The Legionella pneumophila Dot/Icm Type IV Secretion System is Structurally and Functionally Resilient in Absence of Species-specific Proteins Dis2 and Dis3

Abstract

Legionella pneumophila is a pathogenic Gram-negative bacterium that causes Legionnaires’ disease. The main virulence factor of L. pneumophila is the Dot/Icm Type IV Secretion System (T4SS), which translocates effector proteins into the cytoplasm of the host cell, allowing the bacterium to establish a replicative niche. The outer membrane core complex (OMCC), the T4SS machinery localized between the inner and outer membranes, is composed of at least nine proteins organized into various sub-complexes that include the dome, outer membrane cap (OMC), periplasmic ring (PR), and stalk. In this study we describe how two uncharacterized Dot/Icm T4SS components, Dis2 and Dis3, contribute to the structure of the T4SS, the ability of the T4SS to translocate effectors, and the pathogenicity of L. pneumophila. Using cryo-electron microscopy we show that OMCCs purified from a Δdis2 strain are only missing the density for Dis2, while OMCCs purified from the Δdis3 strain lack densities for Dis3 and DotF in the OMC. Despite missing these proteins, the OMC and PR of both mutant OMCCs remain structurally stable. Strains lacking dis2 and or dis3 efficiently replicate in human macrophages; however, they have subtle differences in translocation efficiency for four tested substrates. Combined these data indicate that Dis2 or Dis3 are not required for the stability or global organization of the OMCC, but each protein may contribute to the efficient translocation of specific effectors.

Introduction

Legionella pneumophila (L. pneumophila) is a Gram-negative facultative intracellular bacterial pathogen that is the causative agent of a potentially fatal atypical pneumonia called Legionnaires’ disease [1–5]. L. pneumophila is ubiquitous in the aqueous environment and is usually found in freshwater hosts, such as amoebae [6,7]. Humans can become infected with L. pneumophila when breathing aerosols from contaminated water sources [1,2]. Inhaled bacteria are phagocytosed by alveolar macrophages and subvert the endocytic pathway preventing their elimination through lysosomal degradation [8,9]. As part of this process L. pneumophila creates a Legionella containing vacuole (LCV), a cellular niche where the bacteria safely replicate [3,8–10]. Without the ability to establish an LCV the bacteria are no longer virulent.

The establishment of the LCV is dependent upon translocation of L. pneumophila effector proteins into the host through the Dot/Icm Type IV Secretion System (Dot/Icm T4SS) [10–15]. Many effector proteins have been experimentally validated by measuring the delivery of reporter fusion proteins into host cells [16–18], while still more putative effectors have been identified using bioinformatic analyses [19,20]. Overall, it appears that L. pneumophila can translocate over 330 different effector proteins into the cytoplasm of the host cells [16–24]. Many L. pneumophila effectors alter host cell physiology by interacting with and/or modulating endocytic and secretory pathways, ubiquitination machinery, lipid metabolism, and cell death pathways [5,15,25,26]. While protein translocation is required for L. pneumophila virulence, the mechanism of how effectors move through the Dot/Icm T4SS is not understood.

T4SSs are dynamic multi-protein complexes that span bacterial membranes. In Gram-negative bacteria, such as Escherichia coli, Agrobacterium tumefaciens (A. tumefaciens), Helicobacter pylori (H. pylori), and Coxiella burnetii (C. burnetii), T4SSs move effectors across the inner- and outer-membrane, either into the environment, or into a recipient cell [27–30]. T4SSs can be broadly divided into prototypical (minimized) or expanded systems. Prototypical T4SSs include conjugation systems, which deliver mobile genetic elements between bacteria, and effector translocation systems, which translocate proteins and DNA [27,30,31]. These T4SSs are composed of an outer membrane core complex (OMCC) which spans the bacterial inner- and outer membranes, and an inner membrane complex (IMC) that includes cytoplasmic ATPases [32,33]. The prototypical VirB/VirD4 T4SS of A. tumefaciens OMCC consists of VirB7, VirB9, and VirB10, and most expanded T4SSs also have homologs of these components [27,34–36]. OMCCs in expanded T4SS, like those found in H. pylori and L. pneumophila, are larger in size, contain additional species-specific components, and have a more complex structural organization than prototypical T4SSs [34,37–41]. The Dot/Icm OMCC contains at least nine identified proteins organized into various sub-assemblies (Figure 1A, B) [40,41]. Three Dot/Icm OMCC components, DotD (VirB7), DotH/IcmK (VirB9), and DotG/IcmE (VirB10), have homologs in prototypical T4SSs (Supplemental Table 1) [41–44]. The remaining six proteins, DotC, DotF/IcmG, DotK/IcmN, Dis1, Dis2, and Dis3, are not found in prototypical T4SSs, but are found in other Legionella species [45] as well as the obligate intracellular pathogen, C. burnetii [46]. Interestingly, C. burnetii also has an expanded Dot/Icm T4SS which appears to have a structural organization similar to the L. pneumophila Dot/Icm T4SS when observed in situ by cryo-electron tomography (cryo-ET) [47,48].

Figure 1. Structural organization of the L. pneumophila Dot/Icm T4SS OMCC.

(A) Cartoon of a cross-section of the Dot/Icm T4SS outer membrane core complex (OMCC) and type IV coupling complex (T4CC) in relation to the bacterial outer- and inner-membranes (OM and IM). The color for each protein is used in all panels. The OMCC can be divided into protein sub-assemblies called the Dome, outer membrane cap (OMC), periplasmic ring (PR), Stalk, Plug, and Collar [40,41,49,51,52]. The gray areas represent parts of the OMCC seen by cryo-ET analysis. (B) Cross-section of the atomic models of the dome (7MUQ, chains AG-PG), OMC (7MUC), PR (7MUC) and T4CP (6SZ9) in relation to the OM and IM. C16, 16-fold symmetry; C13, 13-fold symmetry; C18, 18-fold symmetry; C6, 6-fold symmetry. (C–E) Molecular models of the dome (7MUQ, chains AG-PG), OMC (7MUC), and PR (7MUC) rotated around x-axis. (F) Two views of the asymmetric unit of the dome, OMC and PR, with the molecular surface shown for Dis2 (green) and Dis3 (beige). (G) Dis2 sits on the OMC surface facing the outer membrane and shares 768 Ų of surface area with DotC (orange), 350 Ų surface area with DotD (red), and 837 Ų surface area with DotK (light blue). (H) Dis3 extends as a “spoke” from the OMC and shares 969 Ų surface area with Dis1 (blue), 425 Ų surface area with DotD (red), 708 Ų surface area with DotK (light blue), and 688 Ų surface area with DotF (dark purple).

