The L. longbeachae capsule cluster is unique among Legionella species

We first analyzed the presence of the putative capsule cluster and orthologous genes among 58 different Legionella species (Fig 1). This analysis shows that L. longbeachae is the only Legionella species to encode the entire 48 kb capsule cluster. Two other species, L. massiliensis and L. gormanii, carry genes similar to the ABC-type transporter genes ctrBCD and bexD. However, these two species lack most of the glycosyltransferases encoded by L. longbeachae. We did not find similar genes in the genomes of L. pneumophila, except for an orthologous gene of llo3163, a hydroxyacid dehydrogenase similar to serA, indicating that the capsule cluster is specific for L. longbeachae (S1 Table).

Download:

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

Fig 1. Presence or absence of orthologous genes coding for the capsule in 80 Legionella species/strains.

Colored squares represent orthologous genes. Orthologous gene reconstruction was done with PanOCT: amino acid percentage identity cutoff was 30%, BLAST e-value cutoff 10⁻⁵, and minimum percentage match length of subject and query was 65%. For reference, the L. longbeachae NSW150 capsule cluster genes and gene names are shown on the top of the Fig. Color code: blue, pathway genes for nucleotide sugar precursor synthesis; yellow, putative transportation genes; green, putative glycosyltransferases; grey, genes of unknown function.

https://doi.org/10.1371/journal.ppat.1012534.g001

To deepen these analyses, we selected twelve publicly available L. longbeachae strains, comprising nine strains from serogroups 1 (sg1) and three strains from 2 (sg2). All of them contain a similar capsule cluster whose genomic organization and genomic location are conserved (S1 Fig). Only a small region in the central part of the capsule cluster shows few differences among the glycosyltransferases and a duplication of galE in the NSW150 and B41211CHC genomes (Figs 1 and S1). All twelve strains encode genes for the predicted ABC type transporter ctrBCD and bexD (ctrA-like). Furthermore, they all encode glycosyltransferases and nucleotide sugar precursor genes belonging to the same groups of enzymes (S2 Table). This extremely high conservation in gene content, sequence similarity, and in the genomic location suggest that the acquisition of the capsule cluster dated back to a common ancestor of L. longbeachae.

The L. longbeachae capsule cluster genes share homology to soil-dwelling bacteria

The evolutionary history of the capsule cluster in L. longbeachae is not known. Thus, we performed BLAST analysis on all its genes using L. longbeachae strain NSW150 (sg1) as the reference. S1 Table lists the best hits obtained by BLAST search. Across the cluster, we find homologous genes from β-proteobacteria often found in soils, and of γ- and δ-proteobacteria such as the soil-dwelling Burkholderia spp., Geobacter spp., or Pseudomonas spp., but also from Nitrococcus mobilis or Alteromonas spp., which are present in marine environments. Thus, the capsule cluster may have been acquired from soil-dwelling bacteria.

Electron microcopy analyses confirm that L. longbeachae expresses a capsule that is absent from L. pneumophila

To further analyze the L. longbeachae capsule and to study its functional role, we constructed a mutant in the capsule transporter gene ctrC (llo3151) by replacing it with an apramycin cassette. We chose ctrC as a target, since a knockout mutant of a homologous gene in Campylobacter jejuni abrogated capsule expression [39]. When grown in ACES-buffered yeast extract broth (BYE) medium with or without apramycin, the ΔctrC mutant did not display any significant growth differences as compared to the L. longbeachae WT (S2 Fig). To confirm that L. longbeachae indeed expresses a capsule and that the identified gene cluster is responsible for its expression, we used transmission electron microscopy (TEM). The L. longbeachae WT and the ΔctrC mutant as well as L. pneumophila WT (negative control) were grown in BYE until the optical density OD₆₀₀ reached mid-log phase and late-log phase, referred to as exponential (E) phase and post-exponential (PE) phase, respectively. Cells were fixed and stained with cationized ferritin. TEM revealed a capsular layer surrounding the L. longbeachae WT (Fig 2A). The capsule was observed only in PE phase, indicating a growth phase dependent expression (S3A Fig). Its thickness was estimated to be between 50–100 nm, which is similar to ferritin-stained capsules observed in E. coli K30 [40]. In contrast, the ΔctrC mutant strain was devoid of such a layer (Fig 2A) and L. pneumophila WT cells were not stained by the ferritin dye (Fig 2A). Complementation of the ΔctrC mutant strain with a plasmid containing ctrC as well as the downstream genes ctrD and bexD under the control of their native promoter restored capsule expression (Fig 2B). We included the downstream genes as macrocolonies of a ΔctrC mutant strain complemented with ctrC and ctrD only barely restored the WT phenotype compared to those complemented with the longer construct (S3B Fig). However, 50% of the complemented capsule mutant strains expressed a capsular structure indicating that regulation of capsule expression may be more complex (S3C Fig). We confirmed that there was no growth defect of the complemented mutant in liquid culture in comparison to the WT or ΔctrC harboring the empty plasmid (S3D Fig). Taken together, our results confirm that, in contrast to L. pneumophila, L. longbeachae expresses a capsule and that ctrC is mainly responsible for its expression on the cell surface.

Download:

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

Fig 2. The L. longbeachae capsule is expressed in exponential growth phase and contains a highly anionic polysaccharide that cannot be stained by silver staining.

A) L. longbeachae WT, ΔctrC and L. pneumophila WT (negative control) were grown in BYE medium to post-exponential phase. Fixed cells were stained with cationized ferritin. TEM images are representative of n = 3 independent experiments. Scale bar = 200 nm. B) L. longbeachae WT or ΔctrC harboring an empty plasmid (pBCKS) or the complementation plasmid (SSM083) were fixed in post-exponential phase and processed like samples in A). Representative images of n = 3 independent experiments. Scale bar = 500 nm. C) Polysaccharides were extracted from Llo WT, ΔctrC or Lpp WT by enzymatic isolation and subjected to ultracentrifugation. SDS gels were stained with silver nitrate or Stains-all dye. Purified LPS from E. coli O111:B4 (Sigma) was used as a control for silver staining. D) Extracts from Llo WT or ΔctrC harboring the empty plasmid (pBCKS) or the complemented mutant were stained by silver nitrate or Stains-all.

