Development of a Fluorescent Tool for Studying Legionella bozemanae Intracellular Infection

Legionnaires’ disease incidence is on the rise, with the majority of cases attributed to the intracellular pathogen, Legionella pneumophila. Nominally a parasite of protozoa, L. pneumophila can also infect alveolar macrophages when bacteria-laden aerosols enter the lungs of immunocompromised individuals. L. pneumophila pathogenesis has been well characterized; however, little is known about the >25 different Legionella spp. that can cause disease in humans. Here, we report for the first time a study demonstrating the intracellular infection of an L. bozemanae clinical isolate using approaches previously established for L. pneumophila investigations. Specifically, we report on the modification and use of a green fluorescent protein (GFP)-expressing plasmid as a tool to monitor the L. bozemanae presence in the Acanthamoeba castellanii protozoan infection model. As comparative controls, L. pneumophila strains were also transformed with the GFP-expressing plasmid. In vitro and in vivo growth kinetics of the Legionella parental and GFP-expressing strains were conducted followed by confocal microscopy. Results suggest that the metabolic burden imposed by GFP expression did not impact cell viability, as growth kinetics were similar between the GFP-expressing Legionella spp. and their parental strains. This study demonstrates that the use of a GFP-expressing plasmid can serve as a viable approach for investigating Legionella non-pneumophila spp. in real time.


Introduction
Legionella are Gram-negative, rod-shaped bacteria that cause Legionnaires' disease (LD), which is a severe type of pneumonia that remains a global and public health concern. In recent years, cases of Legionnaires' disease have increased at an alarming rate, with outbreaks having been reported across North America, Europe, and Australia [1][2][3]. Although the exact reason for this increase is unknown, it is likely multifactorial, representing increased disease awareness and diagnostic testing, in addition to an increase in the susceptibility of the global population, as well as global warming, and aging infrastructure, both of which favor Legionella growth in the environment. In Canada, Legionella incidence has risen 10-fold since 2000, with 635 cases having been reported to the Public Health Agency of Canada in 2018 [4]. Incidence has also increased 9-fold in the United States (US) and 2-fold in Europe [5,6]. Legionella cases are thought to be under-reported for reasons discussed later. A study by Collier et al. [7] estimated that Legionella species (spp.) caused 10,800 (95% confidence interval (CI) 7280-13,100) patient hospitalizations in the US in 2014, equating to $401 million (95% CI $79-1690) in total direct healthcare costs, highlighting the significant burden that these organisms have on the healthcare system.
In the environment, Legionella spp. can be found in naturally occurring and man-made water systems such as showers, cooling towers, fountains, lakes, and streams [8]. To survive to create fluorescent Legionella strains, which can be used to characterize non-pneumophila Legionella bacterial growth kinetics in vitro (broth) and in vivo (Acanthamoeba castellanii infection model).

Bacteria, Plasmids, Primers, and Growth Conditions
Bacterial strains and plasmids used in this study are listed in Table 1. Primers and amplicon product sizes are listed in Table 2. Escherichia coli DH5α strains were grown in Luria-Bertani broth or on agar, supplemented with kanamycin (40 µg/mL) or ampicillin (100 µg/mL) where appropriate, and incubated overnight at 37 • C. Legionella spp. were streaked from frozen stocks (Buffered Yeast Extract (BYE) + 10% dimethyl sulfoxide) onto Buffered Charcoal Yeast Extract (BCYE) agar at 37 • C with 5% CO 2 for 3-4 days then if required, heavily inoculated into BYE broth [32] for 18-24 h at 37 • C on a tube rotator, supplemented with kanamycin (25 µg/mL).