Cryo-ET studies of vitrified L. pneumophila have provided a low-resolution model of the overall architecture of the T4SS spanning from the outer membrane into the bacterial cytoplasm and single particle cryo-electron microscopy (cryo-EM) analyses of purified L. pneumophila OMCCs have led to a detailed atomic resolution model of the Dot/Icm T4SS OMCC that includes structures of the dome, OMC, and PR [40,41,49–52] (Figure 1A–E). Although there is a stalk, collar, and plug visualized by cryo-ET analysis, there are no high-resolution structures of these OMCC subassemblies [49–52] (Figure 1A, B). Single particle cryo-EM analysis of the Dot/Icm T4SS OMCC identified unexpected symmetry mismatches between the dome, OMC, and PR while also identifying three proteins, Dis1, Dis2, and Dis3, that had not been previously known to be core T4SS components [40,41] (Figure 1A, B). While Dis1 is required for efficient intracellular replication of L. pneumophila in protozoan and mammalian hosts [53], the importance of Dis2 and Dis3 in intracellular replication, T4SS effector translocation, or structural stability of the Dot/Icm T4SS has not been explored.

Dis2 is a 14 kDa protein positioned in the OMC proximal to the bacterial outer membrane, while Dis3 is a 36 kDa protein that forms the 13 radial spokes extending outward from the OMC [41] (Figure 1B–F). Both Dis2 and Dis3 have multiple protein–protein contacts with other OMC components (Figure 1F–H). For example, Dis2 interacts with DotK, DotC, and DotD (Figure 1G), while Dis3 interacts with Dis1, DotK, DotF, and DotD (Figure 1H). For example, the OMC-localized DotF interacts primarily with Dis3, with a shared surface area of 688 Ų. Thus, Dis3 may act as a structural scaffold for DotF inclusion in the OMC. The outer surface of Dis3 is predominantly electropositive, which may promote interactions with the electronegative head groups of the inner leaflet of the outer membrane [41]. The protein interaction networks between Dis2 and Dis3 with other OMC members suggest that both proteins are important structural components of the OMC, which would also make them important for the function of the T4SS and the ability of L. pneumophila to replicate in hosts.

In this study, we examine the contribution of Dis2 and Dis3 to the function and structural organization of the L. pneumophila Dot/Icm T4SS. We generated L. pneumophila strains lacking either dis2 or dis3 and used a combination of biochemistry, mass spectrometry, single particle cryo-EM, and virulence assays to address the role of Dis2 and Dis3 in Dot/Icm T4SS structure and function. Our results show OMCCs purified from the Δdis2 strain only lack density in the OMC for Dis2, while OMCCs purified from the Δdis3 mutant are missing density in the OMC for both Dis3 and DotF, although PR-localized DotF remains. Δdis2 and Δdis3 strains still replicate in human macrophages and each deletion strain has only subtle changes in the level of translocation for some substrates. Combined, these analyses highlight the robustness of the structural organization and function of the Dot/Icm T4SS structure to the loss of these species-specific components.

Results

Dis2 and Dis3 are not essential for Dot/Icm T4SS function

In the previous structural analysis of the Dot/Icm OMCC three proteins, Dis1, Dis2, and Dis3, were discovered to be core components of the Dot/Icm T4SS [40,41]. Although these proteins are not generally conserved across Gram-negative bacteria, they are conserved across Legionella species, with protein homologs of Dis1, Dis2, and Dis3 found in 97%, 63%, and 98% of the reference Legionella species in the Bacterial and Viral Bioinformatics Research Center (BV-BRC) [54] (Supplemental Figure 1). While Dis1 is required for efficient intracellular replication of Legionella in protozoan and mammalian hosts [53], the importance of Dis2 and Dis3 for L. pneumophila intracellular replication in mammalian hosts is not known.

To examine the roles of Dis2 and Dis3 for T4SS function and structure we created Δdis3 and Δdis2 L. pneumophila strains using established protocols [55]. To test for T4SS function we assessed the ability of Δdis2 and Δdis3 cells to grow on charcoal yeast (CYE(T)) agar plates supplemented with 100 mM NaCl (Figure 2A, B). In this assay, cells with functional T4SSs are not able to grow as well on CYE(T) agar with this concentration of salt compared to cells with defective T4SSs [56,57]. Wild type (WT), ΔdotA (T4SS deficient), Δdis3, and Δdis2 cells were grown to post-exponential phase in liquid culture and then plated onto CYE(T) agar plates with or without 100 mM NaCl. Growth was quantified by counting colony forming units (CFUs). WT and ΔdotA strains were included as positive and negative controls [56]. In this assay the WT, ΔdotA, Δdis3, and Δdis2 cells all grew similarly on CYE(T) agar plates with no salt (Figure 2A). As expected, the ΔdotA mutant grew better than WT on solid medium containing 100 mM NaCl, exhibiting an approximately a two-log increase in CFUs (Figure 2B). The Δdis3 and Δdis2 strains, like the WT strain, struggled to grow on CYE(T) agar supplemented with 100 mM NaCl (Figure 2B), suggesting these strains still assemble a functional T4SS.

Figure 2. Functional characterization of the Dot/Icm T4SS in L. pneumophila Δdis3 and Δdis2 strains.

(A–B) Testing for salt sensitivity of wild type (WT, black), ΔdotA (beige), Δdis3 (pink), and Δdis2 (blue) cells by measuring their ability to grow on CYE(T) agar with no salt (A) or 100 mM salt (B). Each column is the average of three independent biological replicates measured in technical triplicate. Error bars indicate standard error of the mean (SEM). *indicates p < 0.05 based on Ordinary one-way ANOVA with Dunnet’s multiple test correction. ns, not significant. (C) Testing the ability of WT (black), ΔdotA (beige), Δdis3 (pink), and Δdis2 (blue) strains to intracellularly replicate in U937-derived human macrophages over 72 h. Each measurement is the average of four independent biological replicates measured in technical triplicate. ± SEM plotted. (D) Testing the ability of wild type (WT, black), ΔdotA (beige), Δdis3 (pink), and Δdis2 (blue) cells to translocate Dot/Icm T4SS substrates in human macrophages. Intracellular cAMP levels were measured in U937-derived human macrophages infected with WT (black), ΔdotA (beige), Δdis3 (pink), or Δdis2 (blue) strains expressing CyaA, CyaA-RaIF, CyaA-SidB, CyaA-SidD, or CyaA-SidF. Each measurement is the average of three independent biological replicates measured in technical duplicate. ± SEM plotted. * Indicates p < 0.05, ** indicates p < 0.01 *** indicates p < 0.0002, **** indicates p < 0.0001 based on 2way ANOVA with Dunnet’s multiple test correction.

Dis2 and Dis3 are not required for intracellular replication in human macrophages

We next tested whether Dis3 and Dis2 are required for L. pneumophila intracellular replication in human macrophages, another assay that assesses T4SS function [13,58]. For this analysis U937-derived human macrophages were infected with either WT, ΔdotA, Δdis3, or Δdis2 L. pneumophila at a multiplicity of infection (MOI) of one and incubated for one hour. After incubation, the cells were washed with phosphate-buffered saline (PBS) to remove extracellular bacteria and fresh medium (with no bacteria) was added. Infected macrophages were harvested every 24-hours over 72 h and lysed to release intracellular bacteria. The lysate was serially diluted onto CYE(T) agar plates, which were incubated for 4–5 days at 37° C before CFUs were counted. In this assay, increasing numbers of CFUs indicate that the L. pneumophila were able to establish an LCV to replicate in the host, a process that requires a functional Dot/Icm T4SS [10–15]. As expected, the L. pneumophila ΔdotA cells were not able to replicate in macrophages over the course of the experiment (Figure 2C). However, WT, Δdis3 and Δdis2 L. pneumophila were able to successfully undergo intracellular replication (Figure 2C). These data indicate that Dis3 and Dis2 are not required to support intracellular replication of L. pneumophila in human macrophages and, thus, these deletion strains still have functional T4SS.