https://doi.org/10.1371/journal.ppat.1012534.g002

The highly anionic L. longbeachae capsule is resistant to silver staining

To isolate the L. longbeachae capsule and to characterize its composition, we first used phenol-extraction for polysaccharide (PS) isolation followed by HPLC analyses. This approach allowed the identification of galactosamine, glucosamine, mannose and possibly quinovosamine in both the WT and the capsule mutant extracts (S4A Fig). After elution of the column, we detected high peaks in the WT sample that might correspond to phosphosugars. However, when visualizing these samples by silver or Alcian blue staining, we observed that phenol extraction seems to collapse the LPS structure, similarly to what has been seen for L. pneumophila [41]. Thus, L. longbeachae LPS neither nor CPS can be extracted by phenol (S4B Fig). Using an enzymatic isolation method [42] and silver staining revealed that L. longbeachae expresses a long O-antigen chain of different lengths between 30–70 kDa. In contrast, a different pattern was present for L. pneumophila with two shorter O-antigen fractions as shown previously, confirming that enzymatic PS isolation preserves the LPS structure in both Legionella species (Fig 2C) [43]. However, no differences between the WT and capsule mutant were observed. Therefore, we tested other dyes to visualize CPS. When using Stains-all dye, a carbocyanine dye that stains highly anionic compounds such as glucosaminoglycans [44], the enzymatically prepared L. longbeachae PS extracts showed a strong signal in the WT, but not the capsule mutant or L. pneumophila (Fig 2C). When we complemented the capsule mutant, Stains-all reveals that the band observed in WT extracts is restored in the complemented mutant as well (Fig 2D). Again, silver staining did not detect any CPS in the complemented strain (Fig 2D). To test whether capsule expression may be temperature-dependent, we grew Llo WT, ΔctrC and complemented ΔctrC at 20°C in liquid medium. CPS extracted from bacteria grown at 20°C showed that also at 20°C the capsule is expressed in PE phase only in the Llo WT and complemented ΔctrC (S4C Fig). The presence of a highly anionic PS corroborates our findings from staining with cationized ferritin for TEM and may explain our unique peak seen in HPLC. Taken together, the L. longbeachae capsule consists of highly anionic PS that cannot be stained by conventional silver staining methods.

Encapsulated L. longbeachae is more sensitive to salt stress

The capsule may protect L. longbeachae from adverse environmental conditions. Thus, we grew the bacteria in liquid culture to PE phase and treated the cells with 0.05% Tween-20 with or without added NaCl or sucrose, as described previously [45]. After treatment with Tween-20, we did not observe a growth defect neither in the WT nor in the capsule mutant. However, in comparison to L. pneumophila, L. longbeachae seems inherently more resistant to detergent, independent of the capsule (S5A Fig). Similarly, both the L. longbeachae WT and the capsule mutant grew at similar rates after treatment with 2 mM or 10 mM H₂O₂. In comparison to L. pneumophila, the two L. longbeachae strains were more tolerant to high H₂O₂ present in the medium (S5B Fig). Thus, we conclude that L. longbeachae is inherently more resistant to detergent and oxidative stress than L. pneumophila.

It has been shown that salt stress in L. pneumophila is linked to virulence [46]. Indeed, a L. pneumophila mutant in the response regulator LqsR was shown to be less virulent but more resistant to salt stress than WT bacteria [46,47]. To test the role of the capsule under salt stress, we grew the L. longbeachae WT and capsule mutant as well as L. pneumophila WT to PE phase in liquid culture and plated the bacteria on BCYE in the presence or absence of 50 mM or 100 mM NaCl, respectively (S6 Fig). At high salt concentrations, both the capsule mutant and L. pneumophila exhibit a 3-log-fold growth difference. Surprisingly, however, there is also a 2-log-fold growth difference between the L. longbeachae WT, and the capsule mutant grown in the presence of 100 mM NaCl. Hence, similar to L. pneumophila, the analysis reveals that the less virulent L. longbeachae capsule mutant is more resistant to salt stress (S6 Fig). This growth difference was not detectable at a lower concentration of 50 mM NaCl, indicating that salt concentration is an important factor for optimal growth of both Legionella species.

The capsule is expressed in vitro and during infection in a growth phase dependent manner

Legionella pneumophila is known to have distinct gene transcription profiles in vitro and during infection depending on the growth phase [48]. Thus, to learn when the genes encoding the capsule of L. longbeachae are expressed and if their expression is growth phase dependent, we performed RNAseq analyses. The L. longbeachae WT and the ΔctrC mutant strain were grown in BYE until the OD₆₀₀ reached E or PE phase. Total RNA was isolated from the bacterial cells, depleted for rRNA, and sequenced to follow their transcriptional profile in both growth phases. This showed that almost all genes of the capsule cluster in the L. longbeachae WT strain were significantly downregulated in PE as compared to E phase (Fig 3A), indicating that the expression of the capsule on the cell surface is growth phase dependent. However, the capsule is visible only in the PE phase (S3A Fig) suggesting that the building of its complex polysaccharides and their export to the cell surface takes time and thus expression of the CPS (Fig 2) is highest in later growth phases. Moreover, when comparing L. longbeachae WT and the ΔctrC mutant strain, apart from the expression of ctrC and the downstream genes of the capsule transporter operon, no significant differences in the transcription of the other capsule genes were observed neither in E phase (S7A Fig) nor in PE phase (S7B Fig, S3 Table). The transcription of the capsule in E phase and its expression in PE phase suggest that L. longbeachae synthesizes its capsule later in infection, probably to be prepared for cell attachment and a new infection of host cells, and/or for survival in the environment.

Download:

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

Fig 3. The L. longbeachae capsule is transcribed in liquid medium and in cellulo, but it is not expressed in a ΔctrC mutant.

A) RNAseq of WT transcripts (n = 4); E vs. PE (E is the reference); consider relevant genes with log2 fold change of ±2 and adjusted p value ≤ 0.05. B) THP-1 cells were infected with wild type L. longbeachae expressing constitutive mKate2 and inducible sfGFP under the control of the capsule promoter (prom_cap). A control plasmid containing constitutive mKate2 and two copies of sfGFP without the capsule promoter was included. Cells were imaged at 22 hours post infection. Scale bar = 50 μm.

https://doi.org/10.1371/journal.ppat.1012534.g003

To test this hypothesis, we analyzed the capsule expression during infection, using a dual reporter to follow the transcription of the capsule transporter operon in cellulo. This dual reporter contains a constitutively expressed red fluorescent protein (mKate2) and two copies of superfolder GFP (sfGFP), each under the control of the native promoter of the capsule transporter. As a negative control we used a plasmid containing constitutively expressed mKate2 and two copies of sfGFP without the native capsule promoter. These plasmids were transformed into L. longbeachae WT and used to infect THP-1 cells or Acanthamoeba castellanii (Figs 3B and S8, S1–S4 Videos). We followed the expression of sfGFP and mKate2 expression over time by live confocal imaging on an Opera Phenix system. We detected GFP expression starting at 12 hours that continuously increased in intracellularly replicating bacteria containing the cap_prom promoter construct, but not the negative control. This revealed that the capsule is transcribed in both eukaryotic hosts. Thus, the capsule is universally transcribed in vitro and upon infection.

