Applying Live Cell Imaging and Cryo-Electron Tomography to Resolve Spatiotemporal Features of the Legionella pneumophila Dot/Icm Secretion System

Summary

Imaging of bacterial cells is an emerging systems biology approach focused on defining static and dynamic processes that dictate the function of large macromolecular machines. Here, integration of quantitative live cell imaging and cryo-electron tomography is used to study Legionella pneumophila type IV secretion system architecture and functions.

Abstract

The Dot/Icm secretion system of Legionella pneumophila is a complex type IV secretion system (T4SS) nanomachine that localizes at the bacterial pole and mediates the delivery of protein and DNA substrates to target cells, a process generally requiring direct cell-to-cell contact. We have recently solved the structure of the Dot/Icm apparatus by cryo-electron tomography (cryo-ET) and showed that it forms a cell envelope-spanning channel that connects to a cytoplasmic complex. Applying two complementary approaches that preserve the native structure of the specimen, fluorescent microscopy in living cells and cryo-ET, allows in situ visualization of proteins and assimilation of the stoichiometry and timing of production of each machine component relative to other Dot/Icm subunits. To investigate the requirements for polar positioning and to characterize dynamic features associated with T4SS machine biogenesis, we have fused a gene encoding superfolder green fluorescent protein to Dot/Icm ATPase genes at their native positions on the chromosome. The following method integrates quantitative fluorescence microscopy of living cells and cryo-ET to quantify polar localization, dynamics, and structure of these proteins in intact bacterial cells. Applying these approaches for studying the Legionella pneumophila T4SS is useful for characterizing the function of the Dot/Icm system and can be adapted to study a wide variety of bacterial pathogens that utilize the T4SS or other types of bacterial secretion complexes.

Introduction

Legionella pneumophila (L. pneumophila), the etiological agent of Legionnaires' disease, inhabits freshwater reservoirs, where the bacteria propagate by infecting and replicating within aquatic free-swimming protozoa. L. pneumophila causes disease outbreaks in humans when inhalation of aerosolized bacteria from potable water sources occurs. In infected cells, subversion of host pathways allows L. pneumophila to delay endocytic maturation of the vacuole in which it resides and to promote biogenesis of a cellular compartment that supports bacterial replication. This process is driven by a specialized bacterial type IVB secretion system (T4BSS) known as Dot/Icm and its repertoire of over 300 "effector" proteins that are translocated into the host cytosol during infection to facilitate manipulation of cellular functions^1,2,3,4,5. Mutants lacking a functional Dot/Icm apparatus fail to deliver effectors into the host cytosol, are defective for intracellular replication, and are avirulent in animal models of disease^6,7.

Many bacterial species have developed extremely complex and dynamic multicomponent machines that are required for infection processes. Other T4BSS like the Dot/Icm system are also essential for intracellular replication of bacterial pathogens such as Coxiella burnetii and Rickettsiella grylli. Although T4BSS are evolutionarily related to prototypical type IVA systems, which mediate DNA transfer and can deliver a limited repertoire of effector proteins, the Dot/Icm system has nearly twice as many machine components and delivers a wide variety of effectors. Presumably, this expansion in the number of components has enabled the Dot/Icm apparatus to accommodate and integrate new effectors easily^8,9.

We recently used cryo-electron tomography (cryo-ET) to solve the structure of the Dot/Icm apparatus in situ and showed that it forms a cell envelope-spanning channel that connects to a cytoplasmic complex. Further analysis revealed that the cytosolic ATPase DotB associates with the Dot/Icm system at the L. pneumophila cell pole through interactions with the cytosolic ATPase DotO. We have discovered that DotB displays a cytosolic movement in most bacterial cells, indicating that this ATPase is present in a dynamic cytosolic population but also associates with the polar Dot/Icm complexes. In addition, DotO forms a hexameric assembly of DotO dimers associated with the inner membrane complex, and a DotB hexamer joins to the base of this cytoplasmic complex. The assembly of the DotB-DotO energy complex creates a cytoplasmic channel that directs the translocation of substrates through the T4SS (Figure 1)¹⁰.

