contain Pten loxP sites (PTEN^M-WT) were included as controls. We used a foodborne model of listeriosis to ensure passage through the intestinal lamina propria where Lm is engulfed by resident macrophages, which are hypothesized to promote dissemination of Lm to peripheral organs [3,4]. 8–10 week old PTEN^M-WT and PTEN^M-KO sex- and age-matched mice were given streptomycin in their drinking water for 48 hours and fasted for 16 hours prior to infection to increase susceptibility to oral infection [43,44]. Mice were then fed 10⁸ Lm in PBS and body weight was recorded over the course of 5 days as an overall measurement of disease severity. To assess bacterial burdens throughout the infection, mice were euthanized at 1, 3, and 5 days post-infection (dpi) and CFU were enumerated from organs. In addition, ilea and colons were fractionated to separate the mucus layer, epithelial cells, and lamina propria compartments and CFU were enumerated from each fraction.

At 1 dpi, we observed significantly increased bacterial burdens in the epithelial cells and lamina propria of the ileum in PTEN^M-KO mice compared to PTEN^M-WT mice (Fig 5A). There was no difference in CFU in the ileum mucus layer (Fig 5A), suggesting that higher bacterial burdens in the epithelial cells and lamina propria were not due to overall increased colonization of the ileal lumen. In contrast to the lamina propria, there were no differences in CFU between PTEN^M-WT and PTEN^M-KO mice in the mesenteric lymph nodes (MLN), spleen, or liver 1 dpi (Fig 6D–6F), indicating that myeloid PTEN is dispensable for initial dissemination of Lm to peripheral organs. We also observed no difference in CFU in the feces or cecum between PTEN^M-WT and PTEN^M-KO mice throughout the experiment (Fig 6B and 6C), demonstrating that PTEN does not affect colonization of the GI tract or fecal shedding. Together, these data show that myeloid expression of PTEN restricts Lm replication in the epithelial cells and lamina propria of the ileum early during infection, however increased bacterial burdens in these compartments does not affect dissemination.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Myeloid PTEN restricts bacterial replication in the intestinal lamina propria.

Mice were orally infected with 10⁸ Lm 10403S or 5 x 10⁸ Lm Li2. Intestinal fractions were isolated from ilea or colons 1 or 5 dpi (MUC = mucus layer, EC = epithelial cells, LP = lamina propria). (A) Ileal fractions and (B) colon fractions from 10403S-infected mice. (C) Colon fractions from Li2-infected mice. Each data point represents a single mouse (1 dpi, n = 7 per genotype; 5 dpi, n = 10 per genotype for 10403S or n = 10 per genotype for Li2). Data are combined from two independent experiments per timepoint. Solid lines indicate geometric means. Dashed lines indicate the limit of detection (l.o.d.). *p<0.05, **p<0.01 as determined by unpaired t tests of natural log-transformed values.

https://doi.org/10.1371/journal.ppat.1011058.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Myeloid PTEN promotes bacterial clearance during murine foodborne listeriosis.

(A) Body weights of infected mice over time, reported as the percent of initial weight before streptomycin treatment. Data are means and SEM of n = 10 mice at each timepoint. (B-F) Mice were orally infected with 10⁸ Lm 10403S and CFU from tissues were enumerated at 1, 3, and 5 dpi. Each data point represents a single mouse (1 and 3 dpi, n = 7 per genotype; 5 dpi, n = 10 per genotype). Data for each time point represents two independent experiments. Solid lines indicate geometric means. Dashed lines represent the l.o.d.. (G) Mice were orally infected with 5 x 10⁸ Lm Li2 and CFU from tissues were enumerated 5 dpi. Each data point represents a single mouse from two independent experiments (n = 10 per genotype). Solid lines indicate geometric means. *p<0.05, **p<0.01 as determined by unpaired t tests of untransformed (A) or natural log-transformed (B-G) values.