Dis3 and Dis2 alter the efficiency of T4SS secretion for some effectors

Although not required for overall T4SS function (Figure 2A–C), conservation of Dis3 and Dis2 across Legionella species (Supplemental Figure 1) combined with their significant protein interaction networks within the OMC led us to hypothesize that they could have substrate specific roles in T4SS translocation. While the mechanism of substrate selection and translocation by the Dot/Icm T4SS is still not fully understood, soluble substrates appear to require a C-terminal stretch of 20–30 amino acids that target the effector for T4SS translocation, which sometimes includes a hydrophobic residue 3 or 4 amino acids from the C-terminus [17,59,60]. Before being translocated by the T4SS, substrates often first interact with the type IV coupling complex (T4CC), that for soluble proteins can either be dependent or independent of the IcmS/IcmW chaperone complex [21,61–64]. The translocation of membrane proteins by T4SSs appears to be even more complex [65,66]. For example, substrates with transmembrane domains and a secretory signal may undergo translocation in two steps, with help from the Sec translocon [67]. Once engaged with the T4SS OMCC, how substrates are moved across the inner- and outer-membranes has not been defined, although there is a proposed model that effector proteins are unfolded before translocation [68].

To test whether Dis3 and Dis2 contribute to the translocation of specific substrates, we monitored the translocation of L. pneumophila substrates by fusing the Bordetella pertussis adenylate cyclase (CyaA) toxin to the N-termini of a panel of Dot/Icm T4SS effectors and measured cyclic adenosine monophosphate (cAMP) production in infected human macrophages [69]. Since CyaA requires calmodulin for enzymatic activity, its cAMP production is only detected once the fusion substrate is translocated into the host cytoplasm. This translocation-monitoring assay is well established for studying L. pneumophila Dot/Icm T4SS effector translocation [59,68,70–72]. For this analysis we monitored the translocation of RalF, an IcmS/IcmW independent substrate, SidB and SidD, IcmS/IcmW dependent effectors, and SidF a transmembrane substrate that requires the Sec translocon for localization to the inner membrane [67,73,74].

CyaA was fused to the N-terminus of RalF, SidB, SidD, and SidF and expression of the fusion proteins in the WT, ΔdotA, Δdis3, and Δdis2 strains were verified by Western blot analysis using an anti-CyaA antibody (Supplemental Figure 2). All the fusion substrates were expressed in the WT, ΔdotA, Δdis3, and Δdis2 strains. While CyaA-SidD in the Δdis2 strain had lower levels of expression it was still detectable (Supplemental Figure 2). U937-derived human macrophages were infected with WT, ΔdotA, Δdis3, or Δdis2 strains expressing either CyaA, CyaA-RalF, CyaA-SidB, CyaA-SidD, or CyaA-SidF at an MOI of 30 for two hours. After two hours, any remaining extracellular bacteria were removed by washing the macrophages with PBS. The infected macrophages were lysed and intracellular cAMP levels were detected by ELISA (enzyme-linked immunosorbent assay). cAMP levels in lysates from macrophages infected by WT, ΔdotA, Δdis3, or Δdis2 strains expressing CyaA alone were low, indicating that CyaA is not an efficient T4SS substrate, (Figure 2D). As would be expected for positive and negative T4SS controls, lysates from macrophages infected with the WT strain expressing the CyaA fusion substrates had ~300–700-fold higher cAMP levels than the ΔdotA strains expressing the same fusion proteins (Figure 2D), validating that this assay can be used to monitor T4SS dependent translocation of effectors into the cytoplasm of human macrophages.

We next measured the ability of Δdis3 and Δdis2 strains to translocate the four CyaA fusion substrates into the macrophage cytoplasm. The Δdis3 strain could translocate all four substrates, with only a subtle, but statistically significant, decrease in its ability to translocate CyaA-SidF as compared to the WT strain (Figure 2D). The Δdis2 strain had a small but statistically significant decrease in its ability to translocate CyaA-RalF (IcmS/W independent) and CyaA-SidD (IcmS/W dependent) as compared to the WT L. pneumophila, although there were no changes in the ability to translocate CyaA-SidB (IcmS/W dependent) and CyaA-SidF (transmembrane substrate) (Figure 2D). While these results suggest that T4SSs lacking Dis3 or Dis2 may not translocate some substrates as efficiently as others, there was no clear trend correlating with IcmS/IcmW chaperone dependence or ability to translocate a transmembrane protein. Thus, Dis3 and Dis2 appear to subtly alter T4SS translocation efficiency of some substrates in human macrophages but, since Dis2 and Dis3 are not localized next to the central channel running through the T4SS, there is not an obvious structural mechanism.

When combined with our results showing Δdis3 and Δdis2 cells can still replicate in human macrophages (Figure 2C), these statistically significant, but relatively small differences in translocation efficiency, do not to alter the function of the Dot/Icm T4SS enough to affect the ability of L. pneumophila’s to infect human host cells.

Loss of Dis3 affects the organization of DotF in the OMC

Structurally, Dis3 forms the 13 radial spokes extending from the OMC and its C-terminus connects with the central portion of the OMC through protein–protein interactions with Dis1, DotK, DotD, and DotF (Figure 1F, H) [41]. The extensive protein–protein interaction network between Dis3 and other OMC components led us to predict that Dis3 could have a larger role than Dis2 in contributing to OMCC structural organization (Figure 1F, G). To more fully understand the contribution of Dis3 to the structural organization of the T4SS, OMCCs were purified from L. pneumophila Δdis3 cells using the same protocol used to isolate WT OMCCs [37,40,41]. Size exclusion chromatography (SEC) of OMCCs purified from Δdis3 cells showed that they eluted at the same volume as WT complexes (Supplemental Figure 3A). Mass spectrometry analysis of the pooled fractions used for structural analysis found that the mutant OMCCs uniformly lacked Dis3 (Supplemental Table 2). The negative stain images and cryo-EM 2D averages of the Δdis3 OMCCs clearly show that the mutant OMCCs are missing the 13 radial spokes seen in WT OMCCs (Figure 3A–D and Supplemental Figures 3B, C and 4A–D).

Figure 3. EM analysis of L. pneumophila Dot/Icm T4SS OMCCs purified from WT, Δdis2, and Δdis3 strains.