Despite these recent advances, little is known about how the Dot/Icm system functions and how each protein assembles to form an active apparatus⁸. Uncovering the regulatory circuitry of the Dot/Icm T4SS is fundamental to understanding the molecular mechanisms of host-pathogen interactions. Therefore, we discuss how to use live cell microscopy and cryo-ET to detect and characterize essential L. pneumophila Dot/Icm system components that are tagged with super-folder GFP (sfGFP). Using quantitative fluorescence microscopy, the polar localization of DotB will be defined in a wild type background or when the type IV system is deleted. Time-lapse microscopy will be used to quantify differences in localization and dynamics between the Dot/Icm cytosolic ATPases.

The combined application of two complementary approaches such as live imaging and cryo-ET provides an advantage compared to other in vitro systems. Both methods are performed in intact cells and preserve the natural environment of the T4BSS, thus minimizing disruption of the native structure during sample preparation. Because overexpression of proteins may impair the stoichiometry of the secretion apparatus, sfGFP fusions are returned via allelic exchange to the Legionella chromosome so that each fusion is encoded in single copy and the expression is driven by the endogenous promoter. Visualization of chromosomally-encoded fusions enables quantification of the exact level of protein being expressed at a defined time point. Cryo-ET also has many advantages for determining the structure of secretion systems. The most notable advantage is that cryo-ET samples are comprised of frozen intact cells that preserve native complexes in the context of bacterial cell architecture. Consequently, cryo-ET may be preferable to biochemical purification approaches, which extract membrane complexes and may strip peripheral proteins from the core apparatus or modify the overall structure. In addition, tagging a protein of interest with a bulky protein such as sfGFP adds a mass that is detectable by cryo-ET and can assist with mapping the different subcomplexes of the Dot/Icm apparatus onto the structure obtained by cryo-ET.

This approach is a powerful tool for uncovering structural information about multimolecular complexes that assemble in the bacterial cell membrane. The interpretation of structures elucidated using these techniques will help the field understand how T4BSS components function, why so many components are required for function, how the components interact within the greater complex, and what functions these subassemblies perform.

Protocol

NOTE: All procedures involving the growth, manipulation, and imaging of L. pneumophila should be performed in a biological safety level 2 laboratory in compliance with local guidelines.

1. Insertion of sfGFP into L. pneumophila Chromosome Using Allelic Exchange and Double Selection Strategy (Figure 2, Figure 3)

1. Clone into the gene replacement vector pSR47S¹¹ the following sequence: the 1,000 bp upstream of the site of interest, then the sfGFP sequence, then the 1,000 bp downstream of the site of interest (Figure 2). The sfGFP sequence should be placed in frame to the N-terminus or C-terminus ends with a linker that contains four to eight amino acids. Transform the resulting vector into E. coli DH5αλpir. Later, streak L. pneumophila (the recipient) for single colonies on charcoal-yeast extract (CYE) agar¹² containing 100 µg/mL streptomycin and grow for 5 days at 37 °C (Figure 3).
2. Streak L. pneumophila on CYE-agar-streptomycin and grow for 2 days at 37 °C (heavy patch)¹⁰. Streak E. coli DH5α transformed with pRK600 helper plasmid (helper)¹³ on LB agar containing 25 µg/mL chloramphenicol. Streak the E. coli DH5αλpir (donor) on LB agar containing 50 µg/mL kanamycin.
3. Perform triparental mating: incubate a colony of the helper, a colony of the donor, and the recipient by overlaying patches of the three strains on a CYE agar plate without selection and incubating for 4–8 h at 37 °C. As negative controls incubate a helper+recipient strain mix and a donor+recipient strain mix for the same periods of time.
4. Resuspend the mating reactions in 500 µL of ddH₂O. Plate 20 µL and 50 µL of the reactions on CYE agar containing 100 µg/mL streptomycin and 10 µg/mL kanamycin and grow for 5 days at 37 °C. Streak four of the resulting clones on CYE agar containing 100 µg/mL streptomycin and grow for 5 days at 37 °C.
5. Streak 16 clones on CYE agar containing 5% sucrose and 100 µg/mL streptomycin and grow for 5 days at 37 °C. Then, streak 32 of these clones on CYE agar containing 100 µg/mL streptomycin and on CYE agar containing 100 µg/mL streptomycin and 10 µg/mL kanamycin and grow for 5 days at 37 °C.