https://doi.org/10.1371/journal.ppat.1011058.g006

There were no differences in weight loss or bacterial burdens in any organs tested at 3 dpi (Figs 6A–6F and S6). By 5 dpi, however, PTEN^M-WT mice lost ~7% of their initial body weight while PTEN^M-KO mice lost nearly 18% of their initial weight (Fig 6A), indicating that PTEN^M-KO mice experienced more severe disease during the later stages of infection. Consistent with increased weight loss, PTEN^M-KO mice displayed significantly increased bacterial burdens in the MLN (~1-log) and spleen (~2-logs) compared to PTEN^M-WT mice at 5 dpi (Fig 6E and 6F). While not statistically significant, CFU were also elevated in the livers of PTEN^M-KO mice compared to PTEN^M-WT mice at 5 dpi (~1 log; p = 0.07) (Fig 6D). In addition, the lamina propria of both the ileum and colon as well as the ileal mucus layer and colonic epithelium of PTEN^M-KO mice also contained higher Lm burdens compared to PTEN^M-WT mice at 5 dpi (Fig 5A and 5B), suggesting that PTEN^M-KO mice contained higher bacterial burdens overall in these organs. Taken together, these data demonstrate that myeloid PTEN limits disease severity and restricts Lm growth during later stages of foodborne listeriosis.

To specifically assess the contribution of PTEN-regulated phagocytosis to controlling Lm infection, we infected mice with the Lm strain Li2, which is phagocytosed using a PTEN-independent mechanism (Fig 2H). We reasoned that any global effects of myeloid PTEN loss would impact both strains equally, while phagocytosis-specific phenotypes would not affect Lm Li2 infection. In contrast to mice infected with Lm 10403S, there were no significant differences in CFU in the MLN, spleen, or colon fractions between PTEN^M-WT and PTEN^M-KO mice infected with Lm Li2 at 5 dpi (Figs 5C and 6G). These data indicate that restriction of Lm 10403S by myeloid PTEN in these organs is largely driven by PTEN-dependent phagocytosis. However, we observed increased bacterial burdens in livers of PTEN^M-KO mice infected with Lm Li2 (S7 Fig), suggesting that PTEN limits CFU in this organ independently of its regulation of phagocytosis.

To determine whether passage through the gastrointestinal tract is necessary for restriction of Lm by myeloid PTEN in peripheral organs, we evaluated the role of myeloid PTEN during i.v. infection. At 2 dpi, PTEN^M-KO mice infected through the i.v. route exhibited increased bacterial burdens in the spleen, but not the liver, compared to PTEN^M-WT mice (S7D and S7E Fig). These data suggest that control of Lm by myeloid PTEN in the spleen is independent of dissemination from the gastrointestinal tract, and provide further evidence that PTEN-dependent phagocytosis contributes to anti-Lm immunity during the systemic phase of infection.

Ethics statement

This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All protocols were reviewed and approved by the Animal Care and Use Committee at the University of Washington (Protocol 4410–01).

Materials & data availability

Further information and requests for resources and reagents, including newly generated bacterial strains, cell lines, and plasmids, should be directed to and will be fulfilled by the lead contact, Michelle Reniere (reniere@uw.edu). Microscopy data reported in this paper will be shared by the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Quantification and statistical analysis

Statistical analyses were performed using GraphPad Prism 9 and Microsoft Excel. Details on the statistical tests used, number of replicates, and definition of center and error are found in the figure legends. Results with p<0.05 are considered statistically significant. Statistical outliers were not omitted in this study.

Bacterial strains and culture conditions

All bacterial strains used in this study are summarized in S2 Table. Bacteria were cultured on brain heart infusion (BHI) agar (Listeria spp.) or LB agar (B. subtilis, E. coli) at 37°C. Unless otherwise stated, bacteria were grown in BHI broth overnight at 37°C with shaking prior to experiments. Antibiotics were used at the following concentrations: streptomycin, 200 μg/mL; chloramphenicol, 10 μg/mL (E. coli) and 7.5 μg/mL (Lm); tetracycline, 2 μg/mL; and carbenicillin, 100 μg/mL. Plasmids were transformed into chemically competent E. coli using heat shock.