(A) Representative EM image of negatively stained Dot/Icm T4SS OMCCs purified from WT cells. Scale bar, 100 nm. A few representative particles are shown below micrograph, scale bar, 20 nm. (B) Representative 2D class averages of WT OMCCs in vitrified ice. Scale bar, 551 Å. Arrow points to the location of the Dis3 “spokes”. (C) Representative EM image of negatively stained Dot/Icm T4SS OMCCs purified from Δdis3 cells. Scale bar, 100 nm. A few representative particles are shown below micrograph, scale bar, 20 nm. (D) Representative 2D class averages of OMCCs in vitrified ice purified from Δdis3 cells. Scale bar, 556 Å. Arrow points to the location of the missing Dis 3 “spokes”. (E) Representative EM image of negatively stained Dot/Icm T4SS OMCCs purified from Δdis2 cells. Scale bar, 100 nm. A few representative particles are shown below micrograph, scale bar, 20 nm. (F) Representative 2D class averages of OMCCs in vitrified ice purified from Δdis2 cells. Scale bar, 654 Å. Arrow points to the location of the Dis3 “spokes”.

To more carefully examine the 3D structure of the mutant OMCC, we used single particle cryo-EM analysis to determine a 3.8 Å structure of the OMC (C13) and 3.6 Å resolution structure of the PR (C18) of OMCCs purified Δdis3 cells (Figure 4A and Supplemental Figures 5–7). The cryo-EM map of the mutant OMCC was clearly missing densities for Dis3 and DotF in the OMC; however, the PR showed no observable differences from WT cells at similar resolutions (Figure 4B). The dome is also present in the 3D reconstruction without applied symmetry and has the same shape and dimensions as the WT OMCC dome (Supplemental Figure 8A, B). The secondary structural features of the dome, composed of 16 copies of DotG, are not well-resolved even in the WT maps [41].

Figure 4. Comparing the structure of the L. pneumophila Dot/Icm T4SS OMCC purified from WT and Δdis3 strains.

(A) Composite EM density map of the Dot/Icm T4SS OMCC purified from Δdis3 cells (light pink) with one asymmetric unit shown in blue. Each map for the OMC and PR was refined with the appropriate applied symmetry (C13, or C18, respectively). The structure is rotated around the x-axis by the indicated degrees. (B) Asymmetric unit of the composite EM density map of the mutant OMCC (blue) rotated 90° on the y-axis. Molecular models of components found in the asymmetric unit of the WT OMCC, have been placed into the composite EM density map of the asymmetric unit (blue) of the mutant OMCC. PDB: 7MUC. Regions missing in the EM map are labeled. (C) Difference map (green) created by subtracting the mutant OMCC EM composite map from the lowpass filtered WT OMCC EM composite map. Structures are rotated 90° around x-axis. (D) Molecular models of components found in the asymmetric unit of the WT OMCC placed into the difference map (green). Structures are rotated 90° around y-axis. PDB: 7MUQ. The color for each protein is used in panels B-D.

Volume subtraction between the OMC and PR maps determined from Δdis3 strain and the filtered maps from the WT purifications revealed significant differences only in the OMC (Figure 4C and Supplemental Figure 9A–C). When compared with the WT OMC the OMC from Δdis3 cells was missing density for Dis3, DotF and an α-helix of Dis1 (Figure 4D). Despite the OMC losing DotF in the absence of Dis3, DotF is still associated with and structured in the PR (Figure 4A, B). This result explains the presence of DotF in the mass spectrometry analysis of the mutant sample (Supplemental Table 2). Thus, while Dis3 serves as a structural scaffold for DotF in the OMC, even the loss of these two proteins does not significantly alter the structural organization of the OMC and these proteins are not essential for translocation of Dot/Icm T4SS substrates or the replication of L. pneumophila in human macrophages. The role of the 13 radial spokes of Dis3 in T4SS function, other than serving as a scaffold for DotF, remain to be determined.

Loss of Dis2 does not impact the overall structural organization of the Dot/Icm T4SS OMC or PR

Next, we investigated the role of Dis2 in the overall structural organization of the OMCC. Dis2 interacts with DotK, DotC, and DotD in the OMC, but does not contact the central channel that runs through the OMC and PR created by DotG and DotH [41] (Figure 1B, F, G). Since the central channel has been hypothesized to support translocation of effectors [36,75,76], the role Dis2 has in altering the translocation efficiency of RalF and SidD is likely more complicated than simply changing the structure of the channel. To visualize the structure of the OMCC in the absence of Dis2, OMCCs were purified from the Δdis2 mutant and compared with OMCCs purified from WT cells. SEC showed that the OMCCs from Δdis2 cells eluted at the same volume as complexes from WT cells (Supplemental Figure 3D). LC-MS/MS analysis of pooled SEC fractions used for structural analysis showed that these OMCCs purified from Δdis2 cells contained all the previously identified Dot/Icm OMCC components with the exception Dis2 [41] (Supplemental Table 2).

Negative stain images and 2D averages of vitrified particles show that the OMCCs isolated from the Δdis2 mutant look very similar to WT OMCCs (Figure 3A, B, E, F and Supplemental Figures 3E, F and 4A, B, E, F). To more carefully characterize the structure of the mutant OMCC, we used single particle cryo-EM analysis. While the OMCCs purified from Δdis2 cells had preferred orientation in the vitrified ice, we were still able to determine a 6.1 Å resolution structure of the OMC (C13) and a 6.3 Å resolution structure of the PR (C18) (Figure 5A and Supplemental Figures 10–12). Overall, the OMCC from Δdis2 cells shares the same overall structural organization as the WT OMCC (Figures 5A, B), Volume subtraction was done between the mutant OMC and PR maps and the filtered maps from the WT purifications (Supplemental Figure 9D–F). This analysis showed that OMCs from the mutant strain lack density for Dis2 as well as density for the N-terminus of DotD (Figure 5C, D). The N-terminal tail of DotD shares 350 Ų of surface area with Dis2 in the neighboring asymmetric unit (Figure 1C–E), so it is not surprising that the tail is organized differently in this mutant OMCs. There were no differences detected in the PR at this resolution (Figure 5C, D). The dome is also present in the 3D reconstruction, either without or with applied symmetry (Supplemental Figure 8C, D), although, as with the dome in in the WT OMCCs (Supplemental Figure 8A), it remains one of the least defined regions of the map. In combination, these analyses show that the structure stability and organization of the OMCC does not require Dis2 even though the lack of Dis2 does have some effect on the translocation efficiency of some substrates.

Figure 5. Comparing the structure of the L. pneumophila Dot/Icm T4SS OMCC purified from WT and Δdis2 strains.