2. Isolation of Clones that Integrated sfGFP into the L. pneumophila Chromosome

1. Streak clones that were sensitive to kanamycin on CYE-agar-streptomycin plates and confirm the insertion of sfGFP into the chromosome with colony PCR. Use primers that are complementary to the sfGFP gene and to the chromosomal region of interest to amplify the insertion junction.
   1. Mix 0.5 µL of each of the 10 µM primers solutions and one colony to a final volume of 12.5 µL and denature for 10 min at 95 °C. Cool on ice for 10 min, add 12.5 µL of 2x PCR master mix solution, and perform a PCR analysis.
2. Grow heavy patches of the isolated colonies on CYE-agar-streptomycin plates for 2 days at 37 °C. Examine the expression levels and stability of the sfGFP fusions by immunoblotting with an anti-GFP antibody.

3. Live Cell Imaging of L. pneumophila with Fluorescently Tagged Dot/Icm Components

1. Preparation of agarose pads
   1. Make about 30 mL of a 1% low-melt agarose solution in water. Microwave in a glass flask for about 90 s, swirling occasionally, until the agarose is completely dissolved.
   2. Place two 22 x 22 x 0.15 mm³ glass slides on the edge of a 25 x 75 x 1.1 mm³ glass slide, one on top of the other. Stack two more 22 x 22 x 0.15 mm³ glass slides on the other edge.
   3. Pipette about 1 mL of the molten agarose into the center slide between the two upper glass slides, then place another 25 x 75 x 1.1 mm³ slide on top of the molten agarose. Try to avoid the formation of air bubbles. Cool the slides at 4 °C for 15 min.
   4. Using a scalpel or razor blade, gently cut the pad into small squares, ~5 x 5 mm². Fix a double-sided adhesive 17 x 28 x 0.25 mm³ frame on a 25 x 75 x 1.1 mm³ glass slide and place several pads on the slide.
2. Image acquisition
   NOTE: The following steps are described for a microscope that is under the control of SlideBook 6.0 and equipped with solid state illuminators, CCD monochrome camera, and a 100x objective lens (1.4 numerical aperture). If needed, use alternative microscopy devices with appropriate hardware and software configurations that can be customized according to the protocol settings.
   1. Dissolve a heavy patch of L. pneumophila in 1 mL of ddH₂O, vortex and pipet 2–3 µL of the dilution onto the pads. Place a 50 x 24 x 0.15 mm³ coverslip gently over the adhesive frame.
   2. In the capture window adjust the ND to 180. Adjust the binning to 2×2 and use the 488 nm channel to expose the sample between 500–1,000 ms. Validate the specificity of the fluorescence signal by imaging untagged L. pneumophila with the same parameters (Figure 4).

4. Quantification of Polar Localization and Dynamics of Dot/Icm Components

NOTE: The following steps are designed for images with 0.129 µm per pixel that were acquired with 2 x 2 binning.