The capsule is crucial for virulence in vivo in mice and in the environmental host Acanthamoeba castellanii

To learn if the L. longbeachae capsule plays a role in virulence in vivo, we infected mice with either the L. longbeachae WT or the ΔctrC mutant at different bacterial loads and followed their survival over 10–15 days. Upon infection with the WT strain with 10⁷ bacteria, all animals succumb to the infection within 5 to 6 days, similar to what has been reported previously [21]. However, all mice infected with the ΔctrC mutant survived the infection (Figs 4A and S9). To confirm that the observed phenotype is due to the missing capsule, we infected mice with the L. longbeachae ΔctrC mutant either complemented or harboring an empty plasmid. All mice infected with the mutant strain carrying the empty plasmid survived the infection. In contrast, 40% of mice infected with the complemented strain died, indicating that complementation partially restored virulence in vivo (Fig 4B).

Download:

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

Fig 4. The capsule mutant is impaired in replication and avirulent in mammalian and protozoan hosts, despite a functional Dot/Icm type IV secretion system.

A) Female C57BL/6 mice were infected with 10⁷ CFU of Llo WT or ΔctrC via the nasal route. Survival was monitored over 15 days. Each group contained 9 mice. B) Infection of female C57BL/6 mice with the capsule mutant harboring an empty plasmid (pBCKS) or the complementation plasmid (SSM083) with 10⁷ CFUs. 10 mice per group. Statistical significance was determined by Log-rank (Mantel-Cox) test. *, p≤0.05; ****, p≤0.0001. C) CFUs from lungs of infected mice were plated at 4, 48, and 96 hours post-infection. Dots represent number of animals per group. D) CFUs from spleens of infected mice were plated at 48 and 96 hours post-infection. Dots represent number of animals per group. Statistical analysis was performed by two-way ANOVA with Tukey’s post-test. E) LDH activity test in converting L-lactate + NAD⁺ to pyruvate + NADH, measured as absorbance at 490 nm. Lung lavage fluid from mice infected with Llo WT or ΔctrC at 72 hours post-infection. Data points correspond to individual mice and are shown as means ± SD. Statistical significance was tested by unpaired t-test. ns, non-significant; *, p≤0.1; **, p≤0.01; **** p≤0.0001. F) Acanthamoeba castellanii trophozoites were infected at MOI 0.1 at 20°C over 7 days. Samples were taken every 24 hours and CFUs are normalized to the input control. Data show means ± SD for n = 3 independent experiments. G) A. castellanii trophozoites were infected with Llo WT or ΔctrC harboring an empty control plasmid (pBCKS) or the complementation plasmid (SSM083) at an MOI of 1. Samples were taken every 24 hours and CFUs normalized to t = 1. Data show means ± SD for n = 3 independent experiments. H) A. castellanii trophozoites were infected at MOI 0.1 with equal amounts of Llo WT or ΔctrC. Samples were taken at 1, 8, 24, 30, 48, 96, 120, and 144 hours post-infection and normalized to the input control. Data show the competitive index of Llo ΔctrC over Llo WT for n = 3 independent experiments. I) Translocation of the known T4SS effector RomA was tested in THP-1 cells infected with Llo WT, ΔctrC, or a T4SS mutant, ΔdotB. Data represent means ± SD of n = 2 experiments. Statistical significance was tested by two-tailed t-test. J) Bacterial replication of Llo WT, ΔctrC and ΔdotB in hMDMs at 20 hours post-infection. Data show median bacteria area ± SD of infected cells of n = 3 independent experiments. Statistical significance was tested by one-way ANOVA. ns, non-significant; *, p≤0.05, **, p≤0.01, ***, p≤0.001, ****, p≤0.0001.

https://doi.org/10.1371/journal.ppat.1012534.g004

L. longbeachae WT replicates to higher numbers in the murine lungs as compared to the capsule mutant, but both seem to reach the blood stream as we measured comparable levels of CFUs in the spleen (Fig 4C and 4D). Furthermore, we measured levels of lactate dehydrogenase (LDH) from bronchoalveolar lavage fluid (BALF) of infected mice to determine whether the capsule plays a role in lung damage. Indeed, the WT induces higher levels of LDH than the ΔctrC strain (Fig 4E), indicating that the capsule is implicated in lung damage.

We then infected A. castellanii at a low MOI of 0.1 and at an environmental temperature of 20°C with L. longbeachae WT, the ΔctrC or the ΔdotB mutant strain. The latter is deficient in the T4SS and was used as a negative control. Replication of the bacteria was monitored by plating the bacteria every 24 hours over seven days. We observed a strong replication defect of the capsule mutant as compared to the WT, which becomes apparent at 72 hours post-infection (Fig 4F). As expected, the ΔdotB mutant failed to replicate in A. castellanii but seems to persist over extended periods of time indicated by the stable CFU counts towards the end of the experiment. Like in mouse infections, A. castellanii infected with the complemented strain restored replication of the bacteria to WT levels while the capsule mutant harboring the empty plasmid was impaired in replication (Fig 4G). We further tested the competitive fitness of the capsule mutant by performing a competition assay in A. castellanii as described previously [49]. This showed that the capsule mutant is completely outcompeted by the L. longbeachae WT strain between 24 to 48 hours of infection (Fig 4H), further underlining the importance of the capsule in virulence of L. longbeachae also in its environmental host.

An important question that arose from the in vivo infections was whether the strong virulence defect of the capsule mutant may be due to impaired effector secretion through the Dot/Icm T4SS in the ΔctrC strain. We thus tested effector translocation using the beta-lactamase (BlaM) secretion assay as described previously [11]. We used a BlaM-fusion to the known T4SS effector RomA [50] and measured BlaM secretion in infected THP-1 cells by flow cytometry after 2 hours of infection. We included the L. longbeachae ΔdotB mutant as negative control and L. pneumophila strain Paris as positive control. Both, the L. longbeachae WT and the ΔctrC mutant translocated RomA successfully into the host cells to similar levels, whereas the ΔdotB mutant failed to translocate the effector (Fig 4I). Thus, the virulence phenotype of the capsule mutant in vivo and in A. castellanii is not due to an impaired Dot/Icm T4SS, as this strain maintains the ability to secrete effectors upon infection.

The capsule provides a replicative advantage in primary human macrophages

To confirm that both Llo WT and ΔctrC replicate inside the LCV, we infected U2OS cells constitutively expressing Sec61β-GFP, a marker of the ER. We followed the replication of the WT, ΔctrC expressing dsRed and as negative control ΔdotB, a strain deficient in a functional T4SS, over time (S11 Fig). Both, WT and ΔctrC replicate inside the cells building large vacuoles. More importantly, both strains recruit Sec61β early in infection, like L. pneumophila for which the recruitment of ER vesicles is a hallmark of intracellular replication.