(A) Composite EM density map of the Dot/Icm T4SS OMCC purified from Δdis2 cells (light blue) with one asymmetric unit shown in yellow. Each map for the OMC and PR was refined with the appropriate applied symmetry (C13 or C18, respectively). The structure is rotated around the x-axis by the indicated degrees. (B) Asymmetric unit of the composite EM density map of the mutant OMCC (yellow). The view of the asymmetric unit is rotated ~220° on the y-axis in relation to the asymmetric unit (yellow) highlighted in the composite OMCC map (light blue) shown in panel A. Molecular models of components found in the asymmetric unit of the WT OMCC have been placed into the composite EM density map of the asymmetric unit (yellow) of the OMCC from Δdis2 cells. PDB: 7MUC. The location of the missing density for Dis2 is marked with a green arrow. (C) Difference map (purple) created by subtracting the mutant OMCC EM composite density map from the lowpass filtered WT OMCC EM composite density map. Map is rotated 90° around x-axis. (D) Molecular models of components found in the asymmetric unit of the WT OMCC placed into the difference map (purple) rotated 90° around the x-axis. PDB: 7MUC. The color for each protein is used in panels B–D.

Discussion

The L. pneumophila Dot/Icm T4SS is an expanded secretion system required for pathogenesis [10–15]. Single particle cryo-EM analysis of the Dot/Icm OMCC showed that it contains DotD, DotH/IcmK, DotG/IcmE, DotF/IcmG, DotC, DotK/IcmN, Dis1, Dis2, and Dis3 [40,41] (Supplemental Table 1). Three of these components, DotD, DotH, and DotG, have orthologs in other bacteria (VirB7, VirB9, and VirB10, respectively), while the remaining proteins are specific to the order Legionellaes with no obvious homologs in organisms with prototypical T4SSs [41,77]. While the specific functions of any of these proteins in the context of the Dot/Icm T4SS OMCC are not known, DotD, DotH/IcmK, DotG/IcmE, DotF/IcmG, DotC, and DotK/IcmN were isolated in genetic screens that disrupted intracellular replication [11,12,14,58,78]. Dis1, Dis2, and Dis3 were only recognized as being core members of the Dot/Icm T4SS through structural analysis of biochemically isolated complexes [40,41]. In this study, we have addressed the roles of Dis2 and Dis3 in the structural organization and function of the L. pneumophila Dot/Icm T4SS OMCC.

Dis2 is a Legionella-specific protein encoded in roughly half of the reference Legionella genomes deposited in the BV-BRC [54] (Supplemental Figure 1C). Many of the Legionella species lacking Dis2 are still pathogenic to humans, indicating there may be a functionally redundant protein in these species or that Dis2 is not critical for human infection. These conclusions are supported by the initial screens investigating L. pneumophila proteins required for intracellular replication in human macrophages which did not uncover Dis2 [13,58,78]. We found that in the absence of dis2, L. pneumophila replicates in human macrophages and has a growth defect on CYE(T) agar supplemented with 100 mM NaCl, both phenotypes that indicate a functional T4SS. Translocation of two effector proteins investigated in this study, RalF and SidD, showed subtle, but statistically significant, decreases in translocation efficiency in the Δdis2 strain. Structural characterization of the OMCCs purified from Δdis2 strains showed both the OMC and PR were very similar to WT OMCCs, and only the major differences was the missing density for Dis2 in the OMC and missing density for the N-terminus of DotD. Although Dis2 also contacts both DotK/IcmN and DotC, neither of these proteins change conformation in the absence of Dis2. Thus, it is not clear why the lack of Dis2 in the structure leads to even the subtle change in translocation efficiency we observed for certain substrates. One possible explanation is that there could be fewer functional Dot/Icm T4SSs in the Δdis2 strain leading to these subtle differences in translocation.

The high degree of conservation of Dis3 among Legionella species indicates that it likely serves an important role in the physiology of Legionella (Supplemental Figure 1). In the present study we found that in absence of dis3, L. pneumophila cells replicated normally in cultured human macrophages, were sensitive to the presence of NaCl on agar plates and translocated all tested substrates with normal efficiency. Altogether, these data indicate that Dis3 is not required for T4SS activity. In characterizing OMCCs purified from Δdis3 strains we found densities for Dis3 and DotF absent in the OMC. Interestingly, density for DotF was still present in the PR of the mutant OMCCs. Although Dis3 is not required for human infections, a ΔdotF mutant has been found to either slightly decrease or be required for efficient intracellular replication in human macrophages, depending on the study [37,79]. Isolated Dot/Icm T4SSs from ΔdotF cells appeared to still form core complexes (or at least OMCs) when visualized by negative stain EM [37] and cryo-ET images of the Dot/Icm T4SS in ΔdotF cells shows what looks like partially assembled complexes, with most of the structural differences being in the PR [80]. Because OMCCs purified from the Δdis3 strain still had a structured PR that retained DotF, we predict that the PR localized DotF is most important for the function of the Dot/Icm T4SS.

Although Dis3 is not vital for intracellular survival in human macrophages, it is required for efficient internalization and intracellular replication in the protozoan hosts Hartmannella veriformis and Naegleria lovaniesis [81,82]. Replication of L. pneumophila in protozoans is crucial for its growth in the environment [83]. Interestingly, cryo-ET analyses showed that the Dot/Icm T4SSs are localized to concave indentations in the OM [51,84]. So far the C. burnetii T4SS is the only other T4SS examined by cryo-ET that also appears to be found with membrane indentations [48]. C. burnetii is intracellular pathogen in the order Legionellae and possesses a homolog to Dis3 that has been postulated to form the characteristic 13 radial spokes of the OMCC visualized in this pathogen by cryo-ET [48]. Since these “frill-like” extensions make extensive contacts with the OM, it is possible that one function of Dis3 proteins is causing and/or stabilizing the indentations seen in the OM in L. pneumophila and C. burnetii, a question that will need to be addressed in the future by cryo-ET analysis of Δdis3 cells. It may also be of future interest to use cryo-ET to visualize whether Dis3 alters the assembly of the Dot/Icm T4SS in the context of the bacterial membrane. However, our finding that L. pneumophila Δdis3 cells can still replicate in human macrophages suggests that if this is a function of Dis3, it might be more physiologically important in the context of amoebas versus human macrophages.

It was surprising that OMCCs isolated from L. pneumophila strains lacking either Dis2 or Dis3 retained their overall structural organization and ability to translocate effectors. The resiliency of the intricately organized Dot/Icm T4SS to the loss of core components contrasts with the sensitivity of the H. pylori Cag T4SS OMCC, another expanded T4SS, to loss of OMC components. In H. pylori, deletion of cagT (VirB7 homolog) or cagM (species-specific) led to the complete collapse of a structured OMC, although the PR remained organized in both deletion strains [85]. Additionally, the lack of a structured OMC in these deletion stains disrupted that ability of the Cag T4SS to translocate the effector protein, CagA [85,86]. The structural resiliency of the Dot/Icm T4SS compared to the Cag T4SS is mirrored in the redundancy (or lack of redundancy) of L. pneumophila’s and H. pylori’s respective effectors. While the L. pneumophila genome encodes a large repertoire of hundreds of Dot/Icm T4SS effectors, with many of them redundant for virulence [19,21,71,87–90], the H. pylori Cag T4SS secretes one protein, CagA [91]. Thus, the promiscuity of effector secretion by the Dot/Icm T4SS coupled with its structural resiliency to perturbation highlights the importance of this system for L. pneumophila’s life cycle and, likely, its ability to thrive in a wide variety of environments and replicate in many types of host cells.