Construction of GFP-Expressing Legionella Strains
In general, L. pneumophila can maintain RSF1010-based plasmids [34,35]. In this view, pKB127 was genetically modified to replace the td∆I marker with a kanamycin resistance marker, sourced from p34S, creating pGFP-Kan for use with the Legionella strains in this study. pKB127 refers to the insertion of a 245 base pair (bp) magA promoter region in the promoterless gfp reporter plasmid pBH6119 [26] and was chosen, since magA is a strong promoter that tends to be more expressed post-exponentially. pBH6119 carries a thymidylate synthase (td∆I) that rescues the thymidine auxotrophy of Lp02, which is a derivative of L. pneumophila Philadelphia-1 [36], thereby providing a convenient non-antibiotic-based plasmid selection. However, since Legionella spp. are generally not thymidine auxotrophs, a kanamycin resistance marker was used as a substitute.
Clinical isolates of L. bozemanae and L. pneumophila, along with L. pneumophila ATCC 33,152 type strain, were confirmed to lack intrinsic resistance to kanamycin by plating on BCYE supplemented with kanamycin and checked for growth daily for 96 h.
To generate GFP-expressing Legionella strains, the P magA gfp reporter plasmid (pKB127) was genetically modified by replacing the td∆I selection marker with a kanamycin resistance marker. Briefly, pKB127 was cut with EcoRI and SacII to drop out the td∆I selection marker, with the PCR-amplified kanamycin resistance marker ligated in its place, creating pGFP-Kan. E. coli DH5α was used as the host strain for cloning approaches. Amplification of the kanamycin resistance marker with engineered restriction sites was achieved with primer set PF KanR EcoRI/PR KanR SacII and template p34S using Q5 High Fidelity Polymerase (New England Biolabs, Canada). The construct of pGFP-Kan was verified by polymerase chain reaction (PCR) using a combination of the internal and external primer sets listed in Table 2. Then, the plasmid was electroporated into the L. bozemanae, L. pneumophila clinical, and L. pneumophila ATCC 33,152 type strains according to the methods described in Tanner et al. [35] to create the strains LP 1031, LB 1032, and LP 1030, respectively, which were then plated on BCYE supplemented with kanamycin to select for the kanamycin resistance phenotype. Plasmid integration was confirmed by screening the kanamycin-resistant GFP-expressing Legionella strains with the PCR confirmation primers ( Table 2).

In Vitro Growth Kinetics and Relative Fluorescence Assays
In vitro growth of the Legionella parental and GFP-expressing strains was assessed in BYE broth as described previously [32,37]. Briefly, each strain was grown on the appropriate BCYE medium for 96 h and then resuspended in BYE, either with or without kanamycin, and inoculated into a 96-well plate at a final concentration of 1.5 × 10 7 cells/mL [35]. The plate was incubated for 24 h at 37 • C with shaking in a BioTek Synergy HTX multimode automated microplate reader and optical density (600 nm; OD 600 ) and GFP relative fluorescence (excitation and emission wavelengths 485/20 nm and 528/20 nm, respectively) was measured every hour. Media background fluorescence and absorbance was subtracted from the relative fluorescence unit (RFU) values and optical density (OD 600 ) data, respectively, and values for each of the GFP strains along with their parental strains were plotted over time.

Intracellular Growth Assays in Acanthamoeba castellanii
For maintenance of A. castellanii ATCC 30234, amoeba were cultured in ATCC 712 medium, Peptone Yeast Extract Glucose (PYG) liquid medium, at 25 • C [26]. Amoeba were split every 48-96 h (depending on the initial seeding dilution). For infections, a 48-h culture of A. castellanii was washed and resuspended in A. castellanii buffer (AC buffer; 4 mM magnesium sulfate, 0.4 mM calcium chloride, 3.4 mM sodium citrate dihydrate, 0.05 mM ferrous ammonium sulfate hexahydrate, 2.5 mM disodium phosphate, and 2.5 mM potassium dihydrogen phosphate) and then seeded into 24-well plates at a density of 1 × 10 6 cells/mL [10]. Once cells were adherent (≈2 h at 25 • C), bacteria (from 3-4day-old BCYE-agar plates) were suspended in AC buffer, and titers were determined at an OD 600 . Then, amoeba were infected with Legionella parental and GFP-expressing strains at a multiplicity of infection (MOI) of 0.1. Following a 1 h incubation at 25 • C, to synchronize the infection, amoeba were washed 3 times in AC Buffer to remove any extracellular bacteria that had not been taken up into the cells [35]. At various time points (1,48,144, and 216 h post-infection), amoeba were lysed using mechanical disruption (i.e., by scraping the monolayer, vortexing the sample for 30 s followed by centrifugation for 8 min at 14,800 rpm) [38], and bacterial counts were determined by serial dilution BCYE plate counts (colony-forming units (CFU)/mL).