Since L. longbeachae can cause the severe pneumonia Legionnaires’ disease in humans, we also tested bacterial replication in human monocyte-derived macrophages (hMDMs). We observed that the WT replicates more efficiently in hMDMs than the capsule mutant as indicated by a larger, bacteria containing vacuole size in the WT infected cells (Fig 4J). To address the question whether the capsule plays a role in modulating cell death, we stained living, infected hMDMs for annexin V, an early marker of apoptosis. However, we did not observe significant difference in annexin V labelling of WT or ΔctrC infected cells (S12A Fig). In addition, we tested whether the capsule may influence the induction of reactive oxygen species (ROS) in hMDMs. We labelled hMDMs with CellROX live cell dye to follow ROS production in infected cells, but we did not observe a significant difference in ROS production (S12B Fig). Interestingly, when we infected human macrophage-like THP-1 cells or murine bone marrow-derived macrophages in vitro, we observed no replication defect of the capsule mutant as compared to the WT, suggesting that successful infection of mice by L. longbeachae depends on the interaction of immune cells with the capsule in vivo (S13A and S13B Fig). These data show that there is a different replicative behavior in cell lines as compared to primary human or protozoan host cells as well as in the murine host.

Taken together, these results show that the capsule is important for replication in both protozoan and mammalian hosts and that the lack of the capsule renders the bacteria avirulent in mice.

The L. longbeachae capsule delays phagocytosis in human cells

Based on data published for other bacteria, we hypothesized that the capsule of L. longbeachae could be implicated in modulating phagocytosis or attachment to eukaryotic cells. To quantify phagocytosis, we incubated the bacteria with THP-1 cells and, after washing off unbound bacteria, we added gentamycin to kill extracellular bacteria and followed the infection by plating bacteria at different timepoints. For attachment studies, we pre-treated half of the cells with 2 μM cytochalasin D (cytoD), a known inhibitor of actin polymerization, and followed the infection by plating the bacteria recovered from the cytoD- as well as the non-treated cells. The overall CFU counts were normalized to the input control. We observed that WT bacteria were phagocytosed more slowly over time than the capsule mutant (Fig 5A). In contrast, both strains attached to THP-1 cells at similar rates, likely due to the synchronization of the infection by centrifugation at the start of each experiment (Fig 5B). This suggests that the capsule is important to delay phagocytosis by host cells.

Download:

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

Fig 5. The capsule delays phagocytosis of host cells.

A) For phagocytosis, differentiated THP-1 cells were infected with Llo WT or the ctrC mutant at MOI 10. Gentamycin was added at the indicated timepoints for 1 hour to kill extracellular bacteria and cells were lysed in water for CFU plating. Data show means ± SD of n = 3 independent experiments normalized to the input control at t = 0. B) For attachment, THP-1 cells were pre-treated with 2 μM cytochalasin D for two hours before infection with Llo WT or ΔctrC at MOI 10. At the indicated timepoints, gentamycin was added to half of the wells to kill extracellular bacteria. Cells were lysed in water and CFUs of gentamycin-treated cells (intracellular bacteria) and non-treated cells (total bacteria) were plated. Intracellular bacteria were deducted from total CFUs, normalized to the input control at t = 0 and data plotted as means ± SD of n = 3 independent experiments. Statistical significance was tested by two-way ANOVA with Tukey’s post-test. ns, non-significant; *, p≤0.05, **, p≤0.01, ***, p≤0.001.

https://doi.org/10.1371/journal.ppat.1012534.g005

Capsules have been shown to surround the bacterial cell wall and to mask LPS or outer membrane proteins like a shield. To further investigate if this could also be the case for L. longbeachae, we tested agglutination by yeast mannan, a complex polysaccharide of the Saccharomyces cerevisiae cell wall that can bind bacterial fimbriae [51,52]. Indeed, the capsule mutant agglutinated, in contrast to the WT and the complemented mutant (S14 Fig). These results indicate that the capsule of L. longbeachae protects fimbriae and other attachment factors that are exposed on the surface of a capsule mutant.

The L. longbeachae capsule impacts the pro-inflammatory cytokine response in primary innate immune cells

Capsules have been shown to mask cell structures like LPS or outer membrane proteins in bacteria. Due to their inherently low immunological recognition, capsules can have detrimental effects in infection settings such as reduced bacterial clearance or failure to mount an effective cytokine response against the pathogen [53]. Pro-inflammatory cytokines such as IL-6, TNFα, or IL-1β can be induced upon engagement of pattern-recognition receptors such as Toll-like receptors (TLRs) or C-type lectin receptors (CLRs) by macrophages. To learn whether the L. longbeachae capsule plays a role in the cytokine response, we infected primary human monocyte-derived macrophages (hMDMs) with the L. longbeachae WT and the capsule mutant strains and measured cytokine levels in cell supernatants at 24 hours post-infection. Based on previous works [21], we expected low cytokine levels upon L. longbeachae infection. We therefore opted for the high sensitivity SP-X array for human pro-inflammatory cytokines (Simoa). We observed a significant increase in IL-6 levels in the capsule mutant compared to the complemented strain (Fig 6A). Moreover, cytokine levels of the complemented strain were as low as those of the L. longbeachae WT strain. In contrast, we did not detect significant differences in TNFα or IL-1β levels (Fig 6B and 6C). These cytokines are known mediators of inflammation in the context of L. pneumophila infections [54,55], and IL-6 and TNFα are highly induced upon infection with the L. longbeachae ΔdotB mutant. However, TNFα and IL-1β secretion in human macrophages does not seem to be driven by the capsule but are rather dependent on the presence of a functional Dot/Icm T4SS.

Download:

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

Fig 6. The capsule influences the cytokine response of primary macrophages and dendritic cells.

A-C) Human monocyte derived macrophages (hMDMs) were infected at MOI 10 and cell supernatants were collected at 24 hours post-infection. High-sensitivity SP-X array was performed on cell supernatants. Data show means ± SD of fold change over L. longbeachae WT of n = 6 independent experiments. Statistical significance was tested by pairwise t-tests; ns, non-significant; *, p≤0.05. D-F) Murine bone-marrow derived DCs were infected at MOI 10 and cell supernatants collected at 24 hours post-infection. Data show means ± SD of n = 3 independent experiments. G-I) Murine bone-marrow derived macrophages were infected at MOI 10 and cell supernatants collected at 24 hours post-infection. Data show means ± SD of n = 3 independent experiments. Statistical significance was tested by one-way ANOVA with Tukey’s post-test. ns, non-significant; *, p≤0.05; **, p≤0.01; ***, p≤0.001; ****, p≤0.0001.

https://doi.org/10.1371/journal.ppat.1012534.g006

L. longbeachae is lethal for mice and fails to induce a strong cytokine response in vitro [21]. Our results show that the capsule plays a crucial role in virulence, and we therefore tested whether cytokine secretion during infection is also impacted by its presence or absence. We infected murine dendritic cells (DCs) with L. longbeachae WT or the capsule mutant. Similar to what was seen in human macrophages, the secretion of IL-6 was highly induced in the capsule mutant and lower in the L. longbeachae WT at 24 hours post-infection (Fig 6D). In contrast to human macrophages, secretion of IL-1β was also induced in murine DCs infected with the capsule mutant as compared to the WT (Fig 6E). Using murine bone marrow derived macrophages (BMDMs), higher IL-6 and IL-1β were detected at 24 hours in cells infected with the capsule mutant compared to those infected with the WT or the complemented mutant (Fig 6G and 6H). However, we did not detect a significant difference in TNFα secretion in murine DCs or macrophages (Fig 6F and 6I). Thus, our results show that in both human and murine infection models the capsule dampens the secretion of pro-inflammatory cytokines.