Mammalian cell lines

Cas9-expressing immortalized bone marrow-derived macrophages (iBMMs) were described previously [22]. THP-1, TIB73, and Caco-2 cell lines were obtained from Dr. Joshua Woodward (University of Washington) [70]. HEK293T cells were purchased from ATCC and obtained from Dr. Jennifer Hyde (University of Washington). Cell lines were cultured at 37°C and 5% CO₂. iBMMs, TIB73, and HEK293T cells were grown in high glucose (4.5 g/L) Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Cytiva), 1 mM sodium pyruvate, 2 mM L-glutamine, and 100 U/mL penicillin-streptomycin (D-10). Caco-2 cells were cultured high glucose DMEM supplemented with 20% FBS, 1 mM sodium pyruvate, 2 mM L-glutamine, and 100 U/mL penicillin-streptomycin (D-20). THP-1 cells were grown in RPMI 1640 supplemented with 10% FBS, 1 mM sodium pyruvate, 2 mM L-glutamine, and 100 U/mL penicillin-streptomycin (RP-10). Cells were plated in medium without penicillin-streptomycin for infections. Cells were determined to be Mycoplasma-free using PCR [71]. Antibiotics were used at the following concentrations: gentamicin, 50 μg/mL; puromycin, 5 μg/mL; blasticidin, 10 μg/mL; and Geneticin (G418), 1000 μg/mL.

Primary cells

Bone marrow-derived macrophages (BMDMs) were cultured as above and grown in high glucose DMEM supplemented with 20% heat-inactivated FBS, 1 mM sodium pyruvate, 2 mM L-glutamine, 10% 3T3 cell supernatant (from M-CSF-producing 3T3 cells), and 55 μM β-mercaptoethanol (BMDM medium). BMDMs were isolated as previously described [72]. Briefly, femurs and tibias from 6–8 week old female C57BL/6 mice were crushed with a mortar and pestle in 20 mL BMDM medium and strained through 70 μm cell strainers. Cells were plated in 150 mm untreated culture dishes, supplemented with fresh BMDM medium at day 3, and harvested by pipetting in cold PBS at day 7. BMDMs were aliquoted in 80% BMDM medium, 10% FBS, and 10% DMSO and stored in liquid nitrogen.

Mice

Animal experiments were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All protocols were reviewed and approved by the Institutional Animal Care and Use Committee at the University of Washington (Protocol 4410–01). Animals were group-housed in University of Washington Department of Comparative Medicine facilities under specific pathogen-free conditions. 6 week old wildtype (#000664) and Tlr2^–/– (#004650) female C57BL/6J mice were purchased from The Jackson Laboratory for isolating BMDMs. To generate Pten conditional knockout mice, male LysM-Cre^+/+ mice (#004781) and female Pten^fl/fl mice (#006440) were purchased from The Jackson Laboratory and bred to produce double heterozygotes. Female heterozygotes were bred back to male LysM-Cre^+/+ mice to generate LysM-Cre^+/+ Pten^fl/wt breeders, which were then bred together to produce LysM-Cre^+/+ Pten^fl/fl knockout mice (PTEN^M-KO) and LysM-Cre^+/+ Pten^wt/wt controls (PTEN^M-WT). Genotyping was performed by isolating gDNA from ear biopsies and following PCR protocols provided by The Jackson Laboratory. 8–10 week old male and female mice were used for infection studies. Animals were assigned to experimental groups by age-, sex-, and littermate-matching and groups were co-housed when possible.