In Vitro and In Vivo Microscopic Imaging of the GFP-Expressing Legionella Strains
All GFP-expressing Legionella strains were cultured on BYCE agar supplemented with kanamycin for 96 h prior to in vitro and in vivo imaging studies. For in vitro imaging, a single isolated colony was resuspended in 100 µL of BYE and mounted on a glass microscope slide. Then, bacteria were heat-fixed (i.e., by placing a drop of bacteria on a slide, letting it dry, and then passing the slide through a flame 2 times) and mounted in ImmunoMount medium (Thermo Scientific, Waltham, MA, USA) before viewing.
To visualize GFP-expressing Legionella within A. castellanii, the day prior to infection, amoeba were placed in a 6-well plate containing sterile poly-lysine coated coverslips (Neuvitro Corp, Vancouver, WA, USA) at a density of 3 × 10 6 cells/well and allowed to adhere overnight in a 25 • C incubator. Then, A. castellanii was infected with each GFPexpressing strain at an MOI of 5, and the infection was incubated for 6 or 12 h. Following infection, the cells were washed 3 times with AC buffer, stained with an ER-Tracker Red dye (BODIPY TR Glibenclamide; ThermoFisher Scientific, Whitby, ON, Canada) for 30 min at 37 • C, and fixed in formaldehyde. Then, the coverslips containing the infected amoeba were mounted onto glass slides with ProLong Gold Antifade Mountant with DAPI (ThermoFisher Scientific, Canada) per the manufacturer's instructions.
For fluorescence imaging, all slides were examined by differential interference contrast (DIC) and confocal microscopy using an LSM 700 confocal microscope with accompanying Zeiss software. GFP-expressing Legionella, the ER-Tracker Red Dye, and DAPI had an excitation and emission spectra of 488/583, 587/615, and 405/461 nm, respectively. Images were generated using a 40× oil immersion objective. All fluorescent images were edited using ImageJ version 1.52A [39].

Statistical Analyses
All experiments were performed in triplicate unless otherwise stated. Data were analyzed using STATA®version 14. Bacterial optical densities and growth (CFU/mL) are reported as the mean and standard error. To determine if optical densities were similar between the fluorescent strains and their parental counterparts at inoculation, a T-test was performed, while a one-way analysis of variance was used to compare optical densities between the Legionella parental strains (L. bozemanae (948), L. pneumophila clinical (950) and L. pneumophila ATCC 33152) as well as those from the GFP-expressing strains (L. bozemanae (LB 1032), L. pneumophila clinical (LP 1031) and L. pneumophila ATCC 33,152 (LP 1030)). To compare bacterial optical densities and fluorescence intensity over time between the GFP-Legionella strains, a repeated-measures analysis of variance was carried out with a Bonferroni post-hoc analysis conducted when data were significant (p ≤ 0.05). All data were plotted using GraphPad Prism version 9 (GraphPad Software, San Diego, CA, USA).

Generation of GFP-Expressing Legionella Strains
Transformation of the Legionella strains by electroporation was carried out with relative ease, requiring one attempt when constructing the GFP-L. pneumophila strains and two attempts for the GFP-L. bozemanae strain. Following electroporation and plating on BCYE supplemented with kanamycin, transformants were streaked out on fresh agar plates for purity, after which frozen stocks were prepared. Fluorescence microscopy confirmed GFP expression ( Figure 1). Colony characteristics of the GFP isolates, such as colony shape and size, did not appear to differ from their parental strains.

Plasmid-Borne GFP Expression Does not Affect the Growth of Legionella In Vitro
To determine if the presence of pGFP-Kan affects Legionella growth profiles, bacterial kinetic studies were performed to compare the growth rates of each of the GFP-expressing Legionella strains with their parental counterparts. Each of the Legionella pairs (with and without pGFP-Kan) displayed similar growth curves over the 24 h period with all of the strains reaching their peak density between 12 and 16 h (Figure 2A-C). No significant differences were seen between the optical densities of the parental strains and those of their GFP-expressing counterparts.