In conclusion, our analyses of the importance of Dis2 and Dis3 for the structure and function of the Dot/Icm T4SS, show that neither of these species-specific proteins are required for the overall structural organization of the OMCC, the ability of the Dot/Icm T4SS to translocate, or for the intracellular replication of L. pneumophila in human macrophages. While L. pneumophila cells lacking either Dis3 or Dis2 exhibited subtle defects in translocation of some effector proteins, these defects were not significant enough to affect intracellular replication, perhaps due to the sheer number and functional redundancy of Dot/Icm effectors. While Dis2 and Dis3 do not have a significant role in the ability of L. pneumophila to replicate in human cells, Dis3 has been shown to be important for intracellular replication in protozoan hosts [81,82]. Thus, it is possible that species-specific T4SS components in L. pneumophila may have more functional relevance in the context of their natural hosts. Additional studies of L. pneumophila internalization and replication in amoebae will be required to explore this hypothesis.

Materials and Methods

Protein sequence conservation analysis

Protein BLAST (pBLAST) and sequence alignments were performed using the Bacterial and Viral Bioinformatics Resource Center (BV-BRC) [54]. The amino acid sequences of Dis1 (Ipg0657), Dis2 (Ipg0823), and Dis3 (Ipg2847) were queried against 63 reference genomes of Legionella (listed in Supplemental file 1); 50% protein sequence similarity was defined as a hit.

Bacterial growth conditions

L. pneumophila laboratory strain Lp02 [78,92], a thymidine auxotroph derived from the clinical isolate Philadelphia-1 was used at WT (Lp02), and the T4SS null strain, Lp03 (ΔdotA) [78] were maintained on charcoal yeast extract thymidine (CYE (T)) agar plates, buffered with N-(2-acetamido)-2-a minoethane-sulfonic acid (ACES, Thermo Scientific) supplemented with 40 μg/mL L-cysteine (Fischer), 10 μg/mL thymidine (Sigma Aldrich), 13.5 μg/mL ferric nitrate (Sigma Aldrich) at 37 °C [93]. Broth cultures were grown in ACES buffered yeast extract media (AYE(T)), supplemented with 40 μg/mL L-cysteine, 10 μg/mL thymidine, 13.5 μg/mL ferric nitrate at 37 °C with aeration [93].

E. coli strains DH5α (Invitrogen), JM109 (Promega), and DY330 were maintained on LB (Luria Bertani, Fisher Bioreagents) agar plates at 37 °C. For blue-white selection of clones of JM109 carrying pGEMT plasmids, LB plates were supplemented with 0.5 mM IPTG and 80 μg/mL Bluo-gal (5-Bromo-3-indolyl-β-D-galactopyrano side, GoldBio). Antibiotics were supplemented as appropriate at the following final concentrations for E. coli and L. pneumophila: ampicillin 100 μg/mL, chloramphenicol 25 μg/mL or kanamycin 50 μg/mL.

L. pneumophila strain construction

Strains used in this study are described in Supplemental Table 3. Primers used in this study are detailed in Supplemental Table 4. DNA and protein sequences were analyzed using Lasergene (DNAStar), using the Clustal method of Lasergene Megalign for sequence alignments. Marked gene deletions strains of L. pneumophila Lp02 (WT) were constructed as described [55]. In brief, dis2 with 850 bp of upstream and downstream flanking regions was amplified, while dis3 with 800 bp upstream and downstream flanking regions was amplified and ligated into the pGEMT-Easy vector, creating the pGEMT-dis2 and pGEMT-dis3 vectors. Then, using lambda red recombineering [94], either the dis2 or dis3 gene was replaced with a kanamycin (kan) resistance cassette, making the pGEMT-dis2Δ::kan and pGEMT-dis3Δ::kan plasmids. Each plasmid was used as a template to amplify dis2Δ::kan and dis3Δ::kan using primer pairs Dis2^a and Dis3^a. At least 1 μg of column-purified PCR product was transformed into L. pneumophila by means of natural transformation at 30 °C. After two days of growth without antibiotics at 30 °C, patches of L. pneumophila were restruck onto charcoal yeast extract thymidine (CYE(T)) agar containing 50 μg/mL kanamycin and incubated at 37 °C for 4–5 days. Gene deletion was confirmed using colony PCR with primer pairs Dis2^a for dis2::kan and Dis3^a for dis3::kan and analyzed by Sanger Sequencing.

Plasmid construction

Plasmids used in this study are listed in Supplemental Table 3. Plasmids used for CyaA fusion protein expression were constructed from plasmid pXDC61 [95]. The plasmid pXDC61, encoding fusion protein Bla(m)-RalF under control of the tac promoter was linearized to remove the beta-lactamase fusion, and add homology to CyaA using primer pair pXDC61^c. The catalytic domain of CyaA (1200 bp) was synthesized by Integrated DNA Technologies (IDT, Newark, NJ, USA) in a pUCIDT cloning vector. CyaA was amplified using primer pair CyaA^e and CyaA-RalF was assembled using In-fusion assembly (Takara Bio). To create CyaA-SidB, CyaA-SidD, and CyaA-SidF, CyaA-RalF was linearized to remove RalF in two pieces, using primer pairs CyaA-RalF part 1^c and CyaA-RalF part 2^c. SidB, SidD, and SidF were amplified from L. pneumophila genomic DNA, with homology to linearized CyaA-RalF using primer pairs, SidB^a,c, Sid^a,c, and SidF^a,c. Plasmids were assembled using In-fusion assembly (Takara Bio).

Transformation of L. pneumophila strains

Plasmids were transformed into L. pneumophila strains using established protocols [55]. In brief, L. pneumophila cells were cultured in AYE(T) media, until OD₆₀₀ = 0.9–1.2 and made electrocompetent as described [55]. Electrocompetent L. pneumophila were electroporated with 2 μL of plasmid preparation at 1.8 kV, 100 Ω, and 25 μF using the BioRad Gene-Pulser Xcell Electroporation system. After a 10–12 h outgrowth in AYE(T) media transformed cells were plated on CYE(T) agar plates containing 25 μg/ml chloramphenicol and incubated for 4–5 days at 37 °C.

Western blotting

Expression of CyaA-fusion proteins was induced by the addition of 1 mM isopropyl ß-D-thiogalactopyranoside (IPTG) to the culture medium for 2 h. Cultures were normalized to OD₆₀₀ of 1.0, the culture pelleted and resuspended in 2X Laemmli sample buffer (BioRad), sonicated, and boiled for 10 min. Molecular weight markers (BioRad) and proteins were separated on a 4–20% polyacrylamide gel (BioRad) under denaturing conditions, transferred to a nitrocellulose membrane (Thermo Scientific), and blocked with 5% BSA in Tris buffered saline (TBS) supplemented with 0.1% Tween. Membranes were probed with mouse anti-CyaA primary antibody 3D1 (Santa Cruz Biotechnology Inc.), or rabbit anti-ICDH primary antibody (Kerafast). Horseradish peroxidase conjugated to anti-rabbit or anti-mouse IgG (Promega) was used as a secondary antibody. Signal was detected using the BioRad Chemidoc imager.