Bacterial strain construction

For the CRISPR/Cas9 screen, we constructed a ΔactA strain expressing GFP constitutively and RFP under control of the actA promoter. RFP expression was used for experiments not described in this manuscript. GFP driven by the constitutive pHyper promoter was cloned into the integrative vector pPL1 [73] by restriction cloning (pPL1.pHyper-GFP). RFP was cloned downstream of the actA promoter into the integrative plasmid pPL2t [74] by restriction cloning (pPL2t.pActA-RFP). To integrate pPL1.pHyper-GFP into ΔactA (DPL3078) [75], this strain was cured of its endogenous prophage at the comK-attBB’ site as previously described [73]. ΔactA was grown overnight in BHI broth at 37°C with shaking. The next day, 8.5 x 10⁸ U153 phage was combined with 4.25 x 10⁷ bacteria (MOI = 20) in a total of 3 mL BHI containing 5 mM CaCl₂ and incubated at 37°C with shaking for 75 minutes. The mixture was then diluted 1:100 and 1:10,000 into 2 mL BHI in separate culture tubes. Each tube was incubated at 37°C with shaking for several hours until the 1:100 dilution reached an OD of approximately 0.1. At this point, the 1:10,000 dilution was diluted and spread plated to obtain single colonies. Colonies were screened by PCR with primers specific for comK-attBB’ for presence of a 417 bp band to indicate successful phage curing. pPL1.pHyper-GFP and pPL2t.pActA-RFP were integrated into phage-cured ΔactA via trans-conjugation from E. coli SM10, generating the strain designated as GFP-Lm.

Measuring bacterial uptake via gentamicin protection assay

Cells were plated the day before infection at the following densities: 5 x 10⁵ (iBMM and THP-1), 6 x 10⁵ (BMDM), and 2 x 10⁵ (Caco-2) per well in 24-well plates, or 5 x 10⁵ (TIB73) per well in 12-well plates. THP-1 cells were plated in the presence of 100 ng/mL phorbol 12-myristate 13-acetate (PMA) 48 hours prior to infection to differentiate cells. Overnight bacterial cultures were diluted in BHI to an OD of 0.05–0.1 and grown at 37°C with shaking to an OD of 1–2. Bacteria were washed twice with PBS, diluted in cell culture medium, and added to cell monolayers at MOI = 1 (iBMM, BMDM, THP-1), 10 (Caco-2), or 50 (TIB73). Medium containing 0.1% FBS was used for TIB73 infections and replaced with D-10 for subsequent steps. For experiments comparing different bacterial strains or species, the inoculation medium was serially diluted and plated to enumerate inocula. When indicated, 24-well plates were centrifuged for 3 minutes at 300 x g after adding bacteria to cells. Macrophages were incubated with bacteria for 30 minutes, washed twice with PBS, and incubated with medium containing gentamicin for an additional 30 minutes. Nonphagocytic cells were incubated with bacteria for 1 hour, washed twice with pre-warmed D-10, and incubated with medium containing gentamicin for an additional 30 minutes. To enumerate intracellular bacteria, cells were washed twice and lysed in 250 μL cold 0.1% Triton-X in PBS. Lysates were serially diluted and plated for CFU.

Lentiviral transduction

HEK293T cells were transfected with pVSV-G (Addgene #138479), psPAX2 (Addgene #12260), and a lentiviral vector at a 2:2:1 molar ratio. The following day, culture media was replaced with fresh D-10 supplemented with non-essential amino acids (Cytiva) and 25 mM HEPES. Lentivirus-containing supernatant was collected 48 hours post-transfection, spun at 300 x g for 8 minutes to remove cell debris, aliquoted, and frozen at -80°C. For transductions, cells were plated with lentiviral supernatants in the presence of 8 μg/mL polybrene, and antibiotics were added 48 hours post-transduction. Transduced cells were passaged in the presence of antibiotics for at least 1 week to allow complete selection.

CRISPR/Cas9 screen

The mouse Brie pooled sgRNA library on the lentiGuide-Puro backbone was obtained from David Root and John Doench (Addgene #73633) [23]. The Brie library contains 78,637 sgRNAs, including 1000 non-targeting controls, and targets 19,674 genes using 4 sgRNAs per gene. The sgRNA library was packaged into lentivirus and transduced into iBMM-Cas9 cells as described above at MOI = 0.3 to ensure single integration events. Transduced cells were selected with puromycin for 1 week and maintained at 1000-fold coverage of the library. Three independent knockout libraries were generated and screened.