Plasmid-Borne GFP Expression Does Not Affect the Growth of Legionella In Vitro
To determine if the presence of pGFP-Kan affects Legionella growth profiles, bacterial kinetic studies were performed to compare the growth rates of each of the GFP-expressing Legionella strains with their parental counterparts. Each of the Legionella pairs (with and without pGFP-Kan) displayed similar growth curves over the 24 h period with all of the strains reaching their peak density between 12 and 16 h (Figure 2A-C). No significant differences were seen between the optical densities of the parental strains and those of their GFP-expressing counterparts.
When comparing the growth profiles of the L. pneumophila parental and transformant strains with that of the L. bozemanae strains, higher densities were seen among the L. pneumophila species compared to those seen with L. bozemanae (OD 600 of 1.0 vs. 0.5, respectively, p < 0.0001). Likewise, the GFP-expressing L. pneumophila strains also displayed higher rates of fluorescence compared to those of the GFP-expressing L. bozemanae (p < 0.0001), with 24-h RFU values of 3385, 3137, and 857 for strains LP 1030, LP 1031, and LB 1032, respectively ( Figure 2D-F). None of the Legionella parental strains were fluorescent as evidenced by the RFU values, which mimicked that of the BYE media control. When comparing the growth profiles of the L. pneumophila parental and transformant strains with that of the L. bozemanae strains, higher densities were seen among the L. pneumophila species compared to those seen with L. bozemanae (OD600 of 1.0 vs. 0.5, respectively, p < 0.0001). Likewise, the GFP-expressing L. pneumophila strains also displayed higher rates of fluorescence compared to those of the GFP-expressing L. bozemanae (p < 0.0001), with 24-h RFU values of 3385, 3137, and 857 for strains LP 1030, LP 1031, and LB 1032, respectively ( Figure 2D-F). None of the Legionella parental strains were fluorescent as evidenced by the RFU values, which mimicked that of the BYE media control.
Similarly, when RFU values were normalized to optical density, L. bozemanae fluorescence was consistently lower than that of the GFP-expressing L. pneumophila strains (Figure 2G). Indeed, the initial GFP activity of the LB 1032 strain began at approximately 2000 Similarly, when RFU values were normalized to optical density, L. bozemanae fluorescence was consistently lower than that of the GFP-expressing L. pneumophila strains ( Figure 2G). Indeed, the initial GFP activity of the LB 1032 strain began at approximately 2000 RFU/OD 600 units, while activity was >3000 RFU/OD 600 units for the LP 1031 and LP 1030 strains. Each of the GFP-expressing strains displayed a similar trend over the 24-h experiment: the RFU/OD 600 values decreased throughout the first 7-10 h of the assay, at which time the RFU/OD 600 values rebounded, increasing gradually until they returned to their approximate starting values. The growth-dependent GFP expression profiles exhibited by all three strains reflect the activity of the magA promoter, which tends to be higher post-exponentially as reported elsewhere [26].

Plasmid-Borne GFP Expression Does Not Affect Legionella Growth within A. castellanii protozoa
To determine if plasmid-borne GFP expression altered Legionella growth and viability in vivo, A. castellanii protozoa were infected with each of the Legionella strains, and intracellular growth kinetics were assessed. Legionella growth profiles appeared to be unaffected by GFP expression. Both parental and GFP-expressing L. pneumophila strains (ATCC 33,152 and clinical) showed similar growth profiles, achieving 10 7 CFU/mL by the 216-h time point (Figure 3). Likewise, the GFP-expressing L. bozemanae strain also exhibited a similar growth profile in vivo compared to its parental counterpart.
strains. Each of the GFP-expressing strains displayed a similar trend over the 24-h experiment: the RFU/OD600 values decreased throughout the first 7-10 h of the assay, at which time the RFU/OD600 values rebounded, increasing gradually until they returned to their approximate starting values. The growth-dependent GFP expression profiles exhibited by all three strains reflect the activity of the magA promoter, which tends to be higher postexponentially as reported elsewhere [26].

Plasmid-Borne GFP Expression Does not Affect Legionella Growth within A. castellanii protozoa
To determine if plasmid-borne GFP expression altered Legionella growth and viability in vivo, A. castellanii protozoa were infected with each of the Legionella strains, and intracellular growth kinetics were assessed. Legionella growth profiles appeared to be unaffected by GFP expression. Both parental and GFP-expressing L. pneumophila strains (ATCC 33,152 and clinical) showed similar growth profiles, achieving 10 7 CFU/mL by the 216-h time point (Figure 3). Likewise, the GFP-expressing L. bozemanae strain also exhibited a similar growth profile in vivo compared to its parental counterpart.
However, as was seen with the in vitro experiments, L. bozemanae growth was less efficient than L. pneumophila, which was highlighted by the lower log titers seen with the L. bozemanae strains (p < 0.001). At each time point, L. bozemanae growth was approximately 1 log lower than the L. pneumophila species, with L. bozemanae achieving 10 5 and 10 6 CFU/mL by the end of the 216-h experiment for the parental and fluorescent strain, respectively. These results indicate that this non-pneumophila spp. may not be as adept as L. pneumophila for intracellular uptake or replication within protozoa.