Cell culture of human macrophages

U937 monocyte cells were obtained from ATCC (CRL-1593.2) and maintained as replicative non-adherent monocyte-like cells in RPMI-1640 GlutaMAX (Gibco) supplemented with 10% (v/v) heat inactivated FBS (Corning), 1,000 U/mL penicillin, and 1,000 U/mL streptomycin. U937 cells were differentiated with 50 ng/μL phorbol 12-myristate 13-acetate (PMA) for 72 h to become adherent, non-replicative cells which exhibit characteristics similar to tissue macrophages [96]. U937 cells were maintained at 37 °C with 5% CO₂.

Intracellular replication assay

4 × 10⁴ U937 differentiated macrophages were seeded in each well of a 96-well plate (Corning Costar). 24 h prior to the start of intracellular replication assays in U937 macrophages the RPMI-1640 GlutaMAX media was replaced with infection media (RPMI-1640 GlutaMAX, 10% v/v heat-inactivated FBS, 100 μg/mL thymidine). On the day of infection, infection medium was aspirated and replaced with infection medium containing 1 MOI of the indicated L. pneumophila strain. The plate was centrifuged at 800g for 10 min at room temperature to synchronize infection and then incubated at 37 °C with 5% CO₂ for 1 h. After 1 h of coincubation, extracellular bacteria were removed using two successive PBS washes and infected cells were incubated with infection medium with no L. pneumophila. At each indicated time point, triplicate wells were lysed with saponin at a final concentration of 0.05% (w/v) in PBS. Lysate was serially diluted in PBS and plated onto CYE(T) agar + appropriate antibiotics. CFUs were counted by hand after 4–6 days of incubation at 37 °C. Reported values are the average of four independent biological replicates measured in technical triplicate.

CyaA translocation assay

1 × 10⁶ U937 macrophages were seeded in each well of a 12-well tissue culture plate (CellTreat). 24 h prior to the start of intracellular replication assays in U937 macrophages, RPMI-1640 GlutaMAX was replaced with infection medium supplemented with IPTG (RPMI-1640 GlutaMAX, 10% v/v heat inactivated FBS, 100 μg/mL thymidine, 1 mM IPTG). L. pneumophila cultures were inoculated from fresh plates 24–48 h prior to infection and grown to OD₆₀₀ > 2.0. On the day of infection, infection media was aspirated and replaced with infection media containing 30 MOI of each indicated L. pneumophila strain. The plate was centrifuged at 800g for 10 min at room temperature to synchronize infection. The infection progressed for two hours at 37 °C, 5% CO₂, then any remaining extracellular bacteria were removed by two successive PBS washes.

Intracellular cAMP in the infected U937 macrophages was measured using a cAMP ELISA (Cayman Chemical Cyclic AMP ELISA kit, Item No. 581001) performed according to manufacturer’s instructions. To prepare cell lysate all excess PBS was removed then 100 μL of 0.1 M HCl in ultrapure DNase/RNase-Free distilled water (Invitrogen) was added to each well. After 20 min at room temperature, cells were dissociated by scraping and the lysate was collected. Lysate was centrifuged at 1,000g for 10 min to pellet cellular debris and the supernatant was stored at −80 °C until the ELISA was performed. Lysates were diluted 10-fold in cAMP reaction buffer included in the ELISA kit before being assayed. Absorbance at 415 nm was measured using the Perkin Elmer Envision plate reader. Reported values are the average of three independent biological replicates measured in technical duplicate.

T4SS OMCC purifications

OMCCs were purified as previously described from post-exponential (OD > 2.5) L. pneumophila Lp02 (WT), Δdis2, or Δdis3 strains [40,41]. In brief, cells were pelleted and resuspended in buffer containing 150 mM Trizma base pH 8.0, 500 mM sucrose, and EDTA-free Complete protease inhibitor (Roche) at 4 °C. The suspension was incubated on the benchtop, with stirring, until it reached ambient temperature. Phenylmethylsulfonyl fluoride (PMSF, final concentration 1 mM), Ethylenediaminetetraacetic acid (EDTA, final concentration 1 mM), and lysozyme (final concentration 0.4 mg/mL) were added to digest the cell wall. Then triton X-100 (final concentration 1%) with AG501-X8 resin (BioRad) was added dropwise to solubilize bacterial membranes, followed by MgSO₄ (final concentration 3 mM), DNasel (final concentration 10 μg/mL), and EDTA (final concentration 10 mM), and then the pH was adjusted to 10.0 using NaOH. The remaining steps were conducted at 4 °C. The cell lysate was centrifuged at 12,000g for 20 min. The supernatant containing the cleared lysate was ultracentrifuged at 100,000g for 30 min to pellet membrane complexes. The membrane complex pellets were resuspended and soaked overnight in a small volume of TET buffer (10 mM Trizma base pH 8.0, 1 mM EDTA, 0.1% Triton X-100). The resuspended sample was centrifuged at 14,000g for 30 min to pellet remaining debris. The supernatant was then ultracentrifuged at 100,000g for 30 min to pellet the remaining membranes. The resulting pellet was resuspended in TET and solubilized membrane complexes were separated using a Superose 610/300 column by SEC in TET buffer supplemented with 150 mM NaCl on an AKTA Pure system (GE Life Sciences). Fractions were collected and screened by negative stain EM. Fractions containing OMCCs were combined. Protein concentration was determined using the MicroBCA assay (ThermoFisher) and a volume corresponding to 50–100 μg was used for liquid chromatographytandem mass spectrometry (LC-MS/MS). Mass spectrometry was performed as described [97].

Negative stain EM

Negative stain EM was performed using established methods [98]. 400-mesh copper grids covered with carbon-coated collodion film (Electron Microscopy Sciences) were glow-discharged for 30 s at 5 mA in a PELCO easiGlow glow discharge unit (Ted Pella). 3.5 μL of purified sample was adsorbed to the grids and incubated for 1 min at 25 °C. The grids were then washed twice with water, negatively stained with 0.75% (wt/vol) uranyl formate solution, and blotted until dry. Negative stain images were taken using the Morgagni TEM microscope (FEI) operating at 100 kV and 22,000× magnification.

Cryo-EM sample preparation and data collection

To prepare OMCC purifications for cryo-EM analysis, 3.5 μL of each purification was applied 2–3 times to a glow discharged Quantifoil 2/2200 mesh copper grid with ultrathin continuous carbon (Quantifoil). Each application to the grid was incubated for 60 s at room temperature and excess liquid was removed by tapping the grid. After the final sample application, the grid was washed in three drops of water. Grids were vitrified by plunging into a slurry of liquid ethane using a Vitrobot (Mark IV, ThermoFisher) at room temperature and 100% humidity.