Knockout libraries were plated at 1000-fold coverage in 150 mm tissue culture dishes. The GFP-Lm strain was grown overnight in BHI broth at 30°C stationary, washed twice with sterile PBS, and used to infect the knockout library at MOI = 10; sterile PBS was added to mock-infected control cells. At 4 hours post-infection cells were washed twice with PBS, and medium containing gentamicin was added to kill extracellular bacteria. At 6 hours post-infection cells were washed twice with PBS, harvested with 0.05% trypsin-EDTA, and fixed with 1% formaldehyde in PBS for 10 min. Fixed cells were washed twice with Flow Buffer (PBS with 5% FBS) and stored at 4°C overnight. The following day, fluorescence-activated cell sorting (FACS) was performed using a BD FACSAria II cell sorter. To isolate uninfected cells, gates were drawn based on mock-infected cells (negative control) and wildtype cells infected with ΔactA pPL1.pHyper-GFP pPL2t.pHyper-RFP (positive control). Double negative cells (GFP^- RFP^-) were collected from the infected knockout libraries. Genomic DNA (gDNA) was isolated from mock-infected and sorted populations from three biological replicates as previously described [76]. The sgRNA locus was amplified and barcoded by PCR using KAPA HiFi polymerase and Ultramer DNA Oligos from IDT (S3 Table). Enough gDNA was amplified from mock-infected libraries to maintain 1000-fold coverage of the library, using an estimation of 6.6 μg gDNA per 10⁶ cells; for sorted populations, all gDNA was amplified. PCR mixes were split into multiple 50 μL reactions, each containing 5 μg gDNA, and cycled 23 times according to manufacturer recommendations. Amplicons were purified using AMPure XP beads, pooled at 1:1 molar ratios, and sequenced on an Illumina HiSeq2500. sgRNAs enriched in the GFP-negative population were identified using MAGeCK as previously described [24]. Gene ontology analysis was performed using Metascape [25].

Construction of knockout cell lines and CRISPR screen validation by flow cytometry

Individual sgRNAs (2 per gene) from the Brie library or designed using Benchling (S3 Table) were cloned into lentiGuide-Puro (Addgene #52963), lentiCRISPR v2-blast (Addgene #98293), or sgOpti (Addgene #85681) as previously described [22,77–79]. Briefly, sgRNAs were annealed, phosphorylated with T4 Polynucleotide Kinase, digested with BsmBI, and ligated into lentiGuide-Puro, lentiCRISPR v2-blast, or sgOpti. Plasmids were packaged into lentivirus and transduced into iBMM-Cas9 cells as described above. To confirm genome-editing, gDNA was isolated using Zymo Quick-DNA Miniprep kits, cut sites were amplified by PCR, and Sanger sequencing performed. Sequences were analyzed using ICE analysis [26] to estimate InDel and knockout efficiency. Clonal cell lines were derived from Pten knockout iBMMs (sgRNA1) for further analysis by single-cell sorting.

Knockout cell lines were plated at a density of 10⁶ cells/well in 12-well plates and infected with GFP-Lm at MOI = 10. After 4 hours, cells were washed twice with PBS and medium containing gentamicin was added for an additional 2 hours. Cells were then washed twice with PBS, harvested with 0.05% trypsin-EDTA, and fixed with 1% formaldehyde in Flow Buffer for 10 minutes. Samples were washed twice with Flow Buffer and stored in Flow Buffer at 4°C overnight. The next day, samples were analyzed on a BD LSRII flow cytometer. GFP fluorescence was measured and quantified by setting a gate for GFP-positive cells on uninfected cells with 5% background.