1. Quantification of polarity for sfGFP fusion proteins (Figure 5)
   1. Adjust the image contrast so the bacteria are clearly visible. Use the region tool to place a 0.25 x 1.3 µm² rectangle starting at the pole and extending into the cytoplasm. The rectangle must remain precisely within the bacterial borders.
   2. Mark at least 200 bacteria and use the Region to Mask button to create masks to the regions of interest. Under Mask Statistics and Mask Scope, choose Object. Then, under Features and Intensity, mark Mean Intensity and Variance.
   3. Export the data and calculate the polarity scores of each bacterium as the ratio between the variance to the mean intensity.
   4. For high-throughput applications use a phase objective and an appropriate condenser setup to acquire images with the phase and 488 nm channels. Make sure to choose fields of view where the bacteria are fully separated.
   5. Adjust the phase channel contrast of the image to a level where the bacteria are clearly visible. Open the dual channel image, launch the Create Segment Mask window and change the channel to phase.
   6. Adjust an appropriate threshold and remove small objects with the Define Objects button. Under Refine Mask choose Remove Edges Objects and later separate masks of bacteria that are adjacent to each other.
   7. Calculate the polarity scores of the signal in the 488 channel for each cell as was previously described in steps 4.1.2–4.1.3.
2. Quantification of dynamics for sfGFP fusion proteins (Figure 6)
   NOTE: Follow the instructions in section 3 to prepare a sample for image acquisition. Dynamics are defined as changes in intensity over time and the following steps are designed for short imaging periods (i.e., several minutes). Add to the pad the appropriate supplements if longer imaging periods are desired.
   1. In the Image Capture window, mark Timelapse, enter the time of intervals in the interval box, and enter 2 in the # of Time Points box. Acquire two successive images of L. pneumophila expressing the fluorescent protein of interest.
   2. Adjust the image contrast until the cells are clearly visible. Follow Figure 6A and the descriptions below to place three different masks. Use the region tool to place a 0.25 x 0.25 µm square in the middle of at least 400 cells.
   3. Use the Region to Mask button to create a mask (mask 1) of the squares of interest. Create a new empty mask (mask 2) and use the pixel tool or the polygon tool to mark the entire cell area of at least 25 random cells, which will be used to calculate fluorescence bleaching. Create a new empty mask (mask 3) and use the large brush tool to mark areas between the cells, which will be used for background subtraction.
   4. Under Mask Statistics and Mask Scope, choose Object for mask 1 and mask 2. Then, under Features and Intensity, choose Mean Intensity and export the data of the two masks. For mask 3, export the mean intensity of the entire mask.
   5. Calculate the change in fluorescence intensity for each object in mask 1 using the following formulas:

where mask 1 is a 0.25 µm × 0.25 µm square placed at the cell center, mask 2 covers the whole cell, mask 3 is the background between cells, t1 is the mean intensity in the first time point, and t2 is the mean intensity in the second time point.

5. Detection of sfGFP Mass Density with Cryo-ET

1. Sample preparation, data collection, and reconstruction
   1. Grow a heavy patch of L. pneumophila expressing a fluorescently tagged Dot/Icm protein on CYE-agar-streptomycin plates for 48 hours at 37 °C. Resuspend the cells in ddH₂O to OD₆₀₀ ~0.7. Add 5 µL of colloidal gold particles (BSA Tracer, 10 nm) to 20 µL of the cell suspension.
   2. Pipette 5 μL of the cell mixture onto freshly glow-discharged holey carbon grid (R 2/1 on Cu 200 mesh) and let stand for 1 min. Blot with filter paper and freeze in liquid ethane using a gravity-driven plunger apparatus as described previously^14,15.
   3. Image the frozen-hydrated specimens with a 300 kV transmission electron microscope equipped with a field emission gun, an energy filter, Volta phase plate, and a direct detection device. Collect single-axis tilt series at 26,000x and 42,000x magnifications, which result in pixel sizes at the specimen level of 5.4 Å/pixel or 3.4 Å/pixel, respectively.
   4. Use the tomographic package SerialEM to collect image stacks at ~0 μm defocus, with a range of tilt angles between -60° and +60° with 3° step increment and accumulative dose of ~60 e^-/Å^2,16. Align dose-fractionated movie images in each stack using MotionCor2¹⁷. Assemble drift-corrected stacks using TOMOAUTO¹⁴.
   5. Align drift-corrected stacks by the IMOD marker-dependent alignment¹⁸. Reconstruct tomograms with the SIRT method¹⁹ for segmentations and direct image analysis and the WBP method²⁰.
2. Subtomogram analysis of sfGFP fusions samples
   1. Use the tomographic package I3 (0.9.9.3) for subtomogram analysis^14,21,22.
      NOTE: The alignment proceeds iteratively, with each iteration consisting of three parts in which references and classification masks are generated, subtomograms are aligned and classified, and class averages are aligned to each other.
   2. Use 4 x 4 x 4 binned subtomograms for an initial alignment. Merge particles that belong to class averages and show electron density corresponding to sfGFP. After sorting particles with sfGFP fusions, use 2 x 2 x 2 binned subtomograms for a focused alignment of a region of interest (like the Dot/Icm cytoplasmic ATPase complex) to obtain a high-resolution structure.