The Intracellular Life Cycle of L. bozemanae is similar to that of L. pneumophila
The intracellular lifestyle of L. pneumophila in protozoa is well established; subversion of the digestive pathway and formation of the replicative vacuole is dependent on the Dot/Icm pathway [40]. However, very little is known about the intracellular lifestyle of non-pneumophila species, such as L. bozemanae.
To gain further insight into Legionella uptake, growth rates, and fluorescence over time, we visualized Legionella infection in A. castellanii using confocal microscopy. At 6 h post-infection, bacteria can be seen within the host ( Figure S1); however, A. castellanii infected with the L. pneumophila spp. tended to have more than 1 bacterium present per cell. However, as was seen with the in vitro experiments, L. bozemanae growth was less efficient than L. pneumophila, which was highlighted by the lower log titers seen with the L. bozemanae strains (p < 0.001). At each time point, L. bozemanae growth was approximately 1 log lower than the L. pneumophila species, with L. bozemanae achieving 10 5 and 10 6 CFU/mL by the end of the 216-h experiment for the parental and fluorescent strain, respectively. These results indicate that this non-pneumophila spp. may not be as adept as L. pneumophila for intracellular uptake or replication within protozoa.

The Intracellular Life Cycle of L. bozemanae Is Similar to That of L. pneumophila
The intracellular lifestyle of L. pneumophila in protozoa is well established; subversion of the digestive pathway and formation of the replicative vacuole is dependent on the Dot/Icm pathway [40]. However, very little is known about the intracellular lifestyle of non-pneumophila species, such as L. bozemanae.
To gain further insight into Legionella uptake, growth rates, and fluorescence over time, we visualized Legionella infection in A. castellanii using confocal microscopy. At 6 h post-infection, bacteria can be seen within the host ( Figure S1); however, A. castellanii infected with the L. pneumophila spp. tended to have more than 1 bacterium present per cell. At 12 h post-infection, the bacteria can be seen clustered together and dividing inside the cell, which is a process that appeared to be similar for both species (Figure 4). At 12 h post-infection, the bacteria can be seen clustered together and dividing inside the cell, which is a process that appeared to be similar for both species (Figure 4).

Discussion
To date, research regarding Legionella has rarely focused on species outside of L. pneumophila SG1. Here, we report, for the first time, the generation of a GFP-expressing L. bozemanae strain derived from a clinical isolate obtained from our local hospital. Transformation did not appear to affect the growth kinetic profiles of the GFP-expressing Legionella strains compared to their parental counterparts.
This work extends our knowledge of non-pneumophila species by comparative studies of the growth profiles and relative fluorescence of L. bozemanae to L. pneumophila. Since the Legionella genus encompasses a diverse group of organisms, it is important to expand research to include non-pneumophila spp., particularly those that can cause disease [41][42][43]. In our study, several similarities and differences between L. pneumophila and L. bozemanae were seen. In broth, although each of the strains grew well in BYE at 37 °C, the L. pneumophila strains featured a faster growth rate with higher optical densities than that observed for L. bozemanae. These results match those observed in earlier studies [14,15,44], which reported that the non-pneumophila spp. often exhibit slower and reduced growth rates compared to L. pneumophila when grown in vitro at 37 °C.
In the protozoan model, the GFP-expressing L. bozemanae and L. pneumophila clinical strains, as well as the fluorescent L. pneumophila ATCC 33,152 strain, were all able to infect and replicate within A. castellanii. As amoeba share many similarities to human phagocytic cells, it was not surprising that the L. bozemanae clinical isolate was able to infect and replicate in the amoeba model. Although the GFP-expressing L. bozemanae strain did not attain growth levels as high as the GFP-expressing L. pneumophila strains, a steady increase in bacterial load was seen throughout the experiment. These findings differ from those of Neumeister and co-workers [45], which reported that L. bozemanae was not able to multi-