Chemical inhibitors

All chemical inhibitors were added to cells 1 hour prior to infection and kept in the culture medium throughout the experiment. To avoid inhibition of bacterial transcription by actinomycin D, cells pre-treated with actinomycin D were washed twice with PBS immediately prior to infection and medium without inhibitor was used for the remainder of the experiment. Inhibitors had no effect on Lm viability, as measured by exposing Lm to each inhibitor for 1 hour in D-10 and enumerating CFU. All inhibitors were dissolved in DMSO except bpV(pic) which was dissolved in water. Inhibitors were used at the following concentrations: actinomycin D (Sigma-Aldrich), 10 μg/mL; cycloheximide (Sigma-Aldrich), 1 μg/mL; bpV(pic) (Cayman Chemical Company), 5 μM; CHIR99021 (Cayman Chemical Company), 10 μM; parthenolide (Sigma-Aldrich), 10 μM; wortmannin (Santa Cruz Biotechnology), 100 nM; INK1117 (GlpBio), 1 μM; TGX221 (Cayman Chemical Company), 10 μM; IC87114 (Cayman Chemical Company), 10 μM; and IPI549 (MedChemExpress), 50 nM.

Phagocytosis of fluorescent beads

Phagocytosis of fluorescent beads was measured using IgG FITC Phagocytosis Assay Kits (Cayman Chemical Company) according to manufacturer instructions. Briefly, FITC-labeled, IgG-opsonized beads were diluted 1:250 in cell culture medium and incubated with iBMMs for 30 minutes. Cells were washed once in assay buffer and incubated in 0.04% Trypan Blue in assay buffer for 2 minutes to quench fluorescence from extracellular beads. Cells were washed once more with assay buffer, harvested using 0.05% trypsin-EDTA, and resuspended in Flow Buffer. Samples were analyzed on a BD LSRII flow cytometer to quantify FITC-positive cells.

Antibody opsonization of Lm

Overnight cultures of Lm 10403S were diluted to an OD of 0.05 in BHI and grown at 37°C shaking until reaching an OD of ~2. Bacteria were diluted to an OD of 1 in PBS, washed once with antibody buffer (PBS +5% FBS) and resuspended in either antibody buffer (un-opsonized) or Listeria O Antisera (BD) diluted 1:10 in antibody buffer (opsonized). Samples were incubated at room temperature for 30 minutes, washed twice with antibody buffer, and added to iBMMs at MOI = 1. Inocula were serially diluted and plated on BHI agar to enumerate CFU. Bacterial internalization was measured by gentamicin protection assay as described above.

Immunoblotting

For PTEN immunoblots, 10⁶ iBMMs per well were plated in 12-well plates. The following day, cell lysates were prepared as described below. For pAkt immunoblots, 1.2 x 10⁶ BMDMs per well were plated in 12-well plates. The following day, overnight bacterial cultures were diluted to an OD of 0.05 in BHI and grown at 37°C with shaking to an OD of 1–2. To eliminate background Akt phosphorylation, BMDMs were serum-starved by incubating in FBS- and CSF-free BMDM medium for 2 hours prior to infection. Bacteria were washed twice with PBS, diluted in FBS- and CSF-free BMDM medium, and added to cell monolayers at MOI = 100. At each timepoint, cells were washed twice in PBS and lysed in 50–100 μL cold RIPA buffer containing Halt Protease and Phosphatase Inhibitor Cocktail (Thermo Scientific). Lysates were incubated on ice for 15 minutes and centrifuged at 15,000 x g for 15 minutes at 4°C to pellet cell debris. Protein concentration of clarified lysates was quantified using a Pierce BCA Protein Assay Kit (Thermo Scientific). 30 μg of each sample were separated by SDS-PAGE and transferred to PVD-F membranes. Membranes were blocked for 1 hour at room temperature in Intercept (TBS) Blocking Buffer (LI-COR) for pAkt blots or Intercept (PBS) Blocking Buffer (LI-COR) for PTEN immunoblots. Blocked membranes were incubated overnight at 4°C with anti-Phospho-Akt (Ser473) antibody (Cell Signaling Technology #9271) or anti-PTEN antibody (Cell Signaling Technology #9559) diluted 1:1,000 in TBS antibody buffer (50% TBS, 50% TBS Blocking Buffer) or PBS antibody buffer (50% PBS, 50% PBS Blocking Buffer), respectively. The following day, membranes were washed with TBST and incubated for 1 hour at room temperature with anti-β-Actin antibody (Cell Signaling Technology #3700) diluted 1:1,000 in PBS antibody buffer. Membranes were washed with TBST and incubated with goat anti-rabbit Alexa Fluor 680 (Invitrogen) and goat anti-mouse IRDye 800CW (LI-COR) antibodies diluted 1:5,000 in PBS antibody buffer for 1 hour at room temperature. Blotted membranes were washed a final time with TBST and imaged on a LI-COR Odyssey Fc imager. For pAkt blots, membranes were stripped by incubation at 55°C in stripping buffer (62.5 mM Tris pH 6.8, 2% SDS, 0.8% β-mercaptoethanol) for 45 minutes followed by 5–10 washes in water. Membranes were reblocked with PBS Blocking Buffer and reprobed overnight at 4°C with anti-Akt antibody (Cell Signaling Technology #9272) diluted 1:1,000 in PBS antibody buffer, followed by goat anti-rabbit Alexa Fluor 680 for 1 hour at room temperature.