Micrographs of OMCCs purified from Lp02 (WT) cells were collected at the Stanford-SLAC Cryo-EM Facility (Menlo Park, CA) using Titan Krios microscopes (Thermo Fisher, Waltham, MA) operated at 300 keV. The images were collected with a K3 Summit DED equipped with a BioQuantum energy filter operating in counting mode, at a nominal magnification of 81,000×, corresponding to a pixel size of 1.1 Å. The energy slit of the BioQuantum was set at a width of 15 eV. The total dose was 50 e/Ų, fractionated over 33 frames. Data were collected using EPU software (Thermo Fisher, Waltham, MA) with a nominal defocus range set from −1.5 to −2.1 μm.

Images of OMCCs purified from Δdis2 and Δdis3 cells were collected using a Titan Krios microscope (ThermoFisher Scientific) operated at 300 kV. The images were collected with K3 Summit Direct Electron Detector (Gatan) equipped with a BioQuantum energy filter (Gatan) operating in counting mode, at a nominal magnification of 81,000×, corresponding to a pixel size of 1.08 Å. The BioQuantum energy slit was set at a width of 20 eV. The total dose for each image was 60 e/Ų fractionated over 60 frames. Micrographs were collected using SerialEM software [99] with a nominal defocus range set from −1.0 to −2.5 μm.

Cryo-EM data processing

Cryo-EM data collection, refinement, and statistics are summarized in Supplemental Table 5. cryoSPARC v.3.0.1 was used to process images of OMCCs purified from WT cells and cryoSPARC v.4.2.1 was used for all other image processing [100].

For OMCCs purified from WT cells, 2,895 movies were aligned and dose-weighted using Motioncor2 [101]. The contrast transfer function (CTF) values were determined using CTFFind4 [102]. A representative aligned and dose weighted micrograph is shown in Supplemental Figure 4A. 51,421 particles were picked from 2,853 micrographs using templates described previously [41] and extracted with box size 510 pix (551 Å). After iterative 2D cleaning 8,181 particles remained. The full panel of 2D classes are shown in Supplemental Figure 4B. 4,248 and 2,166 movies were collected for OMCCs purified from Δdis3 and Δdis2 strains respectively. Movies were dose weighted and aligned using patch motion correction. The contrast transfer function (CTF) values were determined using patch CTF correction. Representative aligned and dose weighted micrographs are shown in Supplemental Figure 4C (Δdis3) and Supplemental Figure 4E (Δdis2).

For analysis of OMCCs purified from Δdis3 cells an initial ~150,000 particles were picked using reference free blob picking which were then used to create templates for template picking (Supplemental Figure 6A). Template picking resulted in ~320,000 particles which were extracted with a 510 pixel (pix) box (551 Å, 1.08 Å/pix). After iterative 2D classification, class averages with clear secondary structural features were retained (Supplemental Figure 4D), corresponding to 15,121 particles. The unbinned images (1.08 Å/pix) were used to generate a reference-free ab initio 3D model with no applied symmetry (C1). 3D classification did not yield any additional structures. The 3.8 Å OMC 3D structure was determined by imposing C13 symmetry in a Non-Uniform (NU) refinement, and then by a local refinement using a mask focused on the OMC (Supplemental Figures 6B and 7A–D). The 3.6 Å resolution map of the PR was determined by applying 18-fold symmetry (C18) in a NU refinement and then by a local refinement using masks specific for the PR (Supplemental Figures 6C and 7E–H). 3D classification did not yield any additional structures for either the OMC or PR.

For analysis of OMCCs from Δdis2 cells, an initial ~300,000 particles were picked using reference free blob picking (Supplemental Figure 11A). Those particles were extracted and 2D classified to yield templates used for improved particle picking. 468,423 particles were picked with templates, extracted with a box size of 600 pix, and binned by 2 (2.16 Å/pix). After iterative 2D classification, class averages with clear secondary structural features were retained (Supplemental Figure 4F), corresponding to 9,017 particles. These particles were extracted with a box size of 600 pix (654 Å, 1.08 Å/pix). The unbinned images (1.08 Å/pix) were used to generate a reference-free ab initio 3D model with no applied symmetry (C1) (Supplemental Figure 11). A 6.1 Å resolution map of the OMC was determined using the following steps: C13 symmetry was imposed on the ab-initio 3D reconstruction using NU refinement. A mask encompassing the PR and dome was created from this map and used in particle subtraction, followed by a final local refinement performed using a mask focused around the OMC (Supplemental Figures 11B and 12 A–D). A 6.4 Å resolution map of PR was determined by applying C18 symmetry on the ab-initio 3D reconstruction using NU refinement. Masks were created to subtract signal corresponding to the dome, OMC and unstructured density using particle subtraction. Local Refinement with C18 symmetry was performed using a mask corresponding to the PR (Supplemental Figures 11C and E–H). 3D classification did not yield any additional structures for either the OMC or PR.

Statistical analysis

Statistical analyses were conducted using GraphPad Prism 10. Analysis of variance (ANOVA) with Dunnett’s multiple comparison test was used to compare results between different strains of bacteria in the salt sensitivity, intracellular replication, and translocation assays. The significance levels are indicated with asterisks: * indicates P < 0.05, ** indicates P < 0.01 *** indicates P < 0.0002, **** indicates P < 0.0001.

Structural analysis and visualization

Structures were visualized using ChimeraX v1.8 [103]. To create the composite map of the OMCC purified from Δdis3 cells, the OMC map was the result of the C13 locally masked refinement (Supplemental Figure 6B), and the PR was the result of the C18 locally masked refinement (Supplemental Figure 6C). To create the composite map of the OMCC purified from Δdis2 cells, the OMC map was the result of the C13 locally masked refinement (Supplemental Figure 11B), and the PR was the result of the C18 locally masked refinement (Supplemental Figure 11C). A composite model for the Dot/Icm OMCC (Figure 1) was created using PDBs 7MUQ (chains AG-PG, dome) and 7MUC (OMC and PR. The models were aligned using the 13 copies of DotC present in the 3D OMCC maps from both Δdis2 and Δdis3 strains). Otherwise, the model of the OMC and PR used for mutant EM density analysis was 7MUC.

For difference map calculations, the WT experimental volumes previously determined [41] for the OMC (C13 symmetry; EMDB 24005) and the PR (C18 symmetry; EMDB24006) were lowpass filtered in cryoSPARC to the resolutions we obtained for our Δdis2 or Δdis3 maps (Supplemental Figure 9). The 3D cryo-EM density of each experimental map was subtracted from the lowpass-filtered WT volume (Supplemental Figure 9). The results were visualized in ChimeraX at a threshold of 0.02 with surface dust hidden (size 10.8 Å) (Supplemental Figure 9C, F).

Appendix A. Supplementary material

Supplementary material to this article can be found online at https://doi.org/10.1016/j.jmb.2025.169310.

Data availability

Cryo-EM maps have been deposited in the Electron Microscopy Data Bank (EMDB) under EMD-70321 (Δdis3 OMC), EMD-70322 (Δdis3 PR), EMD-70323 (Δdis2 OMC), and EMD-70324 (Δdis2 PR).