Representative Results

Homologous recombination with double selection in two steps was used to construct the defined insertion of sfGFP. In the first step, triparental mating was performed, where the pRK600 conjugative plasmid (an IncP plasmid) from the E. coli helper strain MT616 was mobilized to the donor E. coli strain with the suicide vector pSR47S containing the sfGFP gene flanked by the two homologous regions, the origin of transfer oriT and the Bacillus subtilis counterselection gene sacB. Next, the conjugation system from pRK600 assisted mobilization of the pSR47S derivative into L. pneumophila (Lp01, Figure 2A). Clones that successfully integrated pSR47S by a single crossover event were selected due to the antibiotic resistance gene Kan^R and were propagated to allow a second crossover. Clones that lost the sacB gene during the second crossover were selected by plating the cells on the counterselective agent sucrose (Figure 2, Figure 3). Immunoblot analysis with an anti-GFP antibody (α-GFP) was performed to determine whether the sfGFP was cleaved and to assess the stability and expression level of the fusion protein in a wild type or in a complete dot/icm deletion mutant. In addition, before proceeding to microscopy, the correct integration of the sfGFP gene into the chromosome was confirmed by PCR (Figure 3). Once stability and the integration of the fusions were confirmed, live cell fluorescent microscopy was performed, and specificity of the fluorescent signal clone was compared to a parental sfGFP-negative strain. In addition, dual imaging using bright field or phase microscopy was used to examine the proportion of cells that possessed a specific fluorescent signal (Figure 4).

Only a fraction of the cells that produced DotB-sfGFP displayed polar localization of the fusion protein (Figure 5A). To characterize the distribution of DotB fluorescence signal, the intensity of DotB-sfGFP along the longitudinal axis in those cells was quantified. DotB intensity at the poles was about 2x higher than in the cytosol (Figure 5B). Because it is important to determine whether the localization of fluorescently tagged fusion proteins has biological relevance, we asked if DotB-sfGFP polar localization was dependent on a fully assembled T4BSS. Therefore, polarity scores of DotB-sfGFP indicating the ratio of mean fluorescence measured at the poles compared with fluorescence measured near the middle of the cell were determined for individual cells in a wild type and in a T4BSS mutant (Figure 5C–E). Indeed, deletions of the entire Dot/Icm system abrogated polar positioning of DotB-sfGFP, confirming that polar puncta represent association of DotB with the T4SS and the importance of the Dot/Icm machine for recruitment of this ATPase. In most L. pneumophila cells, DotB-sfGFP also displayed cytosolic movement, indicating that it is present in a dynamic cytosolic population but is also capable of associating with the polar Dot/Icm complex. To quantify differences in dynamic motion between the DotO and DotB ATPases, two successive images were acquired and changes in the fluorescence intensity were determined and compared to sfGFP (Figure 6A). While sfGFP and DotO-sfGFP showed little to no dynamic patterns, DotB-sfGFP intensity changes in the cytosol were shifted and were significantly higher. These observations indicate that the DotB ATPase was present in a dynamic cytosolic population and at a late-stage assembly reaction, the spatially dynamic DotB ATPase was recruited to the polar-localized Dot/Icm T4SS (Figure 6B).