Intracellular growth curves

5 x 10⁵ iBMMs per well were plated in 24-well plates overnight. Lm cultures were grown overnight at 30°C stationary. The next day, overnight cultures were washed twice with PBS, diluted in cell culture medium, and added to cell monolayers at MOI = 1. After 30 minutes, cells were washed twice and medium containing gentamicin was added to kill extracellular bacteria. At various timepoints (0.5, 2, 5, and 8 hours post-infection), cells were washed twice with PBS and lysed in 250 μL cold 0.1% Triton-X in PBS. Lysates were serially diluted and plated to enumerate intracellular CFU.

Immunofluorescence microscopy

iBMMs were plated onto 12 mm sterile glass coverslips in 24-well plates at a density of 5 x 10⁵ cells per well. Lm constitutively expressing mCherry (pPL2.pHyper-mCherry) [80] was grown overnight in BHI at 37°C shaking. The following day, overnight bacterial cultures were diluted in BHI to an OD of 0.05 and grown at 37°C with shaking to an OD of 1–2. Bacteria were washed twice with PBS, diluted in cell culture medium, and added to cell monolayers at MOI = 5. Plates were centrifuged for 3 minutes at 300 x g to synchronize infection. At each timepoint, coverslips were washed twice with PBS, removed from the plate, and inverted onto 20 μL of 4% formaldehyde on parafilm. Cells were fixed for 10 minutes and washed by dipping into three 50 mL conicals of PBS, 10 times each. Coverslips were washed in this manner in between each of the following incubations. Fixed coverslips were blocked with 1% BSA in PBS (staining buffer) for 10 minutes and stored in staining buffer overnight at 4°C. To block non-specific binding to Fc receptors, coverslips were incubated with mouse Fc Block (BD) diluted 1:50 in staining buffer for 30 minutes. For immunofluorescence staining, coverslips were incubated for 30 minutes with Listeria O Antisera (BD) diluted 1:100 in staining buffer, followed by goat anti-rabbit Alexa Fluor 488 secondary antibody (Invitrogen) diluted 1:200 in staining buffer. Stained coverslips were mounted on glass slides with ProLong Diamond Antifade Mountant with DAPI (Invitrogen) and imaged on a Keyence BZ-X710 fluorescence microscope at 40X magnification.

Expression of PTEN variants in iBMMs

cDNA was prepared from wildtype iBMMs by extracting total RNA using TRIzol (Invitrogen) and reverse transcription with an iScript cDNA Synthesis Kit (Bio-Rad). PTEN coding sequence was PCR amplified from cDNA and cloned into pLV-EF1a-IRES-Neo (Addgene #85139) [81] using restriction cloning. To prevent CRISPR/Cas9-mediated genome editing in Pten^–/– iBMMs which constitutively express Cas9 and Pten-specific sgRNA, synonymous point mutations were made in each amino acid within the sgRNA recognition site using site-directed mutagenesis to generate a CRISPR-resistant Pten allele (PTEN-CR). PTEN-CR was further mutagenized using site-directed mutagenesis to create G129E or G129R point mutations. Empty pLV-EF1a-IRES-Neo, pLV.PTEN-CR, pLV.PTEN-CR-G129E, or pLV.PTEN-CR-G129R were packaged into lentivirus and transduced into Pten^–/– iBMMs or wildtype iBMMs (empty vector only) as described above. After antibiotic selection, protein expression of each PTEN variant was confirmed by immunoblot as described above.