When cytokine induction by the capsule mutant or the complemented mutant was tested in BMDMs, we observed that complementation of the capsule reduces TNF-α levels in BMDMs as compared to the capsule mutant, but it does not influence IL-6 levels in vitro (S15A and S15B Fig). Further, we tested cytokine levels in BALF from Llo WT or ΔctrC infected mice. Here we observe higher IL-6 induction in the WT as compared to the mutant, but no difference in TNF-α levels in BALF of infected mice at 72 hours post-infection (S15C and S15D Fig). However, we detected higher protein levels in BALF of WT infected mice as measured by Bradford assay (S15E Fig). These data indicate that proinflammatory cytokine induction in vivo is due to higher lung damage caused by the WT as shown by protein levels and lactate measurements.

Ethics statement

All mouse experiments were conducted according to the institutional ethical committees for animal care (Comissão de Ética em Experimentação Animal da Faculdade de Medicina de Ribeirão Preto FMRP/USP), approved protocol number 1248/2023.

Bacterial strains, growth conditions, and cell culture

Bacterial strains used in this study are listed in Table 1. Escherichia coli DH5α subcloning efficiency bacteria were grown in Luria-Bertani broth or on LB agar. Legionella longbeachae strains and L. pneumophila were cultured in N-(2-acetamido)-2-aminoethanesulfonic acid (ACES)-buffered yeast extract broth (BYE) or on ACES-buffered charcoal-yeast (BCYE) extract agar with antibiotics added where appropriate [89]. For the L. longbeachae ΔctrC mutant, 15 μg/ml apramycin (Sigma) were added to the medium. For plasmid maintenance, 5 μg/ml of chloramphenicol was added to the medium. For growth in minimal medium, cells were grown in CDM or MDM according to published protocols [90]. Acanthamoeba castellanii strain C3 (ATCC 50739) trophozoites were grown in PYG medium according to published protocols [13,91]. Infection buffer was prepared as PYG 712 medium [2% proteose peptone, 0.1% yeast extract, 0.1 M glucose, 4 mM MgSO₄, 0.4 M CaCl₂, 0.1% sodium citrate dihydrate, 0.05 mM Fe(NH₄)₂(SO₄)₂•6H₂O, 2.5 mM NaH₂PO₃, 2.5 mM K₂HPO₃] without proteose peptone, yeast extract, and glucose. The human monocyte cell THP-1 cell line (ATCC TIB-202) was maintained in RPMI GlutaMax supplemented with 10% fetal calf serum (FCS) at 37°C and 5% CO₂. For infections, undifferentiated THP-1 cells were seeded with RPMI and 50 μg/ml of Phorbol 12-myristate 13-acetate (PMA) to start differentiation into adherent cells. Cells were grown in the presence of PMA for three days and recovered overnight in fresh RPMI medium before infection.

Download:

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

Table 1. Bacteria and plasmids used in this study.

https://doi.org/10.1371/journal.ppat.1012534.t001

Construction of L. longbeachae ΔctrC

The 0.5 kb flanking regions of L. longbeachae ctrC (llo3151) were amplified using the primer pairs P3/P4 (upstream segment) and P1/P2 (downstream segment) using genomic DNA of L. longbeachae NSW150 as a template (see primers listed in Table 2). The apramycin cassette was amplified with primer pair Apra_s/Apra_as using pTOPO-ApraR plasmid as template. The three PCR fragments were ligated by PCR using primers P1/P3 and cloned into pGEM-T easy vector (Promega). Plasmid was then digested with NotI, and the combined ctrC::Apramycin fragment was ligated into the suicide plasmid pLAW344 [92] cut with the same restriction enzyme. The resulting plasmid, pLGV012, was introduced into L. longbeachae NSW150 as described above, and transformants were plated onto BCYE agar with apramycin. Resulting colonies were grown in BYE with apramycin and plated onto BCYE agar supplemented with apramycin and 5% sucrose. Sucrose-resistant/chloramphenicol-sensitive clones were screened, and successful deletion was verified through PCR and whole genome sequencing.

Download:

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

Table 2. Primers used in this study.

https://doi.org/10.1371/journal.ppat.1012534.t002

Transformation of L. longbeachae

L. longbeachae was grown on BCYE agar plates at 37°C. Bacteria from a fresh plate were washed three times with ice-cold 10% glycerol, and the final pellet was resuspended in ice-cold 10% glycerol. For transformation, a 400 μl aliquot of electrocompetent cells was freshly mixed with 300–600 ng of plasmid DNA and electroporated at 2.5 kV, 1000 Ω, and 25 μF. The cultures were recovered in BYE broth at 37°C with shaking for 16 h and then plated onto BCYE agar with the appropriate antibiotic.

Transmission electron microscopy

Bacterial strains were grown overnight in BYE and OD₆₀₀ was followed constantly to determine E phase (OD₆₀₀ 2.0–2.5) or PE phase (OD₆₀₀ 3.7–4.2). Samples of 10 ml were taken at E and PE phase and centrifuged for 15 minutes at 500 g to remove medium. The medium was discarded, and cells were immediately fixed overnight in 0.1 M cacodylate fixation buffer containing 2.5% glutaraldehyde. Subsequently, fixed cells were washed in 0.2 M cacodylate buffer and stained with 1 mg/ml cationized ferritin (Sigma, F7879-2ML) for 30 minutes at room temperature. After another wash, cells were immobilized in 4% agar before osmium fixation. Agar blocks were sequentially dehydrated in increasing volumes of ethanol and embedded in resin and thin sections were sliced on a Leica UC7 ultramicrotome to 70 μm thickness. Resin slices were mounted and imaged on a Tecnai BioTWIN 20–120 kV transmission electron microscope.