To define the spatial location of DotB relative to the Dot/Icm machine, high-throughput cryo-ET was used to visualize the intact T4BSS apparatus in a L. pneumophila strain expressing DotB-sfGFP. First, reconstruction of a L. pneumophila cell expressing DotB-sfGFP was performed and revealed typical multiple cone-shaped complexes embedded in the cell envelope (Figure 7A, B). We demonstrated previously using epistatic experiments with a variety of Dot/Icm mutants, fluorescent microscopy, and cryo-ET that DotB is spatially located below a unique assembly of the ATPase DotO¹⁰. Subtomogram averaging of the strain expressing DotB-sfGFP determined the overall positioning of DotB in relation to the intact T4BSS machine. This analysis revealed an extra density correlating with the additional mass of sfGFP (Figure 7C). Finally, three-dimensional surface renderings of the entire Dot/Icm T4SS indicates that sfGFP is positioned below the DotB hexamer, which assembles as a disc at the base of the cytoplasmic ATPase complex (Figure 8)¹⁰.

Figure 1: Three-dimensional surface renderings of L. pneumophila Dot/Icm T4SS. OM = outer membrane, IM = inner membrane, PG = peptidoglycan. The image was modified from Chetrit D. et al.¹⁰. Please click here to view a larger version of this figure.

Figure 2: Schematic overview of homologous recombination and allelic exchange used to insert sfGFP into L. pneumophila chromosome. (A) Triparental mating performed between an E. coli MT616 helper strain that harbors the conjugative plasmid pRK600 (an IncP plasmid), a donor E. coli strain that is transformed with the suicide vector pSR47S, and L. pneumophila as a recipient. (B) Schematic model depicting the steps required to insert sfGFP into the L. pneumophila chromosome in order to generate C-terminal sfGFP fusion proteins. Selection of allelic exchange mutants was accomplished in two steps, using kanamycin and the counter selectable marker sacB. Please click here to view a larger version of this figure.

Figure 3: The allelic exchange assay workflow. L. pneumophila was streaked on the appropriate selections and growths times (see the protocol section for details). Later, insertion of sfGFP into the chromosome was confirmed with colony PCR using primers complementary to the sfGFP gene and to the chromosomal region flanking the insertion site. The stability of the fusion protein was examined by immunoblot analysis using an anti-GFP antibody. Bottom left panels: Lane 1 = untagged Lp01, Lane2 = sfGFP expressed from dotB operon, Lane 3 = DotB-sfGFP expressed in a complete dot/icm deletion mutant, Lane 4 = DotB-sfGFP expressed in a wild type background. WB = western blot. The number of clones analyzed in each step are indicated in parentheses. Please click here to view a larger version of this figure.

Figure 4: Live cell imaging of L. pneumophila expressing fluorescently tagged Dot/Icm subunits. (A) L. pneumophila from a two day heavy patch was resuspended in ddH₂O and placed on top of 1% low melting agarose pads. (B) The specificity of the fluorescence signal of a strain that chromosomally encodes DotB-sfGFP was compared to the parental GFP-negative Lp01 strain. Scale bar = 3 μm. Please click here to view a larger version of this figure.

Figure 5: Quantification of polar localization of Dot/Icm system subunits. (A) Real-time visualization of DotB-sfGFP expressed in Lp01 showed polar localization. DotB-sfGFP displayed a diffuse pattern when expressed in a strain where the entire dot/icm gene cluster was deleted (∆T4SS). Unfused sfGFP expressed from the dotB operon showed a diffuse cytosolic pattern and served as a control. Scale bar = 3 μm. (B) Quantification of DotB-sfGFP fluorescence intensity along the longitudinal axis of cells with polar or diffuse patterns. The samples sizes were n = 20 cells for polar and nonpolar cells. *p = 0.02, **p ≤ 0.0001. The significance was calculated by two-tailed Student`s t-test and compared to the intensity at the pole. (C) Schematic model of the masks used to quantify polarity. (D) DotB-sfGFP as shown in Figure 5A with the masks used to quantify polarity. (E) The abundance of DotB-sfGFP at the poles, when it is expressed in a wild type background or when the entire dot/icm gene cluster was deleted, is displayed as frequency curves. The samples sizes were n = 200 cells for each strain. * p < 0.0001. The significance was calculated by two-tailed Mann Whitney test and compared to sfGFP expressed from the dotB operon. Please click here to view a larger version of this figure.