Discussion
To date, research regarding Legionella has rarely focused on species outside of L. pneumophila SG1. Here, we report, for the first time, the generation of a GFP-expressing L. bozemanae strain derived from a clinical isolate obtained from our local hospital. Transformation did not appear to affect the growth kinetic profiles of the GFP-expressing Legionella strains compared to their parental counterparts.
This work extends our knowledge of non-pneumophila species by comparative studies of the growth profiles and relative fluorescence of L. bozemanae to L. pneumophila. Since the Legionella genus encompasses a diverse group of organisms, it is important to expand research to include non-pneumophila spp., particularly those that can cause disease [41][42][43]. In our study, several similarities and differences between L. pneumophila and L. bozemanae were seen. In broth, although each of the strains grew well in BYE at 37 • C, the L. pneumophila strains featured a faster growth rate with higher optical densities than that observed for L. bozemanae. These results match those observed in earlier studies [14,15,44], which reported that the non-pneumophila spp. often exhibit slower and reduced growth rates compared to L. pneumophila when grown in vitro at 37 • C.
In the protozoan model, the GFP-expressing L. bozemanae and L. pneumophila clinical strains, as well as the fluorescent L. pneumophila ATCC 33,152 strain, were all able to infect and replicate within A. castellanii. As amoeba share many similarities to human phagocytic cells, it was not surprising that the L. bozemanae clinical isolate was able to infect and replicate in the amoeba model. Although the GFP-expressing L. bozemanae strain did not attain growth levels as high as the GFP-expressing L. pneumophila strains, a steady increase in bacterial load was seen throughout the experiment. These findings differ from those of Neumeister and co-workers [45], which reported that L. bozemanae was not able to multiply within A. castellanii and rather showed a decreasing bacterial count in amoeba following infection. A possible explanation for these differing results may be attributed to strain differences as variability regarding infection rates, replication efficiency, and pathology has been documented, even within a single species [42,45,46]. Moreover, the lack of standardization regarding experimental conditions for Legionella infections in the A. castellanii infection model may also be a contributing factor.
Next, confocal microscopy was used to visualize Legionella spp. intracellular growth inside A. castellanii over time. L. bozemanae bacteria appear to replicate and cluster together, which may be indicative of vacuole formation, suggesting that the intracellular lifestyle of L. bozemanae may be similar to that of L. pneumophila. However, these findings need to be further defined through the use of detailed microscopy studies of L. bozemanae strains that are dysfunctional for genetic components known to be associated with pathogenesis as characterized by L. pneumophila.
Comparing the fluorescent intensities between the Legionella strains showed that the L. pneumophila strains, as comparative controls, featured stronger fluorescence than L. bozemanae even when normalizing fluorescence by optical density. Heterogeneity in fluorescence signals can occur due to several factors including differences in the copy numbers of the plasmids inside the cells, differences in plasmid protein expression, or poor GFP folding [47,48]. Nevertheless, the matched GFP expression profiles between the Legionella strains in this study with that previously reported by Hiltz and colleagues [49] indicate the recognition of a promoter region sourced from L. pneumophila by transcription and translation machinery in L. bozemanae, thereby allowing GFP expression. Despite the lower levels of GFP expression in L. bozemanae, the bacterial cells were bright enough to be seen clearly during intracellular infections of A. castellanii, and they can be used for further investigations of host-pathogen interactions. Further work is required to refine this approach by optimizing GFP expression through further genetic modifications of the plasmid as well as conducting whole genome sequencing to check for presence of any naturally occurring plasmids.
Taken together, this study shows the feasibility of using plasmid-borne GFP expression as a tool to investigate the intracellular lifestyle of lesser-known non-pneumophila Legionella species reported to cause Legionnaires' disease. Aside from the well-characterized L. pneumophila, and to a lesser extent, L. longbeachae, very little is known of the mechanisms employed by non-pneumophila Legionella species in causing disease in humans. The use of plasmid-borne GFP is advantageous over that of immunofluorescence, as the latter requires cell fixation as well as specific primary antibodies that may not be available for non-pneumophila Legionella species, whereas the former allows real-time monitoring of bacterial growth dynamics in vitro and in vivo. While antibiotic supplementation for plasmid maintenance and metabolic burden of GFP expression may be negative aspects of the use of GFP, these requirements do not seem to significantly affect the growth of L. bozemanae in vitro and in vivo, as observed in this study. Thus, the use of GFP as a fluorescent marker is beneficial to future studies on non-pneumophila Legionella species.