RNA sequencing and analysis

Bacteria were grown at 37°C in BYE and OD₆₀₀ was measured to follow their growth. When bacteria reached exponential phase (OD₆₀₀ 2.0–2.5), 2x5 ml were pelleted and snap frozen in dry ice and ethanol. Bacterial morphology was checked under a microscope. Similarly, pellets were collected from bacteria grown to post-exponential phase (OD₆₀₀ 3.7–4.2) and stationary phase the following day. Bacterial pellets were resuspended in Qiazol (Qiagen) and RNA was isolated according to the miRNA Mini kit (Qiagen). The samples were Turbo DNase digested (Thermo Scientific) and rRNA was depleted using the RiboCop rRNA Depletion Kit for Gram-negative bacteria (Lexogen). Depleted RNA was metal-catalyzed heat-fragmented using the RNA Fragmentation kit (Ambion). RNA quantity was measured using a Qubit 2.0 (Invitrogen) and size distribution was confirmed between 100–200 nt by BioAnalyzer (Agilent Technologies). Fragmented RNA was subsequently processed according to the TruSeq mRNA sample preparation guide by Illumina. RNAseq was performed using Illumina NextSeq 550 multiplex sequencing (Illumina). For analyzing capsule gene expression, we used FASTQ files containing single end reads generated by Illumina sequencing. Sequencing reads were processed with Cutadapt software (version 1.15) to remove adapters. Trimming was performed with Sickle (version 1.33, https://github.com/najoshi/sickle) with a quality threshold (Phred Score) of 20. Reads shorter than 20 nucleotides were discarded. Clean reads were aligned to the Legionella longbeachae NSW150 sequence using Bowtie2 (version 2.3.4.3), and only uniquely mapped reads were kept for the read counts. We used Samtools package (https://github.com/samtools/samtools) to build indexed BAM files from the mapping results. To count the number of reads overlapping each genomic feature, we used featureCounts from the Subread package (version 1.6.3). Only primary alignments were counted. Differential analysis between the different conditions was performed with the R package Sartools (version 1.3.0) using DESeq2 methods. The “median” option was used to compute size factors.

Construction of complementation and dual reporter plasmids

Capsule cluster promoter was amplified using primers SSM_79 and SSM_80, and the region containing ctrC (llo3151) and ctrD (llo3150) was amplified using primers SSM_82 and SSM_85. Both PCRs were performed using genomic DNA of L. longbeachae NSW150 as a template. Fragments were then ligated by PCR, with primers SSM_86 and SSM_87. The obtained amplicon was cloned into pBCKS (Stratagene) by restriction-free cloning [93], resulting in plasmid SSM073. To construct plasmid SSM083, bexD (llo3149) was amplified using primers SSM_113 and SSM_114, and subcloned into SSM073 vector by restriction-free cloning [93]. For construction of the reporter plasmid, the capsule transporter promoter (prom_cap) was PCR amplified using primers SSO_025 and SSO_049 (prom_cap1) or SSO_050 and SSO_049 (prom_cap2) (listed in Table 2). Superfolder GFP (sfGFP, gift from David Bikard) was PCR amplified using primers SSO_051 and SSO_052 (sfGFP1) or SSO_053 and SSO_054 (sfGFP2). A C-terminal degradation tag was added to each sfGFP construct by restriction-free cloning [93] using primers SSO_045 and SSO_046 or SSO_047 and SSO_048. The GFP constructs were fused to the capsule transporter promoter by overlap PCR, one copy with restriction sites for BamHI and HindIII and the second copy with restriction sites HindIII and KpnI. Each copy was first ligated into pBCKS and inserts were confirmed by sequencing. The second copy was subsequently ligated into the first copy pBCKS vector using restriction sites HindIII and KpnI. One copy of mKate2 (Addgene #68441) was fused to a strong promoter from the Anderson collection (Table 2) by overlap PCR using primers SSO_065 and SSO_066 and ligated into the two copy pBCKS vector through restriction sites NotI and BamHI. A control plasmid was constructed by replacing the two promoter regions cap_prom with a triple stop codon by restriction-free cloning using primers SSO_621/SSO_622 and SSO_623/SSO_624, respectively. Plasmid DNA was isolated using Nucleospin Plasmid kit (Macherey Nagel) and all plasmids were confirmed by sequencing.

Imaging of dual reporter in THP-1 cells and A. castellanii

Bacteria expressing the dual reporter constructs (pSS016 or pSS017) were grown to PE phase in BYE medium with 5 μg/ml chloramphenicol. Differentiated THP-1 cells were infected at an MOI of 10 in a μClear 96-well plate (Greiner) for 1 hour. Cells were washed three times in PBS and supplemented with fresh RPMI medium. Subsequently, cells were stained with CellMask Deep Red Plasma Membrane Stain (Invitrogen, C10046) at 5 μg/ml for 20 minutes and washed once. Nuclei were stained with Hoechst dye (Invitrogen, H3570) at 1 μg/ml. Live imaging was performed at 40x magnification using the Opera Phenix confocal microscope (PerkinElmer). Images were obtained every hour and analyzed using the Harmony high-content analysis software (PerkinElmer). A. castellanii were infected in infection buffer at an MOI of 10 for 1 hour at 25°C. After one hour, cells were extensively washed in PBS to remove extracellular bacteria and maintained in infection buffer at 25°C. Cells were imaged at 40x magnification using the EVOS inverted digital microscope (Thermo Fisher).

Animals and in vivo infections

Mice used in this study were bred and maintained in institutional animal facilities of University of São Paulo—School of Medicine of Ribeirão Preto/SP. All mice were at least 8 weeks old at the time of infection and were in a C57BL/6 (Jax 000664) genetic background. For the survival and CFU experiments, approximately 10 and 7 mice per group were used, as indicated in the Figs 4A–4E and S9. For in vivo experiments, the mice were anesthetized with ketamine and xylazine (300 mg/kg and 30 mg/kg, respectively) by intraperitoneal injection followed by intranasal inoculation with 40 μl of RPMI 1640 containing bacteria. For CFU determination, the lungs were harvested and homogenized in 5 ml of RPMI 1640 in a tissue homogenizer (Power Gen 125; Thermo Scientific) [94,95]. Lung homogenates were diluted in RPMI and plated on BCYE agar plates containing streptomycin for CFU determination as previously described. For survival determination, mice were observed once a day with the measurement of their weight [21]. The care of the mice followed the institutional guidelines on ethics in animal experiments.

Bone marrow-derived dendritic cells and macrophages

Bone marrow-derived dendritic cells (BMDCs) and bone marrow-derived macrophages (BMDMs) were generated from C57BL/6 mice as previously described [96,97]. Mice were euthanized and bone marrow cells were obtained from femurs and tibias. BMDCs were harvested from femurs and differentiated with RPMI 1640 (Gibco, Thermo Fisher) containing 20% Fetal Bovine Serum (FBS, Gibco) and 20 ng/ml of recombinant GM-CSF (eBioscience), 2 mM L-glutamine (Sigma-Aldrich), 15 mM Hepes (Gibco) and 100 U/ml penicillin-streptomycin (Sigma-Aldrich) at 37°C with 5% CO₂ for 7 days. The non-adherent/loosely adherent fraction was harvested by collecting the culture supernatant and carefully washing the plate with PBS. After centrifugation of the total volume collected, BMDCs were resuspended in RPMI 1640 supplemented with 10% FBS and plated as indicated. BMDMs were harvested from femurs and differentiated with RPMI 1640 (Gibco, Thermo Fisher) containing 20% FBS and 30% L929-Cell Conditioned Medium (LCCM), 2 mM L-glutamine (Sigma-Aldrich), 15 mM Hepes (Gibco) and 100 U/ml penicillin-streptomycin (Sigma-Aldrich) at 37°C with 5% CO₂ for 7 days. Of note, in some experiments LCCM was replaced for 10% of a conditional medium from 3T3 cells stably expressing mouse MCSF. Cells were detached with cold PBS, resuspended in RPMI 1640 supplemented with 10% FBS and plated as indicated.