Figure 6: Quantification of the dynamics of Dot/Icm fluorescent fusions. (A) A schematic model depicting the masks used to quantify the changes in fluorescence intensity for a dynamic or static sfGFP fusion protein. T_1,2 are the time of the first and second time points and m_1,2,3 are mask 1, mask 2, and mask 3. (B) The dynamics of DotB-sfGFP, DotO-sfGFP, and sfGFP expressed from the dotB operon are displayed as frequency curves. The samples sizes were n = 400 cells for each strain. *p = 3.4 × 10^-12, ***p = 1.0 × 10^-147. The significance was calculated by two-tailed Student`s t-test compared to sfGFP. Please click here to view a larger version of this figure.

Figure 7: Subtomogram averaging is used to determine sfGFP density of DotB-sfGFP. (A) A tomographic slice acquired with 26,000x magnification from a representative L. pneumophila cell expressing DotB-sfGFP, showing multiple Dot/Icm machines embedded in the cell envelope. (B) A tomographic slice acquired with a 42,000x magnification of the same L. pneumophila cell pole shown in A. OM = outer membrane, IM = inner membrane. (C) A reconstruction from a strain expressing DotB-sfGFP shows densities corresponding to DotB and sfGFP (yellow arrows). Please click here to view a larger version of this figure.

Figure 8: Three-dimensional surface renderings of the Dot/Icm cytosolic complex. Three-dimensional surface renderings of the entire Dot-Icm T4SS from ∆dotB (A), wild type (B), and dotB-gfp (C) L. pneumophila strains. (D) The structure of the cytosolic complex shows the positioning of DotB (purple) and sfGFP (green). The X-ray structure of DotB (PDB: 6GEB²³) was docked to the DotB map. IM = inner membrane. Please click here to view a larger version of this figure.

Discussion

Elucidating the functions of bacterial secretion systems is key to a complete understanding of host-pathogen interactions. Secretion systems are complex machines that can inject effectors proteins into host cells, and in some cases promote establishment of a subcellular niche that supports bacterial replication. The above method provides important new tools for studying the Dot/Icm secretion system of the respiratory bacterial pathogen Legionella pneumophila, yielding clues to the mechanisms of effector translocation that are essential for its pathogenicity. This approach can be applied to other secretion systems by employing live bacterial cell imaging and structural studies to dissect their activity at the molecular level. This can be accomplished by studying the spatiotemporal dynamics of fluorescently tagged secretion system components using time-lapse video microscopy.

We constructed clean and unmarked mutations, where a gene was tagged or replaced by a modified allele, which is a fundamental approach for the understanding of pathogenicity at a molecular level as well as for the definition of structure-function relationships that may lead to the production of drug therapies^24,25. Strains of L. pneumophila that have been genetically modified to eliminate individual Dot and Icm proteins can be used to determine the structure of the Dot/Icm system by cryo-ET. An important step in the protocol is to compare the functionality of N-terminus and C-terminus tagged proteins. Therefore, strains encoding the fusions should be tested for replication in eukaryotic host cells, visualized by microscopy to determine whether tagged protein localization in the bacterial cell is dependent on a fully assembled secretion machine, and assessed for stability of proteins expressed from upstream and downstream genes.