Mouse infections

Oral infections were performed as previously described [44,82]. 8–10 week old male and female mice were used for infections. 5 mg/mL streptomycin was added to drinking water for 48 hours to eliminate gut flora. Food and water were then removed 16 hours prior to infection to initiate overnight fasting. Lm was grown overnight in BHI at 30°C statically. The following day, cultures were diluted 1:10 into 5 mL fresh BHI and grown at 37°C with shaking for 2 hours. Bacteria were washed twice with PBS and 10⁸ CFU in 20 μL was fed to mice via pipette. We found that Lm Li2 did not colonize mice as well as 10403S at the same infectious dose, which may be due to the presence of residual streptomycin in the mice at the time of infection [43]. To overcome this, we used a higher infectious dose of 5 x 10⁸, which resulted in bacterial burdens similar to 10403S at 5 dpi. Food and water were returned immediately after infection. For i.v. infections, Lm was grown to mid-log as described above and diluted in PBS. Mice were anesthetized with isoflurane and 1.5 x 10⁵ CFU in 200 μL were administered via retro-orbital injection. Prepared bacterial suspensions were serially diluted and plated immediately after infection to enumerate inocula. Mice were humanely euthanized at 1-5-days post-infection and tissues were collected. Tissues were homogenized in 0.1% NP-40 in the following volumes: MLN, 3 mL; cecum (contents removed and tissue rinsed with PBS), 4 mL; spleen, 3 mL; and liver, 5 mL. Feces were homogenized in 1 mL 0.1% NP-40 with a sterile stick. Gallbladders were ruptured in 500 μL 0.1% NP-40 with a sterile stick. Samples were serially diluted and plated to enumerate CFU in each organ.

Intestinal tissue fractionation

Fractionation of intestinal tissues was performed as previously described [3,83]. Luminal contents of colons and ilea (distal 12 cm of small intestine) were removed by gently squeezing tissues with forceps. Tissues were flushed with 10 mL PBS using a 25 G needle and stored in 15 mL conicals in 5 mL PBS at 4°C overnight. The following day, tissues were cut longitudinally, incubated in 3 mL 6 mM N-acetylcysteine (NAC) for 2 minutes, and shaken vigorously to remove the mucus layer. The NAC wash was repeated twice more for 3 total washes, transferring tissues to fresh tubes each time. To remove epithelial cells, tissues were then cut into small pieces and added to 15 mL conicals with 5 mL RPMI containing 5% FBS, 5 mM EDTA, and 1 mM DTT. Samples were incubated at 37°C with shaking for 20 min, vortexed gently, and filtered through a 70 μm cell strainer. This process was repeated twice more for 3 total EDTA/DTT washes. Tissue pieces were washed once with 10 mL PBS to remove DTT. To isolate lamina propria (LP) cells, tissue pieces were then added to 15 mL conicals with 3 mL RPMI containing 5% FBS, 1 mg/mL collagenase type IV (Sigma-Aldrich), and 40 μg/mL DNase I (Roche). Samples were incubated at 37°C with shaking for 40 min, vortexed, and filtered through 70 μm cell strainers. This process was repeated once more if necessary, until no more large tissue pieces remained. Fractions (NAC washes, EDTA/DTT washes, and LP suspensions) were pooled, centrifuged at 12,000 x g for 20 minutes at 4°C, and pellets resuspended in 150 μL 0.1% NP-40. Samples were serially diluted and plated to enumerate CFU in each fraction.