Replication and competition assays in Acanthamoeba castellanii

Amoeba trophozoites grown at 20°C were washed in infection buffer, seeded at 1x10⁶/ml and left to adhere for one hour prior to infection. Bacteria were resuspended in infection buffer to an MOI of 0.1 and left to infect for 1 hour. Dilutions of an aliquot of input bacteria was plated on BCYE to determine CFUs used for infection (t0). After one hour of infection, amoeba were washed three times in PBS to remove extracellular bacteria and resuspended in infection buffer. Samples of 500 μl were taken at this timepoint (t1), centrifuged at 14000 rpm for 3 minutes and vortexed for one minute to break up amoebae. Samples were subsequently taken every 24 hours for seven days. Experiments were carried out in triplicates and CFUs were counted after three to four days of growth on BCYE at 37°C. Competition assay was carried out as previously described [98]. Briefly, A. castellanii (5 × 10⁶ per flask) in infection buffer was infected at an MOI of 0.1 with a 1:1 mix of wild type L. longbeachae and ΔctrC mutant bacteria. The infected amoebae were grown for seven days at 37°C. Every two days, a sample of lysed amoebae was diluted 1:100 and used to infect a fresh flask of amoebae (100 μl homogenate per flask). Dilutions were plated on BCYE agar plates containing apramycin or not, to determine CFUs.

Infection of U2OS cells

Human osteosarcoma epithelial cells stably expressing Sec61β-GFP (U2OS-Sec61β-GFP) for labelling of the Endoplasmic Reticulum [99] were cultured in DMEM + GlutaMAX (Life Technologies), supplemented with 10% heat-inactivated FBS (Life Technologies). 20000 cells were seeded in DMEM in 96-well Greiner μClear plates and allowed to adhere before infection. Cells were infected with L. longbeachae at MOI 100 for two hours and extracellular bacteria were washed off using PBS. Cells were stained with Hoechst dye (Invitrogen, H3570) at 1 μg/ml to label nuclei. Live imaging was performed at 63x magnification at 37°C and 5% CO₂ using the Opera Phenix confocal microscope (PerkinElmer). Images were analyzed using the Harmony high-content analysis software (PerkinElmer).

Replication assays in THP-1 cells

Infection assays of THP-1 cells (ATTC: TIB-202) were done as previously described [100]. Briefly, cells were seeded and differentiated into macrophage-like adherent cells in 12-well tissue culture trays (Falcon, BD lab ware) at a density of 2 x 10⁵ cells/well. Stationary phase L. longbeachae were resuspended in serum free medium and added to cells at an MOI of 10. After 2 h of incubation, cells were treated with 100 μg/ml gentamycin for 30 minutes to kill extracellular bacteria. Infected cells were then washed before incubation with serum-free medium. At 2, 24, 48, and 72 h, THP-1 cells were lysed with 0.1% Triton X-100. The infection efficiency of the different L. longbeachae strains was monitored by determining the number of CFUs after plating on BCYE agar.

Bacterial replication in BMDMs

For CFU determination, macrophages were seeded at 2×10⁵ cells/well in 24-well plates and cultivated in RPMI 1640 with 10% FBS. Cultures were infected at a multiplicity of infection (MOI) of 10, centrifuged for 5 minutes at 200 ×g at room temperature. After 1 hour of infection, BMDMs were washed twice with PBS, and 1 ml of medium was added to each well. For CFU determination, the cultures were lysed in sterile water, and the cell lysates were combined with the cell culture supernatant from the respective wells. Lysates plus supernatants from each well were diluted in water, plated on BCYE agar plates, and incubated for 4 days at 37°C for CFU determination [94,95].

Beta-lactamase translocation assay

Translocation of T4SS effectors was performed as previously described [101]. Plasmids pXDC61 or SSM012 were electroporated into L. longbeachae strains or L. pneumophila strain Paris as a positive control. Triplicate wells of THP-1 cells were seeded in 96- well plates at 10⁵ cells/well and differentiated into adherent cells with 50 μg/ml PMA for three days. Cells were recovered in RPMI without PMA for another night. Freshly transformed bacteria were induced by addition of 1 mM IPTG and THP-1 cells were infected at MOI 50 with stationary phase bacteria. Cells were subsequently centrifuged to synchronize the infection. At 1h30 after infection cell were incubated with CCF4-AM (Life Technologies) and 0.1 M probenecid at room temperature in the dark. After another 1.5 hours, cells were detached with non-enzymatic cell dissociation solution (Sigma). Flow cytometry was performed using a MACSQuant VYB system (Miltenyi Biotec), with excitation at 405 nm (violet) and emission collection with filters at 525/50 nm (Green) and 450/50 nm (Blue). Flow data were analyzed by FlowJo 10 software (LLC). Gates for green and blue fluorescence were set based on uninfected cells without any treatment. Western blots were performed on protein lysates from induced bacteria to confirm expression of blaM (see S10 Fig for gating strategy and Western blot).

Isolation of human monocyte-derived macrophages (hMDMs)

Ficoll gradient centrifugation (Lympholyte, Cedarlane) was performed to isolate human peripheral blood mononuclear cells (PMBCs) from freshly extracted blood from healthy donors. To isolate CD14+ cells, anti-hCD14 magnetic beads were used and isolated cells were differentiated to hMDMs by addition of 50 ng/ml rhM-CSF (R&D Systems) to XVivo medium (Lonza). After 3 days, medium was exchanged, and cells were differentiated for another 3 days in fresh XVivo supplemented with rhM-CSF. All donors gave written consent under the agreement C- CPSL UNT–No. 15/EFS/023 between the Institut Pasteur and EFS (L’Établissement français du sang), in accordance with articles L1243-4 and R1243-61 of the French Public Health Code and approved by the French Ministry of Science and Technology. Supply and handling of human blood cells followed official guidelines of the agreement between the Institut Pasteur, EFS, and the regulation of blood donation in France.

Bacterial replication, early apoptosis, and ROS production in hMDMs

Human monocyte-derived macrophages were seeded in XVivo medium (Lonza) at 37°C and 5% CO₂. Cells were stained with CellTracker Blue (CMF2HC, Invitrogen) and Hoechst dye (Invitrogen, H3570) for 1 hour pre-infection, before washing the cells in PBS. Cells were subsequently infected with L. longbeachae expressing dsRed (plasmid SW001, kind gift from Hubert Hilbi) at MOI 10 for 1 hour and extracellular bacteria were removed by washing with PBS. To monitor ROS production, cells were stained with CellROX DeepRed (C10422, Invitrogen) for 30 minutes and dye was washed off as indicated by the manufacturer. To monitor early apoptosis, cells were stained with Annexin V AlexaFluor 488 dye for living cells (A13201, Invitrogen). Live cell imaging was performed using the Opera Phenix confocal microscope (PerkinElmer) and bacterial replication, early apoptosis, and ROS production in single infected cells were analyzed using the Harmony high-content analysis software (PerkinElmer).