This protocol utilizes live cell microscopy experiments, which are widely used to study complex biological processes such as signal transduction, self-assembly, active trafficking, and gene regulation. Understanding the dynamics, trajectories, and interactions of single particles in cells represents a fundamental approach in cell biology²⁶. However, given that tracking single particles may not always be possible due to fast molecular movement or low signal-to-noise ratio, evaluating changes in fluorescence intensity may serve as a straightforward approach to quantify differences in dynamics patterns. In addition, because only two time points are required for the analysis, this methodology guarantees only minimal exposure of the sample to fluorescent light and thus prevents major bleaching of the sfGFP fluorescence intensity. Because the agarose pads do not contain any source of nutrients, it is important to note that live imaging should be performed immediately after taking the cells from the CYE agar plates. If longer periods of imaging time are required, then supplements that support L. pneumophila survival and growth should be added to the pads. Determining the polarity of fluorescently tagged proteins can also be modified for high-throughput applications. If these are needed, a phase objective and an appropriate condenser setup should be used to acquire dual channel images (i.e., phase and fluorescent channels). It is crucial to choose fields where the bacteria are fully separated. Use the phase channel to mask entire cells and quantify the ratio of intensities as described above.

Cryo-electron microscopy (cryo-EM) and cryo-ET offer robust visualization of nanoparticles. These techniques involve imaging of the sample in a frozen-hydrated state, allowing visualization of nanoparticles essentially as they exist in their cellular environment²⁷. Because resolution is limited by the thickness of the sample, structures obtained by cryo-ET have lower resolution than the ones that are obtained by cryo-EM. Nevertheless, because in each image tomogram type IV machineries are present in multiple copies, increased resolution can be obtained by increasing the sample size for subtomogram averaging. One major advantage of cryo-ET is that visualization of complex assemblies in intact bacterial cells may better preserve their structure, which can undergo drastic change upon extraction from their natural environment.

In conclusion, the study of dynamic processes and localization of molecules are among the first steps toward understanding the function of biological systems. The strength and focus of the method described here is the direct visualization of Dot/Icm components in living bacteria, which provides the possibility to uncover dynamic processes governing the functions of type IVB machines or other multiprotein assemblies. In addition, coupling live imaging to cryo-ET is a strategy that allows characterization of key components in the Dot/Icm machine in relation to the greater structure of the Legionella pneumophila secretion system, and understanding of how they assemble and direct conformational changes in the apparatus. One limitation of fluorescent protein tags used for studying protein behavior in living cells is their relatively large size, which might influence the function and properties of the tagged protein. Subsequently, assessment of stability and functionality is required. Another limitation of this approach is that currently atomic structures cannot be obtained due to the limited resolution resulting from the thickness of the sample. Modeling and fitting can be an effective way to analyze the densities, thus allowing localization of subunits or evaluation of conformational changes. Future investigations and further perfectioning of this approach will lead to a better understanding of how different subassemblies comprise a functional type IV apparatus.

Materials

Name	Company	Catalog Number	Comments
10 nm colloidal gold particles	Aurion	25486
100x Plan Apo objective (1.4 NA)	Nikon
ACES	Sigma-Aldrich	A9758
Activated charcoal	Sigma-Aldrich	C5510
Agaroze GPG/LMP, low melt	American bioanalytical	AB00981
Bacto dehydrated agar	BD	214010
CoolSNAP EZ 20 MHz digital monochrome camera	Photometrics
Gene Frame, 1.7x2.8 cm, 125 µL	Fisher Scientific	AB-0578
Holey Carbon grid R 2/1 Cu 200 mesh	Quantifoil	Q225-CR1
Iron(III) nitrate nonahydrate	Sigma-Aldrich	216828
K2 Summit camera for cryo-EM	GATAN
L-Cysteine	Sigma-Aldrich	C7352
Microscope cover slides 22x22 mm	Fisher Scientific	12-542B
Microscope cover slides 24x50 mm	Fisher Scientific	12-545K
Microscope slides 25x75x1 mm	Globe Scientific	1380
SlideBook 6.0	Intelligent Imaging Innovations
Spectra X light engine	Lumencor
Taq 2X Master Mix	New England BioLabs	M0270
Titan Krios	Thermo Fisher Scientific
Yeast Extract	BD	212750