Attachment and phagocytosis assays

THP-1 cells (ATTC: TIB-202) were seeded at 4x10⁵ cells/well in 12 well plates in RPMI-10% FCS and differentiated for three days with 50 nM PMA. After 96 hours of differentiation, RPMI without PMA was used to recover the cells overnight. For attachment assays, cells were treated with 2 μM cytochalasin D (Sigma) for two hours prior to infection. Bacterial strains were grown in BYE overnight and bacterial growth was monitored until the bacteria reached post-exponential growth phase (OD₆₀₀ 3.7–4.2). The bacteria were diluted to reach an MOI of 10 in RPMI with or without 2 μM cytochalasin D for attachment assays. At the indicated timepoints, cells were washed three times with PBS to remove non-adhered bacteria and half of the wells were treated with 100 μg/ml gentamycin for one hour to kill extracellular bacteria (control for internalized bacteria). Cells were lysed by addition of ddH₂O for 15’ at 37°C. Dilutions of CFUs were plated for both the gentamycin-treated and untreated wells to determine total CFUs vs. adhered CFUs, respectively. For phagocytosis assays, cells were washed at the indicated timepoints and treated with 100 μg/ml gentamycin for one hour to kill extracellular bacteria. Cells were lysed in ddH₂O for 15 minutes at 37°C and dilutions were plated to determine CFUs. An aliquot of bacteria used for infections was plated for each strain to determine CFUs at t0.

Detergent, oxidative, and osmotic stress assays

To test the resistance of the capsule mutant to salt stress, bacteria were grown in BYE to PE phase (OD₆₀₀ 3.7–4.2) and diluted to 2x10⁹/ml. Ten-fold dilutions were spotted in triplicates onto BCYE plates supplemented with 100 mM NaCl or plain BCYE plates and bacteria were grown at 37°C for 6 days. Tween-20 was tested as previously described [45]. Briefly, bacteria were grown in BYE medium to post-exponential growth phase (OD₆₀₀ 3.7–4.2), washed in ultrapure water and adjusted to an OD₆₀₀ of 2.5. Subsequently, cells were treated with 0.05% Tween-20 ± 300 mM of either NaCl or sucrose for 30 minutes at 37°C with moderate shaking. Ten-fold dilutions were spotted in duplicates on BCYE plates and bacteria were grown at 37°C for 6 days. Similarly, for oxidative stress, post-exponential bacteria were washed once, adjusted to OD₆₀₀ of 2.5, and incubated in the presence of 2 mM or 10 mM H₂O₂ for 30 minutes at 37°C with moderate shaking. Ten-fold dilutions were spotted onto BCYE agar plates in duplicates and bacteria were grown at 37°C for 6 days.

Agglutination assay

Bacteria were grown to post-exponential phase (OD₆₀₀ 3.7–4.2), washed in PBS adjusted to an OD₆₀₀ of 2.5. Subsequently, agglutination was tested by adding 1 mg/ml yeast mannan (Sigma) to the bacterial solutions and incubation at 37°C for 15 minutes. Cells were subsequently observed at 100x magnification on an EVOS inverted digital microscope (Thermo Fisher).

Polysaccharide extractions, gel electrophoresis and HPLC analysis

Bacteria were grown to post-exponential growth phase in BYE medium and washed in PBS. For phenol extraction, cells were treated with 45% hot phenol for 30 minutes according to previous protocols [102]. The aqueous phase was subsequently extracted and extensively dialyzed against water to remove residual phenol. To remove nucleic acids and proteins, extracts were treated with DNase, RNase, and proteinase K, each overnight. After dialysis, extracts were ultracentrifuged using a Beckman Coulter Optima ultracentrifuge at 100 000 rpm for two hours. Samples were freeze-dried and resuspended in ultrapure water for analysis by Dionex and SDS gel electrophoresis. For enzymatic extractions of polysaccharides, bacterial cell pellets were treated with mutanolysin (Sigma) and lysozyme (Sigma) according to published protocols [42]. Cell debris was removed by centrifugation and extracts were treated with DNase, RNase, and proteinase K, and subsequently dialyzed against water. Following ultracentrifugation, extracts were resuspended in ultrapure water.

SDS gel electrophoresis was performed using pre-cast mPAGE 4–20% bis-tris gels (Merck) and 20 μl of PS extracts were mixed in Laemmli buffer and boiled at 95°C for 10 minutes. For silver staining, SDS gels were fixed in 45% ethanol solution containing 0.5% periodic acid (Sigma). After extensive washes in water, gels were submerged in 0.1% silver nitrate solution in water for 15 minutes. Gels were rinsed in water and developed in 3% sodium carbonate solution with added formaldehyde to fix the silver dye. Developer was neutralized by addition of citric acid (Sigma). For Alcian blue staining, gels were stained in 0.1% Alcian blue solution in 40% ethanol/5% acetic acid for 1 hour and destained in acetic acid solution [58]. For Stains-all staining, gels were fixed in 50% ethanol/10% acetic acid solution for 1 hour and extensively washed in water. A 0.1% Stains-all solution was prepared in water and gels were stained in the dark for 30 minutes. Gels were subsequently destained in water. For analysis of monosaccharides. monosaccharides were first released by acid hydrolysis (TFA 4 N, 4 hours at 100°C or HCl 6 N, 6 hours at 100°C). After vacuum drying of the hydrolysate, monosaccharides were identified and quantified by high performance anion exchange chromatography (HPAEC) with a pulsed electrochemical detector and an anion exchange column (CarboPAC PA-1, 4.6 x 250 mm, Dionex) using 18 mM NaOH as mobile phase at a flow rate of 1 mL/min; glucosamine and galactosamine were used as standards [103].

ELISA of human pro-inflammatory cytokines

Differentiated hMDMs were seeded at 3x10⁴ cells per well in XVivo (Lonza) and infected with different bacterial strains at an MOI of 10. Cell supernatants were collected at 24 hours post-infection and centrifuged to remove cell debris and extracellular bacteria. The cleared cellular supernatants were rapidly frozen on dry ice and stored at -80°C until cytokine measurements were performed. ELISA was performed using the SP-X CorPlex Human Cytokine Panel 1 (Quanterix), a sandwich-based multiplex ELISA with higher sensitivity and broad range of detection of 10 human pro-inflammatory cytokines. The SP-X array was performed from 6 independent experiments according to the manufacturer’s instructions, including triplicate standard wells and duplicate wells for each sample. The data was analyzed using the proprietary SP-X analysis software.

ELISA of murine pro-inflammatory cytokines

For ELISA experiments, BMDMs and BMDCs were seeded in 24-well plates (5x10⁵ cells/well). Infections were performed in RPMI 1640 supplemented with 10% FBS. After the indicate time points, the supernatants were collected for cytokine determination using ELISA kits according to the manufacturer’s recommendations (R&D and BD Bioscience).